Is Middle-Class Life Harder Than It Used to Be?

In 1984, the median American household earned about $22,400 a year. A new home cost around $79,600, about three and a half years of gross income. Not easy, but a young couple saving for a few years could realistically get there. A year of public university tuition ran about $1,200, which a student could cover with a summer job. Health insurance, if your employer offered it, was a rounding error on your paycheck.

By 2024, median household income had climbed to $83,700, a 273 percent increase over four decades. But the median new home price hit $420,400, now five years of gross income instead of three and a half. College tuition at a public four-year university crossed $11,000, up over 750 percent. Employer health premiums for a family plan topped $25,000. Income nearly quadrupled, but housing, healthcare, and tuition grew faster. (These are new-home prices throughout. The existing-home median is lower, around $390,000, but follows the same trend relative to income.)

Price growth by category since 1984

Since 1984, the overall cost of living as measured by the Consumer Price Index has risen about 202 percent. Food has kept pace with that average. Medical care is up 427 percent. College tuition is up 753 percent.

The chart separates into two groups. Manufactured goods like apparel and vehicles got cheaper in real terms. Services that depend on human labor or fixed land, housing, education, and medical care, got more expensive. A dollar of college tuition in 1984 costs over eight dollars today.

CPI blends all of these into a single number and presents it as “the” inflation rate. But households don’t spend their income on the CPI basket. The BLS Consumer Expenditure Survey shows that middle-income households spend 40 percent of pre-tax income on just housing and healthcare. For the top quintile, those shares are about half as much. The categories where prices fell or tracked inflation, apparel, electronics, new vehicles, are a small fraction of a middle-class budget. The 202 percent average understates the effective inflation rate for households whose spending concentrates in the categories that rose fastest.

Economist Xavier Jaravel built distributional consumer price indices that weight CPI price changes by each income group’s actual spending. Middle- and lower-income households experienced measurably higher inflation than upper-income households over the past two decades, with the gap driven by housing and transportation costs. A 2024 BLS study in the Monthly Labor Review confirmed the finding: the lowest income quintile faced 0.28 percentage points higher annual inflation than the highest from 2006 to 2023. Different households face different effective inflation rates depending on what they buy.

The divergence has a structural explanation. Economist William Baumol identified the mechanism in the 1960s: labor-intensive services like teaching and medicine can’t improve productivity the way manufacturing can, but their wages must track economy-wide gains or they lose workers to other sectors. Education and healthcare costs structurally rise faster than manufactured goods. Housing adds a separate constraint: fixed land supply in the metros where jobs concentrate.

The divergence started in the late 1980s and accelerated through the 2000s. College tuition was already up 202 percent by 2000, but between 2000 and 2010 it more than doubled to 481 percent above the 1984 baseline. State funding cuts to public universities during and after the 2008 recession drove much of that acceleration.

One asterisk: the Shelter line looks like it tracked general inflation at 223 percent. The BLS Shelter index is built from “owners’ equivalent rent,” a modeled estimate, not actual home prices or mortgage costs. Actual rent paid by tenants in urban areas is up 299 percent since 1984, and home purchase prices relative to income are up 40 percent. The CPI Shelter line is technically correct for what it measures, but it measures something different from the home prices and rents that households actually pay.

Labor cost of major expenses

In 1984, a worker earning the median hourly wage (the middle of all earners, about $8.50) could cover the total cost of attendance at a public university, tuition, fees, room, and board, with roughly 16 weeks of full-time work. By 2024, at a median hourly wage of about $23, the same year of school costs 31 weeks of full-time work. Tuition alone tripled from 3.5 weeks to 12 weeks of labor, but the full sticker price nearly doubled as a share of your working year.

Healthcare stands out: a 7x increase in the labor cost of staying insured. Food was the only category that stayed roughly proportional to wages.

College costs, debt, and earnings by institution type

The standard defense of rising tuition is that college is an investment: pay more now, earn more later. Federal College Scorecard data shows the return varies wildly by institution type. Ten years after enrollment, the gaps are large. Public university graduates earn $45,400 a year and carry $14,600 in debt. Private nonprofit graduates, think traditional universities like NYU or small liberal arts colleges that reinvest tuition into the school, earn $54,200 but paid $46,700 per year for the privilege. Graduates of private for-profits, investor-owned chains like University of Phoenix and DeVry, earn just $33,300 with $12,200 in debt, barely above what they’d make without a degree.

Even at public universities, the best-value option, the math has gotten tighter. A student borrowing $14,600 and earning $45,400 a decade later is better off than someone with no degree, but the margin is thinner than a four-year commitment would suggest. One caveat: these are sticker-price figures. Many students at public institutions receive Pell Grants, state aid, or institutional scholarships that reduce the actual cost by 30 to 50 percent. The CPI tuition figure of 753 percent tracks published tuition; net tuition after aid has risen less, though still substantially.

The New York Fed tracks what a degree actually buys in constant dollars. Since 1990, real wages for recent bachelor’s graduates have grown about 6 percent. Tuition went up 753 percent to buy that. The raw comparison also overstates the degree’s causal effect on earnings, since bachelor’s holders differ from high school graduates in family background, test scores, and other factors that independently predict higher wages. For most of this period, the degree’s value came less from higher earnings and more from avoiding the wage declines that non-degree workers experienced.

The degree’s unemployment protection has also eroded. NY Fed labor market data shows that young degree holders (ages 22-27) still have lower unemployment than same-age workers without degrees, but the gap has narrowed from about 4 percentage points in the 1990s to under 2 today. By 2024, recent graduates hit 4.8 percent unemployment, above the 3.9 percent rate for the overall workforce, a crossover that hadn’t happened before 2019. (Some of that crossover reflects age, not education: young workers always churn more than the full labor force. The same-age comparison is the fairer test, and there the degree still helps, just less than it used to.) About 40 percent of recent graduates work in jobs that don’t require a degree, a share that hasn’t meaningfully changed in three decades.

The federal student loan portfolio now stands at $1.5 trillion in outstanding debt across 43 million borrowers, roughly one in six American adults. The average borrower owes about $36,000. And the debt skews older than the stereotype: borrowers aged 35 to 49 carry more of it than any other age group.

Most of these borrowers are paying off degrees they earned a decade or more ago. About 14 percent are in default. The total portfolio grew 145 percent in just 12 years.

Mortgage rates and home prices

On a $400,000 house, the difference between a 3 percent and 7 percent mortgage rate is $700 a month. Home prices get the headlines, but for most buyers, the interest rate determines whether the monthly payment is manageable.

Falling mortgage rates from 1980 through 2020 helped fuel home price inflation. When rates dropped from 13 percent to 3 percent, a buyer could afford a much more expensive house for the same monthly payment, which meant buyers could bid higher and sellers could ask higher. That 40-year decline in rates kept monthly payments manageable even as median home prices more than tripled in nominal terms.

Between 2020 and 2023, rates more than doubled from 2.96 to 6.81 percent. On a $400,000 house with 20 percent down, that rate change adds roughly $700 per month to the mortgage payment, about $8,400 per year. For a household earning $83,700, that’s an extra 10 percent of gross income from the rate change alone.

A fair objection: mortgage rates were even higher in 1984 (over 13%), so monthly payments as a share of income were actually worse then, about 40 percent of gross income versus 31 percent today. But in 1984, the house itself cost 3.6 years of income. High rates on a small principal hurt, but the underlying asset was reachable, and when rates eventually dropped, refinancing a $64,000 mortgage made a real difference. Today’s buyers face a different situation: the brief 3 percent rate window in 2020-2021 fueled a roughly 40 percent surge in home prices as cheap money let buyers bid far above what homes were previously worth. Then rates climbed back to 7 percent, but the prices never came back down. Anyone who didn’t buy during that window now faces pandemic-era prices and post-pandemic rates simultaneously.

After the 2008 crash, the US entered a decade of underbuilding. Housing starts collapsed from 2.1 million in 2005 to 554,000 in 2009 and didn’t recover to pre-crisis levels for over a decade. Between 2009 and 2019, the country built roughly 3 million fewer homes than population growth demanded. That housing deficit is a major reason prices stayed elevated even as rates were historically low, and why the post-pandemic price spike met almost no slack in supply.

Housing, tuition, and healthcare relative to income

Each of these costs grew faster than income, but at different times and by different magnitudes.

The home price-to-income ratio rose 40 percent since 1984, from 3.6 years of income to 5.0. (A caveat: median new homes have also gotten larger, from about 1,600 to 2,300 square feet. Price per square foot tells a less dramatic story, though still an upward one.) That national ratio averages over wide geographic variation. About half the states still cluster near the 1984 national baseline of 3.6 or below. The affordability crisis is concentrated in the coastal and Mountain West states where housing supply is most constrained.

Health insurance has outpaced both housing and tuition as a share of income. The employer pays about 73 percent of the total family premium, so the household’s direct out-of-pocket share looks modest. But it’s all compensation: every dollar an employer spends on your health plan is a dollar that doesn’t show up in your paycheck. The KFF Employer Health Benefits Survey shows that health premiums consumed about 8 percent of total worker compensation in 1988 versus 18 percent by 2023. Had premiums held at the 1988 share, the average worker would take home roughly $8,800 more per year.

The cost also binds the middle class to employment in a way it doesn’t bind the wealthy. Individual-market coverage for a family runs $20,000 to $30,000 a year. ACA marketplace subsidies phase out for middle-class incomes, so a household earning $80,000 faces near-full premiums on the exchange. A middle-class family with children cannot realistically leave employer coverage, which constrains self-employment, career changes, and early retirement in ways that don’t show up in any price index.

Part of that increase reflects genuinely better medicine, not just higher prices. Cancer five-year survival rates improved from roughly 50 to 68 percent since the mid-1980s. HIV went from a death sentence to a manageable chronic condition. Cardiac mortality dropped by more than half. Some of the 427 percent CPI increase bought real improvements in outcomes. But US healthcare spending has pulled away from every peer country since 2000, and the gap keeps widening.

Administrative costs account for a large share of the gap. The US spends more than double any peer country on healthcare administration alone.

For a deeper look at how intermediaries inflate US healthcare costs, see The $47 Drug That Costs $10,000. Tuition relative to income peaked around 2015 and has declined since, even though the CPI chart shows tuition prices still climbing. Both are true. Nominal incomes grew faster in the last decade thanks to a tight labor market and post-pandemic wage growth. State funding for public universities also stabilized after years of cuts. Paychecks started catching up to tuition. Health premiums kept outrunning wages.

The CPI chart earlier shows tuition sticker prices rising 753 percent. This chart shows tuition’s burden on income rising 171 percent. The difference is income growth. Health premiums are the one cost where income has barely made a dent.

GDP, inflation, and household income

In 2024, US GDP hit $29.8 trillion, up from $4 trillion in 1984, a 662 percent increase. About two-thirds of that growth was prices going up. When you strip out inflation, real GDP grew 191 percent, roughly tripling since 1984. Nominal GDP growth includes the part where groceries cost more, which is the dollar shrinking, not the economy producing more.

Since 1984, real GDP per capita roughly doubled while median household income, adjusted for inflation, grew about 39 percent. Part of that gap is structural: GDP includes business investment, government spending, and returns to capital that were never going to show up in household paychecks regardless of how income is distributed. But the income share data below shows a distributional shift on top of the structural gap: the top 1 percent’s share of pre-tax income doubled over this period, and the middle 40 percent’s share fell by 8 percentage points.

Middle-class families also sit between two cushions. Lower-income households get Medicaid, subsidized housing, and Pell grants. Higher-income households absorb these costs easily. Middle-class families pay full price on healthcare, education, and housing while competing for the same inventory as dual-income professional households and investors.

Productivity, compensation, and income distribution

Between 1980 and 2023, real GDP roughly tripled. Median household income, adjusted for inflation, rose only about 39 percent over the same period. (One caveat: median household income is affected by shrinking household sizes and changing demographics. Per-capita income grew more, though still well below GDP.)

For the first three decades after World War II, productivity and compensation moved in lockstep. Around 1973, the lines split. Output per hour kept climbing while real hourly compensation (wages plus employer-paid benefits like health insurance, adjusted for inflation) lagged behind.

The gap’s magnitude depends partly on how you adjust for inflation: the BLS uses different deflators for output and compensation, which widens the visual gap. Using the same deflator shrinks it by roughly half. But the directional finding holds regardless of method. Total compensation including all employer benefits has tracked productivity more closely than cash wages alone, though a gap remains.

The income share data shows where the gap between productivity and compensation landed.

The middle 40 percent of earners (the 50th to 90th percentile, roughly households making $40,000 to $150,000 in today’s dollars) lost about 8 percentage points of national income since 1980. The top 1 percent gained almost exactly that amount, doubling their share from 10 to 21 percent.

Economists attribute the shift to globalization, automation, and the rising share of returns to capital versus labor. The relative weight of each is still debated.

One qualifier: this is pre-tax, pre-transfer income. The progressive tax system and transfer programs (Medicaid, SNAP, the Earned Income Tax Credit) narrow the gap in lived experience. CBO data shows that post-tax, post-transfer income for middle-quintile households grew roughly 50 percent in real terms since 1979, compared to the 39 percent pre-tax figure. The safety net has expanded since 1984, with the EITC, ACA marketplace subsidies, and wider Medicaid eligibility all offsetting some of the cost pressures documented here, particularly near the lower boundary of the middle class. Economists Bruce Meyer and James Sullivan have also shown that consumption inequality (what households actually buy rather than what they report earning) has risen less than income inequality, partly because of those same transfer programs.

If the middle 40 percent had held their 1980 income share, their households would earn roughly 24 percent more today, enough to absorb much of the housing and healthcare cost growth documented above.

These cost pressures are also shrinking the middle class as a demographic category. Pew Research found that the share of American adults in middle-income households fell from 61 percent in 1971 to 51 percent in 2023, and their share of aggregate income dropped from 62 percent to 42 percent. Some of that movement was upward, but the middle is thinning from both ends.

Homelessness and food assistance

HUD’s annual Point-in-Time count hit 771,000 in 2024, an all-time record and 40 percent above the 2016 low. Chronic homelessness (people without housing for a year or more) nearly doubled over the same period. Family homelessness surged 61 percent in just two years (2022-2024).

The reversal in 2016-2017 lines up with when housing costs started pulling away from wages again after a brief post-recession correction. The 2022-2024 spike has multiple drivers: the mortgage rate doubling documented earlier pushed would-be buyers into the rental market and drove rents up; but the end of pandemic-era eviction moratoriums and emergency rental assistance also removed protections that had kept people housed, and HUD attributes part of the family homelessness surge to immigration-driven shelter demand in major cities. Housing costs are the backdrop, but the recent spike is not housing costs alone.

USDA SNAP enrollment data shows 42 million Americans on food stamps as of 2025, above the pre-pandemic 36 million despite low unemployment. SNAP eligibility thresholds expanded during COVID and were slow to unwind, so part of the elevated enrollment reflects policy, not just declining purchasing power. Still, 1 in 8 Americans qualifying for food assistance during low unemployment is consistent with the cost pressures in the data above.

Household spending in 2024

A median household today earns $83,700. Where does it all go? The BLS Consumer Expenditure Survey tracks the flows. In 2024, the average household took in $104,207 before taxes and spent $78,535. A caveat: these are mean figures, not median. The median household earns $83,700, about 20 percent less. That gap itself is a symptom of inequality: the mean gets pulled up by high earners at the top. Even at the higher average, the margins are thin. For a median household, they’re thinner.

The diagram shows what’s left after each layer of obligations: about $550 a month in net savings. That’s at the average income of $104k. A median household earning $83,700 with roughly similar fixed costs has almost nothing left.

Childcare costs

That budget doesn’t include childcare. For the roughly 11 million households with children under 5, add $15,000 per child to the essential costs column.

Childcare barely existed as a major household expense a generation ago because most households had a stay-at-home parent. In 1975, fewer than half of mothers with children under 18 were in the labor force. Most families had a stay-at-home parent, or relied on informal care networks, grandparents down the street, a neighbor watching five kids on the block, arrangements that cost little or nothing. Formal childcare was a minor budget line for most American families.

That changed as housing, healthcare, and tuition consumed larger shares of household income and a single earner increasingly couldn’t cover the basics. By 2024, three in four mothers with children under 18 were working, and two-thirds of married-couple families with children had both parents employed. Elizabeth Warren documented the structural trap in The Two-Income Trap (2003): the second income didn’t create surplus. It was consumed by bidding wars for housing in good school districts, competing against other two-income families, and by the childcare the stay-at-home parent used to provide for free. Her data showed that a dual-income family earned 75 percent more than a single-income family a generation earlier but had less discretionary income after fixed costs. The second income created a new mandatory childcare cost and drove up the price of the asset families were competing for: housing in good school districts.

At $15,000 a year for one child in infant care, childcare alone consumes 18 percent of the median household income. For a single parent earning $40,000, it’s closer to 38 percent. The variation is extreme: about $6,000 a year in Mississippi, over $23,000 in Massachusetts.

The informal care networks that softened this cost in 1980 have largely evaporated. Geographic mobility means grandparents are often in another state. The stay-at-home neighbor isn’t there when every adult on the street is working too. A cost that was close to zero in 1980 is now $15,000 a year, and it arrives during the years when families are also saving for a down payment and paying off student loans.

The rational response to these costs is to have fewer children, or none. The US total fertility rate fell from 2.12 in 2007 to 1.62 in 2023, according to CDC vital statistics, well below the 2.1 replacement rate. The decline accelerated after 2007, the same year housing costs began pulling away from wages again. Survey data consistently shows that the top reasons young adults cite for having fewer children than they’d like are childcare costs, housing costs, and student debt, the same three categories documented above. Fertility has declined in virtually every developed country, including Scandinavian countries with subsidized childcare and free university tuition, so costs are not the only explanation. But in the US, the stated reasons track the specific cost pressures documented here.

Why this article exists

This article is the reason OpenData exists.

A few years ago I was talking to my dad about how it feels harder to build a middle-class life today than it was for his generation (he’s a Boomer). He gave me the standard answer: “You think it wasn’t hard when we were in our 30s? Pick yourself up by the bootstraps, yadda yadda.” I thought he was wrong, and I wanted to prove it with actual numbers.

So I tried to find the data. Not an op-ed, not a pundit’s talking point, but the actual government numbers. The problem was I didn’t even know where to start. I knew the data had to exist somewhere, but which agency tracks inflation? Which one tracks income? Is tuition data from the Department of Education or the Census Bureau? Are mortgage rates a Federal Reserve thing or something else entirely?

It took hours of Googling just to figure out that inflation data comes from the Bureau of Labor Statistics (BLS), that median income is a Census Bureau dataset published through the Federal Reserve’s FRED portal, that mortgage rates come from Freddie Mac (also via FRED), and that tuition data lives at the National Center for Education Statistics. Four different agencies, four different websites, for one question, and that’s just for US-based data.

Then I actually tried to use the data. The BLS publishes CPI series in fixed-width text files with codes like CUSR0000SEEB01. The Census Bureau’s income data comes in Excel spreadsheets with footnotes embedded in the data cells. Freddie Mac publishes mortgage rates as a CSV that changes column order between years. I spent days just finding the right datasets, figuring out which series codes corresponded to which categories, and getting everything into a format where I could compare numbers across sources. The actual analysis, the part that answers the question, took a fraction of the time.

That experience stuck with me. All of this data is free. It’s open. Our tax dollars paid to collect it. But it’s scattered across so many agencies and published in so many inconsistent formats that a software engineer with over a decade of experience couldn’t answer a basic question: is middle-class life actually harder than it used to be?

That’s what OpenData is for. Instead of knowing that “inflation” means the BLS CPI-U survey, you just search “housing inflation” and get ranked results across every agency. Want to compare income over time? Search “median household income”. Mortgage rate trends? “30 year mortgage rates”. Every dataset is normalized into the same format, queryable through one API, and downloadable as CSV or Parquet. It also plugs directly into ChatGPT or Claude, so you can prompt for data-backed insights in natural language instead of writing queries by hand.

Every chart in this article pulls from datasets you can explore yourself on the platform. The question that started as an argument with my dad turned into the product, and the product gave me the answer.

So here’s the math.

Home prices relative to income are up 1.4x since 1984. College tuition takes 2.7 times the share of income it did then. Health premiums show the sharpest increase at 5.7x (using the total premium including the employer share). The increases are uneven, and weighting them equally overstates the aggregate effect since housing dominates household budgets and grew the least. But each individually represents a larger claim on the median paycheck than 40 years ago.

Meanwhile, real GDP tripled, but the gains went disproportionately to the top. The World Inequality Database shows the top 1 percent’s pre-tax income share doubled from 10% to 21%, and the middle 40 percent’s share fell from 44% to 36%. Median wages, adjusted for inflation, grew about 39 percent.

CPI says inflation ran 202 percent since 1984. For a middle-class household whose budget is dominated by housing, healthcare, and education, the effective rate was considerably higher. The middle class pays sticker price on all three: too wealthy for Medicaid and Pell Grants, not wealthy enough for a $25,000 health premium to be a rounding error.