In an immortal moment in the 1935 Marx Brothers’ classic A Night at the Opera, Groucho tries to explain to Chico why every contract has a “sanity clause,” voiding the agreement in the event of mental defect in either of the parties. “You canna fool me,” answers Chico. “There ain’t-a no Sanity Clause.”

Like Santa Clause, the Easter Bunny and Sasquatch, the Phillips Curve— the supposed trade-off between unemployment and inflation — has gone unobserved for a disturbingly long period of time.

The Federal Reserve, the European Central Bank and the Bank of Japan set a 2% inflation rate as the goal of quantitative easing. According to the Phillips Curve model, 4.1% unemployment in the US (the lowest since 1969), and 5.7% in Germany, the lowest on record, should produce inflation. If labor supplies are scarce, the cost of labor is supposed to rise, and prices are supposed to go up. Except they don’t.

In fact, we rarely see much of a relationship between inflation and unemployment, except in periods of severe recession, where prices fall and unemployment rises. These are brief episodes, though; there are others during which inflation and unemployment fall together. The past sixty years of data show no relationship at all.

1.coreinflation

2.USinflation

Professor Edmund Phelps, a Nobel Laureate in Economics, observes in a recent paper that inflation remains low despite very low levels of unemployment. “What explains the new paradox of low unemployment despite low inflation – or vice-versa? So far, economists – structuralists as well as diehard Keynesians – have been stumped,” Phelps observes.

There is nothing “natural” about the “natural rate of unemployment,” Phelps argues. “It is possible that the ‘natural rate’ is not a constant of nature, like the speed of light. Certainly, it could be moved by structural forces, whether technological or demographic […] For me, a compelling hypothesis is that workers, shaken by the financial crisis and the deep recession that resulted, have grown afraid to demand promotions or to search for better-paying employers – despite the ease of finding working in the recently tight labor market. A corollary hypothesis is that employers, disturbed by the extremely slow growth of productivity, especially in the past 10 years, have grown leery of granting pay raises – despite the return of demand to pre-crisis proportions.”

For one thing, the workforce is aging. More Americans are working past retirement age. The labor force participation rate for workers over 55 has risen since 2009, adding about 3 million prospective workers to the labor pool. The supply of labor is more elastic as older workers choose to stay in the labor force. Older workers, moreover, are likely to be more risk-averse than younger workers, and more likely to trade security for pay increases. The participation rate of Americans aged 55-64 in the labor force is only about 40%, so there is a lot of room for new labor force entrants at the higher end of the age spectrum.

3.fulltimeworkers

That may help explain why average hourly earnings growth has never climbed back to pre-crisis levels.

4.hourlyearnings

Falling prices for goods hold down overall US core inflation.

5.CPI

Part of the reason for lower goods prices is technological: electronics get cheaper every year. Part of the reason is e-commerce. Consumers can comparison shop instantly on the Internet, and are less likely to overpay in impulse purchases at the shopping mall. That is killing the profits of department stores and other traditional retailers, but it keeps prices down. By my estimate, about a quarter of the change in durable goods prices during the past dozen years is explained by the growth of e-commerce.

The Phillips Curve dogma, Prof. Phelps observes, assumes that nothing changes in the economy except demand — the more demand for labor and goods, the higher the prices. That follows postwar Keynesian theory, which taught that “aggregate demand drove everything. High unemployment was caused only by deficient demand, and low unemployment only by abnormally high demand.”

By contrast, Phelps observes, the theory built by Alfred Marshall, Knut Wicksell and Robert Solow “said everything was driven by structural forces: Faster technical progress and a greater preference to work or to save were to be welcomed, because they would boost the supply of labor and capital – and thus employment and investment. But the Keynesians maintained that structural forces were bad, because they cost people their jobs, unless policymakers manufactured enough demand to match the increase in supply.”

Structural factors — demographics, technology, and the organization of retail distribution — turn out to be more influential than demand management. Quantitative easing has been neither a success nor a failure. It’s been largely irrelevant.