외환 거래 수수료 비용

본 논문은 국내 한 외국환은행(이하 은행)의 2012년 1월 2일부터 12월 28일까지 1년 동안 원/달러현물환 대고객 거래자료를 이용하여 대고객 외환시장(customer market)의 거래비용 구성요소를고찰한다. 먼저 정보비대칭비용(asymmetric information costs)이 대고객 외환시장의 거래비용구성요소로 포함되어 있는지를 살펴봄으로써 은행이 정보비대칭 위험이 있는 고객에게 제시하는환율(이하 고객환율)에 정보비대칭비용이 직접적으로 반영되는지를 고찰한다. 만약 고객환율에정보비대칭비용이 포함되어 있지 않다면 재고보유비용(inventory holding costs), 주문처리비용(order processing costs) 등 다른 거래비용 구성 항목들이 포함되어 있는지를 살펴본다. 본 논문이전체표본과 주문크기, 고객유형 및 일중 시간대 별 부분표본을 대상으로, 고객환율 중에서은행간시장의 시장환율에 의해 결정되는 부분인 커버환율을 차감한 FX 수수료에 대해 거래비용구성요소를 분석한 주요 결과는 다음과 같다. 첫째, 거래비용 구성요소 중 정보비대칭비용은전체표본뿐 아니라 모든 부분표본에서 나타나지 않는다. 이는 고객 딜러가 고객이 사적 정보를보유한 고객이라고 기대를 하더라도, 해당고객에게 제시하는 환율에는 정보비대칭비용이 반영되지않음을 의미한다. 이러한 결과는 Glosten and Milgrom(1985)의 이론모형과는 상충되고, Osler, Mende, and Menkhoff(2011)의 “3국면 가격발견” 가설과는 부합하는 측면이 있다. 이 가설은 고객딜러가 고객 중 사적 정보를 보유한 고객과 거래를 하지만, 해당 사적 정보가 딜러가 해당고객에게제시하는 환율에는 즉각적으로 반영되지 않고 추후에 은행간시장 환율에 반영된다는, 즉외환시장에서 가격발견은 대고객시장이 아닌 은행간시장에서 이루어진다는 것이다. 둘째, 재고보유비용 역시 전체표본과 모든 부분표본에서 나타나지 않는다. 이처럼 대고객 거래에서재고보유비용이 나타나지 않는 이유는 국내 은행간시장의 유동성이 높고, 대부분의 거래가전자중개시스템을 통해 이루어짐에 따라 은행의 재고위험이 높지 않기 때문인 것으로 해석된다. 마지막으로, 주문처리비용이 전체표본 및 모든 부분표본에서 뚜렷하게 나타난다. 그리고 주문크기별 부분표본 분석에서는 주문처리비용의 크기가 주문크기와 반비례 관계를 갖는다. 이러한 현상은고객 딜러가 소규모 거래보다는 대규모 거래를 더 유인함으로써 거래 수수료 수익을 최대화하기위함으로 해석된다.

The question of through what mechanism private information is reflected on the exchange rate is a major topic of FX market microstructure research from the perspective of price discovery in FX market. Lyons (1997) points out that “private information in the interbank FX market is attributable more to information obtainable through the flow of customer orders than to information superiority 외환 거래 수수료 비용 on macro variables.” Most previous researches on this topic have been largely limited to interbank markets. Through empirical analysis on interbank markets, previous foreign researches, including those by Lyons (1996), Evans and Lyons (2002) and Payne (2003) as well as domestic researches by Seon and Eom (2006, 2010b), suggest evidence that the asymmetric information cost is reflected on the exchange rate for interbank markets. However, since the FX market maintains a two-tiered structure consisting of the interbank market and the customer market, analysis on the interbank market alone presents limitations in understanding the overall process by which private information is reflected on the exchange rate determined in the FX market. By examining whether the asymmetric information cost is included in the composition of trading cost of 외환 거래 수수료 비용 customer exchange rate in the won/dollar customer FX market, this study seeks to test whether the 3-stage hypothesis of Osler et al. (2011) stands true in the won/dollar FX market. If the asymmetric information cost is included in the trading cost of the customer exchange rate, such would indicate that private information included in the customer order flow is incorporated into the exchange rates in the customer market. However, if the asymmetric information cost is not included in the customer exchange rate, then it would mean that private information included in customer order flow is reflected on exchange rates in the interbank market as championed by the 3-stage hypothesis suggested by Osler et al. (2011) rather than on the customer market. Such a research is expected to improve the degree of understanding on the mechanism through which private information possessed by customers is reflected on the exchange rate, and how price discovery is done within a two-tiered FX market consisting of an interbank market and a customer market. This study is significant in that it is the first research on price discovery in the won/dollar customer market. Moreover, while this study tests the 3-stage hypothesis proposed by Osler et al. (2011), it differs from their research in the following two aspects in terms of methodology. First, pursuant to the trading cost 외환 거래 수수료 비용 analysis method suggested by Foucault et al. (2013), this study analyzes the components of the trading cost of customer exchange rate by categorizing them into three categories: the asymmetric information cost, the inventory holding cost, and the order processing cost. This analytical method is used because, in addition to the asymmetric information cost, the cost of inventory holding and the order processing cost may be also included in the trading cost, whose order flow influences exchange rate fluctuation. Such analysis is advantageous in that, if the asymmetric information cost is not included in the customer exchange rates, it can then suggest whether the inventory holding cost or the order processing cost is reflected on the customer exchange rate and the degree by which the said cost is reflected. As far as the author is aware, this study is the first study that directly analyzes each component of the trading cost of the customer market exchange rate by dividing it into information asymmetry cost, inventory holding cost, and order processing cost. Second, this study categorizes the customer exchange rate into the two elements— the cover rate and the FX fee—and seeks to identify the components of trading cost in the customer FX market that are attributable to each element. Osler et al. (2011) analyzes the asymmetric information cost inherent in the customer exchange rate spread quoted by the dealer to the customers. However, the customer exchange rate is composed of two elements: the cover rate and the FX fee. The cover rate is the price offered by the interbank dealer to the customer dealer and thereby reflects the exchange rate of the interbank market at the times of trading, whereas the FX fee is the price offered by the customer dealer to the customer and may be applied differently depending on the types of customer. If the customer exchange rate is divided thus into the aforementioned two elements, and if each element’s contribution to composition of the trading cost is analyzed, it becomes possible to discover what trading cost is reflected on the price by the interbank dealer and the customer dealer as providers of liquidity. This study analyzed the components of the FX fee and the cover rate trading cost by using materials from a FX bank (henceforth the “bank”) on the won/dollar spot exchange customer transactions over a time span of 1 year from January 2, 2012 to December 28, 2012. The major findings are as follows: First, among the components that make up the trading cost for the FX fee, the asymmetric information cost appears neither in the overall sample nor in any of the partial samples. This indicates that even if the customer dealer expects that the customer possesses private information, the asymmetric information cost is not reflected on the exchange rate offered to the customer. Such result conflicts with the 외환 거래 수수료 비용 prediction by the theoretical model of Glosten and Milgrom (1985) and is consistent with the “3-stage price discovery theory” of Osler et al. (2011) to a certain extent. According to this theory, while the customer dealer may engage in trades with customers who possesses private information, such private information is not reflected on the exchange rate offered by the dealer to the customer but is rather reflected later on the interbank exchange rate. In other words, price discovery is made in the interbank market rather than in the customer market. Second, among the components 외환 거래 수수료 비용 that comprise the trading cost for the FX fee, the inventory holding cost manifests neither in any of the partial samples nor in the overall sample. Such non-appearance of the inventory holding cost in the customer transactions seems to be attributable to high liquidity in the Won/Dollar interbank 외환 거래 수수료 비용 market and low risk of inventory holding, which is itself attributable to the fact that most trades are performed through the electronic brokerage system. Third, among the components that make up 외환 거래 수수료 비용 the trading cost for the FX fee, the order processing cost manifests clearly and differs depending 외환 거래 수수료 비용 on the order size, customer type, and time of day. An analysis of the order processing cost based on trade size—namely, large, medium, and small size of trades—shows an inverse relationship between the size of trades and the order processing cost. This phenomenon is thought to be attributable to the customer dealer’s effort to elicit large scale transactions and thereby maximize profits from fees. In addition, the analysis based on customer type—namely financial firms, large firms, and small and medium firms— shows that the order processing cost of financial firms is larger than that of 외환 거래 수수료 비용 large firms and small and medium firms. This phenomenon seems to be attributable to the fact that in the case of financial firms, the size of orders is relatively small and higher percentage of transactions is processed at certain times of the day when the order processing cost is relatively high. Meanwhile, with regards to time, the order processing cost peaks during the earlier times of the market and gradually declines afterwards. Such phenomenon of higher order processing cost during the early market is thought to be due to the smaller average size of orders during this period. Finally, the components that make up the trading cost with regards to the cover rate are either statistically insignificant or have small values of lower than 0.10 Won, which is the minimum price unit, and are thus economically insignificant. This implies that component elements, such as the asymmetric information cost, the inventory holding cost, and the order processing cost, are not reflected on the cover rate offered by the interbank dealer to the customer dealer.