How Errors in Reading Fundamentals Drive Stock Price Volatility

Knowledge@Wharton:

What led you to do this research into cash flow expectations and asset prices, and what did you find?

There’s a big, outstanding question in finance, which is that prices both for stocks and bonds seem to move around much more than the fundamentals of these assets.

If you look at, say, earnings for the S&P 500 or dividends, these move around but relatively little over time, whereas prices [for those stocks] are volatile. If the fundamentals of these assets are relatively stable, why are prices moving around so much? Given that the price should just be the discounted value of fundamentals, what’s causing these price movements?

Since the price is the discounted value of future fundamentals, there are only two possibilities. One, it’s all about the discount rate. The future fundamentals have low volatility, so maybe the discount rates are moving and that’s why prices have such a high volatility. The second possibility is that perhaps people make mistakes about future fundamentals. So even though the actual future fundamentals are quite stable, maybe people don’t believe they’re as stable. Sometimes they think that future fundamentals will go up and sometimes they

Sean Myers:

think they will go down. And maybe that’s what drives the price volatility.

These two possibilities imply very different things about what investors think about the future. If you look at the history of dividends and earnings and you use a statistical model, your expectations should continue to be fairly stable. If people’s expectations are very volatile, then that means that there’s a disconnect where people disagree with what a statistician would predict for the future. Alternatively, perhaps people agree with the statistician, but their discount rates are very volatile. Typically, the latter story has been what people have focused on. We’ve assumed that people are good at forecasting fundamentals, so they should agree with the statistician, and then we’ve tried to understand the discount rates.

What our research shows is that’s not the case. In fact, almost all of the price volatility seems to be about mistakes people make in their expectations of future fundamentals. When we asked investors and professional analysts about their forecasts of future earnings and dividends, those were much more volatile than a statistical model would imply.

Essentially, the cause of price volatility can only be one — either expectations of future fundamentals or discount rates. Anything that changes in one of those two factors takes away from the other. In a boom, prices are very high. You can split that increase in prices into two: Either people are more optimistic about future fundamentals, or they use very low discount rates.

We’re arguing in our paper that if you measure this optimism in booms and then pessimism [in periods of slump], it accounts for almost all of the price movement, so there’s almost nothing left to be explained by the discount rates. It’s a simple world where people have almost flat discount rates, and they price stocks just based on what they think about fundamentals.

What we find is that people’s actual expectations of returns tend to be quite flat over time. Booms and high prices seem to be associated with overall optimism about dividends and earnings, and busts seem to be associated with pessimistic views about dividends and earnings. Interestingly, even institutional investors have pretty flat discount rates.

For bonds, we find a similar element where bond yields seem to be driven largely by expectations of inflation, which essentially is the fundamental for bonds rather than real interest rate movements or discount rate movements.

Knowledge@Wharton:

we saw prices moving, we thought that people must know something that we don’t. We thought there must be some kind of terrible risk during a recession that makes people demand very high interest rates. And during a boom, we assumed that for some reason the whole world has become very safe, and that’s why discount rates have gone down. Whenever the world doesn’t match our econometric model, we assume that we messed up and we try to figure out that there must be some very important risks that we missed.

We argue in our paper that when our statistical model of how stocks should be priced is different from what we see in the world, this may largely be due to people making mistakes, rather than everybody having good views about the fundamentals and us getting the discount rates wrong. The big theme of our paper is that a lot of stock and bond price movement seems to be because of mispricing rather than some time-varying risk.

Our paper is intended to help improve models by saying that people are making mistakes, and these mistakes are actually very big and seem to account for almost all of the price variation. To an academic audience, the message is essentially that we need to reduce our emphasis on these discount rate models where we’ve been putting almost all our attention and we need to instead shift our focus onto models of how people actually think about fundamentals. How do people in the real world think about future inflation? How do they think about dividends? Why is their guess about the future so different from our statistical models? Why do we find a consistent pattern of mistakes over the business cycle? So, our research is pushing models to look more at how people form beliefs about the future and make mistakes in those beliefs, rather than trying to come up with different risk factors that could drive returns.

from his interview.

Whom did you survey? For stocks, we surveyed professional analysts on their earnings forecasts, and then for bonds and inflation, we looked at the Survey of Professional Forecasters (a quarterly survey of macroeconomic forecasts) and the Federal Reserve’s Survey of Consumer Finances.

Your paper is timely given the recent bull run in the markets.

Knowledge@Wharton: Myers:

In the last, say, 10 years, price-earnings ratios have been extremely high, almost higher than they’ve ever been. A statistical forecast is that this will eventually come back down to normal. We don’t know when, but when it does, that means there will be a long sequence of low returns for these stocks. The key question is, do all these investors know this or are they just very bullish about how these firms will perform? Do they think they’re going to have great dividends and earnings, and they don’t think that prices will come down in the future?

It’s certainly timely in the sense that we are in a period with very unusual prices. Prices have risen enormously, but the fundamentals, or dividends and earnings, have not. It’s just incredibly puzzling. Either dividends and earnings must skyrocket, and then these firms can still produce high returns. Or the fact that prices have diverged so far from fundamentals means prices will eventually come down, which would be the story that these prices are incorrectly high [and that] people are making mistakes about the future.

Everyone is in agreement that prices and fundamentals have to somehow move back together. They’ve diverged quite far. The question is: Do they come back together by the prices falling or by the fundamentals growing very quickly?

We find evidence that at least among professional analysts or inflation forecasters, it seems to mostly be beliefs about the fundamentals. People think these high prices are reasonable because we’re going to experience very high growth, whereas a statistical model says you’re probably going to be disappointed, and that these high prices likely will just be followed by low returns.

Knowledge@Wharton:

How is your paper an advancement over existing research on this subject?

The literature has a lot of theories on what could be happening. There’s the behavioral finance literature that says maybe people are making mistakes and that’s what’s causing all these unusual asset prices. The other theory is that people are rational, and they don’t make these kinds of mistakes, and instead they know some important risk that we don’t, and we need to figure this out.

This paper is essentially saying: Let’s go test this. When we ask people about their expectations, they seem to fall into the behavioral camp where their expectations are making these large, systematic errors that appear to drive most of the prices, rather than people giving us very accurate statistical forecasts where there must be some kind of risk factor that’s explaining the price movements. That is how we’re pushing this forward. Instead of having a theoretical debate, we can directly ask investors what they believe about the future.

Myers: