Arbitrage Equilibrium with Transaction Costs
Niehans, Juerg, Journal of Money, Credit & Banking
This paper explores the significance of quadratic transaction costs in a general-equilibrium model of asset arbitrage. The model is designed to shed light on a specific macroeconomic policy problem. Robert Mundell pointed out twenty-eight years ago that the effectiveness of macroeconomic policy may depend crucially on the degree of mobility in international capital flows (Mundell 1964). Capital mobility, in turn, is often said to depend on transaction costs. The apparent decline in transaction costs in financial markets in the course of recent decades is thus seen as die cause of a progressive increase in die mobility of capital flows and thus of a loss in policy effectiveness. This led James Tobin to the proposal "to throw some sand in the wheels of our excessively efficient international money markets" (Tobin 1982, p. 489) by imposing a tax on all foreign exchange transactions.
It is indeed a widespread notion that transaction costs help to insulate an economy against foreign disturbances while "bottling up" domestic disturbances at home. In the absence of transaction costs, small changes in interest rates are expected to trigger large asset flows, but high transaction costs are supposed to reduce the "churning" of the market. This paper examines the validity of this notion. The central question is how the reaction of market conditions to exogenous disturbances is affected by transaction costs. The main conclusion will be that the transmission of disturbances indeed depends on transaction costs, but more on the details of their relationships than on their absolute level. A general decline of transaction costs, therefore, does not necessarily strengthen the transmission of disturbances, nor can artificial impediments to transactions be relied upon to reduce them.
Capital flows can be considered at two levels. At the first level they are interpreted as changes in a country's net foreign assets, the counterpart to a current account imbalance, which implies changing values of portfolios. In this sense it is not clear that capital mobility has, in fact, increased, and the extensive debate initiated by Martin Feldstein and Charles Horioka (1980) seems to indicate that it may still be quite limited. At the second level of analysis capital flows are interpreted as arbitrage operations between different assets within portfolios of given value. It is these arbitrage operations whose cost appears to have declined so dramatically, thus creating the impression of an almost infinite capital mobility. These are also the transactions to which Tobin's proposal is meant to apply. The present paper is largely restricted to the second, more narrow, perspective. Its analytical framework is one of pure arbitrage. This is defined to mean that each agent is restricted to shifts between assets in a portfolio of given market value. Saving, investment, and asset accumulation are briefly considered at the end. As a consequence, the present analysis can be no more than a preliminary to a study of genuine international capital flows, but it is a potentially significant preliminary.
Transaction costs are defined by the fact that they depend on the volume of assets traded and not on the volume of assets held. In part they consist of explicit brokerage fees as reflected in spreads between buying and selling prices, but these probably constitute not more than a rather small fraction of all transaction costs.(1) In addition, the latter include the search and information costs of the market parties, the costs of defining and transferring property rights, the costs of drawing up, executing and enforcing contracts, the costs of communication of all sorts, and possibly the costs of litigation. The fact that some of these items are quite intangible does not make them less important. A microeconomic analysis of transaction costs would probably assign a dominant role to uncertainty, but for the present purposes they are interpreted simply as the (certain) costs of establishing a required level of certainty. …