Methodology essay
A candid catalog of the inflation vectors we know exist but don't yet quantify. Reverse hedonics, time costs, hidden fees, healthcare access, educational outcomes. What gets missed when you measure prices in dollars.
The Reality Index is built to be more honest than the official Consumer Price Index about what families actually pay. But it is not the complete picture, and pretending otherwise would make us no better than the indexes we criticize.
This essay catalogs what Reality Index does not measure — at least not yet. Some of these vectors are conceptually clear but methodologically difficult. Some are measurable in principle but require data we haven't yet assembled. Some are arguably outside the scope of any price index, but worth naming because their absence shapes how you should read our headline number.
The case for transparency about this is simple: the moment Reality Index claims to capture more than it does is the moment it deserves the same skepticism we apply to CPI. A 32% cumulative gap over 45 years between Reality Index and CPI is what we can defend rigorously. Whether the true gap is larger — and where the additional inflation hides — depends on how seriously you take what follows.
The Bureau of Labor Statistics adjusts CPI for quality changes. The methodology, called hedonic adjustment, treats a faster computer at the same dollar price as a price reduction. The argument: you are paying the same dollars for more value, so real-world inflation is lower than the dollar figure suggests.
The methodology is asymmetric. It is applied aggressively when quality improves. It is rarely applied when quality declines — what economists call reverse hedonics. The result: when the dollar price of a good stays flat but its real-world value diminishes, CPI records no inflation. The dollar at the register did not change. The thing in the bag did.
Specific examples where this is documented:
Eggs. The egg you buy at the grocery store today is not the egg of 1970. Industrial-scale egg production has shifted the typical retail egg toward grain-fed caged hens, which produces eggs with measurably lower omega-3 content, lower vitamin D and A levels, and different yolk-to-white ratios than pasture-raised eggs. A 2007 study by Mother Earth News testing eggs from 14 pastured flocks found that compared to conventional supermarket eggs, pasture-raised eggs contained roughly two-thirds more vitamin A and twice the omega-3 fatty acids. The findings were replicated in a 2010 peer-reviewed study by Karsten et al. in Renewable Agriculture and Food Systems. CPI tracks the dollar price of "a dozen large white grade A eggs" — and that price has indeed tracked CPI roughly. But the eggs themselves are not equivalent goods across time, and CPI is structurally unable to capture this.
Chicken. The supermarket chicken breast of 2025 is mass-produced broiler chicken, typically raised in 47 days from hatch to slaughter on antibiotics and growth-optimized feed. The chicken breast of 1980 was raised over 70-plus days under different agricultural standards. Each pound is the same pound. The thing being purchased is meaningfully different.
Beef. Grain-finished feedlot beef has substantially different nutritional profile and fat composition from grass-finished pasture beef of fifty years ago. The price of "ground beef per pound" has been tracked precisely. The category itself has shifted.
Air travel. The square footage per passenger in an economy seat has decreased materially since the 1970s. Seat pitch has compressed. Free meals, free checked bags, and free changes are now unbundled fees. The price of a flight from Chicago to New York can be lower in nominal dollars than the same flight in 1985 — and CPI records this as deflation in airfare. The actual experience of flying has degraded. Time at the airport increased. Boarding processes added new friction. The product called "a flight" has changed; the comparison across decades is no longer apples to apples.
Where we have specific retail dollar data, Reality Index tracks it. Where the underlying product has changed, we tend to flag this in the individual chart pages but cannot reflect it in the headline number. We don't have a multi-decade nutrient-density-per-dollar series for industrial eggs. We don't have a square-feet-per-passenger-mile index for air travel. These data could exist; they don't.
There is a deeper measurement issue that the Reality Index, like CPI, does not address. Both indexes measure inflation in dollars — the very currency whose purchasing power is being measured. If the dollar itself is losing real-world value relative to durable stores of value, then a dollar-denominated inflation index will systematically understate the loss.
The familiar reference point: a quality men's suit cost roughly one ounce of gold in 1900. A quality men's suit costs roughly one ounce of gold today. A barrel of crude oil cost roughly 1/15th of an ounce of gold in 1970. A barrel of crude oil costs roughly 1/30th of an ounce of gold today. Measured in gold, suits have actually deflated and oil has roughly held its value. Measured in dollars, both have inflated severely.
Whether the dollar is the right unit of account for measuring "inflation" is a serious economic question we do not attempt to resolve. The Federal Reserve operates in dollars. American wages, contracts, taxes, and Social Security adjustments are all denominated in dollars. The Reality Index follows the same convention, both because it is operationally necessary and because doing otherwise would invite the criticism that the project is gold-hard-money advocacy in disguise. Reality Index is not.
But it is worth naming the limitation: Reality Index measures inflation against a yardstick (the dollar) that is itself a variable. Critics who argue that real-world inflation is larger than any dollar-denominated index can show have a methodological point. We don't argue against it. We just don't quantify it.
The price of a good in dollars is not the only cost of acquiring it. Time is the other major input, and time costs have moved meaningfully in some categories without being captured by any inflation index.
Healthcare scheduling has shifted toward longer wait times for primary care appointments in much of the country. Customer service has shifted from staffed phone lines to automated systems with substantial hold times. Retail has shifted from full-service to self-checkout, with the customer providing labor that was previously included in the price of the good. Air travel has added security wait times, boarding procedures, and baggage handling friction that did not exist in 1980.
None of this appears in CPI or in Reality Index. The dollar price of the good or service may be flat or even declining. The hours-of-your-life price has risen.
This is the category Reality Index will likely incorporate in v2, because data does exist if assembled carefully. Shrinkflation — the practice of reducing package size while holding the dollar price constant — has accelerated since 2020 and is empirically tracked by reporters like Edgar Dworsky at ConsumerWorld.org and by ProPublica. A 16-ounce yogurt that became 12 ounces at the same price represents 25% inflation that CPI's BLS APU food series does pick up (because they track per-unit-weight prices). But many products are not tracked at the unit-of-volume level.
Hidden fees — resort fees, "facility fees" in healthcare, "convenience fees" at concert and event ticket purchase, dynamic pricing surcharges — function as inflation that does not appear in the sticker price of the good. CPI tracks sticker prices, not the all-in price including ancillary charges. Reality Index inherits this limitation for the categories where we use BLS subindexes.
Reality Index measures the family health insurance premium, using the KFF Annual Employer Health Benefits Survey, which is the best independent data source on what families and employers actually pay. This is meaningfully better than CPI's medical care subindex, which tracks provider list prices.
But premium is not the same as access. Over the past 25 years, the average deductible on family employer coverage has roughly quintupled. The average out-of-pocket maximum has more than doubled. Insurance denial rates have increased. Network narrowing has reduced the number of in-network providers for specialty care. Prior authorization requirements have expanded.
What families experience as healthcare cost is not just the premium they pay. It is what they pay to actually receive care after the premium. Reality Index measures the premium. The other costs are real and growing, and we do not currently capture them.
Reality Index measures the dollar cost of college tuition using NCES public 4-year tuition data. The price has grown substantially faster than CPI. This is the headline number on our tuition chart.
But what families are buying with that tuition has also changed. The labor market value of a bachelor's degree has compressed in real wage terms relative to the 1970s. The proportion of college graduates underemployed (working in jobs not requiring a degree) has risen. The signaling value of the credential has declined as more workers obtain it.
Tuition price per credit is up. Wage premium per credit is down. Reality Index measures the first. The second is real inflation in the value of what families purchase, and we do not capture it.
Several of the items above are candidates for incorporation in future versions of Reality Index as data and methodology improve. Shrinkflation is the highest-priority addition; the data exists in BLS APU per-unit pricing if we assemble it properly. Hidden fees are tractable in some categories (resort fees, healthcare facility fees) if we partner with industry analysts who collect that data. Healthcare access metrics — deductibles, out-of-pocket maxima, denial rates — are publicly available from KFF and CMS, and a Reality Index Healthcare Access subindex is on the v2 roadmap.
Reverse hedonics is the hardest. It requires multi-decade quality-per-dollar data that does not generally exist for consumer goods. We can document it editorially. We can flag it in individual item pages. We will likely not be able to quantify it rigorously enough to incorporate into the headline number for a long time, if ever. We acknowledge it because the alternative — pretending the eggs have not changed — is the same intellectual error CPI commits.
The point of cataloging what we do not measure is not to walk back what we do measure. The 32% cumulative gap between Reality Index and CPI over 45 years is what we can defend with rigor. The actual gap families experience is, by every honest indication, larger than that. How much larger is exactly the right question. Reality Index v1 establishes the conservative floor. Future versions will narrow the range.
This essay is part of the Reality Index methodology series. For the headline inflation rate and the underlying chart, see the headline rate. For the full methodology paper, see methodology_paper.html. For citation conventions and press inquiries, see for journalists, analysts, and researchers.
Suggestions, corrections, and additional unmeasured-inflation vectors are welcome at info@realityindex.co.