The economy grew last quarter. Your rent went up, your grocery bill went up, and your paycheck did not. Both can be true at once. The gap between the national scoreboard and your bank account has a handful of specific causes.
The Average Is Not the Middle
Most of the confusion starts with one piece of arithmetic. An average, or mean, adds everything up and divides by the number of people. A median lines everybody up from smallest to largest and reads off whoever is standing in the middle.
Take an illustrative town of ten workers. Nine earn $40,000 a year and one earns $400,000. Total income is $760,000, so the mean is $76,000. The median is $40,000, because that is what the people in the middle of the line earn.
Now the top earner's income rises to $700,000 and nobody else gets a raise. The mean jumps to $106,000, up 39 percent. The median has not moved a dollar. A headline reading "average income up 39 percent" would be completely accurate, and nine of the ten people in that town would not recognize it.
Suppose the nine slip to $38,000 while the top earner still climbs to $700,000. The mean is $104,200, still up 37 percent, and the median has fallen. Average up, middle down, both real.
GDP Per Person Is Still an Average
Total GDP rises when the population rises, even if no individual is better off. Dividing by population fixes that, and GDP per capita (total output divided by every person in the country) is a better number for asking whether a typical slice got bigger.
It is still a mean. Splitting output evenly on paper says nothing about where the money landed. A country can post rising real GDP per person for a decade while the median household's income barely moves. The Census Bureau reports both a mean and a median household income, and the mean sits well above the median for the same reason.
A Raise Below Inflation Is a Pay Cut
Real and nominal do the same job for your paycheck that they do for GDP. Nominal is the number printed on the check. Real is what is left after prices are taken out.
Say you earn $50,000 and get a 3 percent raise, taking you to $51,500. Over the same year, prices rise 4 percent. The pile of things you used to buy for $50,000 now costs $52,000. You are earning $51,500 against a $52,000 bill, so in real terms you took a pay cut of about 1 percent. Every conversation you had about it used the word raise.
That is why economists track real wage growth rather than the raise itself. Hiring can be strong and raises can be widespread while real pay stays flat, because the raises and the prices were racing and the prices won.
Slower Inflation Does Not Mean Cheaper Prices
This is the single biggest source of the gap, and it comes from mistaking a level for a rate of change.
The inflation rate measures how fast prices are climbing, not how high they are. When the rate falls from 7 percent to 2 percent, prices are still rising. They are rising more slowly. Nothing has been handed back.
On Bureau of Labor Statistics annual averages, the consumer price index was roughly a fifth higher in 2024 than in 2020, and the slower rate that followed undid none of it. So when the news says inflation has cooled and a shopper says everything is expensive, the two are measuring different things. The rate describes the last twelve months. The shopper is comparing today against a level they remember from years ago, and the index as a whole almost never falls back to an old level, even when individual items do.
The Basket in the Index Is Not Your Basket
The headline index prices a fixed basket, and the weight on each item comes from what the average household spends. Your weights are your own.
Shelter is the heaviest single piece of the US index, more than a third of it by the Bureau of Labor Statistics relative importance weights. Picture a renter in an expensive city with no car and a childcare bill. Rent and childcare are most of where their money goes. Those categories can climb far faster than the overall index while the index still reports a mild number, because it is averaging their rent against televisions and computers, which have been getting cheaper for years.
A student is the sharpest version. Tuition and rent are close to the whole budget, and neither is the average household's biggest line. The index is not wrong. It is answering a question about the average household, and you are not the average household.
Income Is a Flow, Wealth Is a Pile
Income is what arrives each month. Wealth, or net worth, is everything you own minus everything you owe. The two move differently.
Rising asset prices grow wealth for people who already hold assets. A strong stock market or a jump in home values can add enormously to measured wealth without adding a dollar to anyone's paycheck. If you rent and own no investments, a year that made the country richer on paper did nothing for you, and if it pushed your rent up it left you worse off.
Wealth is also far more concentrated than income, so the mean and the median separate much more sharply. In the Federal Reserve's 2022 Survey of Consumer Finances, median family net worth was $192,900 while mean family net worth was $1,063,700. The mean is more than five times the median. Income is nowhere near that lopsided.
How Inequality Actually Gets Measured
Two tools do most of the work. The first is income share by percentile. Sort every household by income, cut the list into five equal groups (quintiles) or into the top 1 percent and everyone else, then ask what fraction of all income each group received. If the top fifth takes half of all household income, that is a sentence anybody can check.
The second is the Gini index, also called the Gini coefficient. It is one number between 0 and 1. The Census Bureau defines 0 as perfect equality, where every household has the same income, and 1 as perfect inequality, where a single household has all of it. Real countries land in between, and higher means more spread out.
Its strength is squeezing a whole distribution into one comparable figure. Its weakness is the same trait. Two very different distributions can share a Gini, and the number cannot say whether the gap opened at the top, at the bottom or in the middle. That is why the Census also publishes plain ratios, such as income at the 90th percentile divided by income at the 50th.
Statistics Turn Before People Do
The last piece is timing. A recession's end date is set by the National Bureau of Economic Research, and it marks the month activity hit bottom, not the month life got better.
The 2007 to 2009 recession officially ended in June 2009. Payroll employment kept falling until February 2010, eight months later, and the unemployment rate did not peak until October 2009. Anybody job hunting through that stretch was told the recession was over while the labor market was still getting worse.
Losses and gains are not symmetric either. Employers cut 8.7 million jobs between the January 2008 employment peak and the February 2010 bottom, about two years. Climbing back to that January 2008 level took until May 2014. Jobs vanish quickly and return slowly, so people spend far longer inside the recovery than inside the downturn, and a recovery feels like nothing until it reaches you.
