‘We overlooked how resilient people are’

Good morning. When The Economist has a picture of a bull on its cover, that should be a sell signal. But the story is about how markets can’t keep rising so fast. So that’s a buy signal. What are we supposed to do? Wait for the cover of Barron’s tomorrow? If you have a better plan, email us: robert.armstrong@ft.com and ethan.wu@ft.com. 

Friday interview: Adam Posen

Adam Posen is president of the Peterson Institute for International Economics. He has been a Federal Reserve economist, a consultant to governments and to the IMF, a fellow at Brookings, and a member of the monetary policy committee of the Bank of England. Below he talks with Unhedged about what we learned from the pandemic, what ails China, what to expect from artificial intelligence and much more.

Unhedged: Over the past few years, we’ve done a massive macroeconomic policy experiment. What have we learned?

Adam Posen: Two things. First is that people and businesses were much more resilient than we gave them credit for. Initially when the pandemic hit, I thought we were going to have, at a minimum, persistent unemployment and a lot of closing businesses. And it didn’t happen. That’s partly because of supportive policy, but I think we overlooked just how resilient people are. By the same token, I spend a lot of my day job worrying about the corrosion of globalisation. That is going on, and it’s not good, but the ability of trade to bounce back from US-China conflict and from supply chain disruptions is also pretty amazing. We’ve seen some shifts in where trade happens, but total trade hasn’t changed very much.

The second thing is about inflation. My generation, and three or four generations of central bankers before mine, feared a repeat of the 1970s. Maybe it wasn’t at the front of our minds, but always easy to bring to mind. What we found was that there really wasn’t much pressure for an upward inflation spiral. That’s really interesting! It may be that the stuff that Ben Bernanke, Thomas Laubach, Frederic Mishkin and I wrote about 25 or 30 years ago on inflation targeting was for real — that if you anchor inflation expectations, you can be credible and still do activist policy. I hope that’s right; it seems a little too good to be true. But I think it may be that if you run credible counter-inflationary policy in a good regime for 30-40 years, people do give you the benefit of the doubt.

Unhedged: That said, in those months when inflation was really screaming, it must have put a bit of a scare into you.

Posen: Absolutely. And because of that, I’m still not entirely team transitory, in the sense that if the Fed or European Central Bank or Bank of England had not started raising rates aggressively, I think things could’ve gotten much worse.

Unhedged: A lot of us are looking at the persistent strength of the economy, and asking: where are the famous lagged effects of monetary policy? And should we be worried about them showing up all at once?

Posen: I will take a little credit here. In the talks I gave starting in the spring of 2021, I kept saying I don’t see a recession as inevitable. I forecasted continued growth in the US, though not quite as strong as we got. A year ago, prior to the SVB banking mess, people were talking about the lagged effects of monetary policy, and I was taking the under. I argued that if we were going to see lagged effects we would’ve seen them by now. That’s been borne out: traditional interest rate-sensitive sectors like, say, residential construction in the US were not responding all that much to Fed tightening. Emerging markets like Mexico were showing resilience, in terms of both growth and financial stability. So it just didn’t make sense that there’s some lagged shoe waiting to drop. The things that were the most vulnerable to higher rates historically were resilient.

Today, you can still tell us a scary story about US commercial real estate, but the overwhelming story is that there’s no reason to expect lagged effects. So now the question is why? And to me, it’s a vindication of the period since the financial crisis — and a damning indictment of US and UK policy prior to the financial crisis. The post-crisis system of capital requirements and good supervision and regulation worked. To a first approximation, the reason rate-sensitive sectors of the US economy and emerging market economies were not affected by aggressive Fed tightening the way they once were is that we went in with so much more financial strength. Balance sheets for non-financial business, core banks and even households were very solid.

It shows that part of the Fed’s monetary policy transmission mechanism is essentially preying on financial vulnerability. People came to that understanding in reverse. You hear talk that the Fed “tightens until it breaks something”. But I’ve started making the flipside of that point: if there’s nothing to break, the Fed isn’t a bull in a china shop; it’s just somebody with a duster. And maybe that means Fed policy is just less effective.

Unhedged: You’ve been supportive of the bullish US productivity story. Explain your idea of a labour-market-to-AI productivity handoff.

Posen: At the moment, I am a reluctant productivity bull, and I do think it’s a two-step process. We’ve had three quarters in a row of pretty strong productivity growth. But is this a headfake like we saw post-Covid, or is this genuine?

My view is that it’s genuine. What happened was that a huge share of workers at the low end of the income distribution basically said you only live once; life’s too short for this. All the while the government was stepping in. So the “reservation wage”, the minimum bundle of pay and working conditions required for a given job, went up. Workers increasingly don’t want to be doing, say, home healthcare for too little money. They’re willing to change jobs.

At the same time, the US had an enormous spike in unemployment, which meant people suddenly switching jobs didn’t need to worry about employers wondering “what’s wrong with you?”. When 20-plus per cent of the workforce is all out of work at the same time, there’s no negative signal. I layer that on top of the common story people are telling about running a hot economy. A bunch of workers moved to larger employers, or threatened to do so to get more out of the current employers. And we know that pairing workers with larger employers makes them more productive. They feel more secure, there’s more investment in training and there’s more upward mobility.

I can’t pretend to have a crystal ball on this, but I think we probably have another six months or so of productivity gains from people settling into new positions.

Unhedged: So that’s the labour market half of the story. What about the AI half? What gives you confidence that AI will be strongly productivity-enhancing?

Posen: I come at this not as a technologist, but as someone who has studied previous waves of productive innovation, as well as those that failed to transpire. Take the late Robert Solow’s exogenous growth theory. What happens is every once in a while, through the generosity of God and the genius of individual researchers, you get a new general purpose technology that affects the average rate of growth and productivity throughout the world — the “next big thing”. So the question is: why do I believe AI is one of those, beyond listening to whichever Sam of the day is featured at Davos this year?

I would point to two big observable factors. First, labour hoarding. On the face of it, the continued hiring of large numbers of workers over the past year and a half doesn’t make much sense. Maybe we were catching up from Covid or employers feared running short of workers. But the payrolls trend of 200,000 new jobs a month ran all the way through 2023. Companies were hiring workers who they wouldn’t necessarily be able to use right now, but thought they may need to use sometime soon. That is necessary, but not sufficient, for an innovation boom. It doesn’t guarantee that innovation is coming, but it’s happened in advance of pretty much every major innovation wave. In a decentralised way, thousands of employers see in their individual business conditions that they’ll be able to use more people.

Second is the crowding in of smart money or private investors into the AI space. Again, this is necessary but not sufficient. We can have bubbles that are not foretelling of a technology boom. But essentially every time that there has been a productivity leap — whether it’s railroads in the late 19th century or the internet in the 1990s — you have what looks like over-investment and crowding in. There’s a logic to that. In a decentralised way, investors decide that there’s really something here, somebody is going to win and I want to bet on it. You saw that in the 1990s. Importantly, you did not see that in the past 20 to 25 years. Things that looked potentially important, like 3D printing or genomics, were mostly avoided by the smart money. But whether it’s private equity or Microsoft’s balance sheet, you’re seeing them get in now.

Again, maybe it ends up being nothing. But as a macro person, this combination of over-investment and labour hoarding looks like a good antecedent for a generative AI productivity boom.

Unhedged: Are we in a new and higher rates regime?

Posen: Economists tend to talk about this as R-star, the neutral interest rate for the economy. The way to think of it is roughly the 10-year real Treasury rate when the economy is at full employment. There were a lot of reasons why this rate was low for about 20 years. One is demographics: ageing societies tend to be more risk-averse and prefer safe assets. Another is the saving glut: huge amounts of savings in China and east Asia had to go somewhere and that pushed down rates. After the financial crisis, you had a combination of regulation and voluntary risk-aversion that pushed a lot of investors into so-called safe assets. And you had a lower rate of productivity growth than you had in the 90s, so underlying real demand was lower.

In my view, one of the big things pushing down R-star has changed fundamentally, and that’s productivity. So if I’m right, then R-star should go up, maybe not exactly one-for-one, but roughly at the same magnitude as productivity growth. So I’m not all the way to McKinsey forecasting 4 per cent productivity growth. But let’s say productivity growth in the US is going to be 2.25 or 2.5 per cent instead of 1 per cent. That’s a big jump. And that puts upward pressure on R-star.

There’s an additional factor, too, which I credit Larry Summers for raising last year. We are in the midst of a sustained fiscal boom. With the possible exception of Germany, the G7 economies and China are about to be or are already spending a lot more on defence, a lot more on green investment and a lot — unfortunately — on industrial policy. That’s a very big part of the world’s savings going into an expansion of structural deficits of 1-2 per cent of GDP. In the US, depending on how you count the shortfalls of revenue over the past few years, the number could even be higher.

You’re basically running the post-Cold War peace dividend in reverse. None of these countries, with the possible exception of Germany, are going to raise taxes to pay for this additional spending. They might cut spending some, but probably not enough to pay for it. So, if you have a sustained erosion of fiscal positions within major economies for the next 10 years, that’s worth three-quarters of a per cent on the 10-year bond, at minimum. And you get a per cent or more from productivity gains. And then R-star is over 2 per cent instead of zero.

Unhedged: You mentioned industrial policy. We recently interviewed Harvard’s Dani Rodrik, who’s on the opposite side of this debate from you.

Posen: I read that! I was pleasantly surprised that Dani said don’t use industrial policy as a jobs programme. That’s certainly true, and it’s very important for a proponent of industrial policy like him to say that. The other surprising thing he said was that you have to consider industrial policy as a portfolio of investments, and you have to let some individual projects fail. I was delighted to see that.

I think he’s being a little naive about the political economy, which is funny given his earlier scholarship. If you start pumping money into a government national champion, are you really going to be allowed to let it fail? But I agree with him completely on both the portfolio approach and not using industrial policy as a jobs programme.

Unhedged: What’s your view on how serious the structural problems in China’s economy are?

Posen: What I call China’s structural problems are very different than what, say, Michael Pettis or Adam Tooze call China’s structural problems. Their focus is on things like China spending too much money on real estate, or too many white elephant investment projects. Both of which are true, but I don’t view them as insurmountable.

On the other hand, what I view as “structural”, meaning persistent and unlikely to change, is what I called in Foreign Affairs several months ago China’s economic long Covid. What I mean is a syndrome besetting the household sector and small business sector, where they are much less willing to invest in illiquid things, like durable goods or other forms of financial investment, and much less willing to respond to stimulus policies. They’re looking for ways to get money out of China. All of these were there before, but are much worse since Covid. To me, Xi [Jinping] and the Communist party’s behaviour during the zero-Covid policy ripped the mask off the party. Suddenly, the average Han Chinese person — not an oppressed Uyghur Muslim or a Hong Kong democracy protester — finds that the party micromanages people’s lives.

In my view, Chinese people always knew the party could expropriate property rights if they wanted to. But there was a pact; I call it the “no politics, no problem” pact. As long as you weren’t a democracy protester or a meddlesome ethnic minority, you could go about your life without anything bad happening. Maybe you have to pay the occasional bribe, and maybe you’re resentful that party officials get special treatment. But by and large, you can run your business and go about your life.

This started to change once Xi consolidated power in 2015, but it was really thrust in the face of the average Chinese person during zero-Covid. To me, that’s structural, because it’s a definite shift in party behaviour. As the FT reported the other day, we’re going back to having Mao-era militias built up at state-owned enterprises. You had arbitrary interventions against video game companies, and on and on.

The only way this changes in my view is if there is some credible way of the party committing to only intervening on very special occasions, limiting its arbitrary decisions about access to property and work. And it’s not credible for them to do that.

I view this as a fundamental regime shift. Paul Krugman once dismissively referred to my argument as being about individuals. I think it’s about an individual, Xi, because of the regime shift. The Chinese economic miracle, between when Deng [Xiaoping] consolidated power in 1979-80 to when Xi did in 2015, was underpinned by the fact that the party lived by the no politics, no problem pact. There was no question they’d kill people in Tiananmen Square, but that was a crackdown on political activity. No one was interfering with the right to make a living, to get rich. This is a fundamental regime shift in China that’s going to be very hard to reverse. And so you’re going to have an ongoing drag on the economy because there’s less investment, more cash-hoarding, more risk aversion, and because the normal stimulus policies may not work.

One good read

Contrarian-signal covers aside, this Economist piece on Russia’s descent into dictatorship is quite good.

FT Unhedged podcast

Can’t get enough of Unhedged? Listen to our new podcast, hosted by Ethan Wu and Katie Martin, for a 15-minute dive into the latest markets news and financial headlines, twice a week. Catch up on past editions of the newsletter here.