“I have come to bury Phillips, not to praise him!”
Paraphrase of Mark Antony in Shakespeare’s Julius Caesar
There seems to be more uncertainty about the current health of the economy and its prognosis for the rest of the year (and beyond) than normal. Major banks and brokers have come out with wildly inconsistent estimates of the probability that the U.S. enters a recession this year and the Federal Reserve itself has recently changed its forward guidance toward a more dove-ish “wait-and-see” attitude regarding future hikes in the Fed Funds Target. The Fed chairman, Jerome Powell, walked back some comments made in December. Earnings and revenue reports have likewise been sending mixed signals (example: see Apple’s January 29 release).
Much of this uncertainty derives from models that use the “Phillips Curve” as a predictor. The Fed, along with other central banks, continues to use the Phillips Curve as a major signaler. The Phillips Curve is an inverse relationship between unemployment and wage growth (and, by extension, inflation) posited by economist William Phillips three generations ago. It’s persistence in light of its under-performance has been nothing short of remarkable. As far back as the 1970s we saw year after year of increases in wages, inflation, and unemployment simultaneously; and then the subsequent decline in the growth of all three in the mid-1980s.
After the 2016 election as unemployment dipped to near-record lows, the response of policy-makers (and many others in both academia and “on the street”) was to prepare for a surge in wages (and the subsequent “cost-push” inflation that would generate). The (Phillips) theory is that low levels of unemployment force companies to bid up wages to attract workers (and, conversely, as unemployment rises workers will forgo wage increases to keep their jobs since companies then have a bigger pool of unemployed workers to choose from). Actually, last year did see nominal wage growth of about 3%, the best seen in almost a decade, but this hasn’t translated into increasing inflation. Why this might be so will be addressed next month.
If the Philips Curve ever did have any efficacy it was in an economy very different from the one we inhabit now. When the idea originated, before the mid-1960s, the welfare state as we know it barely existed, use of debt was far less widespread, and more families were dependent on a single earner. A slowdown in the economy led quickly to softening prices as newly unemployed workers (and their families) stopped spending (unemployment up, wages/inflation down). These unemployed felt the pain of lost income fairly quickly and were ready (required!) to get a job as fast as possible; companies didn’t need to “entice” this natural supply of workers with higher wages, keeping the lid on inflation pressures. But once all these “excess” workers were back to work, companies were forced to resort to higher wages to grow (unemployment down, wages/inflation/up) – and on the cycle went.
In today’s economy, only a fraction in the bottom quintile of income is ever in the workforce, their income a mixture of various government transfer payments. For good or ill (this is an economic and policy debate to have at another time), we have largely removed that “cycling” group of workers from the economy. Those who do become unemployed are cushioned from the pain more than in the past (because of unemployment insurance and other transfers) and can more readily resort to borrowing to fill in missing income. They are also less likely to be the sole breadwinner in a family. All of these factors (and no doubt others) contribute to decoupling the Phillips Curve relationship and it should be retired as a tool of policy management in a modern economy. Perhaps a better quote than Antony’s is Queen Elsa’s in Frozen, “let it go, let it go …”