معاملات نویز سوداگرانه و اعمال نفوذ در بازار ارز

We investigate the possibility that in the foreign exchange market uninformed speculators find it convenient to trade on noise in order to gain an informational advantage they can exploit in future. In a two-period model, we analyze the trade-off between the cost of the “informational investment” and the profits this brings about, studying the optimal manipulation strategy under different hypotheses on the activity of market participants. Our results give a possible explanation for the presence of noise trading in the foreign exchange market.

A large component of transactions in securities markets derives from noise traders. According to Black (1986), noise traders are agents who sell and buy assets on the basis of irrelevant information. These speculators do not possess inside or fundamental information and trade irrationally on noise as though this gave them an edge. Despite its irrational nature, noise trading represents an important aspect of the functioning of securities markets, because it reduces the risk of market crashes and facilitates transactions among agents. Indeed, if all traders were rational it would not be convenient to gather information, because prices would be fully revealing. Conversely, when noise traders are present, rational speculators will gain profits at the expense of the irrational ones. As in this case prices will not be fully revealing, there would still be an incentive to gather information, so that in practice noise trading may be beneficial for the efficiency of securities markets. Different types of behavior are associated with noise and liquidity trading.1 These comprise hedging strategies, such as portfolio insurance and stop–loss orders, popular models of forecasting and trading, such as chart and technical analysis, and so forth. In particular, considering the market for foreign exchange, there is ample evidence of widespread use of chartism by traders. Indeed, on the basis of a survey conducted in the London foreign exchange market, Allen and Taylor (1990) point out that most traders consider chartism at least as relevant as fundamentalism in the formulation of exchange rate expectations in the short run. Likewise, Frankel and Froot (1987), using survey data in the foreign exchange market, find strong evidence of extrapolative expectations and attribute it to the utilisation of chartism by professional traders. A problem with this description of noise trading in securities markets is its profitability. In fact, it is commonly believed that noise traders incur losses, because they buy when prices are high and sell when prices are low. Nevertheless, empirical studies question this thesis. Goodman (1979) compares the profitability of different forecasting techniques in the market for foreign exchange. In the seventies, the worst technically orientated forecasting technique was three times more profitable than the best fundamentally orientated one. Likewise, Schulmeister (1988) finds that large speculative profits of commercial banks in the market for foreign exchange are due to the utilization of technical trading rules. Levich and Thomas (1993) confirm the thesis that chartism is profitable employing a bootstrap approach, while Menkhoff and Schlumberger (1995) extend previous results to a longer period. Therefore, explaining the use of chartism and other forms of noise trading represents an important topic of research. Several explanations have been proposed. Frankel and Froot (1986) suggest that if traders learn slowly the fundamental value of a currency, chartists can dominate fundamentalists and lead the exchange rate on a bubble path. De Long et al. (1990) show that if irrational traders can bear more risk than other traders, they can gain larger profits than risk-averse rational speculators. Froot et al. (1992) show that if traders have short horizons, they may find it convenient to trade on the basis of information completely unrelated to fundamentals. Palomino (1996) proves that in small markets in which investors are not price takers, noise traders might hurt sophisticated traders more than themselves. Our analysis suggests an alternative explanation, based on a particular mechanism of manipulation of expectations and exchange rates. Even if this analysis could be applied to other dealer markets, we refer principally to the foreign exchange market because it is there that most of the evidence on the profitability of noise trading is concentrated. Moreover, it is in this market that the mechanism of manipulation we suggest is most likely to take place. Noise trading may be part of a valid strategy of manipulation and a source of speculative profits, because it can bring about an informational advantage. This happens in a dealer market when noise trading interferes with the learning process of dealers. In fact, assuming that dealers are competitive and cannot distinguish between informed and uninformed trades, noisy market orders placed by an uninformed speculator will reduce their ability to learn the fundamental information contained in the order flow. On the other hand, the uninformed speculator will be able to extract more of the fundamental information present in the flow of orders. Then, in subsequent periods the speculator can use this informational advantage to recoup the losses incurred with the initial noisy market orders and also gain some profits. One should point out that our contribution extends the analysis put forward by Madrigal (1996). In fact, she considers a dealer market in which some speculator observes some of the market orders placed by liquidity traders and acquires an informational advantage with respect to the dealers she can exploit to gain speculative profits. Indeed, the main difference between the two contributions is that whilst in her model the speculator possesses some non-fundamental private information, in our analysis we discuss the case in which the speculator does not have any initial informational advantage. In what follows we consider a model of the foreign exchange market with private information on the fundamental value of a foreign currency. This framework permits studying the impact of noise trading on the learning processes of the market maker and the uninformed speculators and on the stochastic process governing the exchange rate. In addition, we can analyze the trade-off between the cost of noisy market orders and the informational advantage they provide and study under which conditions they may be part of a profitable trading strategy of the uninformed speculators. This paper is organized as follows. In Section 2 we show that an uninformed trader can gain an informational advantage placing random market orders in such a market. In Section 3 we prove that under particular circumstances the speculator can gain positive profits. In Section 4 we extend our basic formulation in several directions to investigate the robustness of our results. A conclusion which summarizes the results of our analysis completes the paper, while Appendix A gives detailed proofs of Propositions and Lemmas.

In this paper we have considered a possible reason which may induce uninformed speculators to trade on the basis of irrelevant information in the market for foreign exchange. We suggest that noise trading may be used to manipulate expectations and exchange rates in order to gain an informational advantage and hence a profit opportunity. To investigate this opportunity, we have developed a two auction model of the foreign exchange market, in which uninformed speculators can exploit the impossibility of the market maker to distinguish between informed and uninformed clients to manipulate her expectations and the exchange rates and hence gain speculative profits. The main results of our analysis are as follows. Injecting noise in the order flow of the dealer, a rational uninformed speculator interferes with the learning process of the market maker and determines her degree of uncertainty on the fundamental value of a foreign currency. The possibility of reducing the resiliency of the market, that is the speed of convergence of the exchange to its fundamental value, permits the speculator to gain and preserve an informational advantage with respect to the dealer. The acquisition of this informational advantage corresponds to an “informational investment”, because the dealer provides liquidity to the market at a cost and hence placing random market orders is expensive. On the other hand, the informational advantage yields profits from future informed trading. The solution of the trade-off between these costs and profits determines the optimal volume of speculative noise trading. The equilibrium of the foreign exchange market with sporadic central bank intervention clearly indicates that speculative noise trading may be part of a profitable trading strategy in the foreign exchange market. In effect, the practice of giving spurious signals to other market participants to condition their expectations is common among professional traders. Investigating the market for foreign exchange, Lyons (1995) finds that a large component of the order flow of a dealer does not contain information. Because most transactions in the foreign exchange market are between professional traders, we can interpret this as evidence of attempts of manipulation. However, our analysis also outlines the limits that exist to this kind of speculative activity. In particular, when central bank intervention is continuous, the scope for manipulation can be severely reduced, because the monetary authorities have the faculty of excluding uninformed speculators from the market. This also suggests that in other dealer markets, in which informed agents trade continuously, the kind of manipulation discussed here is less likely to emerge. Finally, since noise trading is a free good, competition among uninformed speculators will reduce the incentive they have to trade on pure noise.