What the Phillips Curve Actually Says
In 1958, New Zealand-born economist A.W. Phillips plotted UK wage inflation against unemployment for 1861 through 1957 and found a remarkably stable downward curve. Samuelson and Solow replicated the pattern for US prices in 1960 and named it the Phillips curve, and for a decade policymakers treated it as a menu: accept a bit more inflation to buy lower unemployment, or the reverse. The data supported that reading well enough that early 1960s macro models built inflation-unemployment tradeoffs directly into their structure.
The mechanism behind the curve is simple competition for labor and sales. When unemployment runs low, employers bid wages up to attract and keep workers, and firms pass rising costs into prices. When unemployment runs high, workers lose bargaining power, wage growth cools, and pricing power fades. The relationship transmits through the whole economy, which is why one slope parameter can summarize millions of bargaining outcomes.
The modern version conditions that tradeoff on expectations and supply shocks, but the core insight survives: labor market tightness is the demand-side engine of inflation pressure. To see the output side of the same tightness, the Okun law calculator converts unemployment gaps into missing or excess production, the mirror image of what this tool prices on the inflation side.
The Expectations-Augmented Formula
Milton Friedman's 1967 presidential address to the American Economic Association, with parallel work by Edmund Phelps, rebuilt the curve around expectations: inflation = expected inflation - beta x (unemployment - natural rate) + supply shock. The tradeoff survives in the short run for any fixed expectations level, but there is no free lunch across periods, because workers and firms build observed inflation into their next round of wage and price setting.
Run the default numbers to see the decomposition work. Expectations at 2.5%, unemployment at 6.2%, natural rate at 4.5%, and beta at 0.5 produce a 1.7-point slack gap that subtracts 0.85 points, landing predicted inflation at 1.65%. The same labor market with expectations de-anchored at 5.35% would print 4.5% instead, which is the entire difference between an economy drifting back to target and one stuck in an inflationary spiral.
Expectations are the ratchet: once they rise, every point of unemployment slack is spent just standing still against them. That is why central banks obsess over survey and market-based measures of expected inflation, and why the real-rate arithmetic that connects nominal rates to inflation expectations matters so much. The Fisher effect calculator handles that nominal-versus-real conversion when you need the interest rate side of the story.
The Natural Rate Anchor
Everything in the formula pivots on u*, the natural rate of unemployment: the level consistent with stable inflation, built from frictional churn between jobs plus structural mismatches of skills and location. It is unobservable, must be estimated, and drifts with demographics, education, and labor market institutions. Modern US estimates sit near 4.4 to 4.6%, well below the 6% figures common in the 1990s, and estimates from the Fed and CBO continue to move as data revise.
Mismeasuring u* flips the verdict entirely. If you assume a natural rate of 4.5% while the true value is 4.0%, a 4.2% unemployment rate reads as slack when it is actually tight, and the model will underpredict inflation. That error pattern describes much of the late-2010s debate, when unemployment fell to 3.5% in 2019 without the inflation surge that a 5%-plus assumed natural rate would have predicted.
The natural rate is often discussed under its inflation-stability alias, NAIRU, but the underlying idea is a labor-market structure concept that can be decomposed. For separating the frictional, structural, and cyclical components that make up u*, the natural rate of unemployment calculator works through each piece, and its cyclical leg is exactly the gap this Phillips curve tool consumes.
Supply Shocks and the Stagflation Test
The supply shock term is what saved the Phillips curve from the 1970s. When OPEC quadrupled oil prices in 1973-74, inflation hit roughly 11% in 1974 while unemployment peaked at 9% in May 1975. A curve with only a labor market term cannot produce that combination: rising unemployment should have pulled inflation down. Adding the shock term v, and letting expectations drift upward, reconciles the theory with the stagflation that actually happened.
Put 1975-style numbers through the tool: expectations at 6.0%, unemployment at 9.0%, natural rate at 4.5%, beta at 0.5, and a plus-3-point oil shock. The predicted inflation is 6.75%, decomposed as 6 points from expectations, minus 2.25 points from deep slack, plus 3 points of shock. The decomposition is the teaching point: no single input explains stagflation, and neither an unemployment-watcher nor a money-supply-watcher alone would have called it.
The 2021-23 episode ran a similar test with better luck. Supply chain disruptions and energy shocks pushed inflation far beyond what modest labor market tightness implied, then the shocks unwound and inflation fell from its 6.6% core-CPI peak toward 3% with unemployment staying below 4.5% through 2024. To feed the tool accurate current readings, the CPI inflation calculator computes the headline and year-over-year rates from index values.
The Sacrifice Ratio and the Cost of Disinflation
Mode 3 prices the hardest question in monetary policy: how much unemployment does a return to target cost? Cutting inflation from 4.5% to 2.5% at beta 0.5 needs 2 divided by 0.5, or 4 cumulative gap-years. Compressed into one year, unemployment must reach 8.5%; spread over two years, 6.5% suffices. The spreading tradeoff is why central banks move gradually except when credibility itself is burning.
Translated through Okun's 2-to-1 ratio, 4 gap-years equal 8 points of foregone output, a sacrifice ratio of 4 output-points per inflation-point removed. Classic US estimates cluster between 4 and 6. The Volcker disinflation fits the arithmetic: taking CPI inflation from 13.5% in 1980 to about 3% by 1983, a 6-point cut, implied roughly 6 gap-years at a steep de-anchored beta near 1.0, delivered as unemployment peaking at 10.8% in late 1982. At beta 0.5 the same cut would price at 12 gap-years, which is why credibility, which steepens the effective beta, is worth so much.
This cost arithmetic is also why demand management is a blunt disinflation tool and why fiscal restraint during overheating is cheaper than monetary cleanup afterward. For tracing how a demand shock propagates into total spending before it ever reaches prices, the MPC calculator computes the consumption response and spending multiplier, the upstream engine of the inflation this tool forecasts.
Why the Curve Flattened, Then Steepened
Through the 2010s the slope flattened dramatically. A beta near 0.25 fits that decade: unemployment at 3.5% in 2019 pushed core inflation barely above 2%, and cutting inflation through unemployment alone would have cost 8 gap-years per 2 points. Anchored expectations did the stabilizing work instead, holding expected inflation near target while the labor market ran hot, and inflation-fearing policymakers leaned against hikes the flat curve said were unnecessary.
Flat and steep cut opposite ways. A flat beta 0.25 makes a boom nearly free on the inflation side but makes demand-driven disinflation brutally expensive, 8 gap-years versus 2 at beta 1.0 for the same 2-point cut. From 2021 the effective slope steepened visibly, with core inflation responding to a combination of tight labor and supply shocks, and then the 2023 disinflation arrived largely through the shock term unwinding while expectations held, the one route that requires no unemployment gap at all.
Which regime you are in changes every policy conclusion the tool produces, so treat beta as a scenario variable rather than a constant. For the price-side measurement that anchors any regime judgment, the GDP deflator calculator computes the economy-wide inflation measure that covers all domestic production, complementing the CPI's consumption-basket view.
Using the Curve for Forecasting
Mode 2 turns the tool into an expectations auditor. Suppose core inflation prints 6.0% with unemployment at 3.5%, a natural rate of 4.0%, beta at 0.5, and you attribute 1.5 points to supply shocks. The implied expected inflation is 6.0 minus negative 0.25 of tightness minus 1.5 of shock, landing at 4.25%. If surveys instead show expectations anchored near 2.5%, you have found unexplained residual inflation, and the honest response is to widen the model, not to fudge beta.
Good forecasting practice runs the curve as a scenario grid rather than a single line: one row per plausible natural rate, one column per beta, shocks set by your supply-side reading. Cross-check the demand side independently before trusting the inflation side, since both legs of the formula consume the same macro conditions. The GDP gap calculator measures where output sits against potential, and the GDP growth calculator tracks whether the gap is opening or closing at the current pace.
Calibrate expectations inputs from real sources: household surveys, professional forecaster medians, and market-based breakevens each measure a different population and can diverge meaningfully in stressed periods. When the three measures disagree widely, run the tool at each value and treat the spread as your forecast uncertainty. The curve earns its place in a forecast stack by making the labor-market-to-inflation link explicit, not by replacing judgment.
Limits, Revisions, and Honest Use
Every input in the formula is an estimate, and u* is the slipperiest. Natural-rate estimates are revised with each cycle, and the 2022-23 experience, in which estimated slack looked enormous while inflation surged, sent estimates back down as unemployment stayed low without reaccelerating. A tool output is only as current as its natural-rate input, so date-stamp every scenario you build and re-estimate u* rather than defending an old figure.
The unemployment rate itself is an imperfect tightness gauge. Participation swings change the pool without changing pressure, labor hoarding lets firms hold workers through downturns, and demographic shifts move the unemployment rate consistent with any given vacancy level. The labor force participation rate calculator decomposes the employed, unemployed, and not-in-labor-force buckets behind the headline rate you enter here.
Two honest-use rules close it out. First, the curve is nonlinear in practice: it steepens sharply when unemployment runs well below natural, so linear extrapolation from a beta fitted in normal times understates extreme-boom inflation risk. Second, pair every nominal forecast with its purchasing-power meaning, since a 2-point inflation surprise compounds into materially different real outcomes over a decade. The inflation calculator performs that value-of-money translation on any horizon.