Quote revisions in related assets

I present a model of a dealer market with multiple, correlated assets and private information, which generates two principal insights. First, quotes, responding to total order flow, change in the absence of individual order flow. Second, the market structure induces lagged price adjustment at short lags even with perfectly rational market participants. In particular, quotes for assets with low transaction costs lead the quotes for assets with high transaction costs.;In the first empirical analysis, I study how quotes respond to market returns in the absence of individual order flow using daily data for a large sample of stocks. I find that returns, measured using quotes, during non-trading periods are significantly less sensitive to market movements than returns on days with positive trade volume. However, returns when trade resumes fail to correct for low betas during non-trading periods, consistent with quotes correctly reflecting changes in the return generating process.;In the second empirical analysis, I study the relation between transaction costs and price discovery in a sample of dual class issues. The issue with lower average transaction costs provides two-thirds of the price discovery about the common value. Furthermore, at least 20% of the quote revisions in the high transaction cost issue are triggered by trades in the low cost asset, while the reverse occurs almost never. Quote revisions in the low transaction cost asset are followed quickly by quote changes in the high transaction cost issue. Together, these empirical findings strongly support the hypothesis that market makers learn from trading activity in other assets, and revise quotes accordingly.;In the third empirical analysis, I show that information flow and transaction costs are also related in a sample of stocks from a narrowly defined industry. The quotes of stocks with low transaction costs significantly lead the quotes of stocks with high transaction costs. Even when controlling for trade, low transaction cost stocks' quotes predict quotes of high transaction cost stocks. Limit order providers appear to learn from activity in low transaction cost stocks about the fundamental value of similar stocks with higher transaction costs.