Japan: No News Is Bad News

by: Scott Sumner

The recent Japanese stock market rally and “correction” is almost identical to the US experience of 1933. During that period FDR was engaged in a policy of depreciating the dollar. One problem with this policy was that asset prices move on new information. Thus each day the dollar exchange rate had already factored in the expected depreciation for all of 1933. FDR could only drive the dollar lower by doing more than expected. Surprisingly, he was able to do so fairly effectively, partly by pushing much harder than almost anyone expected. By the time the price of gold had risen from $20.67 to about $28, Keynes said enough is enough, and called for a halt to the policy. But FDR listed to George Warren, and persevered until he reached $35/oz.

In October 1933 FDR was dissatisfied with the pace of inflation, so he adopted a new technique, called the gold-buying program. The details are unimportant; the key point is that it gave FDR a way of sending the market signals that he wasn’t satisfied with the pace of dollar depreciation. Both Abe and FDR faced a similar problem—they needed to send signals to the market that they weren’t satisfied, that they intended to do more than the markets expected.

Here’s where the title of the post comes into play. Markets knew that FDR would either provide more signals of monetary stimulus, or be silent. He certainly wasn’t going to call for a stronger dollar. During the gold-buying program of late 1933, the markets knew that on each day FDR would either raise the official price of gold, or leave it unchanged. It would not fall. The EMH predicts that on days where the official price of gold was not raised, news would be viewed as more contractionary than expected, and the free market price of gold would actually fall. And that’s usually what happened. (All the details are available in my book, due out later this year. BTW, the 1933 New York Times commentary on daily movements in the dollar exchange rate were consistent with the EMH; markets moved on more positive than expected policy announcements, not positive policy announcements.

Many pundits are surprised that after rising by 80%, the Japanese stock market fell by 20%. “There wasn’t much news.” But that’s exactly the problem (as the NYT understood back in 1933) the EMH predicts that no news will be bad news, when the only two plausible outcomes are further stimulus signals, or nothing.

Yes, I’ve oversimplified slightly; there arguably was some bad news out of Japan, as Lars Christensen pointed out recently. But the lesson here is still very important. When a market is rocketing upward under a steady drumbeat of good policy news, even a pause in that drumbeat will cause a market setback. After all, markets expected the drumbeat to continue, or at least placed a positive probability on that outcome. When it stops, prices fall. The markets know that the only two plausible outcomes are Abe calling for a weaker yen, or Abe saying he’s happy where things are right now. Abe won’t call for a stronger yen, and investors know that.

This is why I’ve been less optimistic about Japan than even some Keynesians like Krugman and Stiglitz. I regard each day’s level of stock and exchange rate data as the optimal prediction of the long run effect. And I see that market data indicating modest success, but well short of 2% inflation, at least in 2014.

PS. The Dow rose about 80% in the first three months of of the 1933 dollar depreciation, then fell 19% in 3 days. Sound familiar?