I was very annoyed when Larry Summers made a big splash talking about secular stagnation at the IMF’s 2013 Annual Research Conference – annoyed not at Larry but at myself. You see, I had been groping toward more or less the same idea, and had blogged in that general direction (Krugman 2013) – but it wasn’t forceful, and Larry rightly gets credit for making the topic a front-burner issue.
The larger point, of course, is that if you’re following events and looking at the data it’s actually quite natural to raise once again the concerns Alvin Hansen raised 65 years ago, when he worried that low population growth would produce a situation of persistently inadequate demand. In what follows, I’ll lay out four reasons why secular stagnation deserves the buzz it’s now getting.
Observation #1: The zero lower bound matters much more than we thought
Secular stagnation is the proposition that periods like the last five-plus years, when even zero policy interest rates aren’t enough to restore full employment, are going to be much more common in the future than in the past — that the liquidity trap is becoming the new normal. Why might we think that?
One answer is simply that this episode has gone on for a long time. Even if the Fed raises rates in 2015, which is far from certain, at that point we will have spent seven years — roughly a quarter of the time since we entered a low-inflation era in the 1980s — at the zero lower bound. That’s vastly more than the 5% or less probability economists at the Federal Reserve used to consider reasonable for such events.
Suppose that we were to expect the future, on average, to look like the past – specifically, the past since price stability in the modern sense of low stable inflation became the norm. Even then, we would, on current evidence, expect to see a lot of problems with monetary policy at the zero lower bound; that is, we’d expect the world to look a lot more like that envisioned by Hansen than that envisioned by most macroeconomists during the Great Moderation era.
Beyond that, a look at the data suggests that there has been an ongoing trend making ZLB events more likely.
Observation #2: There seems to be a downward trend in real interest rates
It’s not widely remembered now, but there was some discussion of a possible liquidity trap during the 1990-1 recession and the jobless recovery that followed, and much more discussion in the slow recovery after the 2001 recession. And there was a reason: a look at the data suggests that it was getting steadily harder to get monetary traction even before the 2008 crisis. The IMF (2014) has shown that there appears to have been a downward trend in long-term real interest rates over the era of the Great Moderation; the trend is even more visible if you look at short-term rates. Figure 1 shows the Fed funds rate minus core inflation, averaged over business cycles (peak to peak; I treat the double-dip recession of the early 1980s as one cycle). This in turn suggests that my crude calculation above of the odds of hitting the zero lower bound was too optimistic; the downward trend implies that the odds are substantially higher now than they were in the past.
And even that is almost surely too optimistic.
Observation #3: The fundamentals have shifted in a major way since the 2000-7 cycle
As I showed in Figure 1, the average real interest rate over the 2000-7 business cycle was very low by historical standards. Yet the environment of the time was far more favourable for spending than the environment is likely to be looking forward, for at least two reasons.
First, the 2000-7 cycle was marked by a huge and presumably unrepeatable rise in leverage. Household debt rose from 67% of GDP at the 2001 peak to 94% at the 2007 peak, an annual average rise of roughly 4% of GDP. Even if deleveraging comes to an end, we can’t expect this level of debt-supported spending to resume, implying a major hit to aggregate demand – in effect, a 4% of GDP anti-stimulus relative to the last cycle – to become a more or less permanent feature of the economy. This in itself would suggest a substantial fall in the natural rate of interest, and hence a liquidity-trap-prone economy.
On top of this, Hansen’s old concern – slow population growth – is back. It’s not widely recognised just how quickly the demography of growth has changed in western economies. It’s most dramatic in the Eurozone – Figure 2 shows the rate of growth of the working-age population, which has moved rapidly into negative, almost Japanese-style territory. But the US has also seen a sharp drop.
Why is this a problem? For the same reasons Hansen invoked: slow or negative growth in the working-age population means low demand for new investments, both in housing and in productive capital, and therefore reduces the natural rate of interest still further.
So put these items together:
1. A much higher probability of hitting the zero lower bound than we used to think.
2. A secular downward trend in real interest rates even before the 2008 crisis.
3. Changes in fundamentals – an end to ever-rising leverage and a sharp demographic slowdown – that imply still weaker demand looking forward.
Taken together, these factors don’t prove that secular stagnation is here, because other things can happen, but they do make the case for such stagnation alarmingly plausible. And this creates significant problems for policy.
Observation #4: Even unconventional policies have problems dealing with secular stagnation
If you look at the extensive theoretical literature on the zero lower bound since Japan became a source of concern in the 1990s, you find that just about all of it treats liquidity-trap conditions as the result of a temporary shock. Something – most obviously, a burst bubble or deleveraging after a credit boom – leads to a period of very low demand, so low that even zero interest rates aren’t enough to restore full employment. Eventually, however, the shock will end.
The idea that the liquidity trap is temporary has shaped the analysis of both monetary and fiscal policy. And that analysis now looks much more problematic.
Start with monetary policy. The most persuasive story about how monetary policy can work at the zero lower bound is that it can gain traction if you can convince the public that there has been a regime change, that the central bank will maintain expansionary monetary policy even after the economy recovers, in order to generate high demand and some inflation. As I put it a long time ago (Krugman 1998), the central bank must “credibly promise to be irresponsible”.
But if we are talking about Japan, exactly when do we imagine that this period of high demand, when the zero lower bound is no longer binding, is going to begin? And now we are talking seriously about secular stagnation in Europe and the US as well, which means that it could be a very long time before ‘normal’ monetary policy resumes. Now, even in this case you can get traction if you can credibly promise higher inflation, which reduces real interest rates. But what does it take to credibly promise inflation? It has to involve a strong element of self-fulfilling prophecy: people have to believe in higher inflation, which produces an economic boom, which yields the promised inflation. A necessary (though not sufficient) condition for this to work is that the promised inflation be high enough that it will indeed produce an economic boom if people believe the promise will be kept. If it is not high enough, then the actual rate of inflation will fall short of the promise even if people do believe in the promise, which means that they will stop believing after a while, and the whole effort will fail.
Figure 3 offers a way to illustrate this problem, which I have come to think of as the ‘timidity trap’. Of the two curves shown, one is a hypothetical (but I think realistic) non-accelerationist Phillips curve, in which the rate of inflation depends on output and the relationship gets steep at high levels of utilisation. The other is an aggregate demand curve that depends positively on expected inflation, because this reduces real interest rates at the zero lower bound. I have drawn the graph so that if the central bank announces a 2% inflation target, the actual rate of inflation will fall short of 2 %, even if everyone believes the bank’s promise – which they will not do for very long, in any case.
So you see my concern. Suppose that the economy really needs a 4% inflation target, but the central bank says: “That seems kind of radical, so let’s be more cautious and only target 2%”. This sounds prudent, but it may actually guarantee failure. In other words, the problem of getting effective monetary policy, always difficult at the zero lower bound, gets even harder if we have entered an era of secular stagnation.
What about fiscal policy? Here the standard argument is that deficit spending can serve as a bridge across a temporary problem, supporting demand while, for example, households pay down debt and restore the health of their balance sheets, at which point they begin spending normally again. Once that has happened, monetary policy can take over the job of sustaining demand while the government goes about restoring its own balance sheet. But what if a negative real natural rate isn’t a temporary phenomenon? Is there a fiscally sustainable way to keep supporting demand?
In this chapter I’ll leave these questions hanging. The crucial point, for now, is that the real possibility that we’ve entered an era of secular stagnation requires a major rethinking of macroeconomic policy.
Hansen, A (1939), “Economic Progress and Declining Population Growth”, American Economic Review 29(1): 1–15.
International Monetary Fund (2014), World Economic Outlook, April.
Krugman, P (1998), “It’s Baaack: Japan’s slump and the return of the liquidity trap”, Brookings Papers on Economic Activity 29(2): 137-206.
Krugman, P (2013), “Bubbles, regulation, and secular stagnation”, The New York Times blog, 25 September.