Back to Krugman home page
FURTHER NOTES ON JAPAN=S LIQUIDITY TRAP
My note on Japan's liquidity trap has been widely read and cited, but judging both from press reports and my own correspondence there are some aspects of the analysis that have not been fully appreciated. (Partly this is my own fault, because the note was a bit terse - and slightly wrong on one point, as described below). So here are answers to some of the questions that people have asked.
1. What is the policy recommendation?
Many people apparently read my previous note as saying simply that Japan should print money like crazy. I have indeed said this in the past (see What is wrong with Japan? ), and see no harm in such a policy. But I now believe, based on the analysis in AJapan's trap@, that even a very large current monetary expansion will probably be ineffective. What is needed is a credible commitment to future monetary expansion, so as to generate expectations of inflation.
How might the Bank of Japan achieve such a commitment? The natural way is to announce a target rate of inflation over the long term, with the announced intention of doing whatever is necessary to achieve that rate - basically the same as the inflation targeting now followed in, say, New Zealand or the UK, but with the objective being an inflation rate that is Aacceptably high@ rather than Aacceptably low@. The obvious question is how high a rate is needed, for how long. And the short answer is that I don't know, but I am working on it. A guess is that the required inflation rate isn't very high, but that people must expect it to last for a quite long time - we might for example be talking about, say, 4 percent inflation for 15 years.
2. Why would you think this would work?
Consider the following thought experiment: imagine that Japan had not pursued price stability so assiduously in the 1980s, and as a result had entered the 90s with an underlying, expected inflation rate of 5 percent. Given 5 percent expected inflation, monetary policy would have had no problem increasing demand as needed: a zero nominal rate would have been a negative 5 percent real rate, which would have discouraged saving, encouraged investment, and - yes - meant a weak yen, of which more below.
So if Japan had not worked so hard at eliminating inflationary expectations in the 1980s, it would not have its current problems. Indeed, Japan's economic crisis is one of economic history's most spectacular own goals: her policymakers got into this mess partly because they tried to do the right thing.
But while this thought experiment shows how much better off Japan would be if it had inherited inflationary expectations from the past, why do such expectations have to come from the past? Inflationary expectations based on a solemn pledge to provide the necessary inflation in the future would accomplish exactly the same thing.
3. Doesn't Japan have many structural problems that account for its malaise?
Of course Japan has big structural problems. However, these have very little bearing on my argument, one way or the other.
One way to put this is the following: Japan is currently experiencing strong deflationary pressures, which are occurring in spite of what looks like a very expansionary monetary policy. How can this be happening? The proposition known as the Aneutrality of money@ - the assertion that other things equal the price level is proportional to the money supply - is central to every monetary theory I know of. It is not conditional - that is, it doesn=t say that an increase in the money supply raises prices Aif there isn=t too much corporate debt@or Aif the service sector has been deregulated@. The only caveat we sometimes make is the assertion that, because prices may be sticky in the short run, contractionary monetary policy may temporarily cause a recession rather than deflation, expansionary policy a boom instead of inflation. But no structural argument I know explains why an economy should be subject to uncontrollable deflation regardless of monetary policy.
What I pointed out in the previous note, however, is that there is a rarely noticed footnote to the conventional account of the neutrality of money: strictly speaking it applies only to permanent increases in the money supply. A monetary expansion that is perceived as temporary can, under some circumstances - namely when the market-clearing real interest rate is negative - be completely ineffectual at raising the price level, or (if prices are sticky) output. The point is that monetary policy isn=t really impotent; what makes it ineffective is that people don=t believe that it will be sustained. And conversely, a credible commitment to sustained monetary expansion will always be effective as an antidote to deflation.
Now you might say that structural problems created the situation in the first place; and you might be right. If you have a fix for the structural problems quickly at hand, fine. But if you don=t, monetary policy is still needed to fight deflation. And it doesn=t matter why the required real interest rate is negative; as long as it is, monetary policy will be effective if and only if the monetary authority is believed to be willing to allow long-term inflation.
4. Isn=t fiscal policy an alternative answer?
Fiscal expansion is certainly an alternative way of pumping demand into the economy. And given Japan=s grievous plight, coupled with the real uncertainty about what will work, I would advocate fiscal expansion as well as a commitment to inflation.
But I have doubts about fiscal policy=s effectiveness on its own, for three reasons.
First, if Japanese consumers really do exhibit something like ARicardian equivalence@, there will be no multiplier effects from government spending. My previous note actually misstates the implication of the model (that=s one of the marks of a good model - it is sometimes smarter than you are). The true implication there is that current consumption is completely unresponsive to current government spending (it=s tied down by the Euler condition, if you really want to know), so that public works projects will increase spending by exactly the amount the government spends - no more.
Which brings us to the second point: Japan=s long-term financial position is a bit worrisome, so keeping the economy afloat with year after year of massive deficits is a problem. And absent powerful multiplier effects, it will take massive deficits to keep the economy from slipping ever deeper into recession, unless a one-time fiscal push has a long-term effect.
And that is the third point. Although serious discussion of the long-term implications of fiscal stimulus is rare, most people probably have in mind some idea about Apump-priming@. (Actually, very few people either in the United States or in Japan have any idea what that means. AJump-start@ is probably a more useful metaphor). The idea is that once the economy is moving again, confidence will return and spending will no longer need to be supported. Now this might be true, but there is not much evidence for it; it is at least possible that Japan has a structural negative real interest rate that will last for a number of years. That means that the economy must either receive continuing massive fiscal stimulus, or have sustained inflation.
5. Wouldn=t it be enough to eliminate deflation? Why go all the way to inflation?
The answer I am suggesting is that this is not an option. Japan - like any liquidity trap economy - in effect needs inflation, because it needs a negative real interest rate. The slightly paradoxical conclusion which I believe to be true is that the deflation we actually see is the economy Atrying@ to achieve inflation, by reducing the current price level compared with the future. One way or another the future price level must be higher than the price level now, if the economy is to achieve full employment; so if you don=t want deflation, you must supply the expectation of inflation.
I think I should make it clear that I did not start with this conclusion, then make up a model to justify it. What I did instead was start with a very orthodox model - the same sort of model that is favored by people who are vociferously anti-Keynesian and pro-price stability - and ask under what conditions it could generate the apparent ineffectuality of monetary policy we see in Japan. And the need for inflation pops out - to my own surprise, by the way. If you refuse to accept this conclusion, either you must offer some alternative model, or you are saying that your opposition to inflation comes not from analysis but from gut feelings.
6. Isn=t the real answer to clean up the banking mess?
By all means Japan should clean up its banks - close down the really bad ones, or at least clean out their stockholders, recapitalize, and sell them. What the country surely needs is a giant version of the RTC cleanup of our own S&Ls.
That said, will this solve the macroeconomic problem? The trouble with our S&Ls, and presumably with Japan=s banks, was not that they lent too little, but that they lent too much; the reason it was urgent to close bad thrifts was that the further into the red they went, the bigger the incentive to gamble with other peoples= money. But that means that a cleanup of bad banks has at least the initial effect of reducing lending, and hence demand; it will, if anything, worsen the recession (as the aftermath of the thrift crisis did in the US in the early 90s).
The only complication here is that arguably Japan has recently entered a phase in which the government is expected to start closing bad banks in the near future, and in which banks are therefore curtailing lending in an effort to make the cut. Certainly the conventional view is that a credit crunch has emerged in the last year or so; and this credit crunch might be alleviated once the government gets its bank cleanup over with. However, this would still only return us to something like the situation a year ago. I cannot see any reason to believe that bank cleanup will in the end be a strong stimulus to demand, or indeed a stimulus at all.
7. What about the yen?
An inflationary policy is definitely a weak yen policy. So?
The current fear is that if the yen goes, so will the yuan, and then there will be a free-for-all of competitive devaluations and collapsing confidence. Such fears cannot be completely discounted. In the short term, capital flows are indeed highly volatile, and if the 29-year-olds in London who rule the world think that something is true, for a few hours or days it is.
But let=s back up for a moment and ask about the fundamentals. Suppose the yen falls another 30 percent, and that non-Japan Asia devalues by 10-15 percent, keeping its effective rates more or less unchanged. Would that be such a catastrophe? Only if a decline in the dollar exchange rate per se is regarded as a terrible thing by the market - but that is then a case of purely circular reasoning (not necessarily wrong in a world of self-fulfilling crises, but not necessarily right either).
The special concern that seems to motivate U.S. policy is that a weak yen will precipitate a yuan devaluation, with all kinds of confidence-damaging effects. Again, maybe; but if the yen falls by, say, 30 percent, it still remains true that a substantially smaller devaluation by all non-Japan Asia - China included - would be enough to restore cost competitiveness. What=s so bad? The fear, presumably, is that if China devalues at all it will have to be a whopper. But why?
I don=t want to discount these arguments completely. In the end, though, you have to ask how far the policies of the world=s second-largest economy have to be held hostage to the psychology of traders and the convenience of a China that is, remember, still only a fraction of Japan=s economic size at market prices.
8. Isn=t this all a crazy, irresponsible idea?
In approaching the problem of Japan=s trap, I have tried to give as many intellectual hostages as possible - adopting an approach that is as classical, as orthodox as I could. In fact, you could read AJapan=s trap@ as a scholarly essay on the generic issue of liquidity traps in macroeconomics! It turns out that the unavoidable conclusion even of that very buttoned-down analysis sounds very radical: an economy (in principle, any economy - Japan is just the case that motivates the analysis) that is in a liquidity trap needs expected inflation.
Now what do you do when an orthodox economic analysis suggests an unorthodox policy conclusion? You could go with the conventional wisdom - even if it isn=t working. Or you could say that conventional wisdom evolved to meet conventional situations; and that when we are faced with a novel kind of economic malaise, the results of hard thinking deserve to be taken seriously.