Blanchard Warns ‘Zero Chance’ Of ‘Immaculate Conception’ Jobs Miracle

“A critical debate has emerged” around the outlook for the US economy and the Fed’s efforts to vanquish the hottest inflation in a generation.

That debate is “between those who think the current high inflation problem can be resolved without a recession and those who think a significant and sustained rise in unemployment is likely to be needed — and ultimately delivered by the Fed — for inflation to fall back to acceptable levels,” Goldman’s Dominic Wilson and Vickie Chang wrote, in a new note.

Goldman’s own house view is that the Fed can succeed without driving the economy “off a cliff” (as Elizabeth Warren might put it on an irritable day), while the bank identified Bill Dudley, Olivier Blanchard and, of course, Larry Summers, among those who believe the Fed will likely be compelled to push the unemployment rate substantially higher.

Blanchard and Summers collaborated recently to dispense with the notion that the labor market can be brought back into balance simply by a reduction in the number of job openings as tabulated by the JOLTS report. That number is very high, and the ratio to those counted as officially unemployed suggests, colloquially, there are two jobs for everyone who doesn’t have one (familiar figure below).

The Fed believes that tighter policy can help close the gap by compelling employers to eliminate superfluous job openings. Those are “free” job losses on the road to a more balanced labor market — job “losses” that don’t entail anyone actually being fired. That’s the best and, many would argue, only, road to a soft landing.

If that sounds too good to be true to you, you’re hardly alone. During an interview for the latest installment of Goldman’s “Top Of Mind” discussion series, Blanchard assigned “zero probability” to such an outcome. “Some observers seem to be hoping for an immaculate conception,” he said. “The historical relationship between job openings, or vacancies, and unemployment is crystal clear: The job vacancy rate has never substantially declined without a significant increase in unemployment.”

Make no mistake: This isn’t just a “critical debate,” as Wilson and Chang put it. It’s the debate, which is why the bank dedicated an entire “Top Of Mind” installment to it.

Given how important this really is, and considering Blanchard, while speaking to Goldman, said he wished the “basic logic” behind his argument was “was more widely shared,” I think it’s worth highlighting a few short excerpts from his interview with Goldman’s Allison Nathan, who edits “Top Of Mind.”

Below, find what I believe to be the most important passages from Nathan’s discussion with Blanchard, and remember: This is really all that matters currently when it comes to Fed and how they’re thinking about the tradeoff between the labor market, inflation and recession. The passages are abridged and aren’t exhaustive.

Allison Nathan: A key indicator of tightness in the US labor market today is an exceptionally high number of job openings. You recently expressed skepticism that openings can decline without the unemployment rate rising materially. Why?

Olivier Blanchard: Some observers seem to be hoping for an immaculate conception outcome in which job openings decrease and unemployment doesn’t increase, but I see zero probability of that outcome. The historical relationship between job openings, or vacancies, and unemployment — known as the Beveridge Curve — is crystal clear: The job vacancy rate has never substantially declined without a significant increase in unemployment. Admittedly, the current level of vacancies is unprecedented, so we are in uncharted territory, but common sense says that as a firm’s sales decline, it stops or slows hiring and lays off workers. Some firms may start by decreasing hiring because it’s easier to cancel job interviews than to make layoffs, and some firms may do more of one than the other, but it’s absolutely obvious that, if activity declines, firms will do both, which means that vacancies will decline, and unemployment will rise. That’s such basic evidence and logic that I wish it was more widely shared.

Allison Nathan: But isn’t the high level of job openings today largely the result of a pandemic-related decline in match efficiency — the process through which unemployed workers match with open positions — as Covid fears and generous unemployment benefits led fewer people to look for jobs? And if that’s the case, isn’t it possible that as those temporary factors unwind, job vacancies can decline without a substantial rise in unemployment?

Olivier Blanchard: A year and a half ago, I would’ve said that the sharp short-run shifts in demand during the height of the pandemic and the generous unemployment benefits together explained why it was difficult to fill job openings. But the most intense part of Covid has come and gone, unemployment checks are less generous, and yet match efficiency hasn’t meaningfully improved. Intersectoral reallocation is still high as workers are moving across sectors, and many sectors are having a hard time finding workers. The restaurant industry, for example, is unable to find workers, presumably because they’re now employed or looking for jobs in other sectors. And workers have become choosier about jobs — they’re not taking offers from firms if they don’t love the firm. So, while I would not be surprised if some unwinding of these dynamics occurred, much of this mismatch is likely here to stay for some time. In the past, these types of shifts have persisted for a decade or more — until another shock came along. The evidence suggests that we can’t expect improving match efficiency to drive a meaningful decline in vacancies without a sharp rise in unemployment.

Allison Nathan: Isn’t the sharp decline in job openings since their March peak at the same time that the unemployment rate hasn’t changed much evidence that a decline in job openings doesn’t have to be accompanied by an increase in the unemployment rate?

Olivier Blanchard: No. This is a well-known phenomenon known as Beveridge loops — countries “loop” around the Beveridge Curve when the economy turns because firms first react to a slowdown by decreasing hiring rather than laying off workers, so vacancies fall without much movement in unemployment. But these loops don’t last long, because eventually firms start to do both. That said, if the trend of vacancies falling without unemployment rising much persisted for another few months, I would start questioning my conclusions. For the moment, I have little doubt that unemployment will increase as the Fed takes further steps to cool the economy.

Allison Nathan: More broadly, can inflation be tamed without unemployment rising?

Olivier Blanchard: I wish, but no. Three factors are behind the increase in US inflation, all of which argue for an increase in unemployment. One, the decrease in match efficiency means that the natural rate of unemployment has increased, because more unemployment is needed to match workers with any given number of vacancies. Two, the increase in commodity prices has resulted in a wage-price spiral as firms have increased prices to reflect higher input costs, workers have demanded commensurate pay increases and received them due to their strong bargaining position in the current tight labor market, and firms have further raised prices in turn. Whether this spiral continues depends on what happens to commodity and energy prices. If they were to decline sufficiently, the pressure might go away. Otherwise, as is more likely, taming the wage-price spiral will require unemployment in excess of the natural rate. And three, inflation expectations have become slightly unanchored today, and the more unanchored expectations become, the higher the unemployment required to convince firms and workers that form decisions based on expectations that the Fed will achieve its 2% inflation target. The last two factors together mean that unemployment probably needs to rise above its estimated natural rate of 4.5- 5% to tame inflation. How far above is difficult to say, but I wouldn’t be surprised if unemployment needs to rise to 6%, and I put a small but non-zero probability on 7%.