Nominal vs real: the math of feeling poorer after a raise
The second distortion is confusing nominal with real. Nominal wage is the figure printed on your pay slip. Real wage is what that money actually buys — purchasing power. If nominal pay is "the number of bills in your wallet," real pay is "the size of the basket those bills fill." More bills, sure, but if prices climb faster, the basket shrinks.
The formula is shorter than you'd think: real wage change ≈ nominal raise − inflation.
- Pay +3%, prices +5% → about −2% real, purchasing power falls
- Pay +5%, prices +2% → about +3% real, you're better off
- Pay +2%, prices +2% → roughly flat
That first line is exactly "I got a raise, so why is it tighter?" The number on your statement grew, but if groceries, transit, and rent rose faster, what you can buy went backward. The slip smiles; the basket cries. U.S. real average hourly earnings flipping to -0.7% from May 2025 to May 2026 is the measured version of this. The 10th percentile thinned further, to about -0.3% real (around $14.56/hour).
There's a mental layer to it too — what economists call money illusion. People tend to judge money by the visible size of the number rather than its real purchasing power. A fatter paycheck feels like wealth, even if prices rose more and the basket actually shrank. Less a mistake than a natural habit for anyone who lives among price tags every day.
Minimum wage: read old figures next to prices
The same principle detonates most dramatically in the minimum wage. The hourly rate rises 5% in a year while living costs also rise 5% — what happens? A nominal increase, yes, but essentially no real change. To line up nominal against real, you need inflation adjustment: converting amounts from different dates onto the same price basis, usually via the consumer price index (CPI). As a formula: real amount = nominal amount × (base-year price index ÷ comparison-year price index). One point only — measure old money and new money with the same ruler.
So past minimum wages belong strictly in the category of a record. Suppose a country's hourly minimum roughly doubled in nominal terms over some span. You have to subtract how much prices also rose before the real change shows up. A nominal doubling is not a doubling of purchasing power.
Converting between hourly and monthly hides its own illusion. At 10,000 won/hour, 40 hours a week, and about 4.34 weeks a month, monthly hours come to roughly 174 and nominal monthly pay to about 1.74 million won (varies with paid weekly rest and work type). And yet that same 1.74 million fills wildly different baskets in a high-cost city versus a low-cost region.
Putting the three numbers side by side
| Lens | What it reads | Strength | Where it fools you |
|---|
| Mean (nominal) | total ÷ headcount, the slip figure | totals and budgets | extremes pull it up; ignores prices |
| Median (nominal) | the middle person's amount | the "typical" feel | loses the tails of the distribution |
| Real (mean/median) | purchasing power, price-adjusted | honest comparison across time | depends on which index and base year |
The same wage data yields different conclusions depending on which cell you read it in. That's the whole reason +0.8% median real wages in 2025 and -0.7% real average hourly earnings in 2025-2026 only look contradictory: they measure different things (median vs mean) over different windows, each true on its own.
✍️ Operator's note — Don't let an "average salary" stat crush you alone. Honestly, that number might be propped up by a single boss at the top. The "average is 47 million" that floats around drops to the mid-30s as a median, and once you strip out prices and look real, 2025's median was +0.8% while average hourly pay flipped to -0.7% in a year. Same wages, opposite signs depending on the angle. Your position shows up in percentiles, not in one average number.
Caveats and limits
The figures here shift with the measuring body, window, and method. Even "real wages" can flip sign depending on which price index (CPI, core inflation, etc.) and which base year you use. Percentile decompositions are sensitive too — to definitions of regular vs irregular and full-time vs part-time work.
Real wages carry variables beyond prices. Korea's income tax is progressive, so a raise doesn't arrive intact; add the four social insurances (national pension, health, employment, industrial accident) and the pre-tax raise diverges from net take-home growth. A raise thins twice: first through taxes and insurance (nominal to net), then through prices (net to real purchasing power). Cross-country comparisons demand extra care. A simple exchange-rate conversion is half the truth, since the same dollar buys different amounts in different countries, so purchasing power parity (PPP) is often used instead. Add differences in tax and welfare systems, and ranking minimum wages by their headline numbers is almost always an oversimplification. Social feeds, meanwhile, keep surfacing top-end outliers and inflating expectations; layer on survivorship bias — the successful talk, the rest stay quiet — and exposure frequency gets mistaken for actual frequency.
The skill of reading wage numbers sharpens fastest when you estimate them yourself. In PriceGuess's salary estimation game, guessing pay by job and tenure lets you feel the gap between mean and median, nominal and real in your gut. The same price-sense training carries over to the asset price estimation mode. Before "how much do I want to earn," know "what's typical" — and "which ruler measured that number." That's where negotiation and planning actually start.
