More information is not necessarily better information.
That’s the conclusion from an October 2012 New York Federal Reserve study flagged Sunday by Chris Dixon.
In “Have Financial Markets Become More Informative?”, Jennie Bai, Thomas Philippon, and Alexi Savov argue that over the past 50 years, markets have not grown more efficient about how to allocate capital, at least via stocks.
They start out by offering the standard view of what should have happened:
An increase in the supply of information leads to greater price dispersion as markets differentiate among firms. As a result, market prices become stronger predictors of earnings. The size of the predictable component is given by the dispersion [scope of the range] in market prices times their forecasting coefficient.
They next define their variable, which they call “predicted variation.”
This is a score of the ability of stock prices to predict earnings.
Or as they put it, it is “the size of the predictable component of earnings that is due to prices.”
So in other words, they want to know whether the size of the predictable component of earnings has grown.
And it has not.
Here’s the key graph showing that predicted variation declined after 1980 and has remained stuck ever since.
Instead, here’s what they found:
- Stock prices have not increased the magnitude of predicted variation — instead predicted variation has remained “remarkably stable.”
- Volatility has not increased predicted variations. In fact, increased volatility has been caused by more earnings surprises, which have increased by about 2% since 1990
- Higher stock prices are actually a decent predictor of R&D investment
- But R&D investment is not a good predictor of earnings
- And bond spreads do not predict R&D or greater capital expenditure
What happened? The researchers say it may have something to do with the introduction of the NASDAQ in the early ’70s and a proliferation of tech firms that are harder to value.
we examine S&P 500 firms whose characteristics have remained stable. In contrast, running the same regression on the universe of firms appears to show a decline in informativeness. We argue, however, that this decline is consistent with a changing composition of firms: the representative firm today is harder to value. Consistent with this interpretation, we show that the proportion of firms with a high dispersion of analyst earnings forecasts has increased following the introduction of NASDAQ.
But they conclude by arguing that whatever IT improvements markets now enjoy, it has not made markets better:
These results appear to contradict the view that improvements in information technology have increased the availability of low-cost information. A possible explanation is that the relevant constraint for investors lies in the ability to interpret information rather than the ability to record it. If this is the case, a rise in the quantity of data need not improve informativeness or the allocation of resources.
Faster trading is not smarter trading, apparently.
Business Insider Emails & Alerts
Site highlights each day to your inbox.