Globalisation has flattened the world in some ways, but not in others. Many years ago, economists Martin Feldstein and Charles Horioka observed that national savings and investment rates are highly correlated, indicating that investors tended to keep their money at home, rather than diversify globally, even without restrictions on capital flows. Subsequently, other economists documented this ?home bias? for portfolios of shares?investors tilt toward holding shares of their own country?s companies.
This home bias has persisted, even as the telecoms revolution has reduced information asymmetries, calling into question an explanation based on local investors ?sticking with what they know best?. That is, the home bias has been too large to be explained by given differences in information. More recent work, though, came up with an answer?investors have limited capacity for information gathering, and as they choose where to allocate their attention, initial small differences in information get magnified. But how to test that theory?
The Internet not only lets people search for information, it lets researchers measure what people are searching for. My colleague, Thomas Wu, has pioneered the use of these measurements to show that attention matters, and that the allocation of attention affects shareholdings and movements in share prices. In one paper, with Jordi Mondria and Yi Zhang, Wu uses data from AOL searches to show that US investors favour assets of countries they are more familiar with, and this behaviour has a positive feedback to attention allocated to those countries, as measured by searches.
In another paper, with Mondria, Wu uses Google search data to compare local and national searches within the US. The authors indeed find that attention allocation is biased towards local stocks and that companies receiving an increase in asymmetric (local versus national) attention earn higher returns. This result is very striking, because it shows that distance matters, even within a highly developed and sophisticated market economy. Even in the US, local information is important and has significant effects on asset values.
India is an interesting case for thinking about the role of information, because of the connections between hi-tech in the US (particularly Silicon Valley) and India. Indians have also become prominent in the US financial sector (though recently in an undesirable way, in the big insider trading case of the Galleon hedge fund). I have not examined the links between information gathering and share prices, but the patterns of search alone may tell us something.
Consider India?s iconic software company, Infosys. If one examines searches for its stock symbol, INFY, over the last five years, this is what the data show. Google Insights presents normalised numbers, showing the number of searches for the term in question relative to total searches. Furthermore, this ratio is normalised so that the highest number during the period is scaled at 100. Thus, for searches from India, the peak is in June 2005, with the second highest peak (an index of 94) in September 2008. The third highest score is 79, and occurs in July 2006 and April 2009. The monthly average index for the last five years is 45.
The pattern of searches for INFY in the US is very different. The peak is in April 2007, with a score of 99 in January of that year, and scores of 97 in May and June and 96 in October. Clearly, US-based investors were giving Infosys much more relative attention that year than were their Indian counterparts. Infosys, of course, happens to have American Depository Receipts traded on the US NASDAQ stock exchange. Its stock symbol does not figure at all in searches from other countries around the world, even those with large Indian diaspora populations, such as Britain.
One can find differences over time for other companies as well, for example, Cognizant Technology Solutions and HDFC Bank. Local attention varies quite dramatically across the two countries, for companies that have businesses either spanning both locations, or investors who have some connection to both countries. More global companies, like IBM, do not display this kind of local differences in attention. So local presence feeds into local attention.
To say more, one should look at stock returns and see how they correlate with attention. Do Indian investors have a better sense of what is going on with Indian companies? Is any initial local knowledge advantage magnified by increased attention at certain times? As India continues to globalise, as it opens up more to foreign capital, and as more Indian companies become known in the US, and traded on US stock exchanges, understanding the patterns of local knowledge and its transmission to other locations will become increasingly important. This will be yet another factor for financial managers to factor into their calculations.
The author is professor of Economics at University of California, Santa Cruz