For years in my articles and on my radio show I’ve been explaining how US inflation numbers–both CPI and PCE but especially the latter–have been grossly under-reporting inflation in the US. The reasons have to do with the methodologies, and assumptions in those methodologies, calculate price increases. The government keeps indicating prices are rising only a few percentages over the year. The PCE is up only 3.4%. Everyone knows, however, through their personal experience many categories of prices are rising far faster, and not just energy and food but transportation, utility services, insurances and interest rates (i.e. the ‘price’ of money) as government rates (benchmark 10 yr. bond) were up 19% in one month in August 2026 alone! Mortgage rates have risen from 5.9% at the beginning of 2026 to 7.3% today.
Each category of prices are often calculated differently. Using different base year, different weights, different data gathering. Often prices are just ‘inputed’ by government bureaucrats who do the monthly inflation calculation. That is, just made up and plugged into the totals. A good example is the estimation of half of the total price of rents. In other cases, the bureaucrats assume a price for a good has declined when consumers in real time actually pay more. Prices for PCs, Iphones and other electronics and even new cars to some extent are reduced this way.
This bureaucratic manipulation of real inflation is exacerbated by government shutdowns when no data gathering occurs for weeks at a time, as during the fall 2025. When shutdowns are over the government plugs in a number based on pre-shutdown assumptions of trends. Post-shutdown periods price changes are reduced in turn due to imputed plug ins.
Another government shutdown has just begun October 1 in part. Data gathering will be impacted again. More plug in of inflation numbers will follow.
The low-balling of the PCE index has further consequences. For example, it is used to estimate real GDP. The lower the PCE, the higher real GDP and economic growth is reported. US GDP for third quarter 2026 is suspended again due to the current shutdown and suspension of price data gathering. If GDP is eventually reported as rising at a 2% rate, for example, then the low-balled PCE inflation adjustment, off by 1%-1.5% in my opinion, results in an actual GDP of about 0.5% to 1.0%. The economy is nearly stagnating, in other words. The only growth going on, everyone knows, is in AI business investment in data centers, chips, and agents.
A member of my audience recently asked me a very typical question. A colleague of his told him prices were rising only 3.4% according to the government (PCE annual rate). Is that true, he asked, when he himself feels he’s been paying higher prices across the board? Here’s my explanation to him which clarifies some of the questionable ways the government estimates different categories of prices to get the PCE:
Your colleague clearly is satisfied relying 100% on official government statistics re. inflation. Apparently the PCE index around 3%. The CPI is around 4%. Both low ball the actually inflation rate for consumers by 1-2% at minimum. The PCE partly relies on CPI data and cherry picks that and then estimates the rest by mashing together 20 other price sources. It also uses what’s called ‘chain pricing’. That changes the base period every year, which results in lower inflation.
The PCE is not the result of any actual survey of prices. It doesn’t ask consumers what they bought and prices they paid. It asks businesses what they sold. It starts from business sales revenue and extrapolates back to estimate prices employing questionable methodologies and assumptions designed to low ball prices.
For example, did you know the PCE estimates insurance costs by extrapolating from insurance company profits (which the companies under-report)? That it estimates airline prices not by what you and I pay for a ticket but by passenger miles flown. Or utility electrical services by KW hours or therms? Or that it states that rising prices for PCs, Iphones, laptops and even new cars are consistently falling not rising every month? You and I pay the higher price in fact. Forget that. The government bureaucrats say that rising features and functionality mean that prices have fallen. That’s called hedonic pricing. A bureaucrat’s euphemism to say higher prices are actually lower.
And how about interest rates. That’s the ‘price’ we all pay for mortgages, auto loans, credit cards, student loans, etc. right? And that price is escalating isn’t it? But wait. Interest rates (price for money) is not included in the PCE, or CPI. Nor are rising local government fees and taxes.
How about housing inflation? PCE doesn’t include either rising mortgages or prices of home sales. It also reduces rent costs by plugging into total rent inflation an assumed low rent that home owners pay themselves. In other words, it employs the oxymoron idea that 40 million+ home owners allegedly pay themselves a rent. The plug in number they give is arbitrary and lower than the rent actual renters pay. The two estimates are then combined to get a rent price. That lowers the total denominator for actual rent real renters pay their landlords. So the rent price comes out lower. Also, rents hikes are only counted in a new rental contract. If your landlord raised your existing rent by 20% this year, that’s not counted as a rent hike at all in the PCE or CPI.Moreover, the rent inflation numbers the PCE indicates are lagged from six months earlier, not current actual rent hikes. Rent prices are surveyed only twice a year, not monthly.
Are you paying more for beef hamburgers or steaks? Not as much as you think. The bureaucrats assume to change from beef to chicken when the price of beef get too high. That lowers the price of beef they report as a price hike for beef.That’s one of the 20 other price sources they use to calculate the PCE, called the Paasche substitute price index.
PCE assumes different weights than CPI for which goods and services we buy most. And how about drug prices? Drug companies simply move a particular drug into a higher range of their formulary. You and I pay more but it’s not a price increase per PCE. And price of gasoline at the pump? It’s a national average for regular gas, not the unleaded premium we have to buy.
I could go on with numerous further examples of how the assumptions and methodology in both the inflation indexes are constructed so to reduce the general price level reported.
Tell your colleague who believes what the government tells him to start thinking for himself. Investigate the methods and assumptions in the government inflation indexes and he’ll find some surprising stuff that might change his mind that the government always tells the truth and the mainstream media accurately always reports the same!
Dr. Jack Rasmus @drjackrasmus
