The Washington PostDemocracy Dies in Darkness

Do the data support the Federal Reserve’s planned interest rate hikes?

The rationale for them remains somewhat elusive.

The Fed Up movement’s Full Employment Defenders at the Senate hearing where Federal Reserve Chair Janet L. Yellen testified on Tuesday, Feb. 14. (Courtesy of Fed Up)

Both Tuesday and today, Federal Reserve Chair Janet L. Yellen is testifying before Congress on the state of the economy and the Fed’s broad plans on monetary policy. She didn’t make news in the Senate hearing Tuesday, and her pretty upbeat assessment of the U.S. economy, along with her suggestion that further interest rate hikes would probably be forthcoming, were consistent with recent Fed announcements.

And yet, although I don’t think the small rate hikes that I believe the Fed is contemplating will significantly slow the economy’s growth, the rationale for them remains somewhat elusive.

Yellen says March rate hike is on the table, effects from Trump policies remain unclear

First, consider this comment by notorious inflation hawk Jeff Lacker, president of the Richmond Fed, a (thankfully) nonvoting member of the Fed’s interest rate-setting committee: “Taking the range of plausible alternatives into account, my view is that, with unemployment at or below levels corresponding to maximum sustainable employment and with inflation very close to our announced target of 2 percent, significantly higher rates are warranted.”

Hmmm. I agree that the unemployment rate is sending that message, but as I recently stressed, you simply can’t base your labor market assessment purely on the unemployment rate. It’s still biased downward because of weak labor force participation, and no one accurately knows the level of the “natural rate,” or the lowest unemployment rate consistent with stable prices.

Let me show you what I mean using policy rules — which Lacker thinks generally support his conclusion — as an example. If you set the interest rate using the so-called “Taylor rule” that Yellen cites in this speech and plug in the relevant numbers (see data note below for details), Lacker’s got a point, as the rule spits out a target Fed funds rate (ffr) of 1.4 percent, well above the current target of between 0.50 and 0.75 percent. In fact, that’s about where the ffr would end up if the Fed does the three 0.25-basis-point rate hikes this year that it has telegraphed.

But suppose I plug in the underemployment rate instead of the unemployment rate. I’d strongly argue that’s a better measure of slack right now, given the still-elevated number of involuntary part-time workers (they’d rather be full-timers). Then the rule-based ffr comes to 0.65 percent, meaning that rates are fine where they are.

Economist Andy Levin developed a useful measure of slack, which accounts for both un- and underemployment along with the still-depressed labor force. When I plug in an estimate of his number, I get an ffr of zero! If you think that 4 percent is as plausible a “natural rate” for unemployment as the Fed’s 4.8 percent estimate — and given the wide confidence interval around this measure, you really should think that — then the rule-based ffr is 0.75, the high end of the current range.

One last bit of relevant evidence: This whole business is based on the logical contention that, as we squeeze slack out of the economy, inflation will accelerate. This relationship underlies the “Phillips Curve,” a workhorse of macroeconomics that shows the correlation between inflation and slack.

The problem is that there’s evidence to suggest that curve has flattened in recent years, as inflation has been consistently nonresponsive to diminished slack. The figure below shows a statistical analysis of the slope of this curve, wherein I sequentially add observations to the equation and recalculate the result (a “rolling regression”). The negative correlation between prices and slack is stable for a number of decades, but in around 2010, it is reduced by half, from around -0.3 to around -0.15 (that’s how much core inflation would be expected to fall in a year for each added point of slack).

That’s not nothing, but it’s pretty close, and should make Fed officials think hard about whether there’s enough of an inflationary concern to slow growth, jobs and wage gains. In fact, given that the Fed’s inflation target of 2 percent is supposed to be an average and the Fed has been below that target for years now, a period of inflation above 2 percent should be viewed as a desirable outcome.

Of course, Yellen and Co. have to see around corners, play off people’s expectations, and be ready for the curve in the figure above to revert to bigger negatives. So I’m not advocating that they be cavalier about potential pressures building up. But as the recovery is just now reaching folks who’ve been left behind for a very long time — some of whom were out in force at Yellen’s hearing Tuesday to convey the effect that the Fed’s decision-making on interest rates will have on their lives (see their photo above) — I’d urge the Fed to heavily weight the evidence shown above.

Data note: My calculation of the Taylor rule assumed a “neutral” real interest rate of zero, as per the president of the Minneapolis Fed, Neel Kashkari, who recently wrote (in a very thoughtful piece) that “I tend to think [the neutral rate] is around zero today, or perhaps slightly negative.” For my rolling regression of the Phillips Curve, I used the specification from this Ball/Mazumder paper (equation 2), using the core PCE deflator and CBOs estimate of u* (the “natural rate” of unemployment).