Markets are self-aware
By Scott Sumner
Markets are just amazingly wise. But that shouldn’t be surprising, because they must be smarter than us in order to make it tough to get rich. And that’s because if it were easy to get rich we’d quickly run out of steel to build 500 foot yachts. To prevent that steel shortage, markets are incredibly subtle and wise, far ahead of our pea brains.
Of course I’m half joking, but there is a serious point here. Markets are aware that their behavior affects others, and they take that likely affect into account when they respond. Two of the key “others” affected by markets are consumers and the Fed. Thus when markets fall on bearish aggregate demand forecasts, they calibrate that fall to account for the likely response of the Fed. Just imagine if this were not the case. Then after any market crash you should buy stocks, as the market would rebound when the Fed rushed in to cut rates, in order to reassure investors (as they did after the 1987 crash.) Even the recent mini-crash caused the probability of a September rate hike to fall modestly. And of course markets also take account of the fact that consumers will become more bearish if stocks crash. The crash takes account of the likely consumer response, for the same reason it takes account of the likely Fed response.
David Glasner seems to feel that it’s not rational for consumers to change their views on the economy after a stock crash. I will argue the reverse, that rationality requires them to do so. First, here’s David:
Third, EMH presumes that there is a direct line of causation running from “fundamentals” to “expectations,” and that expectations are rationally inferred from “fundamentals.” That neat conceptual dichotomy between objective fundamentals and rational expectations based on fundamentals presumes that fundamentals are independent of expectations. But that is clearly false. The state of expectations is itself fundamental. Expectations can be and often are self-fulfilling. That is a commonplace observation about social interactions. The nature and character of many social interactions depends on the expectations with which people enter into those interactions.
I may hold a very optimistic view about the state of the economy today. But suppose that I wake up tomorrow and hear that the Shanghai stock market crashes, going down by 30% in one day. Will my expectations be completely independent of my observation of falling asset prices in China? Maybe, but what if I hear that S&P futures are down by 10%? If other people start revising their expectations, will it not become rational for me to change my own expectations at some point? How can it not be rational for me to change my expectations if I see that everyone else is changing theirs? If people are becoming more pessimistic they will reduce their spending, and my income and my wealth, directly or indirectly, depend on how much other people are planning to spend. So my plans have to take into account the expectations of others.
An equilibrium requires consistent expectations among individuals. If you posit an exogenous change in the expectations of some people, unless there is only one set of expectations that is consistent with equilibrium, the exogenous change in the expectations of some may very well imply a movement toward another equilibrium with a set of expectations from the set characterizing the previous equilibrium. There may be cases in which the shock to expectations is ephemeral, expectations reverting to what they were previously. Perhaps that was what happened last week. But it is also possible that expectations are volatile, and will continue to fluctuate. If so, who knows where we will wind up? EMH provides no insight into that question.
Let me start with a jellybean example. All citizens are told there’s a jar with lots of jellybeans locked away in a room. That’s all they know. The average citizen guesstimates there are 453 jellybeans in this mysterious jar. Now 10,000 citizens are allowed in to look at the jar. They each guess the contents, and their average guess is 761 jellybeans. This information is reported to the other citizens. They revise their estimate accordingly.
Now imagine a Lucas-type “archipelago economy.” There are 300 million people with access to one island, their local “retail island.” There are another 10 million people with access to two islands, call them a retail island and a much bigger wholesale island. Each person only observes his or her little corner of the world. But the 10 million who have access to two islands trade stocks, the value of which depends on many factors, including the amount of economic activity that is occurring in aggregate. No individual knows the truth, but collectively they do know. The wisdom of crowds.
How should the other 300 million people respond if they wake up and see that stocks have fallen sharply? It would be rational to assume that the more informed traders have seen something that they (collectively) have not seen—worrisome signs in the wholesale sector. (Here you could obviously add a foreign sector, such as China.)
So the type of behavior that David sees as problematic for the rational expectations/EMH model is actually necessary for it to work effectively.
On a related note Tyler Cowen has a new post:
There seems to be genuine uncertainty about what the Fed will do, or not do, this September.
At a superficial glance, the good news scenario is if the Fed’s decision doesn’t matter much for the markets. Woe unto you if your economy is so fragile that a quarter point or so in the short rate, mixed in with some cheap talk, were to matter so much.
So if at first prices were to stay steady, following any Fed decision, then equities should jump in price. That is the “no news is good news” theory, so to speak. It’s a better state of the world if it is common knowledge that the Fed’s actions don’t matter so much in a particular setting.
What if prices jump right away, following a Fed decision? The market might then see that price jumps rely on the Fed making good decisions, whatever those might be. The risk premium might then go up. It is even possible that prices should on net fall in response to that, but in any case it seems that some further price adjustment is in order, perhaps in the downward direction. (That ought to come rapidly.)
Most generally, it seems the initial price move, in response to the Fed’s choice, cannot itself be correct, but that first price move must itself induce further price movements. The price move in response to the Fed’s decision is the economy telling us how much it is relying on the Fed, which right now we do not know.
I think Tyler is also over-thinking the problem. Markets don’t just have views about the world; markets have views on how markets respond to news. Markets are not waiting to find out if markets care about a measly 1/4% rate increase, markets already know how they feel about that issue, they swing wildly on every little facial twitch by Stanley Fischer. What markets want to know is what the Fed plans to do about NGDP. Markets understand that NGDP is really important, and that the Fed’s annoyance at having to hold rates at zero is not important. The markets really hope the Fed will pay attention to the important thing (NGDP) and not the unimportant thing (annoyance at rates being low for so long). The markets are merely waiting to find out what the Fed thinks. If we had conditional stock derivatives we could already know the expected market reaction for each of the two options (no change, up 1/4%.)
PS. Can anyone see any similarity between this post and the previous post in this blog? (Bryan Caplan’s far better post on labor economics.)