On a recent visit to China, we were given access to a tape of Chinese stock price data. To our great surprise, we found that nearly all of the stocks move together—called comovement—either up or down, on any given day. This is not what stock prices are supposed to do. The economic purpose of a stock market is to channel capital to companies that should grow and away from companies that should not. This is because a firm’s cost of capital falls when its stock price rises and vice versa. The efficient markets hypothesis holds that shareholders’ opinions about firms’ future prospects affect stock prices and thereby allocate capital among firms. If the prices of all companies in the economy rise and fall together, the stock market may be failing in its role as an information processor and microeconomic capital allocation device.

Subsequently, we compared the degree of stock comovement across 40 countries using weekly and biweekly stock returns data for 15,920 firms. In rich countries, on average, just over 50 percent of stocks move in the same direction. In poor countries, the fraction moving together is economically and statistically significantly higher.

In other words, comovement of stock prices is negatively correlated with per capita GDP. The result is not an artifact of market size, for the stock markets of small, rich countries, such as those of Denmark, Ireland, and New Zealand, exhibit little comovement; while the stock markets of large, less-developed countries, such as those of Brazil and India, exhibit substantial comovement. Both individually and jointly, country size, various Herfindahl indexes (measuring industrial concentration), and various macroeconomic volatility measures all fail to account for the finding. Comparing firms’ returns on assets also fails to explain why stock prices move or do not move together on the stock exchange.

In contrast, institutional development can explain stock price comovement better than per capita GDP does. In countries with more corrupt governments, stock prices have a greater tendency to move together. (Political corruption perhaps impedes information-based trading, leaving the markets to "noise" traders. They are the opposite of "rational" investors who buy securities based on accurate expectations of future returns.) Further, reduced legal protection for public investors also accompanies high comovement. These findings raise the possibility that inadequate legal and political systems may render the gathering of and trading on firm-specific information relatively unprofitable. Traders may still be willing to invest in diversified country portfolios, but see little point in picking stocks if corporate or political insiders routinely siphon off economic profits. Thus, stock prices in countries with deficient institutional structures exhibit a high degree of comovement.

We also find that a high level of legal protection for public investors is correlated with more firm-specific price variation, consistent with the view that such protection encourages the capitalization of firm-specific information into stock prices.

Our results suggest that simply establishing a stock market in a low-income country is unlikely to improve its microeconomic allocation of capital. Stock markets appear to channel capital selectively across firms only in economies that have relatively honest governments and relatively high levels of protection for public investors’ property rights over their investments.