Portfolio Considerations in Risk Management by Financial Intermediaries: Evidence from Trading Behaviour of London Equity Dealers

Abstract

We use a rich dataset of trades of London equity dealers across all stocks to investigate the importance of portfolio considerations in risk management by financial intermediaries. We examine how the trading and pricing decisions of these dealers in individual stocks are conditioned by their ordinary and inventories in that stock, where the inventory is the dealer's ordinary inventory in that stock corrected for any reinforcing or offsetting effects arising from her inventory positions in other correlated stocks. We examine two types of equivalent inventories: the first based on correlations in total returns (as in Ho and Stoll (1983)), and the other based only on correlations in the unhedgeable component of returns (as in Froot and Stein (1998)). In addition to the mean reversion in ordinary inventory documented in earlier studies, we find strong mean reversion in both total and unhedgeable equivalent inventories indicating that portfolio considerations do influence dealers' overall risk management. However, we find that portfolio considerations are not important in determining whether, and at what price, a dealer executes a particular trade. Dealers with divergent ordinary (rather than equivalent) inventories execute large public and inter-dealer trades; and dealers offer significantly higher price improvement, and charge significantly lower effective spread, for large public trades that reduce the divergence of ordinary inventories rather than the those that reduce the divergence of equivalent inventories. We also find that portfolio considerations at the individual trade level are neither significant for small dealer firms nor for large dealer firms, and also neither significant at the level of desks, nor for the dealer firm as a whole. Cross-sectionally across stocks, we find that the intensity of mean reversion in individual stock inventories depends only on the specific risk of the stock and not on its beta or its industry beta. Our findings are contrary to the expectation that a dealer firm would actively manage their inventory risk exposure by taking off-setting positions in different individual stocks based on correlations between pairs of individual stocks. Instead, these findings are consistent with the decentralization of the market making function within securities firms to individual dealers. In this context, our results appear to be driven by the policy of evaluating and rewarding individual dealers according to the trading profits they generate in the stocks assigned to them, rather than by the difficulties of coordinating and sharing information instantaneously across different individual dealers.