NETs: Time‑Anchored Economics

NETs Core Claim: The 1.5% CPI Drift

Kyle Novack — Sat, 09 May 2026 16:47:37 GMT

Core claim

The Consumer Price Index (CPI) systematically understates true monetary inflation by about 1.5 percentage points per year. This is not a free parameter; it is a triangulated estimate derived independently from multiple mainstream data sources that were not designed to support this thesis.

Step 1: Direction of the error (logic only)

The direction of the CPI error follows from standard mechanisms, without any new data.

Productivity reduces unit production costs.
BLS Total Factor Productivity (TFP) data show agricultural productivity growth of roughly 1.5% per year since 1948, and non‑farm business TFP growth around 2.1% per year.

Competitive markets pass cost reductions into lower prices than would otherwise exist.
Brookings (Dew‑Becker & Gordon) estimate that a 1% increase in trend productivity growth reduces measured inflation by 1.34%, and the San Francisco Fed explicitly links productivity gains to lower inflation via pass‑through.
CPI measures final retail prices after this pass‑through.
By definition, CPI is constructed from the prices consumers actually pay at the end of the production chain.

It follows that CPI measures net inflation (monetary inflation minus productivity‑driven deflation already embedded in prices) rather than gross monetary inflation. The correction must therefore be upward; only the size is an empirical question.

Step 2: Magnitude of the error (triangulation)

Five independent, mainstream sources converge on an annual understatement in the 1.0–2.1% range, with a central tendency near 1.5%.

The conventional bias literature provides a floor: official work already admits 0.5–1.6% bias, though it is framed as “overstating cost of living” rather than missing productivity deflation.
Sectoral productivity (especially agriculture) and the Case-Shiller vs. CPI wedge cluster around ~1.5%, providing a central empirical anchor.
M2 per capita divergence and non‑farm TFP define a plausible upper bound; pushing the correction much above ~2% produces series (e.g., million‑dollar 1970s houses) that are visibly implausible.

Across all five, a constant drift in the neighborhood of 1.5% is the unique range that (a) fits the mechanisms and (b) produces non‑absurd corrected series.

Step 3: What the 1.5% correction explains

Applied cumulatively from the early 1970s, a 1.5% annual understatement is sufficient to reconcile key macro series with both productivity data and lived experience.

Median household income
- Officially: roughly 83,000 dollars in recent data.
- Under NETs correction: closer to 110,000 dollars, consistent with a roughly 50% wage share of GDP rather than the apparent long‑run collapse in labor’s share.NET
Housing and “asset booms”
- Official CPI‑adjusted home prices show a steep upward trend.
- With ~1.5% drift added, long‑run median home prices in 2024 dollars become nearly flat, matching the view that typical houses have not become radically more expensive in real terms, and that much of the apparent boom is mismeasured inflation.
Money supply and prices
- Since the mid‑20th century, M2 per capita has outpaced CPI by roughly 1.7–2.1% per year.
- CPI plus ~1.5% brings the price level into line with the long‑run path of effective money per person (including velocity), rather than leaving a persistent unexplained gap.

These corrections jointly dissolve several standard puzzles: stagnant wages despite record productivity, a housing “crisis” without proportionate construction cost changes, and a money–prices gap often attributed to mysterious velocity shifts.

Subscribe now

Step 4: How to falsify the NETs 1.5% drift

The argument is only meaningful if it is open to being proven wrong. NETs make a set of clear, testable claims.

4.1 Mechanism‑level falsification

Claim: CPI is, by construction, net of productivity deflation.

The framework would be falsified at the mechanism level if rigorous evidence showed that at least one of the following is false:

Productivity growth does not reduce unit production costs in the aggregate (contrary to BLS TFP and standard cost theory).
Competitive markets do not pass cost reductions into lower prices than would otherwise obtain, even over long horizons.
CPI does not in fact measure final retail transaction prices paid by households, but something earlier in the chain.

If any of those foundations fail, the necessity of an upward correction to CPI disappears.

4.2 Triangulation breakdown across independent series

Claim: A single, approximately constant drift (~1.5%) improves the fit of multiple independent series simultaneously.

Falsification route:

Reconstruct key series (housing, wages, GDP per capita, tax revenue per capita, money supply per capita) under alternative CPI drift adjustments using standard, transparent methods.
Test whether any constant correction in the 1.0–2.0% band
- Flattens housing into a plausible band.
- Stabilizes wages and per-person government revenue.
- Aligns the cumulative price level with per-person money.

If no single drift produces this multi‑series coherence, i.e., if housing “prefers” 0.3%, wages 2.8%, taxes –0.5%, and money 0.0%, with no overlap—then the triangulated 1.5% claim fails.

4.3 Sectoral productivity vs prices

Claim: The gap between productivity growth and CPI‑adjusted prices in high‑productivity sectors is of the same order as the proposed drift.

Falsification route:

Run sector‑level regressions of TFP growth on sector price changes (controlling for input costs and monetary expansion) across agriculture, manufacturing, and services.
Estimate the pass‑through and the residual wedge between observed price behavior and what productivity alone would predict.

If those regressions show no systematic residual wedge (i.e., CPI fully reflects productivity‑driven price declines), or a wedge that is small and inconsistent in sign across sectors, the “missed productivity deflation” mechanism is undermined.

4.4 Money–prices relationship

Claim: Over decades, CPI plus ~1.5% tracks the expansion of effective money per person better than headline

Falsification route:

Construct long‑run series for M2 per capita and velocity‑adjusted money and compare them to cumulative CPI and CPI plus various drifts.
If headline CPI already fits money per person as well or better than any drift‑adjusted alternative (including 1.5%), there is no monetary “gap” left for NETs to explain.

4.5 Structural NETs claims beyond CPI

The broader NETs picture adds two further claims that can also be rejected by data.

Flat or nearly flat GDP per capita under the correct adjustment:
If credible reconstructions using alternative, defensible inflation treatments still show strong, sustained growth in real GDP per person over 100–200 years, the “constant value of a human hour” claim is wrong.
Human time as 80–90% of value:
- In NETs, “human time” is broader than formal labor compensation. It includes: (a) labor compensation (wages, salaries, benefits), (b) all corporate profits and proprietors’ income, treated as returns on past and present human time applied to land and capital, (c) government spending on personnel and labor‑intensive services, and (d) the labor component of private investment and overhead that exists to organize, coordinate, and deploy human work. At the level of the whole economy, land and natural resources only generate profit once human beings locate, extract, refine, transport, and sell them, so even resource‑intensive sectors ultimately express human time as the binding constraint; using 2019 BEA and Fed data, summing these categories yields a human‑time share of roughly 87.1% of GDP, versus only about 42% if you look at narrow labor compensation alone.
How to Falsify:
- This broader claim would need to be revised if a careful decomposition of GDP (using BEA income‑side accounts and assigning all profits, mixed income, and public outlays according to the NETs rules above) consistently found that this extended human‑time bundle is well below the 80–90% band and stable in a much lower range over time (for example, clustering near 50–60% rather than around 80–90%).

Step 5: The epistemic fork for critics

Because every ingredient in this argument comes from standard public series and mainstream mechanisms, a critic must either:

Reject those mechanisms (e.g., productivity pass‑through, CPI’s definition, the quantity theory of money), or
Show, via transparent reconstructions, that when you apply the same data with standard tools, the ~1.5% drift no longer yields the best multi‑series fit.

Either path is a valid empirical challenge. But simply asserting that CPI is “roughly right on average” without addressing the specific mechanisms and multi‑series triangulation is no longer enough once the falsification criteria are on the table.

Next Article
NETs: Unmasking Economic Illusions with Time-Based Truth: Master Summary of the Theory

References (partial list)

Selected data and benchmark sources used in the NETs technical summary; a full reference list is available on request.

Bolt, J., & van Zanden, J. L. (2024). Maddison style estimates of the evolution of the world economy: A new 2023 update. Journal of Economic Surveys. Advance online publication. https://doi.org/10.1111/joes.12618
Boskin, M. J., Dulberger, E. R., Gordon, R. J., Griliches, Z., & Jorgenson, D. W. (1996). Toward a more accurate measure of the cost of living: Final report to the Senate Finance Committee from the Advisory Commission to Study the Consumer Price Index. U.S. Senate Committee on Finance. https://www.ssa.gov/history/reports/boskinrpt.html
CoreLogic. (2026). S&P CoreLogic Case‑Shiller U.S. National Home Price Index [Data set]. Retrieved April 22, 2026, from https://www.spglobal.com/spdji/en/indices/indicators/sp-corelogic-case-shiller-us-national-home-price-nsa-index
Federal Reserve Bank of St. Louis. (2024). Gross domestic product (GDP) [Data set]. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/GDP
Federal Reserve Bank of St. Louis. (2024). Median sales price of houses sold for the United States (MSPUS) [Data set]. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/MSPUS
Federal Reserve Bank of St. Louis. (2026). Average hourly earnings of all employees, total private (CES0500000003) [Data set]. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/CES0500000003
Federal Reserve Bank of St. Louis. (2026). Velocity of M2 money stock (M2V) [Data set]. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/M2V
MeasuringWorth. (2025). What was the U.S. GDP then? Retrieved March 14, 2026, from https://www.measuringworth.org/usgdp
U.S. Bureau of Labor Statistics. (2013). Average food prices: A snapshot of how much has changed over a century (Beyond the Numbers, Vol. 2, No. 13). U.S. Department of Labor. https://www.bls.gov/opub/btn/volume-2/average-food-prices-a-snapshot-of-how-much-has-changed-over-a-century.html
U.S. Bureau of Labor Statistics. (2026). Consumer price index for all urban consumers (CPI‑U) [Data set]. U.S. Department of Labor. https://www.bls.gov/cpi/
U.S. Census Bureau. (2024). Historical income tables: Households [Data set]. U.S. Department of Commerce. https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-income-households.html
U.S. Department of Agriculture, Economic Research Service. (n.d.). Agricultural productivity in the U.S.: Summary of recent findings. Retrieved March 14, 2026, from https://www.ers.usda.gov/data-products/agricultural-productivity-in-the-united-states/summary-of-recent-findings
U.S. Department of Agriculture, Economic Research Service. (n.d.). Farm income and wealth statistics: Annual cash receipts by commodity [Data set]. Retrieved March 14, 2026, from https://data.ers.usda.gov/reports.aspx?ID=4055
U.S. Department of Agriculture, National Agricultural Statistics Service. (2023). Crop production 2022 summary (January 2023, ISSN 1936–3737). https://www.nass.usda.gov/Publications/Todays_Reports/reports/croptr22.pdf
U.S. Department of Agriculture, National Agricultural Statistics Service. (2025). Crop production 2024 summary (January 2025, ISSN 1936–3737). Cornell University Library. https://downloads.usda.library.cornell.edu/usda-esmis/files/k3569432s/nk324887m/qn59s0097/cropan25.pdf

Author: Kyle Novack

May 9, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

Food Puzzle Master Summary

Kyle Novack — Sat, 09 May 2026 16:37:30 GMT

A companion summary to the Food Puzzle series (Parts 1–14)

This memo summarizes the core empirical findings from the Food Puzzle series (Parts 1–14) for quick reference.

The Core Paradox

U.S. agriculture has undergone one of the most dramatic productivity transformations in modern economic history. From 1948 to 2021, agricultural Total Factor Productivity (TFP) grew by an average of 1.49% per year, with output nearly tripling while aggregate inputs were flat or declining. The TFP index alone, rising from 0.37 in 1948 to 1.01 in 2021, implies roughly 66.2% expected cost savings at the farm level over that period. Yet CPI‑deflated farm expenses fell only about 20.2%, and per‑capita food revenue in the USDA Food Dollar series has been essentially flat in real terms from 1967 to 2023. The physical economy and the CPI‑measured economy cannot both be right; the Food Puzzle is a systematic attempt to reconcile them.

Parts 1–2 | The Agricultural Anomaly

The anomaly is not confined to the post‑1980 era. From 1948 to 1980, TFP growth implies roughly a 31.37% decrease in real farm expenses, yet CPI‑adjusted per‑capita agricultural expenses increase by 38.37%, a gap of nearly 70 percentage points. Even allowing for the late‑1970s input cost bubble (energy, interest rates, weather shocks), the discrepancy remains too large to dismiss as a cyclical artifact. Post‑1980, TFP implies about 50.8% cumulative cost savings, while CPI captures only about 40.7%, heavily influenced by the same high‑inflation bubble period in the 1970s and early 1980s.

Over the same window, a standard food basket requires 44% fewer work hours in 2024 than in 1980 (2.32 hours vs. 1.30 hours), indicating that productivity translated into higher purchasing power at the household level. Under CPI, that gain is attributed mostly to rising wages, combined with only modest declines in food prices. That attribution is hard to reconcile with the rest of the agricultural monetary data: today’s farm labor force is smaller than in 1948, yet farms produce more food with fewer inputs. On those physical terms, per‑capita food costs should be dramatically lower, not roughly similar or only slightly cheaper than in the mid‑20th century.

Parts 3–5 | The Supply Chain Defense Fails

The most intuitive explanation is that farm‑level savings were absorbed downstream by processing, packaging, transportation, wholesale, and retail. That story fails on its own terms. Each downstream sector experienced its own productivity boom over the same period, which should have compounded the farm‑level savings rather than offsetting them. USDA ERS data show that CPI deflated per‑capita costs in packaging, transport, and food manufacturing were flat or declining, even as the farm share of the food dollar fell to roughly one‑quarter for food‑at‑home, broadly in line with its historical range. A shrinking farm share is therefore better read as evidence that farms have become relatively more productive than the rest of the chain, not that downstream sectors absorbed the entire productivity dividend; the unresolved gap lies in how the dollar aggregates that joint productivity record.

A note on methodology: invoking industry‑specific deflators (USDA Prices Paid/Received indices, BEA sector deflators) to “fix” this discrepancy concedes the core point. If CPI requires a patchwork of sector‑by‑sector corrections to line up with physical quantities in its most data‑rich sector, it fails as an economy‑wide price index. The Food Puzzle standard is stricter: the goal is not to make agriculture’s numbers work in isolation, but to identify a single inflation adjustment that simultaneously reconciles physical and monetary realities across the whole economy.

Parts 6–7 | Corporate Profit Margins Are Not the Culprit

A second common explanation is that concentrated market power in branded food allowed large corporations to capture the productivity dividend. Analysis of 10 major food companies, including Tyson, Coca‑Cola, and PepsiCo, shows stable or declining net profit margins over the relevant period. If corporate concentration were absorbing the productivity gains, margins would be widening over time; they are not. Market power is concentrated in branded and discretionary segments, not in staple food categories, where productivity gains are largest, and the puzzle is sharpest.

Subscribe now

Part 8 | External Shocks Are Transitory, Not Structural

Supply disruptions, droughts, the COVID‑19 pandemic, and the Russia‑Ukraine war are frequently cited as explanations for persistently high food prices. The Russia‑Ukraine conflict provides the cleanest recent natural experiment: global food commodity prices spiked in 2022, then largely reverted toward their pre‑shock trend. Such episodic shocks generate temporary price volatility, not the decades‑long structural divergence between TFP‑implied cost savings and CPI‑adjusted price levels documented in the agricultural data.

Part 9 | Regulation Is Already in the TFP Measurement

Regulatory burden is sometimes raised as an explanation for why productivity gains do not translate into lower consumer prices. This objection misreads what TFP measures. Regulations that raise input costs, additional compliance labor, capital equipment, and materials are already captured on the input side of the TFP accounts. Regulations that restrict output reduce the index directly on the output side. The ~50.8% implied cost savings from 1980 to 2021 are therefore a net‑of‑regulation figure, not a pre‑regulation hypothetical. Regulation cannot be embedded in TFP on both the input and output sides and then be invoked again as an independent explanation for why TFP‑implied savings failed to reach consumers.

Parts 10–15 | The Stress Test

With the standard explanations eliminated, the Food Puzzle turns to the measurement of inflation itself. The official CPI serves as the baseline, but it fails the stress test on its own terms before any alternative is applied. Across 16 common food items, 1980 prices adjusted forward to 2024 using official CPI show an average real price decline of only ~3.12%. Against a productivity record implying roughly 40–50% real cost reductions in many staple categories, CPI’s own numbers leave most of the expected savings unaccounted for. A measuring stick that cannot reconcile a century of documented physical productivity gains with the prices consumers actually pay fails to capture the full economic reality.

The question then becomes whether any adjustment to CPI closes this gap without generating new contradictions elsewhere in the data. Three prominent alternatives are applied retroactively to the same 44 years of price data across 16 common items.

Boskin Commission (~1.1% annual overstatement). Applying the Commission’s estimated bias retroactively produces an average real food price increase of ~60% since 1980 across all 16 items, the opposite of what the productivity record predicts. Under Boskin’s logic, real wages rise ~140% and real GDP per capita ~213% since 1980, which is difficult to reconcile with declining intergenerational mobility, falling under‑35 homeownership, and persistent cost‑of‑living distress in survey data.

ShadowStats / Chapwood (~7% annual understatement from 2000 onward). This adjustment yields an average real food price decrease of ~81% since 1980, far beyond even generous readings of agricultural TFP gains (~40–60%). It implies 1980 wages of about $238,518 in 2024 dollars, a 72% real-wage collapse, and median home prices of around $1.27 million, an economic depression worse than the 1930s, which clearly did not occur.

NETs 1.2–1.8% annual understatement, centered on 1.5%. A modest, directionally inverted application of Boskin’s insight, that CPI systematically misses the deflationary force of productivity, resolves the puzzle cleanly. Real food prices then fall broadly in line with TFP‑implied savings, while real wages, housing, and GDP per capita remain internally consistent with observable living standards.

The final validation comes from classical monetary theory. When each inflation measure is expressed as a cumulative multiplier and plotted against per capita M2, the standard monetary driver of price levels, the NETs adjustment tracks per capita M2 most closely. Official CPI and Boskin sit far below any plausible path of monetary expansion, while ShadowStats/Chapwood exceeds even the monetary base. The NETs adjustment is the only measure that remains consistent with both the food‑price evidence and the monetary aggregates.

Conclusion

The Food Puzzle is not a critique of any single data series. It is a reconciliation test: can the economy’s most data‑rich, most physically measurable sector, agriculture, be made to tell a coherent story when CPI is used as the inflation deflator? After eliminating every standard explanation and stress‑testing every prominent CPI alternative, the answer is no. A systematic CPI understatement of 1.2–1.8% per year, centered on 1.5%, is the only adjustment that resolves the paradox without generating new contradictions elsewhere in the data. The implications extend well beyond food: if CPI understates inflation by this margin, then real wages, real GDP, and real living standards have been persistently mismeasured, and the widespread sense that the economy is failing ordinary households is not a sentiment problem, but a measurement problem.

To keep this summary compact, detailed methodology and source references are omitted. For questions about how any figure was derived, please refer to the article parts cited in each subheading above, where full methodology, data sources, and calculations are documented.

Next,

Why Agriculture is the Perfect Smoking Gun: Food Puzzle, Part 1

Author: Kyle Novack

May 9, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

Why Measuring Inflation Is Like Baking Grandma's Chocolate Cake

Kyle Novack — Fri, 08 May 2026 13:31:36 GMT

If you made it through Part 1, you now have a more accurate picture of CPI than most people ever develop, including many of the people paid to have opinions about it. You know what the basket is, why things get weighted differently, and why the BLS tracks price changes rather than prices themselves. That’s no small thing.

But there is a difference between understanding what CPI does and understanding why getting it exactly right may be impossible. The first is a matter of reading carefully. The second requires sitting with the complexity long enough for it to stop feeling abstract. So before we go any further, we’re going to bake a cake.

Specifically, Grandma’s chocolate cake. The one she’s been making since 1970. The one you always remember tasting amazing.

The question we are going to sit with is this: how would you even know if your measurement of inflation was wrong? Not wrong in an obvious way, but quietly, systematically, consistently wrong in a direction that nothing in the system was designed to detect. That is the conceptual problem this article is building toward. Because if you cannot answer that question satisfactorily, then everything built on top of that measurement, every interest rate decision, every wage negotiation, every inflation-adjusted figure in the economy, is built on a foundation that has never been fully examined.

The Measurements Keep Shifting

In this section of the analogy, the recipe represents the CPI basket itself, the collection of goods and services the BLS uses to track the cost of living. The ingredients are the individual items in that basket. The ratios between them are the weights. Keep that in mind as we go.

The first obstacle for CPI is that the economy it measures never holds still, and neither does grandma’s recipe.

Grandma’s original recipe called for two cups of flour and half a cup of cocoa. But over time, as tastes shifted and ingredient costs changed, the recipe had to be adjusted. Maybe it’s now a cup and three-quarters of flour and three-quarters of a cup of cocoa. The individual ingredients are the same, but the balance among them has been quietly redrawn to reflect people’s changing demands.

This is exactly what the BLS does when it updates the CPI weights. Housing used to carry a different share of the average household budget than it does today. So did transportation. So did food. As spending patterns shift, the ratios between categories are rebalanced to ensure the index continues to reflect reality. Starting in 2023, the BLS began updating these weights annually rather than every two years, precisely to keep the measurement aligned with what consumers are spending money on.

And that updating must happen. If the BLS stopped adjusting the weights, it would start measuring the cost of things people no longer buy in the proportions they once did. Demand for outdated goods drops, and their prices can behave oddly as production shrinks or markets thin out. If CPI kept using its old weights, the index could start to drift away from how people live, effectively measuring yesterday’s shopping cart. The updates are not a flaw. They are what keep CPI credible.

But now take the problem one step further. What if it’s not just the ratios that change, but the entire cake? What if your town stops wanting chocolate cake altogether and switches to vanilla? You can either continue measuring the cost of a chocolate cake that no one wants anymore, or update the recipe to reflect what people are buying.

The answer must be to switch to baking a vanilla cake. Because the whole point of CPI is to measure the cost of living for a typical American household, and if that household has moved on to vanilla, then chocolate is no longer the right thing to measure. You change the recipe not to lose the thread, but to keep it. Every adjustment, every rebalancing, every basket update exists for the same reason: to make sure that tomorrow you are still measuring the same concept you were measuring yesterday, even if the specific goods that define that concept have changed.

That’s a genuinely difficult thing to do well. And for the most part, the BLS does it well.

Which makes what comes next more unsettling. Because the problem we’re about to look at has nothing to do with the ingredients, or the ratios, or even which cake you’re making. It has to do with the measuring cups themselves.

Subscribe now

The Measuring Cups Are Shrinking

In this part of the analogy, the measuring cups represent the dollar itself, the unit we use to measure everything else. The ingredients represent the goods that productivity makes cheaper.

Imagine you’ve been baking Grandma’s cake for decades. You know the recipe by heart. You’ve made all the right adjustments over the years, updated the ratios, switched to vanilla when the town moved on from chocolate, kept everything honest and current. And the cake keeps coming out tasting pretty good. Maybe even better than it used to.

But something has been happening in the background that nobody noticed. The measuring cups have been slowly shrinking.

Not dramatically. Not all at once. Just a little, every year. What used to be a full cup is now seven-eighths of a cup. Then three-quarters. The recipe looks identical. You’re following the same instructions. But the actual amount of each ingredient going into the cake has been quietly decreasing for decades.

And here’s what makes it so disorienting: you start to notice that you are using more cup flour than you used to, but getting less cake. Something feels off. You could swear the recipe used to go further. But the cake still tastes good, better than it ever did, so you talk yourself out of it. Maybe you’re misremembering. Maybe this is just how it works now.

And here is what makes this so difficult to detect. The cake keeps tasting better. Not just fine, genuinely better. The flour is more refined, the cocoa richer, the eggs more consistent. Every year, the ingredients improve, and every year, the cake comes out a little more impressive than the year before. So even as you find yourself using more flour than you used to, even as something in the back of your mind says the recipe used to go further than it does now, the cake on the plate in front of you looks better than anything Grandma ever made. How do you argue with that? How do you tell someone the measuring cups are shrinking when the cake keeps getting better?

This is the dynamic that has hidden the error for so long. Living standards kept improving. The things people could buy with their wages kept getting better and more abundant. So even as the dollar was quietly losing purchasing power, the productivity gains flowing through the economy were delivering real improvements in what that dollar could buy. The cake tasted better. And that made it almost impossible to notice that the measuring cups were shrinking.

And it is exactly what has been happening with the dollar and productivity.

Over the past century, productivity gains have been quietly making everything cheaper to produce. Better machinery, more efficient supply chains, improved agricultural yields, and, more recently, technology and automation have steadily driven down the cost of producing the goods that make up the CPI basket. Those lower production costs have flowed through to lower consumer prices, or at least lower prices than would have prevailed without the productivity gains.

In standard economics, inflation is defined as the rate of change of a price index, such as the CPI. By that definition, CPI has done its job: it tracks those observed price changes. What NETs argues is that those changes already include a powerful, invisible offset from productivity, so the inflation we see is a net number, after the productivity dividend has quietly lowered costs.

Which means CPI has been measuring net inflation all along, and reporting it as if it were gross inflation.

The measuring cups were shrinking. But the ingredients kept getting better. And that was enough to keep everyone from asking the question that needed to be asked.

The Last Person Who Remembered

There is one more thing worth understanding about how we got here.

When Grandma was alive, the recipe had a guardian. Someone who had made the cake a thousand times, who knew instinctively what it was supposed to taste like. Who could tell immediately when something was off, not because she was following a formula, but because she carried the reference point inside her. The recipe existed on paper, but Grandma was the living standard against which everything else was measured.

When Grandma left, we kept the recipe. We updated it carefully, adjusted the ratios, and switched from chocolate to vanilla when the town moved on. We did everything right. But we lost the one thing the recipe could never capture, the ability to taste whether we were still making the right cake.

In the real economy, Grandma was the combination of two things: the gold standard, which gave the dollar a physical anchor independent of the economy it was measuring, and the market itself, which carried the collective knowledge of millions of transactions and could self-correct in ways no central measurement framework ever could. When both of those left together around 1970, one removed by policy choice, the other gradually displaced by the growing belief that economists and policymakers could actively manage the economy, we were left with the recipe and nothing else.”

And the further we get from when Grandma left, the harder it becomes to remember what the cake was supposed to taste like. The people updating the recipe today never tasted the original. They are adjusting based on what they can measure, calibrating against a standard they inherited rather than one they ever experienced directly. The recipe keeps getting updated. The measuring cups keep shrinking. And the cake kept coming out tasting pretty good, until it didn’t. Over time, the cracks started to show up in the data: wage series that don’t match productivity, housing that is unaffordable even on paper, food that never seems to get as cheap as farm efficiency would suggest.

Pretty good is not the same as right. And the further we get from when Grandma left, the more the difference begins to show, not because anyone went looking for it, but because, after long enough, an error that compounds every year eventually becomes too large to explain away.

So What Are We Actually Measuring?

Step back and look at what we’ve established.

The ingredients keep changing. The ratios keep shifting. The entire recipe needs to be updated periodically to ensure we’re still measuring the right thing. The BLS handles all of that with genuine rigor and intellectual honesty. Those are hard problems, and they do a serious job of managing them.

But underneath all of that, running silently in the background for over a century, is a different kind of problem, one that has nothing to do with how carefully the recipe is followed. The measuring cups have been shrinking, and the improving quality of the ingredients has masked it so well that we’ve only ever had faint hints that something was off, nothing screamingly obvious. To be clear, no inflation measure can ever be perfect; money is an abstract measuring stick, not a physical rod. But some errors are random noise, and some are systematic drifts, and NETs is arguing we’ve had the second kind.

The result is that what we call inflation, the number that moves markets, sets interest rates, adjusts Social Security checks, and anchors almost every major economic decision in the American economy, is a net figure. It is monetary inflation minus the productivity dividend. And because productivity has been a persistent, compounding feature of the American economy for over a hundred years, that dividend has been silently subtracted from the measured inflation rate every single year.

That gap, whatever its precise magnitude, has been compounding silently for over a century. And the nature of compounding means that even a small consistent error, sustained long enough, produces an enormous divergence between what we think has happened to the dollar’s purchasing power and what has. It cascades outward into every inflation-adjusted figure in the economy, real wages, real GDP, real investment returns, and real Social Security benefits. Every conclusion we’ve drawn from those figures carries this error embedded within it.

This is not an indictment of the BLS. They have been measuring what their framework was designed to measure, and they have done so carefully and honestly. The error is not in the execution. It is in the foundation, a framework that was never built to account for a sustained, century-long productivity boom of this magnitude.

The cake kept tasting good. So nobody checked the measuring cups.

That is what Part 3 is for. And the conclusion might surprise you, because before we can say CPI is wrong, we must reckon with something that almost nobody in this debate is willing to admit. That CPI is working as designed. The problem runs deeper than the measurement itself.

Sources

U.S. Bureau of Labor Statistics. (2025, January 30). Consumer price index: Overview (Handbook of Methods). U.S. Department of Labor. https://www.bls.gov/opub/hom/cpi/home.htm

Author: Kyle Novack

May 8, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

What Is CPI, and Why Does It Run the Economy?

Kyle Novack — Tue, 05 May 2026 13:31:24 GMT

Every month, the U.S. Bureau of Labor Statistics (BLS) releases a number that moves markets, shapes political careers, and determines how much of a raise millions of Americans will get. The Federal Reserve uses it to decide whether to raise or cut interest rates. Social Security checks are adjusted by it. Government bonds are priced against it. It is embedded in more contracts, policies, and financial decisions than almost any other single figure in the American economy.

That number is the Consumer Price Index (CPI), and it is baked into more economic decisions than most people realize. Other measures exist: the Personal Consumption Expenditure index (PCE), which the Federal Reserve prefers for monetary policy, and the GDP Deflator, calculated by the Bureau of Economic Analysis. But at the level of fiscal policy, public understanding, and the sheer number of decisions it powers, no single measurement rivals CPI. The reason is straightforward: without adjusting for inflation, historical nominal figures are almost useless for identifying trends. CPI is what makes that adjustment possible.

Understanding why CPI matters is only the first step. The next step is learning the methodology behind its calculation, because that methodology reflects a genuine attempt to grapple with one of the hardest problems in economics: how do you accurately measure the cost of living for an entire country when that economy is constantly shifting beneath you? CPI is not a single price or a simple average. It is the result of hundreds of thousands of data points, careful statistical decisions, and methodological choices that the BLS has refined over decades. When someone on the news says ‘inflation came in at 3.2%,’ there is an enormous amount of machinery behind that sentence that never gets explained.

This article is about that machinery. Before we can meaningfully ask whether CPI is accurate, and in later parts of the NETs project, I will ask that hard question, but first, we need to understand what it is trying to do and how it goes about doing it. You cannot critique a ruler without first understanding how it was built.

So let’s open the hood.

Before we do, one quick note on sources: all data, figures, and methodology described in this article are drawn directly from the U.S. Bureau of Labor Statistics Handbook of Methods, the primary document the BLS publishes explaining how CPI is built.

What Is the Consumer Price Index?

According to the U.S. Bureau of Labor Statistics, which calculates and publishes the index, the CPI measures the average change over time in the prices of a market basket of consumer goods and services. In plain terms, it tracks how much more or less expensive everyday life is getting.

There are actually several versions of CPI, each designed to measure inflation for a specific group or purpose. The three main ones are:

CPI-U (CPI for All Urban Consumers): the broadest and most widely cited version, covering about 93% of the U.S. population.
CPI-W (CPI for Urban Wage Earners and Clerical Workers): a narrower version used primarily to calculate Social Security cost-of-living adjustments.
C-CPI-U (Chained CPI for All Urban Consumers): a more flexible version designed to account for the fact that consumers shift their buying habits when prices change.

In this article, we focus exclusively on CPI-U. It is the index most people are referring to when they say “CPI,” the one used for most inflation adjustments, and the one that sits at the center of the debates we will examine in later parts of the Nets Project.

The Basket of Goods

The foundation of CPI is the market basket, a representative sample of the goods and services that American households buy. To account for the fact that prices vary significantly across the country, the BLS collects data from 75 urban areas, ensuring that regional differences in the cost of living are reflected in the sample rather than obscured. From those areas, the BLS constructs this basket using about 100,000 price quotes for goods and services and 8,000 rental housing unit quotes, organized into more than 200 categories across eight major groups:

Food and Beverages: breakfast cereal, milk, coffee, chicken, wine, full-service meals, snacks

Housing: rent, owners’ equivalent rent (an estimate of what homeowners would pay to rent their own home), utilities, furniture

Apparel: men’s shirts, women’s dresses, baby clothes, shoes, jewelry
Transportation: new vehicles, airline fares, gasoline, auto insurance
Medical Care: prescription drugs, physician services, eyeglasses, hospital services
Recreation: televisions, toys, pets, sports equipment, museum admissions
Education and Communication: college tuition, telephone services, computer software
Other Goods and Services: tobacco, haircuts, funeral expenses

This basket is not fixed forever. The BLS regularly reviews and updates the categories and items it tracks to ensure they reflect how Americans actually spend money today. If streaming services have replaced DVD players, or if a new category of spending has emerged, the basket should reflect that reality. The goal is simple: measure what people are actually buying, not what they bought a decade ago.

Subscribe now

Weighting: Not Everything Counts Equally

Once the basket is defined, the BLS conducts surveys across those 75 geographic regions, as determined by the 2010 census, to determine how much of their income households spend in each category. This produces a weight for each group, which determines how much that category can move the overall index.

To make this concrete: in the September 2025 CPI report,

Weights of Different Categories:

Shelter: 35.471
Food: 13.646
Transportation: 7.247

What that means in practice is that even if transportation costs rose 2.3% in a given year, the contribution of that increase to the overall CPI number is only about 0.2 percentage points, because transportation simply does not make up a large enough share of the average household budget to move the needle much. Shelter, by contrast, is so heavily weighted that even a modest rent increase ripples through the entire index.

Starting in 2023, the BLS began updating these expenditure weights annually, adjusting how much each category counts toward the index to better reflect shifting consumer habits. This was itself a response to longstanding criticism that infrequent weight updates allowed the index to drift out of sync with how people spend money. Even with this change, the weighting system is still one of the most consequential and most debated aspects of how CPI is built. We will come back to it.

What CPI Actually Measures: Changes, Not Prices

Here is something that surprises most people: CPI doesn’t report raw shelf prices; it reports how those prices change over time, combined into a single index

The reason is straightforward. A pound of ground beef costs more in San Francisco than it does in rural Mississippi. If the BLS tried to compare raw prices across the country, regional cost differences would create enormous distortions. Instead, they record the percentage change in price from one period to the next within each region. If ground beef costs $4.00 in 2024 and $5.00 in 2025 in a given market, CPI records a 25% price increase — regardless of whether the underlying price is higher or lower than the national average.

This is an elegant solution to a real problem. But it also means CPI is measuring a rate of change, not an absolute level of affordability. That distinction matters more than it might seem, and we will return to it in a later article.

Quality Changes and Hedonic Adjustments

One of the more technically sophisticated and more controversial things CPI does is adjust for quality changes in the goods it tracks.

Consider a simple example: a box of cereal that was 20 ounces last year is now 16 ounces but costs the same price. The sticker price hasn’t changed, but you are getting less product. CPI will recognize this and adjust, recording an effective price increase even though the number on the shelf stayed the same. This is called a quantity adjustment, and it is uncontroversial.

More complex is what happens when a product genuinely improves. Say a manufacturer discontinues a shirt and replaces it with one made from a higher-quality material at the same price. The BLS has to decide: is the consumer getting more for their money? If they conclude the new shirt is better, they may record zero inflation, or even deflation, on that item, because the quality-adjusted price effectively went down.

Technology is where this gets especially pointed. If a laptop goes from $1,000 to $1,200 but the new model has a faster processor, double the RAM, and 250 gigabytes of additional storage, the BLS may determine that the price increase is fully offset, or more than offset, by the quality improvement. In that case, CPI records zero price increase, or possibly a price decrease, even though you paid $200 more at the register. This method is called a hedonic adjustment, and it is one of the most debated tools in CPI’s methodology.

Whether hedonic adjustments are the right approach is a legitimate debate. For now, the key point is that CPI is not simply reading price tags; it is making judgment calls about value, quality, and equivalence constantly.

A Common Misconception About Substitution

One of the most frequently repeated criticisms of CPI is that it allows substitutions that obscure real inflation: if steak gets too expensive, the BLS just assumes people switch to chicken and calls it a day. This makes CPI sound like it is gaming the numbers to report lower inflation than people experience.

This is worth correcting directly, because it is not accurate for CPI-U.

CPI-U does allow within-category substitutions; if consumers shift from buying ground chuck to ground round, for example, the index can reflect that. But it does not allow cross-category substitutions. If beef prices spike, CPI does not assume people are now buying chicken instead and adjust accordingly. Beef and chicken are in different categories, and the index treats them separately.

This is a meaningful distinction. The criticism applies more accurately to the Chained CPI (C-CPI-U), which is specifically designed to account for broader substitution behavior. CPI-U, the standard index, does not work that way.

A Valiant Attempt at an Impossible Task

Step back and look at what CPI is trying to do: take a constantly changing economy, new products, disappearing products, shifting consumer habits, improving technology, and regional price differences, and distill all of it into a single number that represents how much more expensive life is getting for a typical American. Not just once, but every month. And reliable enough that the Federal Reserve will base its interest rate decisions on it.

That is an enormously ambitious undertaking. And the BLS, for what it is worth, does not pretend it is perfect. The methodology has been refined repeatedly over decades, and the economists who build and maintain CPI are aware of its limitations. They have published extensively on where the index struggles.

But understanding what CPI is trying to do and appreciating the genuine difficulty of the problem are the necessary first steps before asking whether it succeeds. Because the next question, the one this series is ultimately building toward, is a serious one: what happens if this number, the number at the foundation of so much of our economic understanding, is consistently and systematically off?

To get there, we first need to understand just how hard the problem of measuring inflation really is, not in abstract terms, but in a way that makes the difficulty feel real.

That is what Part 2 is for: we examine this through Grandma’s amazing chocolate cake recipe.

Sources

U.S. Bureau of Labor Statistics. (2025, October 24). Consumer price index summary. U.S. Department of Labor. https://www.bls.gov/news.release/archives/cpi_10242025.htm
U.S. Bureau of Labor Statistics. (2025). Handbook of methods: Consumer price index. U.S. Department of Labor. https://www.bls.gov/opub/hom/cpi/home.htm

Author: Kyle Novack

May 5, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

Why the 1.5% Drift Validates Your Gut Feeling About the Economy

Kyle Novack — Fri, 01 May 2026 13:31:19 GMT

In the last section, we tested one thing: how well the three main interpretations of inflation compare with the most honest way to assess potential monetary inflation, using M2 per capita, adjusted for velocity of money growth. I also added my thesis that CPI is understating true inflation by 1.5% a year, the 1.5% drift, to the mix to see how it compares. Only the 1.5% drift compared well.

That single alignment doesn’t close the case, but it raises a serious question: if the ruler is distorted by even that small an amount, what else has it been mismeasuring? This article answers that question for food. We’ll take the 1.5% adjustment and run it through the same stress test that all other inflation interpretations went through: farm expenses, wages, supply chain costs, and what you pay at the grocery store, to see if the pieces finally fit.

Before I get into the data, it’s worth being clear about what I am and am not claiming. The Food Puzzle does not prove that CPI is wrong. It does not prove the drift is exactly 1.5%. What it does is reveal gaps, places where the numbers CPI gives us don’t line up with first principles of economics, with physical reality, or with what you feel every time you check out at the grocery store. Those gaps deserve an honest conversation. The 1.5% drift is the hypothesis that best explains them. The rest of the NETs Project is where we put that hypothesis to the test.

Unless otherwise noted, all dollar figures in this section are expressed in “1.5% drift”–adjusted 2024 dollars. The 1.5% drift is defined as the official CPI-U inflation rate plus 1.5 % points per year, compounded annually from each series’ base year. This adjustment reflects the hypothesized systematic understatement of inflation that the Food Puzzle has been testing throughout Parts 1-13.

To keep the main story readable, I limit in-text citations and technical details. Tables and figures include only short captions and a few key references, so you can follow the argument without wading through footnotes on every line. All the underlying data series, transformations (such as per-capita conversions, the construction of the 1.5%-drift inflation multipliers, and percentage-change calculations), and exact formulas are documented in the Methods and Sources section at the end of this part. If you have questions about where a number comes from or how a graph was constructed, that is the place to look for full sourcing and methodology.

The Farm Data

When we adjust agricultural revenue and expenses per person using the 1.5% drift rather than the standard CPI, a trend emerges. The numbers finally start to track the productivity line much more closely and, in some periods, fall even faster than productivity alone would predict. That overshoot isn’t an error. It can be partly accounted for by the agricultural bubble of the 1970s and 1980s, where elevated prices had to revert to normal, pushing costs down further than the long-run trend would suggest on its own. On top of this, all the nominal data tells a different story under the lens of the 1.5% drift. Why the overshoot goes beyond the bubble is a question the broader NETs framework will address; it requires context from several other layers of the analysis that haven’t been laid out yet.

When we look at agricultural output measured in U.S. dollars and per-capita expenses, the data also align much more closely with historical reality. The major price movements throughout the century all have clear explanations:

World War I (1914–1918): Wartime demand and uncertainty drove prices sharply higher, followed by a crash once the war ended.
The Overindebted Farm Boom (Early 1920s through the Great Depression): Elevated farm prices persisted through the 1920s, only to collapse during the Great Depression.
World War II (1939–1945): Uncertainty about food supply dramatically increased revenues and expenses for farms. On top of this, governments spent heavily to feed soldiers in the field. After the war, prices dropped and continued falling as productivity improvements compounded.
The Energy Crisis and the End of the Gold Standard (1970s–Early 1980s): Severe speculation about the dollar’s purchasing power, combined with the Iranian oil crisis, drove heavy investment into agriculture and created a well-known agricultural bubble. Importantly, by this point, crop yields per acre had already improved significantly, and the share of the population needed to produce food had fallen dramatically. meaning the underlying pressure was already deflationary. That’s why this bubble was smaller in scope than earlier price movements.

Today, farming is more productive than ever, and per-capita food prices are the lowest relative to income. Americans now spend a smaller share of their income on food than at any point in recorded history. All the data aligns throughout the century.

Farm profit margins tell the same story. Once you strip out the price spikes from World War II and the 1980s grain crisis, sustained margins were higher during the Great Depression than they are today. This is also part of why total output and per-capita expenses fell more than productivity did, because less value could be obtained from food. The farm sector was never secretly hoarding the productivity gains.

This is what makes farming unique. Demand per capita can’t meaningfully grow; biology sets the ceiling. So, every efficiency gain the farmer made didn’t open new markets; it just made the existing market cheaper. Market forces drove them to innovate so relentlessly, and become so efficient, that they forced themselves to become essentially irrelevant in the overall economy, shrinking from roughly 18% of GDP at their WWI peak to around 2% today.” The market pressure on farming didn’t stop at output prices. The other major cost of production, wages, was being quietly compressed at the same time.

When we adjust farm wages for inflation, they have fallen by roughly 30.9% since 1958. Under the official CPI, those same wages appear to have risen by about 78.4% over the same period, so, on paper, workers look better off. That gap is the drift made visible. Businesses believed they were paying more each year, and in nominal terms, they were. But because inflation was being mismeasured, much of that apparent rise was illusory. The typical farm worker’s purchasing power quietly contracted while the official numbers said everything was fine.

As striking as that number is, farm wages turn out to be the best case in the food industry. When we apply the same 1.5% drift correction across the broader food supply chain, wages collapse even more.

Every single segment of the food industry, from the field to the factory to the warehouse to the checkout lane, shows the same pattern. Workers were nominally paid more every year. Inflation quietly erased it. And because the ruler was bent, nobody could see it happening. This is not a farming problem. It is an economy-wide problem that the food industry makes impossible to ignore.

This matters beyond its implications for workers. As productivity gains compressed input costs, falling real wages further reduced input costs, lowering food prices even more. As I showed in Part 7, the food industry is extremely competitive, and as Part 6 demonstrated, corporate profits in this sector are not rising. That means businesses cannot hold onto these savings; competition forces them through to the consumer. The result is that food prices could fall even further than productivity alone would predict, because cost compression is occurring on multiple fronts simultaneously, leaving nowhere else for those savings to go. With costs falling at the farm, wages compressing across every stage of the supply chain, and competition preventing businesses from holding onto the savings, there is only one place left for all of that to show up: the price you pay at the store.

Subscribe now

What You Pay at the Store

When we apply real inflation to total food revenue per capita, it falls about 45.8% from 1967, somewhat less than the roughly 66% drop at the farm level. That gap exists because the rest of the supply chain doesn’t capture the full farm-level improvement, not because middlemen swallowed the savings. As we showed earlier in this series, every stage of the supply chain got more efficient, too. The farm share of the food dollar shrank because farms improved faster than the rest of the chain, not because downstream costs exploded.

Using USDA data, acres harvested per capita decreased from roughly 0.9 acres in 1980 to around 0.6 acres in 2024, a nearly 29% decrease in land utilization. Total food revenue per person tells a similar story by falling by about 34.4% in real terms over the same period. For a sample of 17 commonly purchased, lightly processed staples, items that require relatively little value-added work beyond the farm, prices fall about 48.72% after adjusting for real inflation. That’s slightly larger than the aggregate estimate, which is exactly what you’d expect for simple, minimally processed foods that sit closest to farm-level productivity gains. Add in the suppressed wages, and the difference becomes almost self-evident.

The table above also tells a broader story that goes well beyond groceries. Median household income, adjusted for the 1.5% drift, has fallen 31.58% since 1980. That decline is happening even as more households have shifted to two incomes just to keep pace, meaning families are working more and still losing ground. The gap between what paychecks should buy and what they buy isn’t a personal failure. It’s the drift, compounding quietly for decades.

Housing tells a similar story. The home that most people assume has grown in value is largely an illusion. Adjusted for real inflation, home prices are down about 8.63% since 1980. For millions of Americans, their home is their single largest asset, the cornerstone of their retirement plan, their financial safety net, and the wealth they hoped to pass on to their children. Under the 1.5% drift, most of that assumed wealth quietly evaporates.

And then there’s GDP per capita, the number politicians and economists point to as proof that the economy is growing, and life is getting better. Adjusted for the 1.5% drift, GDP per capita is essentially flat since 1980, up just 0.29%. Nearly half a century of reported economic growth, and in real terms, output per person has barely moved.

Taken together, these three data points reframe everything. It’s not that the economy stopped working. It’s that the ruler we use to measure it has been making economic decline for the average person just trying to get by, like you, look like progress for a very long time.

What This Means

The Food Puzzle started as a simple question: if farms and food companies can produce so much more with so much less, why haven’t real food prices collapsed? The answer they did. It’s that we couldn’t see it.

The physical economy did its job. Productivity rose. Supply chains became more efficient. Corporate margins didn’t secretly balloon. But our inflation gauge left most of that deflationary boom invisible, and that invisibility has had real consequences for real people. It’s why wages that looked like raises weren’t. It’s why your grocery bill feels wrong even when the official numbers say it shouldn’t. It’s why millennials and Gen Z are on track to be the first generations in modern history not to be better off than their parents, and economists have largely accepted that as an unfortunate fact of life rather than a measurement problem worth investigating.

The Food Puzzle doesn’t prove the drift is 1.5%. It doesn’t prove CPI is broken beyond repair. What it does is open a door. It shows that there are gaps in how we measure the economy that can’t be explained away by supply chains, corporate greed, regulation, or bad luck. Those gaps follow a pattern. And that pattern points toward a distorted ruler, one that has been making genuine progress invisible and genuine erosion look like stability for a very long time.

The food chapter is now closed. But what it leaves behind is bigger than food. If a systematic measurement error has been distorting how we read one of the most data-rich sectors in the entire economy, then that same ruler has been distorting everything else it has ever measured, too: your wages, your home’s value, your retirement savings, the interest rate on your mortgage, and the policies that govern all of it. The rest of the NETs project is about following that thread. Because if the ruler is bent here, it’s bent everywhere. And everything we thought we understood about how the economy works needs to be reconsidered.

Methods and Sources

Graph 1: Expected Price Decreases from Productivity vs. Observed Changes in Revenue and Expenses Per Capita, 1948–2021 (Adjusted for 1.5% Drift)

This graph compares the cumulative deflationary impact implied by agricultural productivity gains against the observed changes in per-capita farm revenue and expenses, all adjusted for real inflation (CPI-U plus 1.5% annual drift).

Data series:

Total Factor Productivity (TFP): USDA ERS Agricultural Productivity in the U.S., annual TFP index 1948–2021. https://www.ers.usda.gov/data-products/agricultural-productivity-in-the-united-states/
Agricultural revenue: USDA ERS Farm Income and Wealth Statistics, “Value of Production” (all commodities), annual 1948–2021.
Agricultural expenses: USDA ERS Farm Income and Wealth Statistics, “Total Production Expenses,” annual 1948–2021.
Population: U.S. Census Bureau historical estimates.
CPI-U: BLS Consumer Price Index, annual averages 1948–2021.

Calculation:

For each starting year from 1948 onward, three percentage changes to 2021 are calculated:

1. Expected price decrease from productivity (blue line):
For a given year (e.g., 1980), the TFP index value in that year is compared to the 2021 TFP index. The percentage increase in productivity is calculated as:
% TFP Increase = [(TFP(2021) / TFP(year)) - 1] × 100 This productivity gain is then converted into an expected price decrease, assuming full pass-through to consumers:
Expected Price Decrease = -[% TFP Increase] (For example, if TFP rose 50% from 1980 to 2021, the expected price decrease is -50%.)

2. Observed revenue per capita change (green line):
Nominal revenue for the starting year and 2021 are converted to constant 2024 dollars using the 1.5%-drift-adjusted inflation multipliers (CPI-U inflation rates plus 1.5 percentage points annually, compounded). Both values are divided by population to get per-capita figures. The percentage change is:
% Revenue Change = [(Revenue per capita(2021) / Revenue per capita(year)) - 1] × 100

3. Observed expenses per capita change (gray line):
Same method as revenue, using total production expenses instead. Signs are flipped so that cost declines plot as positive numbers and cost increases (like the 1948–1980 period) appear as negative.

The graph plots these three series against each starting year (1948, 1949, ... 2020) on the x-axis, with the y-axis showing percentage change to 2021. This reveals whether observed revenue and expense changes track the productivity-implied cost reductions, or fall short of them.

Note: The 2021 endpoint is used (rather than 2024) because it is the last year available in the USDA TFP series at the time of analysis. All dollar values are expressed in constant 2024 dollars for consistency with the rest of the Food Puzzle series, even though percentage changes are calculated to 2021.

Graph 2: Agricultural Revenue and Expenses Per Capita, Adjusted for Real Inflation (1.5% Drift), 1910–2024

This graph provides the historical context referenced in “The Farm Data” subsection, showing how major price movements align with known economic events (WWI, Great Depression, WWII, 1970s–80s agricultural bubble).

This graph displays total U.S. agricultural commodity revenue and total production expenses on a per-capita basis, adjusted for “real inflation” (official CPI-U plus 1.5 percentage points annually), expressed in constant 2024 dollars. It tracks how farm-level costs and revenues have changed relative to population and the corrected inflation measure, with major historical price movements (WWI, Great Depression, WWII, 1970s–80s agricultural bubble) annotated to provide context for volatility.

Revenue and expenses are taken from the USDA ERS Farm Income and Wealth Statistics, “Value of Production” (all commodities) and “Total Production Expenses,” annual series 1910–2024. Population data provide by Macrotrends 2023 and 2024 and Maddison-style estimates for all other years. Official CPI-U annual averages (1913–2024) are from the BLS, with interpolated or proxy estimates for 1910–1912 where needed.

For each year from 1910 to 2024, annual CPI-U inflation rates are computed, then 1.5 percentage points are added to produce the “real inflation rate.” A cumulative inflation multiplier series is constructed starting from 1910 (base = 1.000), compounding the real inflation rates year by year. Nominal revenue and expenses are converted to constant 2024 dollars by multiplying by the ratio [Multiplier(2024) by the nominal price] then divided by population to obtain per-capita figures in 2024 dollars.

Graph 3: Farm Profit Margins, Adjusted for Real Inflation (1.5% Drift), 1910–2024

Agricultural profit margins are calculated as (total value of agricultural production−total farm production expenses)÷total value of agricultural production×100, using the ERS series on the value of agricultural production and total production expenses from the Farm Income and Wealth Statistics. This margin is constructed directly from the same output and expense series used in the per‑capita graphs, to keep the accounting consistent.

Table: Real Wage Changes Across the Food Supply Chain, 1958–2024

Nominal average hourly earnings for three food-sector industries and agriculture are taken from BLS and USDA sources, then converted to constant 2024 dollars using both official CPI-U and the 1.5%-drift adjustment (official CPI inflation rates plus 1.5 percentage points annually, compounded from 1958).

Data sources:

1958–1989 (SIC codes): U.S. Bureau of Labor Statistics, “Employment, Hours, and Earnings, United States, 1909–94: Bulletin 2445” (September 1994). Food Manufacturing (SIC 20): Vol. 1, p. 477; Grocery Wholesalers (SIC 514): Vol. 2, p. 873; Food and Beverage Retailers (SIC 54): Vol. 2, p. 901. https://fraser.stlouisfed.org/title/189/item/5437

1990–2024 (NAICS codes): FRED/BLS series: Food Manufacturing (NAICS 311): IPUEN311W200000000; Grocery Wholesalers (NAICS 4244): IPUGN4244W20000000; Food and Beverage Retailers (NAICS 445): IPUHN445W200000. Retrieved December 27, 2025, from https://fred.stlouisfed.org

Agriculture (1958–2024): USDA NASS “Farm Labor” survey, hourly wage rates for hired farm workers. https://quickstats.nass.usda.gov. Note: In 1972, reporting changed from “per hour with board and room pay” to “all hired farm workers” ($/hour); 1972 definition used from that year forward.

Calculation: For each sector, 1958 and 2024 nominal wages are adjusted to 2024 dollars using: (1) standard CPI-U, and (2) CPI-U + 1.5% annual drift, compounded. Percentage change = [(2024 wage / 1958 wage in 2024$) – 1] × 100.

Comparison of Current (2024) Food Prices vs. 1980 Prices Adjusted for 1.5% Drift

All price and income figures are converted to 2024 dollars using a 1.5-percentage-point annual understatement adjustment to CPI-U before comparison.

Food item prices: For each of the 17 food items, nominal monthly prices for 1980 and 2024 are taken from the BLS Average Price (AP) series and averaged over the 12 months to get an annual price for each year. The 1980 annual prices are then rolled forward to 2024 by taking the official CPI-U inflation path and adding 1.5 percentage points to the annual inflation rate from 1980 onward, reflecting the hypothesized minimum understatement of CPI. This produces 1.5%-drift-adjusted 2024-dollar values, which are compared with actual 2024 prices to compute percentage differences. The “average price change” row is the simple mean of these 17 percentage differences.

Income and housing: Median household income is taken from Census historical income tables; median home prices from the FRED MSPUS series. The 1980 values are inflated to 2024 dollars using the same 1.5%-drift adjustment (official CPI-U inflation rates plus 1.5 percentage points per year from 1980 onward) before computing percentage differences, so income and housing are treated on the same basis as the food basket.

GDP per capita: GDP levels are assembled from MeasuringWorth and FRED, with population from Maddison-style estimates and Macrotrends. I compute GDP per capita as real GDP ÷ population, then express the 1980 value in 2024 dollars using the same 1.5%-drift-adjusted CPI path where needed, to match the treatment of other income variables.

Acres Harvested per Capita, 1980–2024

Number of harvested acres summed across barley, corn/maize, cotton, oats, rice, soybeans, sugarbeets, sugarcane, tobacco, and wheat. Sources: USDA National Agricultural Statistics Service, Crop Production 2022 Summary and Crop Production 2024 Summary (historical tables on yields and harvested acreage); population series from Bolt and van Zanden (2024) and Macrotrends.

Acres harvested per capita: Total Acres harvested by 10 crops divided by total U.S. population; Sources Maddison style population estimate and Macrotrends.

Sources

Bolt, J., & van Zanden, J. L. (2024). Maddison-style estimates of the evolution of the world economy: A new 2023 update. Journal of Economic Surveys, 38(1), 1–41. https://doi.org/10.1111/joes.12618
Federal Reserve Bank of St. Louis. (2024). Median sales price of houses sold for the United States (MSPUS) [Data set]. FRED. https://fred.stlouisfed.org/series/MSPUS
Federal Reserve Bank of St. Louis. (2024). Real gross domestic product per capita (A939RX0Q048SBEA) [Data set]. FRED. https://fred.stlouisfed.org/series/A939RX0Q048SBEA
Macrotrends. (2025). United States population 1820–2024 [Data set]. https://www.macrotrends.net/global-metrics/countries/usa/united-states/population
U.S. Bureau of Economic Analysis. (2024). Gross domestic product [Data set]. https://www.bea.gov/data/gdp
U.S. Bureau of Labor Statistics. (1994). Employment, hours, and earnings, United States, 1909–94: Bulletin of the United States Bureau of Labor Statistics, No. 2445 [Employment and Earnings, United States]. https://fraser.stlouisfed.org/title/189/item/5437 (Accessed December 22, 2025. Volume 1, Page 477; Volume 2, Pages 873, 901)
U.S. Bureau of Labor Statistics. (2024). Average price data (AP), U.S. city average. https://www.bls.gov/charts/consumer-price-index/consumer-price-index-average-price-data.htm
U.S. Bureau of Labor Statistics. (2024). Current Employment Statistics (CES), average hourly earnings by industry [Data set]. https://www.bls.gov/ces/
U.S. Bureau of Labor Statistics. (2026). Consumer Price Index for All Urban Consumers (CPI-U) [Data set]. https://www.bls.gov/cpi/
U.S. Bureau of Labor Statistics. (n.d.). Employment for manufacturing: Food manufacturing (NAICS 311) in the United States [FRED series IPUEN311W200000000]. Federal Reserve Bank of St. Louis. Retrieved December 27, 2025, from https://fred.stlouisfed.org/series/IPUEN311W200000000
U.S. Bureau of Labor Statistics. (n.d.). Employment for retail trade: Food and beverage stores (NAICS 445) in the United States [FRED series IPUHN445W200000]. Federal Reserve Bank of St. Louis. Retrieved December 27, 2025, from https://fred.stlouisfed.org/series/IPUHN445W200000
U.S. Bureau of Labor Statistics. (n.d.). Employment for wholesale trade: Grocery and related product wholesalers (NAICS 42444) in the United States [FRED series IPUGN4244W20000000]. Federal Reserve Bank of St. Louis. Retrieved December 27, 2025, from https://fred.stlouisfed.org/series/IPUGN4244W20000000
U.S. Census Bureau. (2024). Historical income tables: Households [Data set]. https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-income-households.html
United States Department of Agriculture, National Agricultural Statistics Service. (2022). Crop production 2022 summary. https://www.nass.usda.gov/Publications/Todays_Reports/reports/croptr22.pdf
United States Department of Agriculture, National Agricultural Statistics Service. (2025). Crop production: 2024 summary. Cornell University Library. https://downloads.usda.library.cornell.edu/usda-esmis/files/k3569432s/nk324887m/qn59s0097/cropan25.pdf
U.S. Department of Agriculture, Economic Research Service. (n.d.). Agricultural productivity in the U.S.: Summary of recent findings. Retrieved March 14, 2026, from https://www.ers.usda.gov/data-products/agricultural-productivity-in-the-united-states/summary-of-recent-findings
U.S. Department of Agriculture, Economic Research Service. (n.d.). Farm income and wealth statistics [Data set]. https://data.ers.usda.gov/report.aspx?ID=4059
U.S. Department of Agriculture, Economic Research Service. (n.d.). Farm income and wealth statistics: Value of production and cash receipts tables [Data set]. https://data.ers.usda.gov/report.aspx?ID=4055
U.S. Department of Agriculture, Economic Research Service. (n.d.). Food expenditure series [Data set]. https://www.ers.usda.gov/data-products/food-expenditure-series/
U.S. Department of Agriculture, National Agricultural Statistics Service. (1988). Farm labor (November 14, 1988). https://esmis.nal.usda.gov/sites/default/release-files/x920fw89s/dz010r78x/6d56zz45b/FarmLabo-11-14-1988.pdf
U.S. Department of Agriculture, National Agricultural Statistics Service. (n.d.). Farm labor survey: Wage rate [Data set]. QuickStats. Retrieved December 27, 2025, from https://quickstats.nass.usda.gov/?referrer=grok.com#517CF532-00CD-37E5-A724-DA9A106E59CE
Williamson, S. H. (2025). What was the U.S. GDP then? MeasuringWorth. http://www.measuringworth.org/usgdp/

Author: Kyle Novack
Date: May 1, 2026
A Monumental Venture, LLC research project Novack Equilibrium Theory (NETs)

The Final Straw: Comparing Inflation Metrics to Monetary Expansion

Kyle Novack — Tue, 28 Apr 2026 13:31:14 GMT

The extreme critiques in the last section highlight a genuine issue: the official CPI may indeed under‑capture some cost pressures in specific urban or lifestyle contexts, but the magnitude of these pressures leads to conclusions that defy empirical reality. ShadowStats applies a rough, constant adjustment to official data rather than a full methodological recalculation. The Chapwood Index relies on informal surveys of frequently purchased items in major cities, which overweights high‑cost urban areas and volatile essentials. Both approaches overcorrect, producing numbers that do not line up with observable productivity gains, wage trends, or overall economic stability.

Like the Boskin Commission, these alternatives identify real measurement flaws but misjudge their scale and direction. The Food Puzzle, and the broader perception of stagnation, remains unresolved until we adopt a more modest adjustment, such as my roughly 1.5% annual understatement proposed here (the 1.5% drift). This allows for unmeasured deflationary forces from productivity without drifting into absurdity.

How to Read the Data in This Post

To keep the main story readable, I keep in‑text citations and technical details to a minimum. Tables and figures include only short captions and a few key references, so you can follow the argument without wading through footnotes on every line. All of the underlying data series, transformations (such as per‑capita conversions and velocity adjustments), and exact inflation formulas are documented in the Methods and Sources section at the end of this part. If you have questions about where a number comes from or how a graph was constructed, that is the place to look for full sourcing and methodology.

Why the Monetary Test is Necessary

To put these inflation measures under one more stress test, we need to compare them with monetary expansion. In the classical monetary tradition, going back at least to the quantity theory of money, sustained inflation is largely seen as a monetary phenomenon: over long periods, the price level tends to track the amount of money circulating in the economy, once you allow for real growth. That does not mean every short‑term price move is caused by money, but it does give us a simple benchmark: if an inflation index claims large changes in the value of the dollar, those changes should bear some reasonable relationship to how much the money stock itself has grown.

First, we need to clarify the key money‑supply metrics.

Monetary Base (high‑powered money): Currency in circulation plus reserve balances held by banks at the Federal Reserve; in 2024–2025, this stood at roughly 5.4–5.9 trillion dollars.
M1: Currency held by the public plus transaction deposits such as checking accounts.
M2: The broadest and most relevant for consumer spending; includes M1 plus small‑denomination time deposits (under 100,000 dollars) and retail money‑market mutual fund shares.

Alongside these levels, economists track the velocity of money, how many times, on average, each dollar of M2 is used in transactions over the course of a year. If M2 measures the size of the pool of liquid dollars, velocity measures how quickly that pool circulates through actual purchases. A higher velocity means the same stock of money is doing more work in generating nominal GDP. A lower velocity means more of it is sitting idle in reserves or savings. Because many critics argue that “extra M2 never reached the real economy,” velocity is crucial here: it already discounts dollars that do not turn over into measured spending and therefore is the natural way to adjust raw money growth before comparing it to long‑run price changes.

To ensure a fair and honest comparison, there is one final step: we need to adjust for population size. When more people share a stable stock of money, the amount available per person falls, which mechanically raises the value of each dollar and pushes the system toward monetary deflation. Too few dollars chasing goods and services. Persistent deflation is just as destabilizing as high inflation in any economy. This means that a growing population naturally requires a larger money supply over time. That means the relevant question is not simply “How big is M2?” but “How much money is there per person?”, and, once we bring velocity back in, “How much effective money per person is actually turning over into the economy each year?”

That is why, in this section, I take the two extra steps of adjusting M2 by velocity of money, so only the portion of broad money that turns over into measured GDP each year is counted, and I convert this “effective” money stock into per‑capita terms so that simple population growth does not exaggerate inflationary pressure. These adjustments do not make the data flawless, but they deliberately incorporate the standard objections to using M2 as an inflation proxy and provide the fairest monetary benchmark we can construct from public series before comparing it to different inflation interpretations.

How Different Inflation Interpretations Compare to the Money Supply

The graph above shows how many dollars today it takes to equal one dollar from the past under different measures of inflation. For example, the M2‑baseline adjusted for velocity of money indicates that $1 in 1980 is roughly equivalent to about $7 in 2024, depending on the exact dates and revisions. Each curve is an inflation “multiplier”: it takes a past dollar and compounds it forward to 2024 to show how much nominal cash you would need today to have the same purchasing power if that index were the true measure of inflation. The graph plots these CPI‑based multipliers alongside two monetary benchmarks: one based on the monetary base per person and one based on M2 per person adjusted by the velocity of money, which reflects only the portion of broad money that actually turns over into measured spending each year.

Standard CPI (green) and the Boskin‑adjusted index (orange) show relatively little cumulative growth, far below any measure of monetary expansion, an unlikely outcome if prices broadly reflect more dollars chasing goods.
ShadowStats/Chapwood (yellow, about 7 percentage points above CPI since 2000) at times outpaces even explosive growth in the monetary base, implying an inflation level that cannot be reconciled with money creation or observed economic stability.
My 1.5‑percent annual understatement thesis (red) tracks the velocity‑adjusted M2‑per‑capita line most closely. This inflation interpretation is the most consistent with the classical view that broad money growth is the main driver of long‑run price levels of the inflation measures tested. It does mildly overshoot this benchmark, but the gap falls within the range of calculation error for GDP, M2, and an inflation adjustment based on an average drift. In practical terms, the money‑supply path is consistent with a CPI understatement of about 1.4–1.5% per year. I use 1.5% as a simple, rounded estimate rather than a claim of exact precision.

This alignment is not accidental. This version of M2 is the best proxy for liquid money that can readily enter the real economy and influence consumer prices, whereas the monetary base, while crucial for banking policy, includes reserves that may never reach households directly. My proposal does not require a perfect one‑for‑one match, but it does provide the clearest connection between our inflation gauge and monetary reality.

It is also important to note that I arrived at the 1.5‑percent drift estimate before I ever compared it to monetary expansion. In other words, I did not tune the 1.5‑percent figure to match M2 per capita adjusted for the velocity of money; the close fit in this graph emerged only after the fact, as one more link in a growing body of evidence consistent with the drift thesis. At the same time, neither the Food Puzzle results nor this version of the M2 benchmark, nor any single chart, is enough on its own to “prove” the theory. Taken together, they motivate the 1.5‑percent interpretation as a serious, testable alternative, but fully validating or refuting it will require further work across other sectors, countries, and data sources.

The Results of the Food Puzzle.

At this point, every path we have tested points in the same direction. The physical data from farms and food baskets say prices should have fallen much more than CPI admits; Boskin’s downward correction makes the puzzle worse; the ShadowStats–Chapwood adjustments overshoot into fantasy; and when we line all of them up against the growth of money per person, only a modest 1.5‑percent annual understatement keeps prices and money moving in step. In other words, the Food Puzzle is not a mystery of missing savings in the real economy but a measurement problem in our main inflation gauge. The final step is to show, piece by piece, how this 1.5‑percent drift not only reconciles food with productivity, but also brings wages, housing, and everyday living costs back into a story that matches what households actually feel, which is the work of the next section, “Why the 1.5% Drift Validates your Gut feeling about the Economy.”

Methods and Sources

How Many 2024 Dollars Equal One Dollar of the Past (1970–2024)

This graph shows the cumulative inflation multiplier—how many 2024 dollars are required to match the purchasing power of one dollar in a given past year—under four different inflation adjustment methods. It also plots three measures of money‑supply growth, adjusted where noted, which serve as benchmarks for the classical view that long‑run price changes are largely driven by increases in money available to consumers.

Inflation multipliers (forward‑adjusted to 2024):

Boskin Commission adjustment (orange): Subtracts about 1.3 percentage points of annual overstatement before 1996, then 1.1 points after, following the 1996 Advisory Commission report.
Author’s 1.5‑percent understatement thesis (red): Adds 1.5 percentage points to reported CPI across the period to capture unmeasured deflation from productivity, as developed in the NETs framework.
ShadowStats/Chapwood‑style adjustment (yellow): Uses official CPI through 1999, then adds about 7 percentage points of annual understatement from 2000 onward, based on published divergences from Shadow Government Statistics and the Chapwood Index.

Money-Supply Multipliers:

Monetary base per capita (gray): Currency in circulation plus reserve balances at the Federal Reserve (BOGMBASE), divided by U.S. population, then expressed as a ratio to its 2024 level.
M2 per capita adjusted for velocity (blue): M2SL multiplied by the velocity of M2 (M2V), then divided by population and expressed as a ratio to its 2024 value; this series approximates “effective money per person” that actually turns over into measured nominal GDP each year.

For each inflation series, a dollar from a past year is compounded forward to 2024 using that method’s cumulative factor. Money‑supply levels (monetary base, M2, and M2×velocity) are taken from FRED as annual year‑end values, then divided by population (Macrotrends for recent years and Maddison‑style estimates from Bolt and van Zanden for earlier years) and expressed as ratios to their respective 2024 per‑capita levels (for example, 1980 M2×velocity per capita ÷ 2024 M2×velocity per capita).

Sources

Board of Governors of the Federal Reserve System (US). (2026). M2 (M2SL) Dataset. Federal Reserve Bank of St. Louis. Retrieved April 19, 2026, from https://fred.stlouisfed.org/series/M2SL
Board of Governors of the Federal Reserve System (US). (2026). Monetary base: Total (BOGMBASE) Dataset. Federal Reserve Bank of St. Louis. Retrieved April 19, 2026, from https://fred.stlouisfed.org/series/BOGMBASE
Bolt, J., & van Zanden, J. L. (2024). Maddison style estimates of the evolution of the world economy: A new 2023 update. Journal of Economic Surveys, 38(1), 1–41. https://doi.org/10.1111/joes.12618
Boskin, M. J., Dulberger, E. R., Gordon, R. J., Griliches, Z., & Jorgenson, D. W. (1996). Toward a more accurate measure of the cost of living: Final report to the Senate Finance Committee from the Advisory Commission to Study the Consumer Price Index. U.S. Senate Committee on Finance. https://www.ssa.gov/history/reports/boskinrpt.html
Bureau of Labor Statistics. (2026). Consumer price index for all urban consumers (CPI‑U) Dataset. U.S. Department of Labor. https://www.bls.gov/cpi/
Chapwood Index. (n.d.). The Chapwood Index: Our solution. Retrieved January 4, 2026, from https://chapwoodindex.com/the-solution/
Federal Reserve Bank of St. Louis. (2026). Velocity of M2 money stock (M2V) Dataset. Federal Reserve Bank of St. Louis. Retrieved April 19, 2026, from https://fred.stlouisfed.org/series/M2V
Macrotrends. (2025). United States population 1820–2024 Dataset. Macrotrends LLC. https://www.macrotrends.net/global-metrics/countries/usa/united-states/population
Williams, J. (2023, June 14). Shadow Government Statistics: Analysis behind and beyond the economic reporting. Retrieved January 4, 2026, from https://www.shadowstats.com/alternate_data/inflation-charts

Author: Kyle Novack

April 27, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

Why Severe Understatement of Inflation Goes Too Far

Kyle Novack — Fri, 24 Apr 2026 13:31:32 GMT

Having seen that the Boskin Commission’s “CPI is too high” correction makes the Food Puzzle worse, the natural next question is what happens when CPI understates inflation instead. That is the story told by ShadowStats and the Chapwood Index, which have become two of the most widely cited sources for the claim that official inflation has been running far below reality since the early 2000s. ShadowStats argues that methodological changes, substitution, hedonic adjustments, and geometric averaging have caused CPI to miss several percentage points of inflation each year. The Chapwood Index makes a similar claim from another angle, reporting double‑digit “true” inflation in many major cities based on a fixed basket of everyday items, with no quality or substitution adjustments (Williams, 2023, & Chapwood Index, n.d.).

In this section, I use those critiques as a stress test for the “CPI understates inflation” story because they are the most widely cited examples of that view. Both ShadowStats and the Chapwood Index point to a similar bottom line: that CPI has understated inflation by at least 7 percentage points a year since around 2000. For consistency with earlier parts of the Food Puzzle, I apply that implied understatement, an extra 7 percentage points a year, to the same food basket, incomes, and housing data used before. As we will see, this pushes the numbers into a different kind of implausibility. They imply 73-87% real price declines for staple foods and 60-75% real declines in wages, home prices, and GDP per capita since 1980, patterns we simply do not observe in either the physical data or everyday life.

How to Read the Data in This Post

Unless otherwise noted, all dollar figures in this section are expressed in 2024 dollars using either the official CPI‑U or a ShadowStats/Chapwood‑style upward adjustment to CPI‑U, as specified in each table or figure. The charts and tables draw on standard public datasets, USDA (NASS and ERS), BLS, Census, FRED, MeasuringWorth, Macrotrends, and similar population series so nothing here depends on exotic or private sources. Where I use a ShadowStats/Chapwood‑style adjustment, the goal is to test that framework against the Food Puzzle, not to present it as the default method for measuring inflation.

To keep the main story readable, I limit in‑text citations and technical details; each figure has a short caption explaining how it was calculated, which datasets it uses, and whether it relies on official CPI‑U or a ShadowStats/Chapwood‑style path. A Methods and Sources section at the end lists all underlying series, transformations, and inflation adjustments, so anyone who wants to audit or recreate the charts can do so.

Why ShadowStats and Chapwood Say CPI Is Too Low

Before we dive into the results of the stress test, it is important to understand why ShadowStats and the Chapwood Index believe CPI is understating inflation by so much in the first place.

ShadowStats: Re‑creating an Older CPI

ShadowStats argues that methodological changes to the CPI since the early 1980s, such as allowing consumers to substitute cheaper goods, adjusting for quality improvements, and using geometric rather than arithmetic averaging, have caused the official index to understate true inflation. Its SGS‑Alternate CPI series is presented as a reconstruction of pre‑1980 methods (or approximations of them) and typically shows inflation running several percentage points higher than the official CPI each year (Williams, 2023).

In practice, this translates into a claimed understatement of roughly 7 percentage points per year in recent decades, with some estimates suggesting closer to 9 percentage points since 2000, and cumulative price increases exceeding 600%. In his own methodological notes, Williams explains that the ShadowStats‑Alternate CPI series is built by reverse‑engineering the BLS CPI‑U‑RS research series and adding estimates of the inflation effects of various methodological changes “on an additive basis,” and that “the series is not recalculated” from underlying price data. In other words, the SGS‑Alternate CPI is an adjusted version of the official index, not a fully independent reconstruction from raw prices. For this analysis, I set those debates aside and simply take the published divergence at face value, adding 7 percentage points to reported CPI inflation from 2000 onward (Williams, 2008 & Williams, 2023).

The Chapwood Index: A Fixed Everyday Basket

The Chapwood Index, created by investment adviser Ed Butowsky, makes a similar case from a different angle. It claims that official CPI misses the true rise in the cost of living because it does not fully capture price increases in everyday essentials across U.S. cities. Chapwood tracks prices for 150 frequently purchased items, groceries, utilities, insurance, entertainment, and local services in the 50 largest metropolitan areas. They use a fixed basket with no quality or substitution adjustments and simply report the raw percentage change each year (Chapwood Index, n.d.).

The index routinely shows inflation 7–10 percentage points, or more, higher than CPI, with some city‑level readings in the low‑ to mid‑teens (for example, around 13–14 percent in some recent years versus roughly 3–4 percent in official CPI). Butowsky and his supporters present this as the “real” erosion of purchasing power, arguing that “the CPI no longer measures the true increase required to maintain a constant standard of living” (Chapwood Index, n.d.).

The Stress Test

Now that we’ve seen why ShadowStats and the Chapwood Index think CPI is missing so much inflation, we can ask the key question for this series: how do their numbers hold up when we run them through the Food Puzzle stress test? The results are striking. Applying their adjustments to CPI produces extraordinary apparent price declines since 1980. Real food prices fall 73-87% across categories, far beyond even the most optimistic reading of agricultural productivity gains.

USDA data indicate agricultural Total Factor Productivity growth of about 1.49 percent per year from 1948 to 2021; compounded over four decades, that might reasonably imply real food price reductions of 40–60 percent for many staples. A 73–87 percent drop overshoots this dramatically, implying a kind of hidden hyper‑deflation that we simply do not see in either living standards or actual market behavior.

How Wages and House Prices Compare after the Adjustment.

The distortion becomes even clearer when we look beyond food prices to wages, housing, and overall output. Under a 7‑point understatement adjustment, the 1980 median household income of about $17,710 would equate to roughly $337,569 in 2024 terms. This implies real household incomes have fallen around 75% since 1980. Median home prices (about $66,400 in 1980) would translate to roughly $1.27 million today, suggesting real home prices have declined by about 67%, and GDP per capita shows a similar real drop of roughly 64%.

If any of this were accurate, we would be living through an economic depression far worse than the 1930s: widespread poverty, collapsing living standards, and mass unemployment on a scale that would dominate history books. Instead, the U.S. has seen ongoing, if uneven, growth in real consumption and rapid technological progress, with nothing resembling the collapse implied by these numbers. ShadowStats and Chapwood don’t just fail the stress test. They require a hidden economic catastrophe that somehow left no trace in the physical world.

The Catastrophic Results of the Stress Test

Taken together, the Boskin correction and the ShadowStats/Chapwood adjustments push us into opposite but equally unrealistic worlds. One turns a productivity‑rich food system into a sector with rising prices. The other implies wage and price collapses so extreme that we would be living through a depression worse than the 1930s. Neither picture matches the physical data or everyday experience.

Before we discard them entirely, it is worth giving all three CPI interpretations one final, neutral test. Classical monetary theory says that, over time, inflation should largely reflect the expansion of the money supply. In the next section, we line up official CPI, the Boskin‑adjusted index, the 7‑percent‑understatement path, and my proposed 1.5‑percent drift against M2 money per person (adjusted for velocity) to see which, if any, tracks the monetary inflation we have created.

Methods and Sources

Comparison of Current (2024) Food Prices vs. 1980 Prices Adjusted for ShadowStats and Chapwood CPI Critiques

All price and income figures are converted to 2024 dollars using a ShadowStats/Chapwood‑style understatement adjustment to CPI‑U before comparison.

Food item prices: For each of the 17 food items, nominal monthly prices for 1980 and 2024 are taken from the BLS Average Price (AP) series and averaged over the 12 months to get an annual price for each year. The 1980 annual prices are then rolled forward to 2024 by taking the official CPI‑U inflation path and adding 7 percentage points to the annual inflation rate from 2000 onward, reflecting the minimum understatement implied by ShadowStats and the Chapwood Index. This produces ShadowStats/Chapwood‑adjusted 2024‑dollar values, which are compared with actual 2024 prices to compute percentage differences. The “average price change” row is the simple mean of these 17 percentage differences.

Income and housing: Median household income is taken from Census historical income tables; median home prices from the FRED MSPUS series. The 1980 values are inflated to 2024 dollars using the same ShadowStats/Chapwood‑style adjustment (official CPI through 1999, then CPI plus 7 percentage points per year from 2000 onward) before computing percentage differences, so income and housing are treated on the same basis as the food basket.

GDP per capita: GDP levels are assembled from MeasuringWorth and FRED, with population from Maddison‑style estimates and Macrotrends. I compute GDP per capita as real GDP ÷ population, then express the 1980 value in 2024 dollars using the same ShadowStats/Chapwood‑adjusted CPI path where needed, to match the treatment of other income variables.

Sources

Bolt, J., & van Zanden, J. L. (2024). Maddison style estimates of the evolution of the world economy: A new 2023 update. Journal of Economic Surveys, 38(1), 1–41. https://doi.org/10.1111/joes.12618
Bureau of Labor Statistics. (2026). Consumer price index for all urban consumers (CPI‑U) Dataset. U.S. Department of Labor. https://www.bls.gov/cpi/
Chapwood Index. (n.d.). The Chapwood Index: Our solution. Retrieved January 4, 2026, from https://chapwoodindex.com/the-solution/
Federal Reserve Bank of St. Louis. (2024). Gross domestic product (GDP) Dataset. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/GDP
Federal Reserve Bank of St. Louis. (2024). Median sales price of houses sold for the United States (MSPUS) Dataset. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/MSPUS
Macrotrends. (2025). United States population 1820–2024 Dataset. Macrotrends LLC. https://www.macrotrends.net/global-metrics/countries/usa/united-states/population
U.S. Bureau of Labor Statistics. (2024). Average price data (AP), U.S. city average Dataset. U.S. Department of Labor. https://www.bls.gov/charts/consumer-price-index/consumer-price-index-average-price-data.htm
U.S. Census Bureau. (2024). Historical income tables: Households Dataset. U.S. Department of Commerce. https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-income-households.html
Williamson, S. H. (2025). What was the U.S. GDP then? MeasuringWorth. http://www.measuringworth.org/usgdp/
Williams, J. (2008, September 10). Response to BLS article on CPI misconceptions. Special comment. Shadow Government Statistics. http://www.shadowstats.com/article/special-comment
Williams, J. (2023, June 14). Shadow Government Statistics: Analysis behind and beyond the economic reporting. Retrieved January 4, 2026, from https://www.shadowstats.com/alternate_data/inflation-charts

Author: Kyle Novack

April 28, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

Walking the Bridge to Your Dream Life

Kyle Novack — Thu, 23 Apr 2026 13:31:10 GMT

Every Labor Day in Michigan, thousands of people get the chance to do something unusual: they close the Mackinac Bridge to cars and let you walk it.

Five miles over open water. Steel, wind, and a lake on both sides. It’s a tradition, but it’s also something else: a surprisingly accurate picture of what it feels like to build your dream life.

Not the fantasy version where everything “clicks” overnight. The real version of progress is slow, awkward, and invisible for a long time.

The strange thing about the bridge

Since it is a suspension bridge, the bridge itself is not flat. Instead, it is one long arc. At first, it doesn’t feel like much is happening.

You start out with energy. There’s excitement. You’re surrounded by people. You take your first few hundred steps and look up…

And it feels like you’re in the exact same place.

The view hasn’t changed much. The shore behind you looks only slightly farther away. The shore ahead of you doesn’t look any closer. The bridge surface rises so gradually that you’re climbing without really noticing that you’re climbing.

You are putting in effort. Your legs are working. But your brain can’t connect that effort to visible progress yet. There is no satisfying “I’m getting closer” sensation. Just more walking, more bridge, more of the same.

That’s where most people quietly start to think, “Wow, this is longer than I expected.”

The invisible half of the journey

This first stretch of the Mackinac Bridge is the part no one romanticizes.

You’re technically moving forward, but it doesn’t feel like it. The incline is there, but you only notice it as a kind of low-grade resistance—enough to tire you out, not enough to feel dramatic. You can’t see the full distance you’ve already covered, and you can’t clearly see the finish line either.

This is exactly what it feels like when you decide to change your life.

You make a choice:

I’m going to build a business.
I’m going to change careers.
I’m going to get out of this schedule I don’t control.
I’m going to design my days around what matters to me.

At the moment of decision, it feels big, clear, and powerful.

But then the work begins.

You start waking up earlier. You carve out time to write. You study. You build systems. You say “no” to things that used to be automatic “yeses.” You make calls, send messages, build relationships, and learn difficult skills.

And for a while, nothing looks different.

The income looks the same.
The calendar looks the same.
The stress might even look worse.

You’re walking the bridge, but your brain hasn’t yet gotten the payoff of seeing the other side get closer. The effort is real, but the progress is still mostly hard to measure.

This is the most dangerous phase, because it’s where it’s easiest to tell yourself a story that isn’t true:

“If I were on the right path, it would be working by now.”
“Other people seem to move faster than this.”
“Maybe I should go back to what I was doing before.”

You’re not failing. You’re just on the incline.

The moment everything changes

Eventually, if you keep walking, you reach the midpoint of the Mackinac Bridge.

At this point, because of the wall of individuals walking the bridge with you, you’re about an hour into the walk. Then you notice something at the end of the bridge. For the first time, you see the giant banner across the bridge that says FINISH. For the next two-plus miles, you can see the finish line the entire time.

Every step now gives you a visible reward. The shore is getting closer. The details get sharper. Buildings separate from the skyline. What used to be “somewhere out there” becomes “right there.”

You’re still walking, but the psychology has flipped. You no longer have to rely purely on faith that your steps matter. You can see that your steps matter.

Life works the same way.

There’s a point in the process where the grind starts to generate visible results. It could be:

The first time someone you don’t know reaches out because your work resonated.
The first month your “side” income meaningfully covers a real expense.
The moment your calendar reflects your priorities more than your obligations.
The first time you realize, “If I keep doing what I’m doing, this will work. It’s just math and time.”

That’s the top of the bridge. You haven’t arrived yet, but you can see that arrival is inevitable if you keep going.

The work after that isn’t easy, but it’s different. It’s fueled by evidence rather than just hope.

The life I’m walking toward

For me, the “other side of the bridge” is pretty specific.

I’m building a life where I can focus deeply on developing this framework that rethinks how we measure inflation and value, without having to ask permission to use my own time or squeeze it into the margins of everything else.

I’m also building a life where I can choose to be with my daughter and my wife when I want to, not just when a schedule allows me to.

That’s the picture I hold in my mind when the walk feels long.
That’s my shoreline.

And right now, I’m still on the incline.

I’m doing the work that doesn’t always look impressive from the outside: writing, refining ideas, slowly building an audience, and putting structures in place that future-me will rely on. There are days when it feels like just more bridge, more steps, more of the same.

But that doesn’t mean nothing is happening.

It means I haven’t reached the hump yet.

Where are you on your bridge?

If your life feels like effort without obvious progress, it’s easy to interpret that as a sign you’re doing the wrong thing.

Sometimes that’s true. Course corrections are real. But very often, you’re just in the same place as everyone who has ever walked from where they are to where they want to be:

On the invisible half of the bridge.

You can’t feel the finish line getting closer yet. Your legs are tired. The shore behind you doesn’t look that far away. Turning back would be easy.

But the only way to get to that moment, where you can finally see your new life getting closer with every step, is to keep walking when everything in you says, “This isn’t working.”

If you’re there right now, on that long, unromantic stretch, don’t underestimate what you’re doing.

You’re not stuck.
You’re not failing.
You’re not behind.

You’re just not at the top of the bridge yet.

Keep going. The view from the midpoint will change everything.

By Kyle Novack

April 23, 2026

Believing CPI Overstates Inflation Only Deepens the Food Puzzle

Kyle Novack — Tue, 21 Apr 2026 13:31:31 GMT

We’ve already seen that the official Consumer Price Index (CPI) fails the Food Puzzle stress test, so the natural next question is whether a major critique of CPI does any better: the 1996 Boskin Commission. The Commission famously concluded that CPI most likely overstates inflation by about 1.1 percentage points per year, with a plausible range of 0.8–1.6. In its own words, changes in the CPI “have substantially overstated the actual rate of price inflation, by about 1.3 percentage points per annum before 1996 (the extra 0.2 percentage point is due to a problem called formula bias inadvertently introduced in 1978 and fixed this year),” and it is “likely that a large bias also occurred looking back over at least the last couple of decades” (Boskin et al., 1996)

How to Read the Data in This Post

Unless otherwise noted, all dollar figures in this section are expressed in 2024 dollars using CPI‑U or a Boskin‑style adjustment to CPI‑U, as specified in each table or figure. The charts and tables draw on standard public datasets, USDA (NASS and ERS), BLS, Census, FRED, and widely used population estimates, so nothing here depends on exotic or private sources. Where I use Boskin’s adjustment, the goal is to test his framework against the Food Puzzle, not to present it as the default method for measuring inflation. To keep the main story readable, I limit in‑text citations and technical details; each figure has a short caption explaining how it was calculated and which datasets it uses, including whether it relies on official CPI or on the Boskin‑adjusted path. A Methods and Sources section at the end lists all underlying series, transformations, and inflation adjustments, so anyone who wants to audit or recreate the charts can do so.

Re‑testing CPI with the Boskin Commission Thesis

To test this view against the Food Puzzle, we apply Boskin’s estimated bias retroactively: subtracting 1.3 percentage points from reported CPI inflation each year from 1980 to 1995, and 1.1 percentage points from 1996 onward. This isn’t just Boskin’s number in isolation; most mainstream “CPI is too high” critiques cluster in that same 0.8–1.6 percent band. Meaning that if Boskin’s adjustment fails this test, any similar adjustment will fail with it. As we will see, this correction makes the Food Puzzle worse, not better. Adjusted numbers make real food prices look even higher over time, despite the massive productivity gains documented in earlier parts of this series (Boskin et al., 1996).

What Boskin’s Fix Does to Food Prices

Using the same 17 staple foods from Part 10 that we used to critique the official CPI, we now take their 1980 prices and carry them forward to 2024 dollars under the Boskin methodology. Under this adjustment, every single item ends up more expensive in real terms in 2024 than in 1980. Standard CPI at least shows modest declines in some categories, such as chicken and potatoes. Boskin’s correction erases those gains and turns them into increases. In a sector where productivity has clearly exploded, that is the opposite of what we should expect.

Boskin’s Story vs. the Cost of Living

This is especially striking when you consider the Boskin Commission’s main narrative, which rests on a simple idea about how people shop. Their argument rests on structural changes that should lower the cost of living: consumers shifting to discount stores, new outlets, and cheaper substitutes, so that a given standard of living should require less income than the CPI implies. Yet when we apply their own correction to basic foods, every price in the test basket rises in real terms. That is a strange way to demonstrate an improvement in living standards (Boskin et al., 1996).

It also collides with how the Boskin findings have been used in policy debates. If CPI has been overstating inflation, then Social Security, pensions, and other CPI‑linked benefits have been rising faster than necessary to maintain a given standard of living, so policymakers have argued that future cost‑of‑living adjustments can safely grow more slowly. In plain terms, that means slower growth in inflation‑adjusted incomes for millions of households. If the Boskin-adjusted numbers indicate that food is getting more expensive across the board, then it is hard to reconcile that with the idea that smaller cost-of-living increases are enough to keep those households whole.

Incomes, Generations, and the Squeezed Middle

The contradiction becomes even starker when we extend the adjustment to broader economic indicators and incomes. Under Boskin’s logic, real average wages have risen 113% since 1980, and real GDP per capita has grown 213%. To see what that means in everyday terms, consider how their adjustment affects the income numbers: the 1980 median household income of $17,710 becomes roughly $39,000 in 2024 terms. On paper, that implies that a reasonable real income level today should buy a clearly better standard of living than it did in 1980. In that story, a typical worker or median‑income household should feel far more comfortable than their parents did; homes should be more affordable, raising children should be easier, and moving up the ladder should be more common.

Economists often respond that a household on 39,000 dollars today enjoys far better technology and conveniences than a similar‑income family in 1980, and that is true: smartphones, streaming, and modern appliances are real gains. But in a society that keeps accumulating knowledge and engineering know‑how, it would be strange if the average home didn’t have better technology than in the past; that is the natural trajectory of a progressing economy, not evidence that everything else is fine.

Those visible gadgets also make an easy scapegoat for stagnating living standards, even though most of them are occasional or one‑time purchases, not the recurring costs that determine whether a budget balances each month. When we look at the essentials, housing, food, energy, and childcare, a $39,000 income leaves very little room to cover the basics, let alone save or build security. In practice, most households need a much higher income than that just to participate in today’s society without constant financial strain.

Yet the opposite is widely observed. Homeownership rates for under‑35s have fallen sharply, fertility rates are at historic lows, and a recent Pew survey found that 74% of Americans describe current economic conditions as only fair or poor. Intergenerational mobility data tell the same story. Roughly 90% of children born in the 1940s earned more than their parents, compared with less than half of those born in the 1980s. In many parts of the country, even households near today’s median income report that housing, childcare, and groceries leave little room for savings or security, which is hard to reconcile with charts that say real incomes have more than doubled (Feiveson, 2024, Lautz, 2025, Walsh 2026, & Copeland, 2025).

Boskin’s adjustment also quietly rewrites the story of how life supposedly got better. It implies that most of the improvement we have seen in hours worked to buy a standard basket of goods comes not from correctly measured price relief, but almost entirely from rising incomes. There is little evidence for that. Even under standard assumptions, a world of strong productivity and rapid real wage growth should produce broad, felt gains in affordability, not a generation that feels squeezed while the charts say they are winning.

What Boskin Gets Right—and Wrong

If the Boskin Commission were correct about the direction of CPI’s bias, we would not be living through the cost‑of‑living crisis that defines today’s economy. Their discussion of specific measurement issues, substitution bias, quality change, and new goods was insightful, and many of their technical recommendations were eventually adopted by the BLS. But their core diagnosis, that these flaws cause CPI to overstate inflation, points the wrong way. In a sector where productivity is powerfully deflationary, the evidence from food, wages, and housing suggests that the real bias in CPI is understatement of deflation, not overstatement of inflation.

This analysis does not dismiss the Boskin report outright; it builds on its foundation while flipping the interpretation. The evidence from food prices, wages, housing, and mobility suggests that the Commission identified real measurement issues but misdiagnosed their net effect on the cost of living. That misdiagnosis leaves the Food Puzzle, and much of the perceived stagnation of recent decades, unresolved. And if trimming a little inflation off the top makes the paradox worse, the natural next question is whether going to the opposite extreme, claims that CPI has understated inflation by as much as 7 percentage points a year since 2000, can do any better.

Methods and Sources

Comparison of Current (2024) Food Prices vs. 1980 Prices Adjusted for Boskin Commission’s Overstatement of CPI

All price and income figures are converted to 2024 dollars using a Boskin‑style adjustment to CPI‑U before comparison.

Food item prices: For each of the 17 food items, nominal monthly prices for 1980 and 2024 are taken from the BLS Average Price (AP) series and averaged over the 12 months to get an annual price for each year. The 1980 annual prices are then multiplied by an adjusted CPI ratio that subtracts the Boskin Commission’s estimated bias from reported inflation (1.3 percentage points per year from 1980–1995, 1.1 points from 1996 onward) before rolling 1980 forward to 2024. This produces Boskin‑adjusted 2024‑dollar values, which are compared with actual 2024 prices to compute percentage differences. The “average price change” row is the simple mean of these 17 percentage differences.

Income, housing, and rent: Median household income is taken from Census historical income tables; median home prices from the FRED MSPUS series. The 1980 values are inflated to 2024 dollars using the Boskin‑adjusted CPI series before computing percentage differences, so that income and housing are treated on the same basis as the food basket.

GDP per capita: GDP levels are assembled from MeasuringWorth and FRED, with population from Maddison‑style estimates and Macrotrends. I compute GDP per capita as real GDP ÷ population, then express the 1980 value in 2024 dollars using the Boskin‑adjusted CPI series where needed, to match the treatment of other income variables.

Sources

Bolt, J., & van Zanden, J. L. (2024). Maddison style estimates of the evolution of the world economy: A new 2023 update. Journal of Economic Surveys, 38(1), 1–41. https://doi.org/10.1111/joes.12618
Boskin, M. J., Dulberger, E. R., Gordon, R. J., Griliches, Z., & Jorgenson, D. W. (1996). Toward a more accurate measure of the cost of living: Final report to the Senate Finance Committee from the Advisory Commission to Study the Consumer Price Index. U.S. Senate Committee on Finance. https://www.ssa.gov/history/reports/boskinrpt.html
Bureau of Labor Statistics. (2026). Consumer price index for all urban consumers (CPI‑U) [Data set]. U.S. Department of Labor. https://www.bls.gov/cpi/
Copeland, J. (2025, October 3). Most Americans continue to rate the U.S. economy negatively as partisan gap widens. Pew Research Center. Retrieved January 7, 2026, from https://www.pewresearch.org/short-reads/2025/10/03/most-americans-continue-to-rate-the-us-economy-negatively-as-partisan-gap-widens/
Federal Reserve Bank of St. Louis. (2024). Gross domestic product (GDP) [Data set]. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/GDP
Federal Reserve Bank of St. Louis. (2024). Median sales price of houses sold for the United States (MSPUS) [Data set]. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/MSPUS
Feiveson, L. (2024, December 18). How does the well‑being of young adults compare to that of their parents? U.S. Department of the Treasury. Retrieved January 6, 2026, from https://home.treasury.gov/news/featured-stories/how-does-the-well-being-of-young-adults-compare-to-their-parents
Lautz, J. (2025, January 7). Young buyers continue to be fenced out of homeownership. National Association of Realtors. Retrieved January 7, 2026, from https://www.nar.realtor/blogs/economists-outlook/young-buyers-continue-to-be-fenced-out-of-homeownership
Macrotrends. (2025). United States population 1820–2024 [Data set]. Macrotrends LLC. https://www.macrotrends.net/global-metrics/countries/usa/united-states/population
U.S. Bureau of Labor Statistics. (2024). Average price data (AP), U.S. city average [Data set]. U.S. Department of Labor. https://www.bls.gov/charts/consumer-price-index/consumer-price-index-average-price-data.htm
U.S. Census Bureau. (2024). Historical income tables: Households [Data set]. U.S. Department of Commerce. https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-income-households.html
Walsh, R. (2026, January 6). Does the U.S. have a fertility crisis? Johns Hopkins Bloomberg School of Public Health. Retrieved January 7, 2026, from https://publichealth.jhu.edu/2026/is-the-us-birth-rate-declining
Williamson, S. H. (2025). What was the U.S. GDP then? MeasuringWorth. http://www.measuringworth.org/usgdp/

Author: Kyle Novack

April 21, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

CPI at Odds with the Food Economy

Kyle Novack — Fri, 17 Apr 2026 13:31:52 GMT

By the end of the last section, we had run out of plausible suspects. Farms, supply chains, corporate margins, consumer behavior, external shocks, and regulation all failed the stress test and were dismissed from the witness stand. Each explained small pieces of the story, but none could absorb the huge cost savings implied by a century of agricultural productivity gains. The food system is not innocent in every respect, but the central question is: why have prices not fallen nearly as far as production efficiency suggests they should? The evidence points away from the physical economy and toward the measurement system itself.

The main yardstick we use for inflation, the Consumer Price Index (CPI), tracks how the prices of a typical consumer basket change over time. It only ever sees the prices households pay, after markets have already absorbed both money‑driven inflation and productivity‑driven cost reductions. Agriculture gives us the best experiment we have for testing whether CPI has a productivity blind spot, because we can see, in physical units, how much input it now takes to produce a given amount of output. The conclusion is that CPI tends to understate how much the dollar has really weakened, because part of the inflation is “hidden” behind efficiency gains that never show up as lower measured prices.

To see how big that bias might be and in which direction, the next articles will put three major CPI interpretations, as well as the official index itself, under the same stress test we have used throughout this series. The first is the Boskin Commission’s view, the most influential official critique, which argues that CPI overstates inflation. The second is a more extreme set of claims from ShadowStats and the Chapwood Index, which say CPI wildly understates it. The third is a more modest 1.5‑percentage‑point‑per‑year understatement, my own thesis, developed outside the food series but tested against it. When we run that 1.5% drift through the same food price and productivity data, it behaves as a good measuring stick should: it turns the food puzzle into a consistent story rather than further breaking the data.

How to Read the Data in This Post

Unless otherwise noted, all dollar figures in this section are expressed in CPI‑adjusted 2024 dollars. The charts and tables draw on standard public data sets, USDA (NASS and ERS), BLS, Census, FRED, and widely used population estimates, so nothing here depends on exotic or private sources. To keep the main story readable, I limit in‑text citations and technical details; each figure has a short caption explaining how it was calculated and which datasets it uses. A Methods and Sources section at the end lists all underlying series and transformations, so anyone who wants to audit or recreate the charts can do so.

Why CPI Fails the Food Puzzle.

The first test is simple. We compare how much total commodity revenue per person and total agricultural expenses per person fell, after CPI adjustment, with how much they should have fallen given measured Total Factor Productivity (TFP) gains. In Graph 1, the blue line shows the cumulative decrease in cost implied by productivity, while the green and gray lines show the observed changes in revenue and expenses per capita. From the 1970s into the 1980s, the observed series sometimes come close, but they usually lag well behind the productivity benchmark.

That apparent catch‑up needs context. The 1970s and early 1980s were a massive bubble period for farm costs, driven by energy crises, high inflation, high interest rates, rising debt burdens, and weather‑related supply problems. Graph 2 shows that CPI-adjusted total agricultural expenses per person almost doubled from 1910 to 1980; even from 1948 to 1980, they rose 38%, despite an expected 31% decline due to productivity, leaving a roughly 70-point gap that no single bubble can explain. The surge leading up to 1980 also distorts the baseline, since prices were already elevated by the agricultural bubble; thus, the post-1980 decline (40.69% under CPI) still falls well short of what sustained TFP should deliver.

Once we set aside the 1970s and early 1980s bubble, a consistent pattern emerges. For most starting years across the twentieth century, the CPI‑adjusted declines in revenue and expenses per person reach only about half to three‑quarters of what the TFP data say they should, typically ending up roughly 25 to a bit over 40 percentage points short of the productivity‑implied cost reductions. Taken together, those gaps show that CPI systematically misses a large share of the savings from productivity gains in agriculture.

This still leaves us with abstract aggregates, so the next step is to look at the prices you pay at the grocery store and compare them with those in 1980. Across 17 common staples in a typical household basket, CPI‑adjusted data show only a 3.12% decline in average prices since 1980, even after the massive bubble in agricultural expenses and revenue during that period. That figure also closely matches CPI’s own inflation picture, because CPI relies on percentage changes rather than price levels to limit calculation errors. Yet when we total the cost of all 17 items and compare it across years, the basket is only about 11.5% cheaper, a strikingly small payoff for the underlying productivity boom and for coming down from such an extreme bubble. These are exactly the kinds of foods that should be most exposed to cost reductions from farm and supply‑chain productivity gains: they are basic staples that households buy over and over again, creating strong pressure for large, efficient supply, and they are only lightly processed, so the post‑farm supply chain adds relatively little value and its own productivity improvements should reinforce, not offset, the cost savings from the field.

We also must step back and ask a deeper question. Over the same period, CPI says median household income is up about 29.3%, and GDP per capita is up almost 90%, yet many Americans feel more squeezed than ever as they try to balance their monthly budgets. At the same time, the median home price has risen roughly 72.6%, so wages have clearly not kept pace with the cost of a basic place to live.

In an economy that can produce food, goods, and services far more efficiently than in 1980, how is it possible that higher measured incomes coexist with a growing sense that everyday life is harder to afford? When we put CPI’s handling of the food puzzle alongside its picture of household finances, it becomes difficult to believe that the index is fully capturing the economic reality most families face. In that sense, CPI has failed the stress test; the numbers and the physical world no longer line up. At this point, CPI must show that its interpretation of the economy is still reliable, rather than remaining the standard, mostly because it got there first and was built into everything else. The natural next question is whether an alternative interpretation of inflation can do better. We turn first to the Boskin Commission’s view, to see whether its claim that CPI overstates inflation fits the hard data any more closely than the standard index does.

Methods and Sources

Estimated Cumulative Deflationary Impact of Agricultural Productivity Gains vs. Observed Changes in Per Capita Revenue and Expenses (1948–2021, CPI-Adjusted)

All series are converted to per‑capita values in 2024 dollars using the official CPI, then expressed as percentage changes from each base year to 2021, the last year available in the USDA TFP series. In practice, that means the 2021 values are first restated in 2024 dollars for consistency with the rest of the Food Puzzle series, and the percentage changes shown in the graph are calculated using these CPI‑adjusted figures; this introduces a small timing mismatch, but it keeps all dollar amounts in a common 2024 frame while preserving a valid comparison of trends.

Blue line (TFP‑implied decrease): Uses the USDA ERS TFP index to compute cumulative productivity gains from each year to 2021, then converts those gains into the expected percentage decline in costs if productivity were fully passed through (for example, a 50.8% expected expense reduction from a 1980 base to 2021).
Green line (revenue per capita): Uses USDA ERS all‑commodity farm revenue, deflated by CPI into 2024 dollars and divided by population, then shows the percentage change in real revenue per capita from each base year to 2021.
Gray line (expenses per capita): Uses USDA ERS total production expenses, deflated by CPI into 2024 dollars and divided by population, then shows the percentage change in real expenses per capita from each base year to 2021, with signs flipped so that cost declines plot as positive numbers and increases (like the 38.37% rise from 1948–1980) stand out as negative.

Graph 2: Total Agricultural Production Expenses Per Capita, Adjusted for CPI (1910–2024)

This graph uses total U.S. agricultural production expenses from the USDA ERS Farm Income and Wealth Statistics (production expenses series), converted to real per capita values in 2024 dollars.

Nominal expenses for each year 1910–2024 are deflated with the official CPI (all‑items U.S. city average) to obtain constant‑dollar values.
These CPI‑adjusted expenses are divided by total U.S. population in each year to produce real expenses per capita.
The plotted series is simply this real per‑capita expense level over time, from which changes like the 38.37% real increase between 1948 and 1980, and the subsequent post‑1980 decline, are calculated directly as percentage differences between years.

Comparison of Current (2024) Food Prices vs. 1980 Prices Adjusted CPI

All price and income figures are converted to 2024 dollars using the official CPI‑U before comparison.

Food item prices: For each of the 17 food items, nominal monthly prices for 1980 and 2024 are taken from the BLS Average Price (AP) series and averaged over the 12 months to get an annual price for each year. The 1980 annual prices are then multiplied by the CPI ratio (CPI-2024 ÷ CPI-1980) to obtain CPI-adjusted 2024-dollar values, which are compared with actual 2024 prices to compute percentage differences. The “average price change” row is the simple mean of these 17 percentage differences.
Basket totals and hours of work: The “basket” cost in each year is the sum of the 17 food items’ prices; hours of work required are calculated as basket cost ÷ average hourly wage for that year (BLS wage series), with 1980 wages also CPI‑adjusted to 2024 dollars for consistency.
Income, housing, and rent: Median household income is taken from Census historical income tables; median home prices from the FRED MSPUS series. The 1980 value is inflated to 2024 dollars using CPI before computing percentage differences.
GDP per capita: GDP levels are assembled from MeasuringWorth and FRED, population from Maddison‑style estimates and Macrotrends; I compute GDP per capita directly as real GDP ÷ population and then CPI‑adjust it to 2024 dollars where needed.

Sources

Bolt, J., & van Zanden, J. L. (2024). Maddison style estimates of the evolution of the world economy: A new 2023 update. Journal of Economic Surveys, 38(1), 1–41. https://doi.org/10.1111/joes.12618
Bureau of Labor Statistics. (2026). Consumer Price Index for All Urban Consumers (CPI-U) [Data set]. U.S. Department of Labor. https://www.bls.gov/cpi/
Federal Reserve Bank of St. Louis. (2024). Gross domestic product (GDP) [Data set]. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/GDP
Federal Reserve Bank of St. Louis. (2024). Median sales price of houses sold for the United States (MSPUS) [Data set]. FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/MSPUS
Macrotrends. (2025). United States population 1820–2024 [Data set]. Macrotrends LLC. https://www.macrotrends.net/global-metrics/countries/usa/united-states/population
U.S. Bureau of Labor Statistics. (2024). Average price data (AP), U.S. city average. https://www.bls.gov/charts/consumer-price-index/consumer-price-index-average-price-data.htm
U.S. Census Bureau. (2024). Historical income tables: Households [Data set]. U.S. Department of Commerce. https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-income-households.html
U.S. Department of Agriculture, Economic Research Service. (n.d.). Agricultural productivity in the U.S.: Summary of recent findings. Retrieved March 14, 2026, from https://www.ers.usda.gov/data-products/agricultural-productivity-in-the-united-states/summary-of-recent-findings
U.S. Department of Agriculture, Economic Research Service. (n.d.). Farm income and wealth statistics [Data set]. U.S. Department of Agriculture. https://data.ers.usda.gov/report.aspx?ID=4059
U.S. Department of Agriculture, Economic Research Service. (n.d.). Farm income and wealth statistics: Value of production and cash receipts tables [Data set]. U.S. Department of Agriculture. https://data.ers.usda.gov/report.aspx?ID=4055
Williamson, S. H. (2025). What was the U.S. GDP then? MeasuringWorth. http://www.measuringworth.org/usgdp/

Author: Kyle Novack

April 17, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

How Regulation Touches, But Doesn’t Explain, Food Prices

Kyle Novack — Tue, 14 Apr 2026 13:31:29 GMT

The last major suspect we can examine before questioning the measuring system itself is regulation. No one disputes that rules and compliance requirements add cost to doing business. The narrower question is whether those burdens are big enough, and structured in the right way, to absorb the massive cost savings that agricultural and supply‑chain productivity should have delivered. Regulation, please step onto the witness stand.

Recent work by the Organization for Economic Co-operation and Development (OECD) shows just how heavily businesses feel this load. In its Simplifying for Success survey, business organizations across member countries report that regulatory requirements and compliance now rank as their single most significant challenge. It was found to be ahead of challenges related to finding workers with the right skills, tax pressure, or geopolitical risk, and that this burden has been mounting over time. Yet when the OECD digs into the numbers, it finds that the share of employment devoted to regulatory tasks, compliance, reporting, and related functions, has indeed risen, but only modestly, in the economies it tracks (Andrews, Duran-Franch, Turban, & OECD Economics Department, 2025).

What the Data Say About Compliance Burdens

Approximate estimates from that work suggest that between 2012 and 2024, the share of jobs tied to regulatory tasks rose from about 3.7% to 3.9% in Europe (with data to 2023), from roughly 3.55% to 3.7% in Australia, and from around 2.85% to 3.2% in the United States. Those increases are meaningful at the firm level. They represent real people and hours being diverted into non‑production work, but when set against the scale of long‑run productivity growth in sectors like agriculture, they are relatively small. That contrast becomes clearer when we look at how economists measure productivity using Total Factor Productivity (TFP). Increases in compliance staff, reporting, and monitoring appear as higher measured labor and capital inputs in the denominator, rather than as unrecorded costs outside the system. (Andrews, Duran-Franch, Turban, & OECD Economics Department, 2025).

How Regulation Shows Up in TFP

To see why, it helps to be clear about how TFP is measured. Total Factor Productivity is calculated as total output divided by a quality-adjusted index of all inputs, including labor, capital, land, and materials. Regulations affect both sides of that ratio. Compliance requirements that add staff, monitoring equipment, or paperwork increase measured inputs, lowering TFP from the denominator side. Rules that restrict what can be produced from a given bundle of inputs lower the measured output, thereby reducing TFP from the numerator side. In both cases, any positive TFP growth we still observe is already net of regulatory drag, not evidence that regulation somehow “used up” the productivity gains (Wang, Nehring, Mosheim, & Njuki, 2024; USDA Economic Research Service, n.d.).

That is exactly what the historical record shows for U.S. agriculture. From 1948 to 2021, agricultural TFP still grew about 1.49% per year even as regulatory burdens increased, and farm output nearly tripled while measured inputs were flat or declined. Whatever costs regulation imposed are already baked into those TFP figures; the productivity dividend the Food Puzzle works with is a post‑regulation reality, not a pre‑regulation fantasy.

A Concrete Test: FSMA and Crop‑Sector TFP

A concrete example helps make this more tangible. The Food Safety Modernization Act (FSMA) of 2011 tightened standards to reduce the risk of foodborne illness, especially for produce growers, and clearly raised costs. Studies estimate that FSMA increased compliance costs by roughly 0.9% to 6–7% of annual sales for small and very small farms, and by about 0.3% to 0.9% for larger operations, numbers that are not trivial for individual businesses. Yet in the decade following FSMA’s implementation, crop‑sector TFP still grew at an average of around 1.2% per year as yields and overall output continued to rise. That rate is lower than the longer‑run average of about 1.5%, which is exactly what we would expect if regulation acts as a drag on measured productivity, yet it still falls far short of erasing the underlying gains (Bovay, Ferrier, & Zhen, 2018, & USDA Economic Research Service, 2025).

Why Regulation Cannot Resolve the Food Puzzle

This matters for the Food Puzzle because when we talk about the cost savings implied by productivity growth, those gains are at least partly net of regulatory burden. Agricultural TFP has continued to rise even as rules have tightened and compliance costs have increased, which suggests that, all else equal, correctly measured real food prices should show a clearer downward drift over time than we currently see in CPI‑adjusted data. The fact that CPI‑adjusted food prices mostly flatten or edge down only slightly is therefore not strong evidence that regulations fully absorbed the savings. Instead, it points to a gap in how we translate physical and cost realities into the price index. Put differently, regulations certainly raise costs for farms and food companies at any given moment, but many of those costs already appear in measured inputs, and TFP has nonetheless risen strongly. The long‑run productivity dividend traced throughout this series should be treated as a conservative, after‑regulation figure rather than a pre‑regulation ideal.

That is why, having carefully examined regulations, we must reject them as the missing piece of the Food Puzzle. They matter, but they do not bridge the gap between what the physical and cost data imply food should cost and what CPI says it does cost. Regulation, you are dismissed from the witness stand.

Where the Evidence Now Points

By this point in the Food Puzzle, we have put a long list of suspects on the stand.

On the farm side, we tested whether the savings were hiding in farm wages, capital costs, total production expenses, or higher farm profit margins, and found that none of these could absorb more than a fraction of the productivity gains.
In the supply chain, we examined packaging, transportation, labor, and the farm share of the food dollar, and found flat or falling real per‑capita costs and rising productivity, not an explosion of middleman markup.
We then turned to corporate behavior, showing that major food companies’ operating and gross margins, and their share of at‑home food spending, do not display the kind of sustained break you would expect if they had quietly captured the missing savings.
We looked at market dynamics and consumer behavior and found that the combination of high functional utility, recurring purchases, and easy substitution makes it extremely hard to keep food prices far above competitive levels for long.
We examined external shocks: pandemics, wars, droughts, energy spikes, and found that while they cause sharp but temporary price surges, per capita availability of key crops and calories has continued to rise, with prices reverting after the shocks rather than remaining permanently high.
Finally, we called regulation to the stand and saw that, while compliance burdens are real and do show up in measured inputs, the available evidence suggests they are too small to account for the full dividend implied by long‑run TFP growth.

Taken together, the evidence from farms, supply chains, corporate margins, consumer behavior, shocks, and regulations all points in the same direction: the physical and economic reality of food has become far cheaper to produce and distribute than our official price indexes are willing to admit. If the goods are simpler, the data cleaner, the productivity gains larger, and the competitive pressures stronger than in almost any other sector, and we still cannot see the full savings in CPI‑adjusted prices, then the problem is not the food system. It is the yardstick. With all the obvious suspects now cleared, it is time to call the ambiguous suspect, the measurement instrument itself, to the witness stand. This is required to determine how a systematic blind spot in CPI has quietly distorted our view of inflation, wages, and the dollar’s true value for more than a century. A task we turn to in the next article when we examine the measurement system itself.

Sources

Andrews, D, Duran-Franch, J, Turban, S, & OECD Economics Department. (December 16, 2025) Time for a Regulatory Reset? Clearing the path for productivity and Dynamism. Retrieved January 4, 2026, from https://oecdecoscope.blog/2025/12/16/time-for-a-regulatory-reset-clearing-the-path-for-productivity-and-dynamism/
Bovay, J., Ferrier, P., & Zhen, C. (2018). Estimated costs for fruit and vegetable producers to comply with the Food Safety Modernization Act’s Produce Rule (EIB-195). U.S. Department of Agriculture, Economic Research Service.
Wang, S. L., Nehring, R., Mosheim, R., & Njuki, E. (2024). Measurement of output, inputs, and total factor productivity in U.S. agricultural productivity accounts (Report No. TB-1966). U.S. Department of Agriculture, Economic Research Service.
U.S. Department of Agriculture, Economic Research Service. (n.d.). Agricultural productivity in the U.S.: Summary of recent findings. Retrieved March 14, 2026, from https://www.ers.usda.gov/data-products/agricultural-productivity-in-the-united-states/summary-of-recent-findings

Scarcity Was Never the Problem

Kyle Novack — Fri, 10 Apr 2026 13:31:58 GMT

At this point in the Food Puzzle, we have already called agriculture, the supply chain, and corporate market power to the witness stand to see whether they swallowed the cost savings that should come with massive productivity gains. Each has been dismissed. This is not hyperbole when it comes to productivity. It is what ought to happen in a well‑functioning economy. When inputs stay broadly stable while output surges, those input costs are spread over many more units, pushing the average cost per unit down for consumers.

Now it is time to call a new witness: external shocks and offsets. Droughts, wars, pandemics, and energy price spikes can clearly push prices up in the short run by creating temporary imbalances. Supply contraction causes uncertainty while demand stays the same or even rises. The question is whether a long enough string of such events could plausibly keep overall food prices permanently elevated, erasing the savings that productivity says should be there.

How to Read the Data in This Section

All numbers and charts in this part of the Food Puzzle draw on standard public data sets, including USDA crop and calorie‑availability series, BLS inflation indexes, international energy statistics from the U.S. Energy Information Administration, and global production data from Our World in Data and related sources. To avoid clutter, I keep in‑text citations brief; each figure includes a short note on how it was constructed, and a dedicated “Methods and data sources” section at the end lists every underlying series and transformation (such as inflation adjustments or per‑capita conversions) so readers can audit or replicate the calculations. Unless otherwise noted, all dollar amounts are expressed in CPI‑U‑adjusted 2024 dollars, while physical quantities (such as tons or pounds of crops and calories per person) and ratios (such as per‑capita production) are reported in their natural units.

What the Supply Data Actually Show

The broader data argue strongly against that explanation. The USDA and related agencies maintain detailed, long‑running accounts of the U.S. food system, including daily per‑capita calorie availability after adjusting for food loss. From 1970 to 2010, that measure has risen by about 22%. This is crucial because the series is already after harvest and accounting for food loss.

If widespread, persistent shortages were truly eroding the productivity dividend, we would expect total available calories per person to fall, or at least become unstable, over time. Instead, total availability and availability by major food group remain remarkably steady, even through large regional weather shocks. For example, the 1977 drought hit much of the western United States, including California.

History tells a similar story in more extreme episodes. During the Great Depression and the Dust Bowl, conditions on the ground were brutal for farmers in the hardest‑hit regions: crops failed, soil blew away, and families lost both land and livelihoods. Yet when we step back to the national data, per‑capita crop production rises from its 1929 lows through 1939, when the drought finally eased (USDA, 2022).

In other words, local hardship and visible distress did not translate into a collapse in national food supply. As transportation, storage, and logistics have improved, the geography of production has diversified. Over time, the U.S. food system has become increasingly resilient, able to absorb most natural disasters without threatening overall availability, even when certain regions or specific crops are hit hard.

How the Food System Has a Built‑In Backstop

There is also a quiet safety valve in the system that rarely shows up in standard discussions: food waste. Estimates suggest that up to 40% of food in the United States is ultimately discarded, with an implied value of around $218 billion per year. This waste cannot explain why prices have not fallen more, for two reasons. First, much of it occurs after purchase, once the food has already cleared the supply chain. Second, historical spoilage during production and distribution was higher than it is today. This means technical progress has reduced waste upstream in the supply chain, even as consumer‑side waste has increased (RTS, n.d.).

What it does tell us, however, is that there is a built‑in backstop against genuine, long‑run scarcity. In a true supply crunch where shortages are felt directly, households have a powerful incentive to cut back on what they throw away. This would stretch available supply without needing a proportional jump in production, making it even harder for external shocks to sustain chronically high prices on their own.

When Psychology Manufactures Shortages

If availability has been broadly stable or rising, why do prices still spike in certain periods? One answer is that many of the most visible “supply shocks” are really episodes of fear and uncertainty rather than large, lasting collapses in physical supply. The early months of COVID‑19 are a vivid example. People walked into stores and saw empty shelves. Your social media feed filled up with photos of stripped‑bare aisles, and the now‑famous toilet‑paper panic became a kind of national joke with a real edge of anxiety behind it.

Faced with a frightening unknown, households and companies both stopped behaving like calm optimizers. Shoppers grabbed whatever they could, and firms hesitated to run down inventories because they did not know how long the disruption would last. As Jonathan Haidt notes in The Happiness Hypothesis (2006), under stress, we lean heavily on emotion and intuition. The resulting rush to hoard toilet paper, despite no underlying shortage in production capacity, captured that shift perfectly.

Industry estimates from IndexBox show that the U.S. market for toilet paper, napkins, towels, and tissue was about 23 billion dollars in 2020, up only 2.3% from the year before, compared with an average growth rate of 4.4% from 2007 to 2020. In other words, sales grew less than usual in the year of peak panic despite the viral images of scarcity.

Why Shocks Don’t Stay in Charge

The 2022 invasion of Ukraine offers another test case that many people felt directly at the gas pump. Almost overnight, headlines warned of threats to both energy and key agricultural exports. This was most notable as U.S. drivers watched gas prices at the pump climb week after week. Russia was the world’s largest wheat exporter at the time, with Ukraine also a major player. For crude oil, Russia was the third‑largest producer globally, supplying about 11.2% of global output according to the U.S. Energy Information Administration. Sanctions on Russian oil tightened markets and helped push the national average price of regular unleaded gasoline to about $5.02 per gallon in June 2022, compared with an average of $2.91 per gallon in 2021, a roughly 72.5% increase from the annual average to the peak (U.S. Energy Information Administration, 2026 & U.S. Energy Information Administration, n.d.)

Even here, the story is less about a permanent collapse in supply than about a violent adjustment period. Russian oil did not vanish; it was rerouted to buyers such as China and India, and once new trade patterns settled in, prices eased back. In CPI-adjusted dollars, U.S. gasoline prices have since drifted toward levels consistent with mid-2000s real prices, signaling that the war-related spike was sharp but temporary, not a new, permanent plateau.

Wheat and other crops followed a similar arc globally. Wheat prices surged under the combined pressure of the Ukraine invasion and COVID‑era logistics snarls, with average annual prices rising about 28% from 2021 to 2022, yet global wheat production increased as Ukraine’s output fell, Russia’s rose, and total world tonnage moved higher rather than collapsing (Macrotrends LLC, n.d.).

By 2025, after adjusting for overall inflation, wheat prices had retreated to levels in line with the lowest points of the prior quarter-century. Famine conditions never materialized in the United States or in any location not actively engaged in conflict. From the vantage point of a shopper standing in a bread aisle, the episode felt like “supply shock inflation.” In the data, it looks like a temporary price bulge on top of a system whose underlying capacity continued to expand.

When we zoom out even further, the resilience of the system becomes even clearer. Looking at eight key global crops from 1961 to 2023, barley, maize (corn), potatoes, rice, soybeans, sugarcane, sunflower, and wheat, combined per‑capita output follows a steady upward trend rather than a pattern of repeated shortfalls. There are no long‑lasting declines large enough to support the idea that recurring supply shocks have kept world food supplies chronically tight. If sustained, multi-decade high prices were driven by physical scarcity alone, we would expect to see persistent dents in this kind of per-person production data. Instead, per-person production has increased.

Scarcity Was Never the Problem

Putting these pieces together, we reach a clear verdict. External shocks and offsets can absolutely cause sharp, short‑term surges in food and energy prices and inflict real hardship on specific regions, crops, and producers. But the national and global production data, the behavior of per‑capita availability, and the way prices eventually normalize all point in the same direction. These shocks do not hold prices high enough, for long enough, to consume the missing savings that productivity gains say should have shown up. As a long‑run explanation for why food prices have not fallen more, “supply shocks did it” fails under scrutiny and joins the supply‑chain and corporate‑greed stories as a third suspect to be dismissed from the witness stand in the Food Puzzle trial.

Before we move on from standard suspects, there is one last suspect that often lurks in the background of conversations about food prices: regulation. From food‑safety rules and environmental standards to labor laws and labeling requirements, it is easy to imagine a growing thicket of mandates quietly layering costs onto every meal and soaking up productivity gains. In the next section of the Food Puzzle, we will put that story on the witness stand and ask a narrower question: not whether regulation is costly, it clearly is, but whether those costs are large and persistent enough to explain why food prices have not fallen nearly as far as the physical world says they should.

Methods and Sources

Crop production during the Great Depression and Dust Bowl

I use historical production data from the U.S. Department of Agriculture, National Agricultural Statistics Service (NASS), as reported in Crop Production 2022 Summary and its underlying tables. The analysis covers ten major U.S. field crops—barley, cotton, maize/corn, oats, rice, soybeans, sugar beets, sugarcane, tobacco, and wheat. For each crop, I take total annual production for 1929 and 1939, convert quantities to a common unit where needed, and sum across crops to obtain aggregate output by year. The resulting comparison shows that total crop production increased between 1929 and 1939, despite the Great Depression and Dust Bowl, supporting the claim that national output rose even as regional conditions deteriorated.

Inflation‑adjusted U.S. gasoline prices

To compare recent gasoline spikes with earlier periods, I use annual average U.S. Regular Conventional Gas Price data from the U.S. Energy Information Administration, series GASREGCOVW, accessed via FRED, Federal Reserve Bank of St. Louis, for 1991–2025. I convert these nominal annual prices into constant 2024 dollars by deflating them with the CPI‑U (All Items) index from BLS, using 2024 as the base year. The resulting real price series shows the COVID‑19 and Russia‑Ukraine spikes as temporary surges, with inflation‑adjusted gasoline prices later receding toward levels consistent with the mid‑2000s rather than remaining permanently elevated.

Wheat production in Ukraine, Russia, and globally

To assess how the Russia–Ukraine war affected wheat supply, I use country‑level and global wheat production data from Our World in Data’s Agricultural production database (Ritchie, Rosado, & Roser, 2023). I extract annual wheat production (in tonnes) for Ukraine, Russia, and the world for 2021 and 2022, then compute percentage changes over this interval for each series. These calculations show that Ukraine’s wheat output fell sharply while Russia’s rose and global wheat production increased overall, supporting the claim that world supply did not collapse even though regional production was severely disrupted.

Global per‑capita production of staple crops

Annual production data for eight crops—barley, maize (corn), potatoes, rice, soybeans, sugarcane, sunflower, and wheat—come from Our World in Data’s Agricultural production database (Ritchie, Rosado, & Roser, 2023). Production data in tonnes were summed across the eight crops, converted to pounds using the factor 1 metric tonne = 2,204.62 pounds, and then divided by the annual world population from Worldometer’s World population by year series, which elaborates United Nations estimates, to obtain per‑capita values.

Sources

Haidt, J. (2006). The happiness hypothesis: Finding modern truth in ancient wisdom. New York, NY: Basic Books.
IndexBox. (2025, January 10). Market for toilet, towel, and tissue paper in the United States. Retrieved January 3, 2026, from https://app.indexbox.io/report/4803h9/840/
Lee, J. H., Duster, M., Roberts, T., & Devinsky, O. (2022). United States dietary trends since 1800: Lack of association between saturated fatty acid consumption and non‑communicable diseases. Frontiers in Nutrition, 8, 748847. https://doi.org/10.3389/fnut.2021.748847
Macrotrends LLC. (n.d.). Wheat prices – 60 year historical chart. Macrotrends. Retrieved April 9, 2026, from https://www.macrotrends.net/2534/wheat-prices-historical-chart-data
Richter, F. (2022, March 9). Ukraine crisis likely to push up wheat prices. Statista. Retrieved January 3, 2026, from https://www.statista.com/chart/26928/largest-wheat-exporters/
Ritchie, H., Rosado, P., & Roser, M. (2023). Agricultural production. Our World in Data. https://ourworldindata.org/agricultural-production
RTS. (n.d.). Food waste in America in 2026. Retrieved January 2, 2026, from https://www.rts.com/resources/guides/food-waste-america/
U.S. Bureau of Labor Statistics. (2025). Consumer Price Index for All Urban Consumers (CPI‑U): All items in U.S. city average [Data series CPIAUCSL]. Retrieved from https://www.bls.gov/cpi/
U.S. Energy Information Administration. (n.d.). Petroleum and other liquids: Annual refined petroleum products consumption, world. International data browser. Retrieved April 9, 2026, from https://www.eia.gov/international/data/world/petroleum-and-other-liquids/annual-refined-petroleum-products-consumption
U.S. Energy Information Administration. (n.d.). U.S. regular conventional gasoline price [GASREGCOVW]. In FRED, Federal Reserve Bank of St. Louis. Retrieved January 3, 2026, from https://fred.stlouisfed.org/series/GASREGCOVW
United States Department of Agriculture, National Agricultural Statistics Service. (2022). Crop production 2022 summary (ISSN 1936‑3737). https://www.nass.usda.gov/Publications/Todays_Reports/reports/croptr22.pdf
Worldometer. (n.d.). World population by year. Elaboration of United Nations data. Retrieved January 4, 2026, from https://www.worldometers.info/world-population/world-population-by-year/

Author: Kyle Novack

April 10, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

Why Markets Won’t Let Food Stay Expensive

Kyle Novack — Tue, 07 Apr 2026 13:30:47 GMT

To understand why food corporations struggle to keep prices elevated, we need more than corporate financials. We need to look at how the food market works when ordinary people make choices. The key idea is that food has two different kinds of utility: its functional utility (you need it to live) and its priced utility (what you are willing and able to pay per unit, given your budget and alternatives). The tension between those two layers is what makes food such a difficult sector to sustain high prices.

Functionally, food is at the top of the hierarchy. Without it, nothing else in the economy matters. But because food must be purchased repeatedly, often multiple times per week, the system cannot tolerate very high prices per unit for long without causing severe strain on households. Markets and politics push hard in the opposite direction: they work to suppress the priced utility of food, keeping each unit as cheap as possible so that families can survive the enormous lifetime cost of eating. Well still have money left over for housing, services, and everything else that makes a modern economy possible. Food does not cost less than luxuries because it matters less; it costs less per unit precisely because it matters so much and shows up constantly.

Once we see that distinction, consumer behavior in food markets becomes easier to interpret. When people stand in front of a shelf, they are not choosing between “food” and “no food.” They are choosing among many versions of “functionally acceptable” food, and their attention shifts to price and perceived value. A study by 84.51°, a Kroger‑owned analytics firm, found that about 60% of respondents said their purchase decisions were driven by whether a brand was a good value for the money. Only around 5% said they would buy one brand and nothing else. In food, “brand loyalty” usually means “my first choice if the price is close,” not “a product I will defend at any price” (84.51°, 2023).

That is why private labels and substitutes matter so much for pricing power. As a Sevendots analysis notes, private‑label products have been gaining share in more commoditized food categories where perceived differentiation is low, and brand equity is weak or eroding. Such as canned goods, everyday bakery items, and snacks. If a bag of branded tortilla chips costs one dollar more than the store brand, a growing share of shoppers simply shifts to the cheaper bag. Functionally, both chips serve the same purpose: they satisfy hunger, fill the bowl, and pair well with salsa. Once functional utility is “good enough,” the decision pivots to priced utility—what combination of taste and price feels like the best deal (Bielli & Mcasllister, 2025).

Experimental work on willingness to pay for different food options shows the same two‑layer pattern. In one set of studies on plant‑based protein alternatives, researchers offered respondents several sandwiches (beef burger, bacon beef burger, chicken sandwich, and Beyond Meat burger in one group; beef burger, bacon beef burger, chicken sandwich, and a chicken wrap in another) and varied prices across nine scenarios. People did not abandon their preferred options immediately because those options still carried a higher perceived functional utility (familiar taste, satisfaction, habit). But when the chicken wrap was made nearly $3 cheaper than the others, or the Beyond Meat burger about $ 1.10 cheaper, measurable shifts in choices appeared. Price did not override the function immediately; it did so only after the value gap widened sufficiently (Tonsor, Lusk, Schroeder, 2022).

Another study they conducted focused on ground protein choices: two beef options, two plant-based options, a chicken option, and “other” category, while varying only the price per pound at the grocery store. At the intermediate price level, Tyson’s chicken breast was chosen about 45% of the time because it offered a strong combination of functional utility (perceived health, familiarity) and price utility (acceptable cost). When the price of Beyond Beef alone fell by one dollar, its projected share almost doubled, from 3.63% to 7.04%, mainly at the expense of the Impossible Burger. When store‑brand 80% lean ground beef became cheaper, its share rose while more expensive beef options lost ground. The functional utility of chicken and branded beef did not change; what changed was their price position relative to the alternatives (Tonsor, Lusk, Schroeder, 2022).

The most striking finding was that lowering the price of a perceived “healthier” plant‑based option not only increased its own share but also nudged more consumers toward Laura’s Lean Natural Ground Beef, even though Laura’s Lean sticker price stayed the same. A shift in one product’s priced utility altered the value landscape for another product whose price did not move. And when both ground beef options dropped by one dollar, the projected selection rate for chicken fell from 45.24% to 37.17%. Almost eight percentage points of demand moved away from chicken purely because beef’s priced utility improved, even though chicken’s functional utility was unchanged (Tonsor, Lusk, Schroeder, 2022).

These results put pressure back on the corporations in an important way. In normal times, the functional utility of food is non‑negotiable. People must eat, and many products easily clear the “good enough” bar on taste and quality. That makes the market unusually sensitive to small changes in priced utility, relative price, and perceived value, because once the function is satisfied, price differences become the main lever. A seemingly modest price increase can cause a brand to lose volume to private labels or rival products that now appear to offer better value. Even holding your own price steady while competitors cut theirs can hurt your sales, because consumers evaluate you in a relative, not absolute, frame.

For food companies, that is a difficult environment in which to keep prices high, just because they would like to. Every unit they sell carries real costs for labor, ingredients, packaging, and logistics, but every attempted price increase collides with consumers who have both the incentive and the ability to trade down to cheaper yet functionally acceptable alternatives. In this sector, high functional utility coexists with frequent repetition and abundant substitutes. The result is strong downward pressure on utility prices over time.

Any story that blames persistently high food prices on corporate power alone must reckon with this two‑layer reality. The functional importance of food does not give corporations unlimited pricing power. It invites the opposite: a system in which competition, politics, and consumer behavior work together to keep unit prices as low as they can be sustained. That is why, when we are looking for the missing savings from a century of productivity gains, we cannot simply point to food companies and assume they have captured them. The structure of food demand itself makes that very hard to do. Market Power and corporate capture you are dismissed from the witness. Next we are calling External Shocks and Offsets to the witness stand to see if natural disaster and supply constraints can consume the cost savings.

Sources

Bielle, A., & McAllister, C. (July 24, 2024). The return of Private labels: Why retailer brands are reclaiming growth in the global CPG market. Retrieved January 2, 2026, from https://sevendots.com/the-return-of-private-labels-why-retailer-brands-are-reclaiming-growth-in-the-global-cpg-market/

84.51°. (2023). The loyalty shift decoded: Key insights for winning shopper devotion in an uncertain economy [White paper]. https://www.8451.com/loyalty-shift-decoded/

Tonsor, G. T., Lusk, J. L., & Schroeder, T. C. (2022). Market potential of new plant‐based protein alternatives: Insights from four US consumer experiments. Appl Econ Perspect Policy, 2022: 1-18. https://doi.org/10.1002/aepp.13253

Author: Kyle Novack

April 07, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

Big Food and Greedflation: What the Margins Actually Show

Kyle Novack — Fri, 03 Apr 2026 13:31:58 GMT

In previous sections, we walked through the post-farm supply chain and found little evidence that packaging, transportation, or grocery-sector labor quietly swallowed the farm-productivity dividend. Real per capita costs in those categories tended to be flat or falling after adjusting for CPI to create 2024 dollars, and productivity outside the farm gate generally improved rather than deteriorated. If the savings did not disappear into trucks, boxes, or checkout lanes, the next logical turn is to the largest remaining suspect: market power and corporate capture

It is now time for the corporations themselves to take the witness stand. The charge is straightforward: large food companies use their size and market power to squeeze suppliers on input prices while keeping shelf prices elevated for consumers. The gap between falling internal costs and rising final prices translates into permanently higher profits for executives and investors. This is the narrative commonly called “greedflation.”

That charge has real emotional force. Shoppers see grocery bills rising and read about consolidation in meatpacking, soft drinks, snacks, and packaged goods. It is natural to suspect that concentrated corporate power is quietly capturing the productivity dividend that should have reached households. But an emotional narrative is not the same as evidence. We need to examine the financial records directly (Ryan, 2024).

How to Read the Data in This Section

All numbers and charts in this part of the Food Puzzle draw on standard public data sets, including USDA Food Expenditure and Food Dollar tables, BLS labor and wage series, official CPI indexes, and the published annual reports and Form 10‑K filings of major food companies. To avoid clutter, I keep in‑text citations brief; each figure has a short note on how it was constructed, and a detailed “Methods and data sources” section at the end lists every underlying series so that readers can audit or replicate the calculations. Unless otherwise noted, all dollar amounts in this section are adjusted for inflation using the CPI‑U and expressed in constant 2024 dollars, while ratios such as profit margins and revenue shares are calculated in nominal terms.

Tyson, the Meat Industry Giant

Meatpacking is often presented as the prime example. Commentators note that a handful of meatpacking giants control much of the fed‑cattle market and argue that they can manipulate prices, labeling, and supply in ways that hurt both ranchers and consumers. They conclude that these pressures cause beef prices to increase, including double-digit rises in 2025, as apparent proof that concentrated corporate power is driving inflation. On the surface, this appears to be a strong case for treating large food companies as the main suspects (Rechenberg, 2025).

To see whether that intuition holds up under scrutiny of the data, we start with one of the most visible firms in the sector: Tyson Foods. If a company with substantial scale is quietly keeping the productivity dividend for itself, we would expect to see the evidence in its operating margins, the percentage of revenue left after paying for cost of goods sold (COGS) and operating expenses. Over the 1997–2025 period, Tyson’s operating margins averaged in the mid‑single digits, with a clear but temporary spike around 2021–2022 when inflation was elevated across the economy. After that spike, margins fall back toward the lower end of their historical range, and the company even records an operating loss in 2023.

Those patterns matter because they sit alongside persistent food inflation. According to the CPI for food in U.S. city averages, food prices rose by almost 32% between January 2020 and December 2025, and many households feel the increase is larger in their own budgets. If Tyson were using that environment to permanently widen the gap between prices and costs, we would expect to see a clear step‑up in profitability that stays in place. Instead, we see a temporary surge followed by a reversal toward past norms. Tyson does not appear to be a firm that has locked in a new era of structurally higher margins at the expense of consumers (U.S. Bureau of Labor Statistics, 2026).

Examining Financials for 10 Large Food Companies

To check whether this is an isolated case or part of a broader pattern, we widen the lens to ten major food processing and producing corporations: The Coca‑Cola Company, PepsiCo, Nestlé, Mondelez International, General Mills, Conagra Brands, Campbell Soup Company, Hormel Foods, Tyson Foods, and the legacy Kellogg Company before its split into Kellanova and WK Kellogg Co. These companies feel omnipresent in the supermarket. Their brands fill shelves, coolers, and freezers, and it is easy to assume they collectively dominate the at‑home food market.

The revenue data tell a more nuanced story. In 2022, total at‑home food expenditure in the United States was roughly 1.02 trillion dollars, while the combined revenue of these ten corporations was about 375 billion dollars, around 37% of at‑home food spending. In the late 2000s, the same group accounted for just over half of at‑home food spending. Over time, their combined share has fallen from above 50% to below 40%, even as their products have remained highly visible. Some of this decline reflects spin-offs and corporate restructuring, but the direction of the trend is still informative: the group’s share of the at-home-food expense pie has been shrinking rather than expanding.

When we examine operating margins across these ten companies from the early 1990s through the mid-2020s, we see a pattern like Tyson’s. Each firm tends to operate within a stable profitability band. Margins rise and fall with commodity cycles, input cost shocks, and demand changes, but there is no broad, sustained upward break that would signal a long-running campaign to capture the missing savings from a century of productivity gains. If anything, the data look like normal business volatility, not the fingerprint of a structural shift in bargaining power strong enough to absorb a 50% system‑wide cost reduction.

One common counterargument is that even if operating margins look stable, executives might be absorbing the cost savings through higher salaries, bonuses, and stock‑based compensation, so the gains never appear in operating income. To probe that concern, we can look at gross margins instead, the share of revenue left after paying for the cost of goods sold, before overhead and administrative expenses. Gross margin gives us a more direct sense of whether companies are paying less for inputs than they receive from customers.

Across this group of ten food companies, gross margins have generally not marched higher, and for many firms they have drifted down. That means it has become more expensive, not cheaper, to acquire the raw materials and intermediate goods needed to make their products, relative to the prices they can charge. If these companies were quietly harvesting large cost savings from lower production costs while using their market position to keep prices high, we would expect their gross margins to widen over time. Instead, the evidence points to input-cost pressures rather than to a hidden surplus being skimmed away.

Recapping the Case on Big Food Companies

Taken together, the financial records of the large food companies we have examined do not support the idea that they have captured a huge, unshared productivity dividend. We do not see operating margins breaking into a new, permanently higher range. We do not see the group’s share of at‑home food spending rising relentlessly. We do not see broad, sustained increases in gross margins that would signal a large spread between what companies pay suppliers and what they charge consumers.

None of this means market power and corporate behavior are irrelevant. Concentration can still harm specific groups, such as ranchers facing a small set of dominant buyers, and some corporate practices may well deserve regulatory scrutiny. What the evidence does show is that, in aggregate, corporate margins in the food sector do not explain why official price measures fail to reflect the full impact of long‑run productivity gains. The “corporate greed did it” story feels satisfying, but it does not resolve the central puzzle in this chapter.

In that sense, the corporate sector is a suspect we had to examine carefully. Once its financial statements are on the record, however, the case for treating it as the main culprit behind the missing food savings looks weak. If food is still more expensive than a century of innovation suggests it should be, the explanation lies deeper than the profit margins of brand‑name companies.

In the next part of this chapter, we turn to the dynamics of markets and consumers themselves and ask a different question: why do competitive pressure and household behavior make it extremely difficult for food corporations to keep prices elevated for long? While you are waiting for that section, it is worth reading “Why Food Has to Be Cheap for Everything Else to Exist,” which explains why food functions as a constant survival subscription and why keeping it affordable is essential for the rest of the economy to thrive.

Methods and Sources

Calculation for Food Inflation

Food inflation over 2020–2025 is calculated from the Consumer Price Index for All Urban Consumers: Food in the U.S. City Average (series CPIUFDSL), using the percentage change in the seasonally adjusted index from January 2020 to December 2025, as reported by FRED and the U.S. Bureau of Labor Statistics.

Gross and Operating Margin Calculations for Major Food Companies

All company-level revenue, cost of goods sold (COGS), and operating income figures used in the margin analysis are drawn directly from the annual reports and Form 10‑K filings of ten large food producers and processors: Campbell Soup Company, The Coca‑Cola Company, Conagra Brands Inc., General Mills, Inc., Hormel Foods Corporation, Kellogg Company (pre‑split legacy entity), Mondelez International, Nestlé S.A., PepsiCo, Inc., and Tyson Foods, Inc. For each firm and year, I record total revenue, COGS, and operating income as reported in the consolidated income statement, without adjusting for inflation, since margins are ratios rather than dollar levels.

Gross margin is calculated as (Revenue − COGS) ÷ Revenue × 100, using each company’s reported COGS (or equivalent “cost of sales” line where labels differ). Operating margin is calculated as Operating Income ÷ Revenue × 100, where operating income is the subtotal reported after production costs and operating expenses (selling, general and administrative, marketing, and R&D) but before interest, taxes, and one‑time items. In cases where firms undergo restructurings, spin‑offs, or changes in segment reporting (for example, the legacy Kellogg Company before its split, or Tyson corporate reorganization years), I retain the published totals for the consolidated entity in that year and note major structural breaks in the graph captions and endnotes rather than adjusting or splicing series.

Revenue Share of Major Food Companies in At-Home Food Spending

For each year from 2004–2022, I sum the nominal annual revenues of ten large publicly traded food and beverage companies (Campbell Soup Co., The Coca‑Cola Company, Conagra Brands Inc., General Mills, Inc., Hormel Foods Corp., Kellogg Company pre‑split, Mondelez International, Nestlé S.A., PepsiCo Inc., and Tyson Foods) using figures reported in their annual reports and Form 10‑K filings. I then divide this combined revenue total by nominal U.S. at-home food expenditures for the same year, as reported in the USDA Economic Research Service Food Expenditure Series, to obtain each year’s percentage share. Because both the numerator and denominator are in current dollars, no inflation adjustment is applied for this ratio; spin‑offs and restructurings (such as Conagra’s Lamb Weston spin‑off or Kellogg’s later split) are treated as structural breaks and noted in captions and endnotes rather than retroactively adjusted.

Sources

Campbell Soup Company. (1996–2025). Annual reports and financial publications. AnnualReports.com. Retrieved December 31, 2025, from https://www.annualreports.com/Company/campbell-soup-company
Coca-Cola Company. (1993–2024). Annual reports and Form 10-K filings. AnnualReports.com. Retrieved December 31, 2025, from https://www.annualreports.com/Company/coca-cola
Conagra Brands, Inc. (and predecessor Conagra Foods, Inc.). (2000–2025). Annual reports and Form 10-K filings. AnnualReports.com. Retrieved December 31, 2025, from https://www.annualreports.com/Company/conagra-brands-inc
Consumer Price Index for All Urban Consumers: Food in U.S. City Average [CPIUFDSL], retrieved from FRED, Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/CPIUFDSL, April 2, 2026.
General Mills, Inc. (2002–2024). Annual reports and financial publications. AnnualReports.com. Retrieved December 31, 2025, from https://www.annualreports.com/Company/general-mills-inc
Hormel Foods Corporation. (1996–2025). Annual reports and financial publications. AnnualReports.com. Retrieved December 31, 2025, from https://www.annualreports.com/Company/hormel-foods-corp
Mondelez International, Inc. (and predecessor Kraft Foods Inc.). (1997–2024). Annual Reports and Form 10-K Filings. Investor Relations Publications. Retrieved December 31, 2025, from https://www.mondelezinternational.com/investors/financials/annual-reports/. (Note: The 2001 Annual Report includes comparative financial data back to 1997; earlier years are from legacy Kraft Foods filings accessible via SEC EDGAR.)
Nestlé S.A. (2000–2024). Annual reports and financial publications. Nestlé Investor Relations. Retrieved December 31, 2025, from https://www.nestle.com/investors/publications
PepsiCo, Inc. (1998–2024). Form 10-K: Annual Reports for the fiscal years ended December 31, 1998 through 2024. U.S. Securities and Exchange Commission (EDGAR, CIK 0000077476). Retrieved December 31, 2025, from https://www.sec.gov/edgar/browse/?CIK=77476
Ryan, K. (October 5, 2024). This Infographic Shows How Only 10 Companies Own All The World’s Food Brands. Retrieved from https://www.good.is/this-infographic-shows-how-only-10-companies-own-all-the-worlds-food-brands/
Rechenberg, R. (November 3, 2025). Beef prices: Blame the Packers, Not American Ranchers. Retrieved January 1, 2026, from https://prosperousamerica.org/beef-prices-blame-the-packers-not-americas-ranchers/
Total At-Home Food Expenditures: U.S. Department of Agriculture, Economic Research Service. (2025). Nominal Food and Alcohol Expenditures, with Taxes and Tips, for All Purchasers, by Outlet Type [Excel spreadsheet]. Food Expenditure Series. Retrieved December 31, 2025, from https://www.ers.usda.gov/data-products/food-expenditure-series/ (data extracted from the “Nominal expenditures” sheet for food-at-home categories).
Tyson Foods, Inc. (1997–2025). Annual Reports and Form 10-K Filings. Investor Relations Publications. Retrieved December 31, 2025, from https://ir.tyson.com/reports/annual-reports/default.aspx.
U.S. Securities and Exchange Commission. Kellogg Company. (2001–2022). Form 10-K: Annual Reports for the fiscal years ended December 31, 2001 through 2022. Archived via EDGAR (CIK 0000055067). Retrieved December 31, 2025, from https://www.sec.gov/edgar/browse/?CIK=55067. Operating income and revenue figures extracted from the consolidated income statements in each annual filing. 2003-2022.

Author: Kyle Novack

April 3, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

The Supply Chain Heist Part 2

Kyle Novack — Tue, 31 Mar 2026 13:31:24 GMT

In the last section, we put the modern food supply chain on the witness stand to test a popular story: that all the savings from America’s farm productivity boom were quietly absorbed between the farm gate and the checkout lane. Using USDA Food Dollar and marketing-bill data, along with BLS labor and wage series adjusted into constant 2024 dollars. We tracked the per-capita costs of key post-farm activities, such as packaging, transportation, and labor. Instead of exploding upward, packaging costs trended down, and transportation stayed roughly flat from 1967-2023, even after a 1997 data revision that should have pushed measured transportation costs higher.

Labor, the most intuitive suspect, also failed to explain away the missing savings once we measured it correctly. While total post‑farm employment in food manufacturing, grocery wholesaling, and food and beverage retailing grew, each worker now serves more people than in 1958. On top of this, real labor costs per person barely rose over seven decades, roughly equivalent to a single bad year of inflation. Real hourly pay in these sectors also rose modestly, but those higher wages were largely offset by the number of people served by workers, leaving no sign of a structural labor-cost explosion large enough to swallow an agricultural revolution.

Finally, we looked beyond costs to productivity across the rest of the chain. Total Factor Productivity in wholesale and retail trade rose strongly from 1987–2023, with more modest but still positive gains in food manufacturing, meaning these “middlemen” were generally becoming more efficient, not less. When we weight each sector’s productivity-driven cost changes by its share of the food dollar, the result is an overall cost reduction on the order of 19%, which makes it hard to argue that the supply chain simply skimmed off the farm’s savings—and points instead to a deeper problem in how our price indexes are measuring (or masking) those gains.

How to Read the Data in This Section

All numbers and charts in this part of the Food Puzzle come from standard public data sets, including USDA marketing‑bill and Food Dollar tables, BLS labor and wage series, official CPI indexes, and documented population estimates. To avoid clutter, I keep in‑text citations brief; each figure has a short note on how it was constructed, and a detailed “Methods and data sources” section at the end lists every underlying series so that readers can audit or replicate the calculations. Unless otherwise noted, all dollar amounts in this section are adjusted for inflation using the CPI‑U and expressed in constant 2024 dollars, so differences over time reflect real changes rather than general price inflation.

The Silent Reality of the Farm Share

CPI has one more key witness to call in defense of the standard story of the supply chain: the farm share of the food dollar. On the surface, the chart looks damning. In the early 1950s, farms accounted for close to 40 percent of what consumers spent on food; by 2023, that share hovers near 11 percent. This is widely taken as proof that the supply chain gradually swallowed the savings from the agricultural productivity boom. If the farmer’s slice keeps shrinking, the thinking goes, the middlemen must be eating the difference.

But that conclusion follows only if we understand what “farm share” measures and how the underlying data have changed over time. The legacy USDA Marketing Bill series is not directly comparable to the modern Food Dollar series. They define the pie differently, and the modern version also includes all forms of food spending, such as restaurants and other away-from-home options, which have always had very low farm shares. As more of our food budget flows through those channels, the overall farm share is mechanically pulled down, even if nothing “bad” is happening in the grocery supply chain.

That is why the Food Dollar food‑at‑home farm share, about 24.3 percent in 2023, gives us a much better apples‑to‑apples lens for this puzzle than the all‑food number, and why we should be cautious about treating the long‑run drop from older Marketing Bill estimates to today’s aggregate Food Dollar as proof that the supply chain simply ate the farm’s gains.

What, exactly, is the farm share telling us? At its core, it is a story about how the final dollar paid by the consumer gets split among all the stages that add value along the way. If consumers shift toward more restaurant meals and more heavily processed products, it is expected, not suspicious, that the non‑farm stages account for a larger share of the final price. The question is not “Did the farm share fall?” but “Does the fall necessarily mean the supply chain became bloated and ate all the savings?” To test that, we can run a simple thought experiment using the estimated cost savings from Section 4 to see how shares would shift if sectors simply became cheaper at different rates, without anyone padding their margins.

When we start with a 32 percent farm share under a legacy‑style definition and apply the estimated cost changes, agriculture’s share falls to about 24 percent, while food manufacturing rises from 35 percent to roughly 44 percent. Transportation, packaging, wholesale, and retail all move a few points in one direction or another, but the key point is that the total still adds up to the same amount: we have simply recut the pie based on who became more or less expensive relative to everyone else. This table is not a replacement for the official Food Dollar series; it is a thought experiment whose purpose is to show that even if every sector were playing fair, and even if the supply chain were becoming more efficient, the farm share could still drift downward simply because it was reducing its own costs at the same time other parts of the system were also cutting costs or holding their value steady.

Seen in that light, the long‑run decline in the farm share looks less like a smoking gun and more like a composite of three forces: definitional changes in the data, a shift toward more spending on away‑from‑home and more processed food, and genuine changes in who does what along the chain. None of those automatically implies that the supply chain has absorbed all the farm’s productivity dividend as higher costs. To answer that claim, we need to zoom in further—first by looking at farm share for specific categories like fresh fruits, vegetables, and dairy, and then by testing everything we have learned against a simple, concrete basket of staple foods.

Farm Share of Category Specific Food Groups

The USDA’s own numbers give us a way to see how different food categories perform. For 2010–2024, the Food Dollar data include sector‑specific farm‑share estimates for various product types. When we look specifically at fresh fruits, fresh vegetables, and dairy, we are looking at categories where the processing story should matter least, and the farm’s role should be most visible. If the supply chain were truly adding unnecessary costs, we would expect the farm share of these categories to trend downward over time.

Instead, the pattern is much more stable. Farm share for fresh fruit has been rising over the last decade and now sits at about 39 percent. Fresh vegetables account for around 26 percent of the farm share in 2023, and milk and dairy products account for around 25 percent in 2024. Those numbers align much more closely with the 1960–1980 Marketing Bill era once we make the apples‑to‑apples adjustments. In other words, when we strip away restaurants and heavy processing and just look at basic staples, the farm’s slice of the dollar has not mysteriously vanished.

This matters because the items in your everyday grocery cart look much more like these categories than like ultra-processed snack aisles. For the kinds of foods people buy regularly, milk, cheese, produce, and basic proteins, the farm share today is still in the same ballpark as it was when the earlier Marketing Bill data were compiled. That does not mean nothing has changed in the food system; it means the “processed food” story cannot bear the weight it is often asked to. If farm shares for simple, minimally processed items remain healthy and the supply chain for those items is more efficient than before, we should expect to see meaningful price relief in the final prices of a basic basket of staples.

Examining an Illustrative Household Food Basket

That brings us back to the puzzle that will anchor the rest of this section: when we build a simple, largely unprocessed food basket and compare 1985 to today, the CPI‑adjusted total cost barely moves—yet the number of hours the average worker needs to buy that basket falls by about 30 percent. With everything we now know about productivity, farm share, and supply chain efficiency, that flat inflation-adjusted price is exactly what no one should expect to see.

1985 and 2025 food prices are taken from BLS Average Price Data (AP), U.S. city averages, using the same source for all items except 1985 milk, which is drawn from a historical review of nominal milk prices by year. Average hourly earnings come from FRED’s ‘Average Hourly Earnings of All Employees, Total Private’ (CES0500000003), with monthly values averaged to obtain an annual wage; hours of work required equal total basket cost divided by this wage.

In Section 3, we compared the cost of a basket of staple items in 1985 to the same basket in 2025, using BLS average price data and adjusting 1985 prices into 2025 dollars. The list is deliberately boring: bread, flour, milk, cheese, butter, coffee, potatoes, rice, a couple of meats, eggs, and sugar. Apart from bread, bacon, flour, and sugar—which require some processing—most of these are straightforward farm products that must be packaged, transported, portioned into consumer sizes, and stocked on shelves.

When you look at the table item by item, some prices go up, and some go down in CPI‑adjusted terms. Bacon and eggs are noticeably higher; cheese, butter, rice, and pork chops are lower; other items move only a little. What matters is the total. In 2025 dollars, the entire basket costs almost the same: about 64 dollars in 1985 and about 64 dollars today. On paper, in CPI-adjusted terms, it looks as if forty years of agricultural innovation and supply-chain improvement have left the price of this basket essentially unchanged.

This basket is not meant to replicate the full CPI; it is meant to strip away complexity and focus on basic foods where we would most expect productivity gains, on the farm and in the supply chain, to show up. For these items, we know three things: farms are vastly more productive than they were in 1985, the post-farm supply chain has not become a runaway cost center in real per-capita terms, and the farm’s share of the dollar for similar fresh categories remains healthy. Everything points in the same direction: this basket should be meaningfully cheaper in real terms than it appears in CPI‑adjusted dollars. The fact that the hours of work fall sharply while the CPI‑adjusted price stays flat is the clearest signal yet that the problem lies not with farmers or middlemen, but with the way our price indexes translate real‑world savings into “inflation‑adjusted” numbers.

Taken together, the farm‑shares evidence, and the simple basket tells a consistent story. For basic foods, farms are far more productive than they used to be, the supply chain has not turned into an out‑of‑control cost vacuum, farm share for comparable fresh items remains healthy, and the hours of work needed to buy a staple basket have fallen sharply, even though CPI‑adjusted prices appear flat. That leaves very little room for the idea that “too many middlemen” or rising distribution costs somehow swallowed the gains.

Before we can fully blame CPI’s methods for this disconnect, there are more familiar suspects we must call to the stand. Maybe it is not measurement error at all, but corporate behavior: powerful food companies using their market power to squeeze farmers on one side, raise prices on the other, and quietly pocket the difference. In the next section, we will put that story under the same kind of pressure test. Following the data on profits, margins, and concentration to see whether “market power and corporate capture” can really account for the missing savings, or whether it too starts to crumble once the numbers are on the table.

Methods and Sources.

Farm Share

Farm Share of the Food Dollar, 1953–2023. This graph shows the long‑run decline in the share of consumer food spending attributed to farms and agribusiness. For 1953–1966, farm‑share percentages are taken directly from USDA historical tables (Table 15 in “Total Farm and Marketing Totals”). For 1967–1996, farm share is calculated as total farm value divided by total consumer food expenditures in the USDA marketing‑bill data. For 1997–2023, farm share is defined as the combined value‑added share of agribusiness and farm production in the USDA ERS Food Dollar Series (“FoodDollarDataNominal (1).xlsx”). Because these series use different concepts and coverage, comparisons across periods should be treated as approximate trends rather than precise year‑to‑year changes.

Thought Experiment: Simulated Food-Dollar Shares After Differential Productivity Gains

I start from the 1967 USDA Marketing Bill breakdown of the food dollar and construct a simplified sector map. Agriculture, transportation, packaging materials, and food manufacturing shares are taken directly from the 1967 marketing‑bill data; the remaining portion of the dollar is assigned to wholesale and retail, which splits that residual share evenly. I then assume an initial total of 1 billion dollars in consumer food spending and allocate it across sectors according to these baseline shares. Next, I apply the estimated real cost changes for each sector by multiplying its initial dollar amount by (1−expected price decrease), producing a new dollar cost for every sector under the “after productivity” scenario. Finally, I sum those adjusted sector totals to obtain a new overall food‑dollar amount and recompute each sector’s share as its adjusted cost divided by this new total. This exercise does not attempt to re‑estimate the official Food Dollar series; it is purely a thought experiment to show how relative cost changes across sectors can shift farm share over time, even when every part of the supply chain is becoming more efficient.

Total Factor Productivity in Post‑Farm Food Supply Chain Industries, 1987–2023

Uses total factor productivity (TFP) indexes from the U.S. Bureau of Labor Statistics for four NAICS‑defined industry groups that approximate the post‑farm food supply chain: food and beverage manufacturing (NAICS 311–312), wholesale trade (NAICS 42), retail trade (NAICS 44–45), and transportation and warehousing (NAICS 48–49). TFP measures the efficiency with which combined inputs—labor, capital, energy, materials, and purchased services—are converted into real output, so higher index values indicate more output per unit of total input.

Source. U.S. Bureau of Labor Statistics. (2025). Major industry total factor productivity (KLEMS): Annual data [Excel spreadsheet]. Office of Productivity and Technology. Retrieved December 28, 2025, from https://www.bls.gov/productivity/tables/major-industry-total-factor-productivity-klems.xlsx (data extracted from the “Annual” sheet for NAICS 311–312, 42, 44–45, and 48–49).[

Note. TFP series are available only for NAICS-based industries beginning in 1987; the aggregates include some non-food activities (e.g., tobacco within 311–312, non-grocery wholesalers within 42), but they provide close proxies for productivity trends in the post-farm food supply chain. The series ends in 2023, the latest year available at the time of extraction

Sources.

Federal Reserve Bank of St. Louis. (2026). Average hourly earnings of all employees, total private (CES0500000003) [Data set]. FRED. https://fred.stlouisfed.org/series/CES0500000003
Taste of Home. (2023, November 27). Here’s the price of milk the year you were born. https://www.tasteofhome.com/article/heres-the-price-of-milk-the-year-you-were-born/
U.S. Bureau of Labor Statistics. (2013). Average food prices: A snapshot of how much has changed over a century (Beyond the Numbers, Vol. 2, No. 13). U.S. Department of Labor. https://www.bls.gov/opub/btn/volume-2/average-food-prices-a-snapshot-of-how-much-has-changed-over-a-century.htm
U.S. Bureau of Labor Statistics. (2026). Table B‑3. Average hourly and weekly earnings of all employees on private nonfarm payrolls by industry sector and selected industry detail [Data table]. The employment situation — January 2026. U.S. Department of Labor. https://www.bls.gov/news.release/empsit.t19.htm
U.S. Bureau of Labor Statistics. (2024). Average price data (AP), U.S. city average. https://www.bls.gov/charts/consumer-price-index/consumer-price-index-average-price-data.htm
USDA Economic Research Service. (1999). Food cost review, 1950–97 (Agricultural Economic Report No. 780). U.S. Department of Agriculture, Economic Research Service.
United States Department of Agriculture, Economic Research Service. (n.d.). Food dollar series: Nominal food dollar data [Data set]. Retrieved March 27, 2026, from URL-of-download-page
United States Department of Agriculture, Economic Research Service. (n.d.). Food dollar series: Nominal food dollar archived data [Data set]. Retrieved March 27, 2026, from URL-of-download-page
U.S. Bureau of Labor Statistics, Office of Productivity and Technology. (2025). Major industry total factor productivity (KLEMS): Annual data [Data set]. U.S. Department of Labor. Retrieved December 28, 2025, from https://www.bls.gov/productivity/tables/major-industry-total-factor-productivity-klems.xlsx
U.S. Department of Agriculture, Economic Research Service. (2025). Price Spreads from Farm to Consumer: Highlights and Interactive Charts. Retrieved December 31, 2025, from https://www.ers.usda.gov/data-products/price-spreads-from-farm-to-consumer/highlights-and-interactive-charts. (Chart adapted from the interactive tool showing farm share for fresh fruit, fresh vegetables, and milk and dairy baskets.)

Author: Kyle Novack

March 31, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

The Supply Chain Heist Part 1

Kyle Novack — Fri, 27 Mar 2026 13:32:05 GMT

In previous sections, I showed how U.S. agriculture became vastly more productive and how food became much more affordable in terms of the hours of work required to buy it. Yet CPI‑adjusted food prices often reflect only a fraction of the decline that this productivity boom should have produced, given basic economic principles. Fewer people now work in agriculture while serving a much larger population, which should lower the labor cost per meal. At the same time, the production process itself requires fewer inputs per unit, so as output scales up, fixed and variable costs are spread over more units, further lowering prices.

That leaves us with an obvious next question: if those savings were real, where did they go? The most common answer is that they were absorbed by the modern food supply chain. So let us call our first suspect to the witness stand.

For readers joining here, this section builds on the first three sections of the series, which cover the farm productivity boom, the strange behavior of CPI-adjusted farm costs, and the gap between food affordability and official price measures.

How to Read the Data in This Section

Unless otherwise noted, all dollar amounts in this section are adjusted for inflation using the CPI‑U and expressed in constant 2024 dollars, so differences over time reflect real changes rather than general price inflation.

How Supply-Chain and Non-Farm Costs Shape Food Prices

The supply-chain explanation begins with something ordinary people can see for themselves: the modern grocery store is not just a place where raw farm output changes hands. It is a highly organized distribution system that sorts, packages, stores, moves, brands, and displays an enormous range of products in forms that are easy for households to buy and use. If food savings disappeared somewhere between the farm and the checkout lane, this is the part of the economy where many readers would expect to find them.

That intuition becomes even stronger once we remember how different grocery shopping used to be. In the early 1900s, most stores kept goods behind the counter, and shoppers typically handed a clerk a list rather than walking the aisles themselves. Most of those stores carried fewer than 100 items, and even well-stocked ones may have had only around 250. The modern self-service model began to take shape with Piggly Wiggly in 1916, which offered aisles, checkout lanes, shopping carts, and more than 1,000 items. By 1923 the chain had expanded to more than 1,200 locations (Blazeski, 2016 & ArcGIS StoryMaps, 2020).

From there, the system only became more elaborate. By the 1950s, supermarkets were carrying around 10,000 items while relying on low margins and high volume. Today, the average supermarket carries about 31,795 items and generates roughly $1 trillion in annual sales. That increase in variety and convenience implies a much more complex network of packaging, warehousing, transportation, stocking, and inventory management than existed a century ago. On the surface, then, the standard story makes sense: perhaps farms got cheaper, but the system surrounding them got more expensive (Ganzel, 2007; Food Marketing Institute, n.d.).

That story, however, must survive a basic test. If the supply chain really absorbed the missing savings, then major post-farm costs should show a clear tendency to rise in real per-capita terms over time. But when the USDA food-dollar data are adjusted for CPI and population, packaging materials and transportation do not behave like runaway cost centers. In the long run, packaging trends down overall, and transportation remains roughly flat to slightly down from 1967 to 2023. Even though the transportation category was broadened in 1997 in a way that should have pushed measured costs higher. This does not imply CPI is ‘right’; it simply means that under standard CPI‑based adjustments, these categories are already flat or falling rather than rising.

This does not prove that the supply chain is irrelevant. It does show that two of its most visible components do not fit the simple idea that post-farm costs swallowed the entire productivity dividend. At minimum, the first pass through the evidence gives us a mixed verdict rather than a conviction. The suspect still looks plausible, but the data already suggest we will need a more careful look, especially when we turn to labor.

Labor’s Role in the Food-Price Puzzle

Labor is the part of the supply‑chain story that initially looks most persuasive. In a modern food system, millions of people clearly work beyond the farm gate, and total post‑farm employment in food manufacturing, grocery wholesalers, and food and beverage retailers rose from roughly 3.5 million in 1958 to about 5.9 million. This sounds like exactly the kind of expanding cost structure that could soak up farm‑level savings.

But headcount alone is the wrong metric. What matters is not just how many people work in the supply chain, but how much labor cost each consumer ultimately bears once those costs are spread across the population. Measured this way, the story looks very different. The number of people served by each post‑farm worker has improved by almost 16 percent since 1958. On top of this, USDA Food Dollar and Marketing Bill data, converted to per capita terms, show that real labor costs per person in food manufacturing have been trending downward over the long run. Even after accounting for the 1997 break between the older Marketing Bill series and the newer Food Dollar series.

To dig deeper, we can move from totals to paychecks, extending the analysis beyond food-manufacturing labor to the rest of the chain. Using inflation‑adjusted hourly wage data for food manufacturing, grocery wholesalers, and food and beverage retailers, and weighting those wages by the size of each industry’s workforce. We find that the average real hourly cost rises only 19.8% over nearly seven decades.

Taken together, accounting for both wage changes and the fact that each worker now serves about 16 percent more people. The implied labor cost per person rises by only about 3.8 percent over the entire period—roughly one bad year of inflation. Not a structural force capable of swallowing seven decades of farm‑level savings.

The labor evidence points in the same direction as the packaging and transportation evidence. The supply chain clearly became larger and more complex, but that does not automatically mean it became more expensive in real per-capita terms. A bigger system can still be more efficient, and the data indicate that post-farm labor became more productive rather than simply more burdensome. Labor, you may step down from the witness stand.

So far, the supply chain case has begun to weaken. Packaging and transportation do not appear to be runaway cost centers in the long run. Also, labor looks even less threatening when measured on a per-person basis. That leads to the next question: are the industries between the farm and the checkout lane creating their own productivity gains, or were they just living off the gains created on the farm?

How Supply‑Chain Productivity Should Reinforce Farm Cost Savings

The productivity data give a clear answer. Across the core post-farm parts of the food system, food manufacturing, wholesale trade, retail trade, and transportation and warehousing. Total Factor Productivity generally held steady or rose, rather than collapsing. The gains were slower and less spectacular than those in agriculture, but they were still real. This means these sectors were adding to overall efficiency rather than simply cancelling out farm-level progress.

That point matters more than it may seem. Total Factor Productivity measures how efficiently industries convert combined inputs, such as labor, capital, energy, materials, and purchased services, into outputs. When those industries learn to use their resources more efficiently over time, they push down the real cost of getting food from the farm to the consumer. Even if their progress is far slower than the spectacular gains on the farm itself.

The wholesale and retail evidence is especially important because those are the sectors people most often picture as bloated middlemen. Total Factor Productivity rose by roughly 37 percent in wholesale trade and about 47 percent in retail trade between 1987 and 2023. Those gains align with what we know about the evolution of these industries. Key technologies and practices include:

Better inventory tracking
Barcodes and RFID
Larger, taller warehouses
Standardized cargo handling
More efficient picking systems

All these improvements helped move more goods with less waste and less labor per unit of output. (Boysen & De Koster, 2025, NOVA, 2020, Barcodes Inc., n.d., & U.S. Department of Homeland Security, 2024).

Food manufacturing looks more restrained, and that is fine because it faces its own challenges that make productivity harder to achieve. These gains can also be obscured by the added complexity in modern processing, which is a discussion for another time. The long‑run pattern here is one of gradual improvement, not explosive change. This is not a story about forcing every part of the food system to mimic the agricultural boom. It is a simpler point: even where progress was slower, the post‑farm system was still moving toward greater efficiency, not deeper inefficiency.

That is because productivity gains across the chain work like a weighted average. As agriculture’s share of the food dollar shrinks, its contribution to total savings naturally falls, while gains in other sectors matter more. Put differently, what counts is not just how much each sector improves, but how big a slice of the bill it represents. The table below illustrates this with rough estimates: three sectors show price reductions of about 30 percent, while food manufacturing becomes somewhat more expensive. When we weight each sector’s estimated change in cost by its share of the food dollar, the combined effect is an overall cost saving of about 19 percent. This is not evidence that any one sector is “skimming” the gains; it is simply the arithmetic of how different parts of the chain interact once you add them up.

This suspect’s case is now hanging by a thread. We are no longer looking at a story where the supply chain simply swallowed farm‑level savings. Instead, the evidence points to a system that, on net, became more efficient even as it grew more complex. But before we let this suspect walk, there is one more place we need to check: how all of this shows up in the prices of a simple, unprocessed food basket, leaving the usual “processed‑food” excuse with very little room to hide.

This distinction matters because it explains why the standard narrative feels half‑right. It is like noticing that the first runner just set a world‑record lap. Then, looking at the final team time and seeing only a tiny improvement, you conclude the other runners must have slipped—even though, on closer inspection, each of them also ran the fastest laps of their careers.

In the food system, something similar is going on. There really are more stages, more brands, and more services wrapped around our food today. As a result, those layers do absorb part of the farm’s cost reductions. But once you allow for the productivity gains we have just walked through, the evidence no longer points to “too many middlemen” as the main reason CPI-adjusted food prices did not fall further. Instead, it points to measurement systems that make genuine savings look smaller than the physical and productivity records suggest.

At the same time, we are not done with this suspect yet. The next step is to follow these aggregate patterns down to the level where prices hit you: the individual items in a simple basket of staple foods. Critics often argue that supply‑chain costs are rising because we buy more processed, branded products. But the basket I use deliberately avoids those complications, focusing on basic, unprocessed items that are easy to compare over time.

We have already seen a first pass at this basket in Section 3 of the Food Puzzle, but now we can add more context by examining how its CPI-adjusted cost behaves alongside the productivity story. When we do, the usual escape hatch for the supply‑chain narrative gets much narrower, and the puzzle posed by CPI becomes even harder to ignore.

Sources and Methods

Methods

Food Dollar and Marketing Series (Labor, Packaging Material, Transportation)

Data for 1967–1996 come from the USDA’s historical marketing‑bill series for domestically produced farm foods. Which report detailed input components such as packaging materials, transportation, and fuels/electricity. Data for 1997–2023 use the revised USDA ERS Food Dollar Series, which expands coverage to include imports and away‑from‑home food expenditures and reports value added by broad industry groups. The 1997 transition reflects a methodological expansion and therefore produces an apparent level shift in totals rather than a sudden real jump in costs. To maintain comparability over time, I combine agribusiness inputs with farm production value after 1997 and collapse downstream service categories—wholesale trade, retail trade, food services, finance and insurance, advertising, and legal/accounting—into a single “Other” aggregate representing marketing, distribution, and overhead costs. Corporate profits are not reported separately in the Food Dollar Series; they are embedded in each industry’s value added, and fuels/electricity are bundled into “Other” for 1967–1969. Population denominators for per‑capita figures are drawn from Bolt and van Zanden’s Maddison‑style historical estimates (up to the latest overlapping year) and Macrotrends’ United States population series for 1950–2025, which I splice into a single annual U.S. population series before dividing CPI‑adjusted dollar values by population.

Total labor data for food manufacturing, grocery wholesaler, and food and drink retailer.

Employment totals for the post‑farm food system are constructed by summing three industry groups: food manufacturing, grocery wholesalers, and food and beverage retailers. For 1939–1989, I use annual employment counts from the U.S. Bureau of Labor Statistics bulletin Employment, Hours, and Earnings, United States, 1909–94 (Bulletin No. 2445), which reports sector‑level series under the older SIC classification. For 1990–2024, I use seasonally unadjusted annual employment series from the BLS Current Employment Statistics program, accessed via FRED, for Food Manufacturing (NAICS 311), Grocery and Related Product Wholesalers (NAICS 4244), and Food and Beverage Stores (NAICS 445). Because BLS switched from SIC to NAICS around 1990, there is a classification break; I handle this by reporting long‑run totals and trends rather than treating 1989–1990 as a continuous year‑to‑year change. The “3.5 million for 1957” and “5.9 million for 2024” employment figures in the text are simply the sum of these three-sector series.

How many people does one worker serve?

To estimate how many people each post-farm food worker serves, I divide the total U.S. population for each year by the combined employment in food manufacturing, grocery wholesalers, and food and beverage retailers. Population comes from the spliced Maddison/Macrotrends series described above, and employment totals come from the BLS bulletin (1939–1989) and the three NAICS‑based BLS/FRED series (1990–2024).
wholesaling, and food and beverage retail only and therefore exclude restaurant and other foodservice employment, which makes the “people served per worker” ratios a conservative measure of supply‑chain labor productivity.

Inflation‑Adjusted Hourly Pay in Post‑Farm Food Jobs, 1958 vs. 2024 table

Nominal hourly earnings for food manufacturing, grocery wholesalers, and food and beverage retailers come from BLS Current Employment Statistics series on average hourly earnings of production and nonsupervisory employees for each sector (CEU3231100008, CEU4142400008, and CES4200000008). For each year, I deflate these nominal wages using the CPI‑U all‑items index so both 1958 and 2024 wages are expressed in constant dollars. The values shown in the figure are simple sectoral averages (no employment weighting), intended to illustrate broad changes in real hourly pay across major post‑farm food industries.

Total Factor Productivity in Post‑Farm Food Supply Chain Industries, 1987–2023

Uses total factor productivity (TFP) indexes from the U.S. Bureau of Labor Statistics for four NAICS‑defined industry groups that approximate the post‑farm food supply chain: food and beverage manufacturing (NAICS 311–312), wholesale trade (NAICS 42), retail trade (NAICS 44–45), and transportation and warehousing (NAICS 48–49). TFP measures the efficiency with which combined inputs—labor, capital, energy, materials, and purchased services—are converted into real output, so higher index values indicate more output per unit of total input.
Source. U.S. Bureau of Labor Statistics. (2025). Major industry total factor productivity (KLEMS): Annual data [Excel spreadsheet]. Office of Productivity and Technology. Retrieved December 28, 2025, from https://www.bls.gov/productivity/tables/major-industry-total-factor-productivity-klems.xlsx (data extracted from the “Annual” sheet for NAICS 311–312, 42, 44–45, and 48–49).[
Note. TFP series are available only for NAICS-based industries beginning in 1987; the aggregates include some non-food activities (e.g., tobacco within 311–312, non-grocery wholesalers within 42), but they provide close proxies for productivity trends in the post-farm food supply chain. The series ends in 2023, the latest year available at the time of extraction

Sources

ArcGIS StoryMaps. (2020, December 23). Grocery stores in 1910s. Retrieved December 21, 2025, from https://storymaps.arcgis.com/stories/c3c065b10dae4c1eb689564e55e2aad8
Barcodes Inc. (n.d.). History of barcodes. Barcodes Inc. Retrieved December 28, 2025, from https://www.barcodesinc.com/articles/history.htm
Blazeski, G. (2016, December 25). Piggly Wiggly – The first true self-service grocery store. The Vintage News. Retrieved December 21, 2025, from https://www.thevintagenews.com/2016/12/25/piggly-wiggly-the-first-true-self-service-grocery-store/
Bolt, J., & van Zanden, J. L. (2024). Maddison style estimates of the evolution of the world economy: A new 2023 update. Journal of Economic Surveys. Advance online publication. https://doi.org/10.1111/joes.12618
Boysen, N., & de Koster, R. (2025). 50 years of warehousing research—An operations research perspective. European Journal of Operational Research, 320(3), 449–464. https://doi.org/10.1016/j.ejor.2024.04.031
Canning, P. (2011). A revised and expanded food dollar series: A better understanding of our food costs (Economic Research Report No. 114). U.S. Department of Agriculture, Economic Research Service. http://www.ers.usda.gov/publications/pub-details/?pubid=44827
Food Marketing Institute. (n.d.). Food industry facts. Retrieved December 21, 2025, from https://www.fmi.org/our-research/food-industry-facts
Ganzel, B. (2007). Supermarkets dominate. Wessels Living History Farm. Retrieved December 21, 2025, from https://livinghistoryfarm.org/farming-in-the-1950s/making-money/supermarkets-dominate/
Macrotrends. (n.d.). United States population 1950–2025. Retrieved December 12, 2025, from https://www.macrotrends.net/global-metrics/countries/usa/united-states/population
NOVA. (2020, May 14). When was the ASRS system invented? NOVA Racking. Retrieved December 28, 2025, from https://www.novaracking.com/When-was-the-ASRS-system-invented-id3513603.html
U.S. Department of Agriculture, Economic Research Service. (1999). Food cost review, 1950–97 (Agricultural Economic Report No. 780). U.S. Department of Agriculture. https://ers.usda.gov/sites/default/files/_laserfiche/publications/41035/15334_aer780h_1_.pdf
U.S. Department of Agriculture, Economic Research Service. (2024, November). Food Dollar Series [Data set]. https://www.ers.usda.gov/data-products/food-dollar-series
U.S. Department of Agriculture, Economic Research Service. (2026). Food Dollar [Data product]. https://www.ers.usda.gov/data-products/food-dollar
U.S. Department of Homeland Security. (2024, August 1). Radio frequency identification (RFID): What is it? U.S. Department of Homeland Security. Retrieved December 28, 2025, from https://www.dhs.gov/archive/radio-frequency-identification-rfid-what-it
U.S. Bureau of Labor Statistics. (2025). All employees, food manufacturing (NAICS 311), United States [IPUEN311W200000000]. In FRED, Federal Reserve Bank of St. Louis. Retrieved December 27, 2025, from https://fred.stlouisfed.org/series/IPUEN311W200000000
U.S. Bureau of Labor Statistics. (2025). All employees, food and beverage stores (NAICS 445), United States [IPUHN445W200000000]. In FRED, Federal Reserve Bank of St. Louis. Retrieved December 27, 2025, from https://fred.stlouisfed.org/series/IPUHN445W200000000
U.S. Bureau of Labor Statistics. (2025). All employees, grocery and related product wholesalers (NAICS 4244), United States [IPUGN4244W200000000]. In FRED, Federal Reserve Bank of St.
United States Bureau of Labor Statistics. (1994). Employment, hours, and earnings, United States, 1909–94 (Bulletin No. 2445). U.S. Government Printing Office. https://fraser.stlouisfed.org/title/employment-earnings-united-states-1909-94-4039
U.S. Bureau of Labor Statistics. (2025). Consumer Price Index for all urban consumers (CPI‑U), U.S. city average, all items [Data set]. U.S. Department of Labor. Retrieved December 20, 2025, from https://www.bls.gov/cpi
U.S. Bureau of Labor Statistics, Office of Productivity and Technology. (2025). Major industry total factor productivity (KLEMS): Annual data [Data set]. U.S. Department of Labor. Retrieved December 28, 2025, from https://www.bls.gov/productivity/tables/major-industry-total-factor-productivity-klems.xlsx
U.S. Bureau of Labor Statistics. (2025). Average hourly earnings of production and nonsupervisory employees, food manufacturing, not seasonally adjusted (Series CEU3231100008) [Data set]. U.S. Department of Labor. Retrieved December 27, 2025, from https://data.bls.gov
U.S. Bureau of Labor Statistics. (2025). Average hourly earnings of production and nonsupervisory employees, merchant wholesalers, nondurable goods, not seasonally adjusted (Series CEU4142400008) [Data set]. U.S. Department of Labor. Retrieved December 27, 2025, from https://data.bls.gov
U.S. Bureau of Labor Statistics. (2025). Average hourly earnings of production and nonsupervisory employees, retail trade, seasonally adjusted (Series CES4200000008) [Data set]. U.S. Department of Labor. Retrieved December 27, 2025, from https://data.bls.gov

Author: Kyle Novack

March 27, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

From Farm Gate to Cart: The Missing Miracle

Kyle Novack — Tue, 24 Mar 2026 13:31:43 GMT

Up to this point, we have stayed on the farm side of the story: more food, fewer people, less land, and a money world that somehow refuses to show the full extent of the savings. In this section, we follow those crops past the farm gate and into the grocery aisle. Food is one of the best stress tests for CPI because it is not optional. It hits every household every week, and staple items are far easier to compare over time than products whose quality, branding, and features constantly change. If our inflation yardstick is working, a century of productivity gains should show up as clearly lower real costs for an ordinary basket of groceries.

This is exactly why the earlier farm evidence matters so much. We know U.S. agriculture experienced an extraordinary productivity boom: from 1910 to today, the average number of people served per farm rose from about 14.5 to 182. Yet in CPI-adjusted dollars, total farm output per person fell by only about 26%. This closely matches the roughly 28% reduction in profit margins, while total farm expenses per person rose by about 7%. In other words, once we run everything through CPI, it still appears more expensive to farm today than in the past, even after a century of huge efficiency gains.

But farm data by itself cannot settle the issue. Consumers do not buy raw agricultural output; they buy food after processing, packaging, transportation, wholesaling, and retailing have all added value and cost. So even though the monetary data already suggest that farm‑level savings are not showing up the way we would expect, we still must ask a separate question: did those productivity gains ultimately translate into lower costs for consumers at the grocery store?

When we run a simple grocery‑basket test, the results are mixed. We can see this by using an illustrative household basket built from staple grocery items that appear in both 1913 and today. Using this standard basket of goods, we find that in 1913, the CPI‑adjusted cost would have been $87.08. Today, that same basket costs about 26% less than it did in 1913, adjusted for inflation. That is hard to believe: only a 26% decrease in food prices when every part of the process from farm to consumer has seen dramatic improvements. However, one piece of evidence shows that productivity gains have made food more affordable. When we examine the hours of work required to buy that same basket, we find they have been reduced by roughly 86%. This means food has clearly become more affordable over time in terms of human labor, even if CPI does not fully reflect the underlying productivity story.

If we compare the same basket of goods from 1985 to today, the story gets even more complicated. USDA data show that farm productivity rose from 0.69 to 1.10 by 2021, implying a potential 34% decrease in food production costs over that period. This comparison matters because the 1985 economy looks much more like our modern economy than the 1913 economy does. Yet when we compare the total food basket cost adjusted for CPI, it is roughly the same price in 2025 as it was in 1985. That is not a typo: according to CPI, households see essentially no difference in the cost of this food basket despite the massive changes in the farm economy underneath.

This is confusing because economists often say the cost of living has improved. Under this reading of the data, it takes about the same inflation‑adjusted income in 2025 as in 1985 to cover the same basket of goods. That does not sound like an improvement in the cost of living, which is why so many people feel the pinch in the modern economy. Economists will quickly point out that, compared with wages from that year, it now takes about 30% fewer hours of work to pay for the same basket, and they wear that figure like a badge of honor. But there is more to the story than the fact that the hours went down; what matters is how we achieved that cost reduction.

This illustrative household food basket is like blood spatter at a crime scene. It tells us something happened, but it does not, by itself, tell us exactly how. We know the hours of work needed to buy the basket fell by 30%, indicating improved affordability, but hours alone cannot tell us why. Within any single year, comparing prices to wages is straightforward: if both are in the same dollars, the hours come out the same.

The trouble starts when we compare prices across decades. Then the inflation adjustment becomes the ‘ruler’ we use to reconstruct the scene. This is more like piecing together a crime from grainy cameras and faded memories than watching it unfold in real time. If that ruler is off, we can easily mistake wage growth for cheaper food, or the other way around.

Economists are quick to say that if the hours of work required to buy food are falling, living standards must be improving, and the rest is just quibbling over indexes. But in a genuinely productivity‑driven economy, an hour of work should buy more over time because the real inputs needed to create each unit of output keep shrinking.

When productivity is doing the heavy lifting, we should see both the CPI‑adjusted price of the basket and the hours of work required move down together. Not one flat while the other does all the work. When hours fall, but CPI-adjusted prices stay stubbornly level. The pattern stops looking like a clean productivity win and starts looking like a problem with the yardstick we are using to measure inflation in the first place.

If CPI is roughly right, then one of the most dramatic productivity booms in the economy translated into only slight real declines in many staple food prices. This is hard to square with the scale of the agricultural transformation shown by the data. If CPI is missing part of the story, then this food basket is not just a table of grocery prices. It becomes evidence that productivity‑driven cost reductions were real but were masked by a deeper devaluation of the dollar.

Getting this right matters because the inflation measure we choose does not just change the story of food; it changes the story of wages, housing, and living standards more broadly. Even small differences in the index can reshape how we interpret past economies and how we judge the one we are living in now.

Because this pattern clashes with what both common sense and basic economic theory would predict, commentators and researchers have developed a set of familiar explanations for why grocery prices have not fallen more. These explanations are often presented as if they fully resolve the issue. In the next section, I lay them out clearly and then test the evidence to see whether they really explain why decades of productivity gains have not translated into lower food prices.

Before we turn to the standard explanations, a brief note on my use of the data. Wherever possible, I trace each series back to the earliest year the underlying source is reliable and internally consistent, even though not every series appears in a chart. I do not apply any exotic statistical adjustments. Instead, I use standard published series and stitch them together in a new way, with full citations so you can independently verify every step if you have any concerns.

The Standard Narrative for Food Prices.

When we ask why grocery prices have not fallen more despite massive gains in agricultural productivity. It is easy for commentators and researchers to fall back on a familiar set of explanations to dismiss the puzzle. None of these explanations is frivolous, and that is precisely why they are so persuasive at first glance: each contains enough truth to feel satisfying before the deeper questions are asked. Here are the five familiar suspects.

1. Suspect: Supply-Chain Dilution and Non-Farm Costs

The first and most intuitive answer is that the savings generated on the farm were absorbed by everything that happens after the farm gate: processing, packaging, transportation, wholesaling, and retailing. Even if raw agricultural production became far cheaper, consumers do not buy raw farm output; they buy food that has been moved, stored, branded, processed, and sold through a far more elaborate system than existed in the past.

2. Suspect: Market Power and Corporate Capture

A second explanation is that large food and beverage companies gained enough scale and bargaining power to keep consumer prices elevated even as their own input costs improved. In this view, concentration in branded and packaged foods allows dominant firms to squeeze suppliers, defend shelf prices, and capture a larger share of productivity gains rather than pass them through to households.

3. Suspect: External Shocks and Offsets

A third explanation indicates that a steady stream of disruptions affects food markets, including weather events, droughts, floods, animal disease outbreaks, energy price spikes, wars, and policy shocks. Even if productivity is quietly pushing costs downward in the background, repeated disruptions like these can keep prices elevated, create volatility, or obscure the long-run trend that would otherwise be easier to see.

4. Suspect: Regulation

A fourth explanation is that the modern food economy carries more rules, compliance burdens, and legal requirements than it once did. These added costs absorb part of the productivity dividend. Under this argument, farms, processors, distributors, and retailers may all be more efficient than they used to be. That thicker layer of regulation keeps those savings from fully reaching the consumer.

5. Suspect: Measurement Flaws in Inflation.

The final objection is that the problem may not lie entirely in food itself, but in the yardstick we use to compare prices across time. If the Consumer Price Index misstates inflation or causes us to misread past prices, then food may have fallen more than we think, and the puzzle may be partly a measurement problem rather than a production problem.

Taken together, these five suspects form the standard narrative for why grocery prices have not fallen more in real terms, despite agriculture’s extraordinary productivity boom. The real test, however, is not whether each explanation sounds plausible on paper, but whether the data show that any of them absorbed the missing savings that should have reached consumers.

In the next article, The Supply Chain Heist, we put the first suspect on the stand and follow the food dollar step by step from the farm gate to the checkout line to see how much of the farm-level miracle was quietly consumed by processing, packaging, transportation, wholesaling, and retailing.

Sources and Methodology

Note on agricultural figures: Summary statistics on farm productivity, costs, and profit margins over 1910–2024 in this section are derived from USDA and related data using methods documented in Section 1 (‘Why Agriculture Is the Perfect Smoking Gun’) and Section 2 (‘The Money-World Paradox: Where Did All the Savings Go?’). Readers wishing to audit or replicate those calculations should consult the data sources and methods described there.
Federal Reserve Bank of St. Louis. (2026). Average hourly earnings of all employees, total private (CES0500000003) [Data set]. FRED. https://fred.stlouisfed.org/series/CES0500000003
Rees, A. (1966). Real wages in manufacturing, 1890–1914. In D. S. Brady (Ed.), Output, employment, and productivity in the United States after 1890 (pp. 83–130). National Bureau of Economic Research. https://www.nber.org/system/files/chapters/c2287/c2287.pdf
Taste of Home. (2023, November 27). Here’s the price of milk the year you were born. https://www.tasteofhome.com/article/heres-the-price-of-milk-the-year-you-were-born/
U.S. Bureau of Labor Statistics. (2013). Average food prices: A snapshot of how much has changed over a century (Beyond the Numbers, Vol. 2, No. 13). U.S. Department of Labor. https://www.bls.gov/opub/btn/volume-2/average-food-prices-a-snapshot-of-how-much-has-changed-over-a-century.htm
U.S. Bureau of Labor Statistics. (2024). Average price data (AP), U.S. city average. https://www.bls.gov/charts/consumer-price-index/consumer-price-index-average-price-data.htm
U.S. Bureau of Labor Statistics. (2026). Table B‑3. Average hourly and weekly earnings of all employees on private nonfarm payrolls by industry sector and selected industry detail [Data table]. The employment situation — January 2026. U.S. Department of Labor. https://www.bls.gov/news.release/empsit.t19.htm

Author: Kyle Novack

March 24, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

The Money World Paradox: Where Did All the Savings Go?

Kyle Novack — Fri, 20 Mar 2026 12:30:25 GMT

On the physical side, the story is crystal clear: more food, fewer people, less land. So, the natural next question is simple: how does this transformation show up in agricultural datasets measured in U.S. dollars? If productivity is doing what every textbook says, the cost of farming per person should have collapsed in real, inflation‑adjusted dollars.

How to Read the Data in This Post

Unless otherwise noted, all dollar figures in this section are expressed in CPI‑adjusted 2024 dollars. All the charts and numbers in this article come from standard public data sets: USDA (NASS and ERS), and official population estimates. Citations in the main text are kept to a minimum; each figure has a short caption explaining how it was calculated and which datasets it uses. At the end of the post, a “Source and Methods” section lists all underlying series, so anyone who wants to audit or recreate the charts can do so.

First Suspect: Total Commodity Output Measured in U.S Dollars

Let’s start with the big picture. When we look at total farm commodity output per American, we should see a dramatic drop. After all, we are getting more food per person from far fewer inputs. Instead, the series barely budges. In 1910, total farm commodity output amounted to about $2,027. By 2024, after more than a century of miraculous efficiency gains, that figure is down to about $1,496 per capita, a decline of roughly 26%. That’s it! Tripling productivity and roughly doubling food production per head translates into only a quarter‑off in the official money world.

Total farm commodity output is measured as the U.S. Department of Agriculture Economic Research Service (USDA ERS) value of agricultural sector production from the Farm Income and Wealth Statistics data set, converted to CPI‑adjusted 2024 dollars and then divided by U.S. population. I use value of production (not cash receipts) as the measure of commodity output. Population figures through 2022 come from the Maddison Project (Bolt & van Zanden, 2024), and values for 2023–2024 come from Macrotrends, which closely track the recent Maddison/Census trend.

Second Suspect: Agriculture Labor Expense

You might think: “Okay, maybe the savings just show up in lower labor costs.” After all, farm labor as a share of the population has collapsed by about 94–95%. That should be an enormous source of cost reduction. Let’s test it. I take total farm labor expenses from the USDA and divide by the U.S. population. The result? Per‑capita farm labor costs have fallen by only about 18% since 1910. Not 80%. Not 50%. Eighteen. Agriculture, however, has a peculiar mix of paid and unpaid labor that almost no other industry does. It was true in 1910, and it’s still true today—which is exactly why the story behind that 18% is anything but simple.

At this point, the numbers don’t quite add up. If there are far fewer workers, why haven’t labor costs per person collapsed? The data push us toward four complications:

Family labor doesn’t show up as wages.
In 1910, most farm work was done by owner-operators and their families. They were “paid” through food, housing, and farm profits rather than a paycheck, so a large share of the labor never appears in the official wage bill.
A similar bias still exists today.
Modern farm accounts also undercount family labor. When you divide today’s total labor expenses by all workers on farms—owners, spouses, kids, and hired hands—you get an average cost per worker well below the published wage for hired employees. That’s a sign that many people are still working for less cash and more in‑kind compensation.
We’re mixing different kinds of workers.
Historical data often takes a single pool of labor expenses and spreads them across everyone connected to a farm. That means we’re averaging together highly paid hired specialists, underpaid family helpers, and owner-operators who earn both wages and returns on their land and equipment.
Headcounts and wages don’t fully cancel.
Even after accounting for these complications, total paid labor cost per person has only fallen by 18%. One worker now produces food for far more people, which should spread the cost across more consumers and push per‑capita figures down, but the offsetting forces are strong enough that the drop is small rather than dramatic.

Put together, the evidence says something subtle. Yes, we use far fewer people to produce food than we did a century ago. Yes, the remaining workers are paid much more. But because so much farm work—then and now—happens off the formal payroll, total labor cost per American has only nudged down, not collapsed. Labor savings help, but they can’t be the main driver of cheap food. The real cost revolution must be coming from somewhere else, which means the investigation keeps going.

Third Suspected: Capital Expenses

Maybe the savings are being eaten by machines and buildings instead. The usual story is that tractors, combines, irrigation systems, and grain bins replaced human sweat with expensive capital. If that were true, we would expect capital expenses per person to explode upward over time.

Again, the data do not cooperate. When we look at agricultural capital expenditures per capita, it averaged about $146 in 1910 and around $90 per person in recent years, even dipping to about $68 in some years—a decline of roughly 38% over the long run. The category that is supposed to be ballooning is smaller in real per‑capita terms than it was a century ago. That not only defies the textbook story, but it also falls far short of explaining the missing cost savings we are looking for.

Fourth Suspect: Total Production Expenses

If the savings aren’t in labor, and they aren’t eaten by capital, maybe they are hiding in other inputs—seeds, fertilizer, chemicals, fuel, and so on. This is the last place for the official story to rescue itself. Here, the numbers do move, but not in the way anyone rooting for CPI would hope. When we look at total production expenses per capita—labor, capital, and intermediate inputs together—farm costs per person rise from about $1,238 in 1910 to around $1,333 today, an increase of roughly 7.6%. Put plainly, after all the technological progress, all the bigger tractors, better seeds, and smarter practices, the official statistics tell us it now costs more per American to farm than it did before tractors were common.

Total agricultural (farm production) expense per capita is calculated as total farm production expenses from USDA ERS Farm Income and Wealth Statistics, converted to CPI‑adjusted 2024 dollars, divided by U.S. population. This aggregates hired labor, capital, and intermediate input costs into a single per‑capita measure

Fifth Suspect: Farm Profit Margins

At this point, many readers (and economists) reach for one last explanation: “Ah, the gains must have been captured as higher profits. Farmers and agribusiness simply kept the surplus for themselves.” That would at least be internally consistent: if we see no big drop in consumer‑facing costs, maybe someone is skimming the cream at the top.

The problem is that profit margins went the wrong way, too. Historical USDA accounts show that farm profit margins averaged about 39% around 1910, but in recent decades they have fallen to around 11%. In CPI‑adjusted terms, the modest 26% decline in per‑capita commodity value mostly reflects the fact that farmers’ slice of the pie got thinner, not that they hoarded efficiency gains. The “greedy farmer” escape hatch slams shut, because if farmers were pocketing the cost savings from productivity gains, profit margins would be rising, not shrinking.

Agricultural profit margins are calculated as (total value of agricultural production−total farm production expenses)÷total value of agricultural production×100, using ERS series on value of agricultural production and total production expenses from the Farm Income and Wealth Statistics. This margin is constructed directly from the same output and expense series used in the per‑capita graphs, to keep the accounting consistent.

So we are left with a strange picture. In the real world, we see:

But in the CPI‑adjusted money world, we see some decrease, but they don’t add together; they average out since they are all part of the same metric:

We checked labor. We checked capital. We checked inputs. We checked profits. The savings never show up anywhere they are supposed to. Yet the physical data insist those savings are real and enormous. At that point, it becomes hard to avoid the uncomfortable conclusion: the problem is not with the farms, or the tractors, or the farmers’ honesty. The problem is with the thermometer we are using to measure “real” costs and “real” prices.

Economists already sense that broad price indexes like CPI and standard deflators can muffle what is happening inside specific industries, which is why they often switch to sector‑specific deflators for areas like agriculture—but that “fix” brings its own problems, and that is where we turn next.

What Deflators Actually Are

Before we go further, it helps to define a word economists use casually, and almost no one else ever hears deflator. A deflator is just a mathematical tool we use to translate between nominal dollar values and inflation‑adjusted values over time. When the news says “GDP in constant dollars” or an analyst talks about “inflation‑adjusted profits,” they have taken the original dollar numbers and run them through a deflator to get back to “real” values.

That process works in both directions. Deflators are used not only to deflate current prices to a real base, but also to inflate old prices into today’s dollars. Every time NETs shows a 1910 or 1967 price expressed in 2024 dollars, it uses a deflator built from year-over-year inflation rates to convert. If this sounds like a silly game, it is, but it is a necessary game. As long as inflation is part of how our current monetary system works, we have to adjust for it to make any honest comparison across decades.

CPI is a type of deflator, based on a consumer basket. The national accounts also use other deflators, for GDP as a whole, for specific types of spending, and for individual industries. On paper, each of these tools is trying to answer the same basic question: How many real goods and services sit behind this pile of dollars, once we filter out the changing value of money?

Sector‑Specific Deflators: Shrinking the Mile

Now we can be precise about the problem. Economists will say, “You shouldn’t use CPI to deflate agriculture. Use the agricultural deflator.” That sounds like a technical improvement. Instead of one broad index, we use a sector‑specific one.

But the deeper issue is simpler. Whether a dollar is spent by an individual or a business, its underlying value does not change. One dollar of farm revenue and one dollar of restaurant wages are still claims on the same pool of real goods and services in the wider economy. Industry‑specific deflators quietly presume otherwise. They treat the “real value” of a dollar inside one sector as if it can drift away from the value of a dollar everywhere else, which is why a general inflation deflator is the only honest starting point when you want to compare across industries.

Deflators also cannot change the basic year‑to‑year relationships that tell us what is going on. Imagine a company whose revenue grows 50% while its expenses grow just as fast and profit margins stay stuck at 9%. In nominal dollars, it looks bigger. Once you deflate everything by the same inflation index, you see that the firm did not really grow at all; prices went up, but its underlying business strength did not. No matter what sector‑specific deflator you invent, you cannot turn flat margins into an efficiency miracle.

Sector‑specific deflators are like changing the miles for each vehicle on the road. We all know a semi‑truck and an SUV do not get the same miles per gallon. The truck burns more fuel because it is hauling a much heavier load. That comparison is meaningful because a mile is a mile, and a gallon is a gallon. Now imagine we decide that it is embarrassing for the truck, so we quietly shrink the length of “a mile” whenever we measure its fuel economy. On paper, the truck suddenly looks as efficient as the SUV. In reality, nothing about physics has changed; we just broke the ruler.

Giving every industry its own deflator does the same thing. Agriculture gets one ruler, manufacturing another, and services a third. Each ruler can be bent until the chart for that sector looks tidy, but the cross‑industry comparisons that matter for real people and real policy become meaningless. The agricultural facts do not care which deflator you pick. Farmers produce far more food per person with far fewer people and less land, and farm profit margins are lower, not higher. A deflator cannot make those realities disappear or reappear; it can only change how they look in a spreadsheet.

In a sound measurement system, we should not need a different “mile” for every sector. A single, well-behaved inflation yardstick should let us compare farms, factories, and fast-food chains directly while keeping the basic economic logic in line. The moment we need to rescale the dollar differently for each industry just to make the charts look comfortable, we are no longer measuring the economy; we are editing it.

Up to this point, we have stayed on the farm’s side of the ledger, tracking how much it costs to grow food in the first place. But consumers do not buy direct from the farm; they buy groceries. If the savings from agriculture’s productivity boom really did make their way to ordinary people, we should see it in the prices families pay at the store over time. In section 3, we leave the farm gate and follow the food dollar all the way to the checkout line to see whether those missing savings ever showed up for consumers.

Methods and Sources (Non‑Graphed Series)

Labor expense per capita.
Labor expenses are calculated as total hired farm labor expenses from USDA ERS Farm Income and Wealth Statistics, converted to CPI‑adjusted 2024 dollars, divided by the U.S. population. This measure captures only paid hired labor; unpaid family and operator labor are not included in this expense series, which is part of why official labor costs understate total labor input on farms.

Capital expense per capita.
Capital expenses are calculated as total farm capital expenses from USDA ERS Farm Income and Wealth Statistics, converted to CPI‑adjusted 2024 dollars, divided by the U.S. population. Capital expenses include items such as machinery, buildings, and other farm capital costs as defined in the ERS accounts.

Sources:

Bolt, J., & van Zanden, J. L. (2024). Maddison style estimates of the evolution of the world economy: A new 2023 update. Journal of Economic Surveys, 1–41. https://doi.org/10.1111/joes.1261
Macrotrends LLC. (n.d.). United States population 1950–2024 [Data set]. Macrotrends. https://www.macrotrends.net/global-metrics/countries/usa/united-states/population
U.S. Department of Agriculture, Economic Research Service. (n.d.) Farm income and wealth statistics [Data set]. U.S. Department of Agriculture. https://data.ers.usda.gov/report.aspx?ID=4059
U.S. Department of Agriculture, Economic Research Service. (n.d.) Farm income and wealth statistics: Value of production and cash receipts tables [Data set]. U.S. Department of Agriculture. https://data.ers.usda.gove/report.aspx?ID=4055

Author: Kyle Novack

March 20, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

Why Agriculture Is the Perfect Smoking Gun

Kyle Novack — Wed, 18 Mar 2026 01:01:00 GMT

Agriculture is as close as economics ever gets to a laboratory experiment. We have over a century of exceptionally detailed USDA data on output, acres, labor, capital costs, intermediate inputs, and total factor productivity. All are built on consistent methods and carefully maintained time series. Most sectors can only dream of that kind of statistical backbone.

Even better, the product itself barely changes. A pound of wheat is still a pound of wheat. A bushel of corn is a bushel of corn. Calories from basic staple crops deliver essentially the same nutrition today that they did in 1910. We do not need to argue about whether a 1910 smartphone is “comparable” to a modern one, or whether new features justify higher prices. The output is measured in raw physical units.

On the demand side, everyone in the country participates in the food economy. Every American either helps produce food or buys it, so converting the entire system to per-capita terms does not distort the story the way it might in a niche industry. When we say, “pounds of major crops per capita” or “farm expenses per capita,” we are describing how the food system interacts with every citizen, not just a sliver of the population.

That combination makes agriculture the ideal stress test for our inflation yardstick. The data are clean, the product is stable, and the market is universal. If an inflation measure cannot accurately capture the impact of a century-long productivity revolution in farming, where the goods are simple, and the inputs and outputs are precisely tracked, then the problem lies with the measure, not with the farms. Put bluntly: if our inflation gauge cannot tell the truth about food, it cannot be trusted anywhere.

How to Read the Data in This Post

All the charts and numbers in this article come from standard public data sets: USDA (NASS and ERS), official population estimates, historical GDP series, and labor‑force statistics. Citations in the main text are kept to a minimum; each figure has a short caption explaining how it was calculated and which datasets it uses. At the end of the post, a “Data sources” section lists all underlying series, so anyone who wants to audit or recreate the charts can do so.

The Physical Reality of Modern Farming

Over the past century, American farming has quietly become one of the most efficient systems humans have ever built. Economists track this using something called Total Factor Productivity (TFP), which measures how much output we get from all the main inputs together: land, labor, capital (machines and buildings), and materials like seed and fertilizer. In simple terms, if TFP goes up, it means we are producing more food from the same—or even fewer—resources.

On that score, agriculture has blown past almost every other sector. The USDA’s TFP index for U.S. farms rises from about 100 in 1948 to nearly 300 in recent years. Indicating that overall productivity has roughly tripled since the mid‑20th century. That abstract index shows up in the hard numbers when we compare 1910 to today. From here, I use 1910–2024 data for output, land, and labor because the official TFP series only begins in 1948. The reason it is important to go back as far as possible is that it makes the progress impossible to wave away. The further back we look, the more overwhelming the productivity gains become. This makes agriculture an even tougher stress test for how well CPI measures inflation.

Graph 1. Total Factor Productivity index (1948 = 1). Calculated from USDA Economic Research Service, Agricultural Productivity in the U.S. data series (as of January 12, 2024).

Here is what changed between then and now:

Food produced per person: For the ten major U.S. crops, total output jumped from about 250 billion pounds in 1910 to over 1.4 trillion pounds today. On a per‑person basis, that is an increase from roughly 2,790 pounds a year to more than 4,150 pounds—about 49 percent more staple food per American.
People working in farming: In 1910, about 12 percent of Americans—roughly one in eight—lived and worked on farms. Today, it is about 0.7 percent, roughly one in 150 people, a 94–95 percent drop in the share of the population tied directly to agriculture.

Graph 2: Calculated as the agricultural/farm workforce divided by the total U.S. population, rather than by the labor force, to avoid breaks in how “labor force” is defined over time. A series break remains in 1948 due to a change in farm‑worker definitions (age 14+ pre‑1948; age 16+ post‑1948). Sources: Statista (2025) for historical farm/non‑farm labor force, 1910–1948; U.S. Bureau of Labor Statistics (2025) for agriculture and related employment, 1949–2024

Land used per capita: In 1910, the U.S. harvested about 2.4 acres of cropland per capita for these major crops. Today, it is about 0.65 acres per person, a decline of roughly 73 percent, even though total harvested acres for these crops have stayed near 220 million for more than a century.

Graph 3: Calculated as total harvested acres for the ten major crops divided by U.S. population. Sources: Harvested acreage — USDA National Agricultural Statistics Service (2023, 2025), Crop Production summary reports; population — Bolt and van Zanden (2024) and Macrotrends (n.d.)

People fed by each farm: Around 1910, the average farm fed about 14.5 people. Basically, enough to feed the farm family itself and roughly one other household. Today, the average farm feeds roughly 182 individuals, a more than twelve‑fold increase.

Share of the overall economy (as a percent of GDP): Farm commodity output fell from about 17 percent of U.S. GDP in 1910 to around 2 percent today, even as the physical supply of food surged.

Graph 4. Farm commodity output as a share of GDP. Calculated as total farm commodity output/value of production (current dollars) divided by nominal GDP. Sources: USDA Economic Research Service, Farm Income and Wealth Statistics (annual cash receipts by commodity); Johnston and Williamson (2025), MeasuringWorth “What Was U.S. GDP Then?” for 1910–2023; U.S. Bureau of Economic Analysis (2025) via FRED series GDP for 2024.

Food has never been this abundant or this cheap relative to the rest of what we produce and buy. In a world without this agricultural miracle, the United States would be almost unrecognizable. Using 1910 technology to feed today’s population would require roughly 3.8 times as much cropland, some 831 million acres for the major crops alone. An area comparable to the entire landmass of India and about a third of all U.S. land. We would need on the order of 41 million farm workers, compared with about 2.2 million today. Meaning one in eight Americans would still be tied to farm work. Instead of freeing people to become engineers, teachers, nurses, and software developers, the country would be one vast farm barely keeping up with basic caloric needs.

To see the whole century in one snapshot, here’s what changed between 1910 and today.

Now that we’ve seen just how massive the productivity gains in farming are, we can ask a sharper question: how does this transformation show up in agricultural datasets measured in U.S. dollars? We know the physical reality of the last century, and our intuition says we’d expect the “real” price of farming to have collapsed right along with the land and labor it uses. Instead, once we put CPI on the case, the savings practically evaporate. In the next section, we’ll follow the trail of dollars—and see why every obvious place those savings should show up comes back empty.

Methods and Sources (Non‑Graphed Series)

Crop production in pounds: Yields in pounds per harvested acre (converted from bushels per acre using standard USDA weights where applicable) multiplied by harvested acreage in acres, summed across barley, corn/maize, cotton, oats, rice, soybeans, sugarbeets, sugarcane, tobacco, and wheat. Sources: USDA National Agricultural Statistics Service, Crop Production 2022 Summary and Crop Production 2024 Summary (historical tables on yields and harvested acreage); population series from Bolt and van Zanden (2024) and Macrotrends.

Crop pounds per capita: Aggregate crop production in pounds divided by total U.S. population; same sources as above

Harvested acres: Aggregate harvested acreage (barley, corn/maize, cotton, oats, rice, soybeans, sugarbeets, sugarcane, tobacco, wheat) from USDA NASS Crop Production summaries

Sources

Bolt, J., & van Zanden, J. L. (2024). Maddison style estimates of the evolution of the world economy: A new 2023 update. Journal of Economic Surveys. Advance online publication. https://doi.org/10.1111/joes.12618
Macrotrends. (n.d.). United States population 1950–2025. Retrieved March 14, 2026, from https://www.macrotrends.net/global-metrics/countries/usa/united-states/population
MeasuringWorth. (2025). What was the U.S. GDP then? Retrieved March 14, 2026, from https://www.measuringworth.org/usgdp
· Statista. (2025). United States agricultural and nonagricultural labor force, 1900–1970 [Data set]. Retrieved March 14, 2026, from https://www.statista.com/statistics/1316855/us-farm-nonfarm-labor-force-historical/
· U.S. Bureau of Economic Analysis. (2025). Gross domestic product, billions of dollars, annual (Series GDP) [Data set]. Federal Reserve Bank of St. Louis (FRED). Retrieved March 14, 2026, from https://fred.stlouisfed.org/series/GDP
· U.S. Bureau of Labor Statistics. (2025). Employment level – Agriculture and related industries (Series LNS12034560) [Data set]. Federal Reserve Bank of St. Louis (FRED). Retrieved March 14, 2026, from https://fred.stlouisfed.org/series/LNS12034560
· U.S. Department of Agriculture, Economic Research Service. (n.d.). Agricultural productivity in the U.S.: Summary of recent findings. Retrieved March 14, 2026, from https://www.ers.usda.gov/data-products/agricultural-productivity-in-the-united-states/summary-of-recent-findings
U.S. Department of Agriculture, Economic Research Service. (n.d.). Farm income and wealth statistics: Annual cash receipts by commodity [Data set]. Retrieved March 14, 2026, from https://data.ers.usda.gov/reports.aspx?ID=4055
U.S. Department of Agriculture, National Agricultural Statistics Service. (2023). Crop production 2022 summary (January 2023, ISSN 1936–3737). https://www.nass.usda.gov/Publications/Todays_Reports/reports/croptr22.pdf
U.S. Department of Agriculture, National Agricultural Statistics Service. (2025). Crop production 2024 summary (January 2025, ISSN 1936–3737). Cornell University Library. https://downloads.usda.library.cornell.edu/usda-esmis/files/k3569432s/nk324887m/qn59s0097/cropan25.pdf
U.S. Department of Agriculture, National Agricultural Statistics Service. (n.d.). Quick Stats database: Farms – Number of operations – United States [Data set]. Retrieved March 14, 2026, from https://quickstats.nass.usda.gov/#EA91AED4-C137-3251-B480-328DFF9768B3

Author: Kyle Novack

March 17, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)

Why Food Has to Be Cheap for Everything Else to Exist

Kyle Novack — Fri, 13 Mar 2026 18:00:21 GMT

When an iPhone gets more expensive, most of us shrug.
When rent, groceries, gas, and utilities go up, we panic.

That difference is the doorway into a better explanation of value.

In a recent Substack post, I wrote about the current war involving Iran and why, despite the headlines, it’s unlikely to cause sustained inflation. We care about the war, but most of us are more concerned with how its consequences will affect our lives. We ultimately care about the oil.

We don’t fear “oil” in the abstract. We fear being forced to spend more of our paycheck on the same recurring essentials we can’t stop buying: fuel, groceries, rent, electricity, water. When those bills rise, it feels like our life is getting more expensive even if the latest gadget isn’t.

That feeling is not a psychological glitch. It’s a clue that standard explanations of value are missing something important.

What textbooks get right (and where they stop)

Standard microeconomics says: the first slice of pizza is amazing, the fifth is fine, the tenth makes you sick. Each extra slice gives you less additional satisfaction than the previous one. That’s diminishing marginal utility. You’re willing to pay a lot for early units, less for later ones.

The classic fix to the Diamond–Water paradox builds on this. Water is essential but abundant, so each additional unit of water yields diminishing marginal utility. Diamonds are useless but scarce, so the marginal stone is very valuable. Put another way, water can’t really be used as a wealth indicator, but a diamond can.

It sounds like a reasonable solution, but when you stop to think about it, the marginal‑utility explanation sidesteps Adam Smith’s real puzzle:

· Why does something essential to life carry a low price, while something almost useless carries a high price?

· Why, in normal times, do we accept cheap water and expensive diamonds, but in a crisis we instantly flip and trade diamonds for water?

The marginal‑utility story says, “price is tied to marginal units, not total importance.” True, but incomplete. It describes how you feel about slice one vs. slice five of pizza. It doesn’t explain why any slice of life‑sustaining food is treated as cheaper than a useless diamond in the first place.

To answer that, we need one more layer.

Good economics needs three main ingredients: micro, macro, and human psychology. Micro tells us how individuals and firms trade off needs and wants; macro tracks money, credit, and the big aggregates we care about; psychology explains how real people actually behave under calm and crisis. Leave out any one of them and the theory might look elegant on paper, but like a cake without flour, sugar, or eggs, it will never really set in the real world.

Two layers of value: functional vs. priced utility

We need to distinguish between:

Functional utility
- How much your survival and day‑to‑day functioning depend on the good.
- On this dimension, food and water obviously dominate diamonds.
Priced utility per unit
- How the market ends up pricing a single unit (a slice of pizza, a gallon of water, a carat of diamond) at a given moment.

Textbooks mostly live on layer two and say: “Given abundance and scarcity, the marginal unit of water is cheap, the marginal unit of diamond is expensive.” That’s not wrong. But it doesn’t ask how these prices interact with life’s recurring reality.

Food and water are not one‑time purchases. They are constant survival subscriptions.

The recurring burden of essentials

A typical family of four might easily spend 13,000–19,000 dollars a year on food at home and away from home, in dozens of small trips—20 dollars here, 100 dollars there—just to keep everyone fed (U.S. Department of Agriculture, 2025). Historically, American households often spent a quarter or more of their budgets on food; in some periods and income groups, it was closer to 40–50 percent (Bureau of Labor Statistics, 2006). However, after COVID, the BLS Consumer Expenditure Survey began experiencing response‑rate issues, making the more recent data unreliable. For a more accurate picture today, using the USDA food plans for a family of four and comparing them with current median household income, that food share still lands somewhere around 15–22 percent for many families—hardly a trivial line item (U.S. Census Bureau, 2025).

If that share had remained near those higher levels, almost all household budgets would have been trapped in survival mode. Very little would be left to support demand for anything else:

· Better housing

· Education

· Entertainment

· New technologies or services

As the food share falls, something critical happens: you free up a large slice of the paycheck to be spent on other goods and services. That liberated spending is what allows the rest of the economy to exist and grow. Cheap essentials are not a quirk of the system; they are a prerequisite for a complex, innovative economy.

So:

· Functional utility: Food and water are vastly more important than diamonds.

· Priced utility per unit: Because food and water must be purchased constantly, the system must keep their unit price low enough that the lifetime bill is survivable. Diamonds can be priced absurdly high because they are rare, optional, and non‑recurring.

· Saving: Essential spending never stops, while diamond (non‑essential) spending can be spread over time or skipped entirely to reduce the cost burden.

Food doesn’t cost less than diamonds because it matters less. It costs less per unit because it matters too much and shows up too often. Markets and politics conspire to keep the price of essentials down so households can survive their enormous lifetime cost and still have enough left over to drive demand for the rest of the economy.

Why do people trade diamonds for water in a crisis

Now look at a crisis: drought, siege, contamination, war.

We regularly see people behave in ways that make the simple “abundance vs. scarcity” story feel thin:

People will trade jewelry, gold, or anything portable for water and food.
The willingness to pay for the next gallon of drinkable water explodes, even before the last gallon disappears.
Fear of not finding more tomorrow becomes just as important as actual scarcity today.

Water’s functional importance hasn’t changed. What has changed is that the normal system—productivity, infrastructure, logistics—that keeps water priced cheaply per unit has broken down or is perceived as fragile. Once that protective shell cracks, the underlying hierarchy asserts itself:

Food and water jump to “priceless.”
Diamonds suddenly look like rocks you’re willing to dump to secure the basics.

Standard marginal utility can always say “the marginal utility curve shifted in a crisis.” But it doesn’t tell you why essentials should be able to flip the entire price hierarchy during bad times and then return to being “cheap” during good times.

The two‑layer view does:

In normal times, productivity and stable access let us suppress the priced utility per unit of essentials, even though their functional utility is highest.
In a crisis, that suppression fails, and prices snap toward functional reality. People trade diamonds for water because they have always lived in a world where water’s true importance has dwarfed that of diamonds; a crisis just removes the illusion.
Subscribe now

Where the 1.5% inflation drift hides

Once you see essentials as constant survival subscriptions whose per‑unit prices are deliberately kept low, a different question pops up: what if our main yardstick for “how expensive life has become” is quietly undercounting that process?

In Novack Equilibrium Theory, I argue that official CPI has understated real inflation by roughly 1.5 percentage points per year for over a century. That small‑sounding gap compounds into something huge:

The dollar has lost closer to 99.36 percent of its value since 1913, not “just” ~96.8 percent.
A century‑long deflationary boom in productivity has been misread because CPI was calculating inflation on numbers that have already been adjusted by the market, reducing measured inflation
Essentials look more expensive than they were in the past, but they really aren’t in time-adjusted terms.

Now combine that with the dynamics above.

If technology and productivity are continually driving down the real cost of essentials (food, basic energy, etc.), then any persistent mismeasurement of inflation will show up most painfully in exactly those categories. It’s also where the strain runs in both directions: the food industry needs higher prices to keep quality and stay solvent, while households—whose real incomes have quietly eroded—feel like even basic items are slipping out of reach. We see the sticker price of a pound of ground beef jump from about $ 2 25 years ago to around $6 today and think “nothing changed and the product is worse” (Bureau of Labor Statistics, 2025). In reality, a broken inflation yardstick and falling real wages are colliding right at the grocery checkout.

Why?

Essentials are recurring; you can’t avoid them.
Their true productivity‑driven price path is downward.
Any measurement error that hides deflation becomes a “food puzzle,” and an affordability crisis in the very goods people buy every week.

So if you want to find the distortion from a 1.5 percent CPI drift, don’t start with luxury goods or speculative assets. Start with groceries, gas, and rent. That’s where you are forced to live with the broken measuring stick every month.

Oil shocks as a live case study

This is exactly why we emotionally overreact to oil shocks.

When a war threatens oil supply, your brain doesn’t picture a barrel of crude in a textbook diagram. It pictures:

· The cost to fill your tank

· The impact on shipping and trucking

· The ripple into the price of everything transported, especially food

In my earlier Substack on the current war involving Iran, I argued that even a serious conflict is unlikely to generate sustained inflation in the near term, because the monetary conditions are not right to allow inflation. Businesses will be forced to change their purchasing decisions to sustain current prices. This doesn’t mean short‑term price spikes can’t happen, but once you correct the measuring stick, this new spike is only a bump; the long‑run downward pressure on essentials still comes from productivity.

But the public reaction to those spikes—the panic, the political noise—is a real‑time referendum on this functional vs. priced utility story. People know, at a gut level, that when oil rises, it threatens the affordability of the recurring essentials they care about most. They don’t want to spend more of their finite human time and income just to stand still.

Bringing it together

Adam Smith’s Diamond–Water paradox was really about this question:

Why does a good with extremely high real‑world importance (water) carry a lower price than a good with almost no real‑world importance (diamonds)?

The textbook marginal‑utility fix—“price is about the marginal unit, not total importance”—is logically correct but emotionally unsatisfying. It never really connects to how people live, budget, and panic.

A more complete answer is:

Food and water have the highest functional utility and are bought constantly, year after year. To keep households and the wider economy from being suffocated by survival expenses, the system must keep its per‑unit prices low.
Diamonds have trivial functional utility and are rarely bought. Their per-unit price can fluctuate to extreme levels without threatening anyone’s survival or the broader economy.
In normal times, productivity and stable access suppress the priced utility of essentials; in crises, that suppression cracks, and their true priority is revealed when people trade diamonds for water.

If you want to see how this plays out in real time, with a shooting war and an energy shock that still don’t add up to runaway inflation, go read my Iran war piece after this one. It’s the same story from a different angle: oil as the gateway to understanding why we hate paying more for the things we can’t stop buying.

In the next piece, I dig into the Food Puzzle: why, given everything we know about productivity, official food prices haven’t collapsed nearly as much as they should have—and what that says about the inflation numbers we’ve trusted for a century.

Reference

· Bureau of Labor Statistics. (2006). 100 years of U.S. consumer spending: Data for the nation, New York City, and Boston (BLS Report 991). U.S. Department of Labor

· Bureau of Labor Statistics. (2025). Average retail food prices: Ground beef, 100% beef, per pound, U.S. city average [Data set]. U.S. Department of Labor. https://data.bls.gov/pdq/SurveyOutputServlet

· U.S. Department of Agriculture, Food and Nutrition Service, Center for Nutrition Policy and Promotion. (2025). USDA Food Plans: Cost of food at home, monthly reports [Data set]. https://www.fns.usda.gov/research/cnpp/usda-food-plans/cost-food-monthly-reports

· U.S. Census Bureau. (2025). Income in the United States: 2024 (Current Population Reports, P60‑286). U.S. Department of Commerce. https://www.census.gov/library/publications/2025/demo/p60-286.html

Author: Kyle Novack

March 13, 2026

A Monumental Venture, LLC: research project (Novack Equilibrium Theory – NETs)