Unemployment & Inflation

Types of Unemployment

Economists split unemployment into three buckets, and knowing which bucket someone falls into matters a lot for policy. You can't fix structural unemployment with a stimulus check.

Frictional unemployment is the easiest to wrap your head around. Someone just graduated from UGA with an accounting degree and is sending out applications. A marketing director at Coca-Cola quit in February 2024 because she got a better offer at PepsiCo but doesn't start for six weeks. Both of these people have skills employers want. The jobs exist. The matching just takes time, and that's true even when the economy is booming.

Structural unemployment is a completely different animal. When U.S. coal employment dropped roughly 40% between 2011 and 2020, miners in places like McDowell County, West Virginia weren't between jobs. They were stranded. The General Motors plant in Lordstown, Ohio shut down in 2019, and those assembly-line workers couldn't exactly retool overnight into software developers. Their skills or their geography or both stopped matching what employers needed, and retraining programs take months or years when they work at all, which they often don't. Moving is expensive. Whole communities get hollowed out.

Cyclical unemployment is the one that tracks recessions directly. Aggregate demand tanks, businesses slash payrolls, layoffs cascade through sector after sector. The 2008-2009 financial crisis and the early weeks of COVID in 2020 are textbook examples. Once GDP climbs back to potential, cyclical unemployment goes away.

The natural rate of unemployment is frictional plus structural, meaning it's the unemployment rate when the economy sits at potential GDP and cyclical unemployment equals zero. The CBO pegged it at about 4.4% for 2024. That number will never be zero because people will always be between jobs and some skills will always be mismatched with what the market demands.

Measuring Unemployment

The Bureau of Labor Statistics puts out monthly unemployment numbers, and the way they define "unemployed" is way narrower than most people realize.

The labor force counts everyone age 16 and up who is either working or actively looking for work. If you're retired, a full-time student who isn't job hunting, a stay-at-home parent, or you just plain gave up looking, you're not in the labor force at all.

Unemployment rate = (Unemployed / Labor Force) x 100

To count as unemployed you have to be both jobless AND actively searching. That creates two big holes in the data, and both come up on basically every AP exam.

Discouraged workers are people who stopped looking because they figure nothing's out there for them. Once they quit searching, poof, they drop out of the labor force entirely, and the official unemployment rate actually goes down even though the job market didn't improve one bit. The BLS just stopped counting them. During 2009-2010 this made the recession look less awful than it was. The headline U-3 rate topped out at 10.0% in October 2009, but broader measures that account for discouraged workers put the real damage closer to 17%.

Underemployment is the other blind spot. A software engineer who can't find tech work and is pulling espresso shots part-time at Starbucks shows up as "employed" in the official data. The BLS does publish a U-6 measure that captures discouraged workers plus people stuck in involuntary part-time jobs, but nobody puts U-6 in newspaper headlines. On the AP exam, if a question describes someone working part-time who wants full-time work, that's underemployment and the U-3 rate misses it completely.

Inflation and the CPI

Gas hit $5.02 a gallon in June 2022, AAA's highest recorded national average ever. Grocery bills were up something like 10% year-over-year. The government boils all of that price movement down to one number through the Consumer Price Index (CPI), which tracks the cost of a fixed market basket of stuff a typical urban household buys.

The BLS has published the CPI since 1913, making it one of the oldest continuous economic data series in the country. They pick the basket (housing, food, transportation, medical care, education, etc.) and weight each category by how much a typical budget allocates to it (housing eats up roughly a third). Then they compare what the basket costs now versus a base year, multiply by 100, and that's your CPI.

The formula for turning two CPI values into an inflation rate is straightforward:

Inflation rate = ((CPI_new - CPI_old) / CPI_old) x 100

The CPI has some well-known problems that push the number higher than actual cost-of-living increases, and the 1996 Boskin Commission laid them all out. Substitution bias: beef prices spike so people switch to chicken, but the fixed basket pretends everyone kept buying beef at the higher price. New product bias: the basket gets updated on a lag and misses cheaper or better products entering the market (smartphones weren't in the original basket, obviously). Quality change bias: a $900 laptop this year runs circles around last year's $900 laptop, so calling that "unchanged price" doesn't capture the improvement. The Boskin Commission estimated these biases together overstated inflation by about 0.8 to 1.1 percentage points per year. And that matters a ton for things like Social Security COLAs, which are pegged directly to CPI.

The Phillips Curve

A.W. Phillips published his paper in 1958 after crunching nearly a century of British wage and unemployment data, 1861 through 1957. The pattern was consistent: low unemployment went with rising wages and prices, high unemployment went with low inflation.

The short-run Phillips curve (SRPC) slopes downward with unemployment on the x-axis and inflation on the y-axis. Up and to the left means lower unemployment, higher inflation. Down and to the right, the reverse. Lyndon Johnson demonstrated this relationship beautifully (if accidentally) in the late 1960s. He funded both Vietnam and Great Society programs at the same time, unemployment dropped below 4%, and inflation crept past 5%.

The long run is a totally different story, and this is the part students trip on constantly. Milton Friedman and Edmund Phelps, working separately in the late 1960s, both argued the tradeoff wouldn't hold permanently. The long-run Phillips curve (LRPC) is vertical at the natural rate of unemployment. Their logic: eventually workers and firms adjust their inflation expectations, so a central bank that keeps trying to hold unemployment below the natural rate through endless money creation just gets accelerating inflation with no lasting employment benefit. The 1970s proved them right in spectacular fashion. Paul Volcker's Fed finally killed the inflationary spiral in 1981-1982 by shoving the federal funds rate above 20%, which triggered a brutal recession (unemployment hit 10.8%), but it broke the cycle.

What shifts the SRPC? Expected inflation, mainly. When people start expecting higher inflation, because they've been living through it or because the Fed signals loose policy, the whole curve shifts up so that at every unemployment rate actual inflation runs higher. When expectations cool back down the curve drops. After Volcker, anchoring inflation expectations became basically the central obsession of central banking worldwide.

Stagflation and Supply Shocks

On October 17, 1973, the Organization of Arab Petroleum Exporting Countries slapped an oil embargo on nations that backed Israel in the Yom Kippur War. Oil prices quadrupled within months. Inflation surged. Unemployment surged. Both at the same time.

The Phillips curve said that wasn't supposed to happen.

Economists coined the term stagflation (stagnation plus inflation), and the standard Keynesian framework, which assumed policymakers could always pick their spot on a nice clean inflation-unemployment tradeoff, basically fell apart. What actually happened was an adverse supply shock: oil price spikes jacked up production costs across nearly every industry all at once. Firms cut output and laid off workers while simultaneously raising prices to cover their higher costs. On the Phillips curve diagram an adverse supply shock shifts the SRPC upward and to the right, meaning higher inflation at every level of unemployment.

The policy bind was genuinely awful. Fight inflation with contractionary policy and you make unemployment worse. Fight unemployment with expansionary policy and you make inflation worse. Arthur Burns ran the Fed through most of the 1970s and tried to do a little of both. It didn't work. Inflation was still running near 13% when Volcker replaced him in August 1979.

Favorable supply shocks flip the whole thing around. The late-1990s tech boom combined with cheap energy shifted the SRPC downward and to the left, delivering something rare: lower inflation AND lower unemployment at the same time. The U.S. hit 3.9% unemployment in 2000 with inflation staying under 3.5%. That doesn't happen from demand management alone. It takes supply-side conditions cooperating.

Worked Example

CPI was 248 last year and 255 this year.

Inflation rate = (255 - 248) / 248 x 100 = 7 / 248 x 100 = 2.8%.

The labor force is 160 million, with 6.4 million unemployed.

Unemployment rate = 6.4 / 160 x 100 = 4.0%.

Now connect this to the Phillips curve. If the natural rate is 5.0% and the current unemployment rate is 4.0%, the economy is running below its natural rate. On the short-run Phillips curve that puts you in the upper-left zone, lower unemployment than the natural rate with inflationary pressure building. The standard policy prescription would be contractionary policy to nudge unemployment back toward 5.0% and cool inflation off. This was more or less where the U.S. sat in late 2021, right before the inflation spike of 2022 forced the Fed into its most aggressive rate-hiking campaign since Volcker.

Different scenario: a supply shock pushes unemployment to 7.0% while inflation jumps to 5.0%. That's stagflation, and the SRPC shifted upward and to the right. No demand-side tool can fix both problems at once. Expansionary policy brings unemployment down but pours gasoline on inflation. Contractionary policy tames prices but deepens the jobs crisis. The entire decade of the 1970s was a painful demonstration of exactly this dilemma, and the AP exam loves asking about it in free-response format.