We further find that a household’s sensitivity to neighbors’ investment choices increases with the population of the household’s community. Such diffusion in stock trading affects individual investors’ asset allocation decisions. For example, although residents in larger metropolitan areas have substantially more diverse investment opportunities and tend to invest more in local stocks, we find that their local stock investments tend to remain just as concentrated as those made by residents of less populated communities (who have a significantly smaller pool of potential local investments). This tendency is consistent with the notion that residents in more populous geographic areas might be exposed to word-of-mouth effects to a higher degree than residents in less highly populated areas.
Finally, to disentangle the contributions of correlated preferences and the structure of the local economy to the observed correlation between individual investors’ stock purchases and those of their neighbors from “word-of-mouth” effects, we conduct two tests. First, we consider the level of sociability of the state to which the household belongs and find that the relation between industry-level household purchases and neighborhood purchases is substantially stronger among households in the more sociable states. Second, we consider the households’ own preferences (as revealed by the composition of their respective portfolios across industries at the beginning of each quarter), preferences of the households’ respective neighborhoods (as revealed by the composition of the neighborhoods’ aggregate portfolios), as well as the composition of local firms and workers by industry. We find that one-quarter to one-half of the overall diffusion effect among both local and non-local investments cannot be attributed to these sources. We regard the remaining portions of the diffusion effect as a conservative lower bound on the impact of word-of-mouth communication effects on household trading decisions. Disentangling the overall information diffusion effect into word-of mouth communication and
5
other diffusion effects potentially yields further insight as to how correlated trading among individuals may influence stock prices.
Our results complement and extend those of Hong, Kubik, and Stein (2005), suggesting that word-of-mouth effects are a broad phenomenon that affects financial decisions made by both mutual fund managers and individual investors. The two studies provide evidence supportive of word-of-mouth effects using different techniques, thereby adding to the robustness of the overall finding. Hong, Kubik, and Stein (2005) rule out alternative explanations for correlated trading patterns by examining trading activity before and after Regulation FD and by focusing on trades in stocks for which investor relations are unlikely to be a contributing factor (stocks not local to the managers and small stocks). In this paper, we disentangle possible explanations for correlated trading patterns by exploiting differences in sociability of communities across the U.S., as well as introducing several controls for similarity in investment preferences within the community (as manifested by previous household investment decisions) and the composition of the local economy.

Download Address: http://www.nber.org/papers/w13201.pdf    
Download Restriction: Access to the full text is generally limited to series subscribers, however this transition economy free access is provided. More information about subscriptions and free access is available at http://papers.nber.org/papers/w13201