조세피난처는 적은 금융규제와 철저한 익명성 보장을 특징으로 한다. 따라서 이곳으로 부터의 자본흐름은 불순한 의도를 가지고 있고 시장에 부정적인 영향을 미칠 것으로 의심받지만, 이들의 영향에 대한 실증적인 증거는 부족하다. 본 연구는 조세피난처로 부터의 자본흐름이 한국 주식시장에 미치는 영향에 대해 가격충격과 변동성 측면에서 실증분석하였다. 검증에 앞서 먼저 외국 자본흐름을 국가별로 살펴본 후, OECD가 지정한 조세피난처 중 가장 규모가 큰 5개국(케이만제도, 버뮤다, 바하마, 미국령 버진 아일랜드, 영국령 버진아일랜드)을 조세피난처로 정의하였다. 이들은 전체 외국인 거래 금액의 약 10%를 차지하였다. 조세피난처 외국인들은 시장수익률이 상승(하락)한 후 매수(매도)하고, 수익률이 상승(하락)한 종목을 매수(매도)하는 추세추종매매(positive feedback trading) 행태를 보였다. 본격적인 분석 결과는 다음과 같다. 첫째, 가격충격을 살펴보기 위해 사건연구 방법을 통하여 조세피난처 외국인의 대규모 거래일을 선택하여 그 거래일 5일 전후의 수익률을 살펴보았다. 조세피난처 외국인의 대규모 순매수와 순매도거래는 사건일 당일 주가에 유의적 가격충격을 주었으며, 특히 순매수거래의 경우 그 가격충격이 사건일 이후에도 지속되는 현상을 보였다. 둘째, 패널 회귀분석을 통해 조세피난처 외국인의 거래가 증가할수록 주식의 변동성이 증가하는 현상을 발견 하였다. 위의 결과들은 조세피난처의 자본흐름이 한국 주식시장에 유의한 영향을 줌을 의미한다. 조세피난처가 아직도 비밀의 영역으로 남아있음을 생각할 때, 본 연구 결과는 조세피난처로부터의 자본흐름이 더욱 투명하게 개선되어야 함을 시사하며 OECD와 G20이 추진하는 조세피난처 개선 조치들을 지지하고 있다.
The Organisation for Economic Co-operation and Development (OECD) (1998) defines a tax haven as having the following characteristics: (a) no or only nominal taxes, (b) a lack of effective exchange of information, (c) a lack of transparency, and (d) no requirement that the activity be substantial. As these guarantee anonymity and little regulation in financial transactions, capital flows from tax havens are suspected to derive from dishonest motives and to have a negative impact on the local financial market. In Korea, there also has been a debate about local investors’ round-tripping through accounts in tax havens. Specifically, local investors set up paper companies in tax havens and camouflage themselves as foreign investors by trading local securities through these companies. The Korean media often refer to these investors as ‘black- haired foreigners’. There is, however, little empirical evidence of the effect of capital flows from tax havens because relevant data are not available.
This paper empirically investigates the impact of capital flows from tax havens on the Korean stock market using the 246 country codes provided by the Korea Exchange (KRX). The primary data source is the TAQ (trade and quote) dataset, which includes the full history of all trades made in the Korea Composite Stock Price Index (KOSPI) market between January 2006 and June 2008. Each trade record contains the price, quantity, time-stamp, investor class, and, most importantly, the country of origin of both buyers and sellers.
First, I compute the dollar trading volume of both purchases and sales across nations and list all the foreign countries investing in the KOSPI market. The summary statistics indicate that investors from the U.K. constitute a quarter of the total trading activity of all foreign investors in Korea. U.S. investors account for roughly 14% of all trades from foreign countries. The Cayman Islands, which is a tax haven, is the third largest foreign country investing in Korean stocks, accounting for 9% of all foreign trading. I consider a country a tax haven if it was listed in the 1998 OECD report on tax havens. Finally, I define the following 5 jurisdictions as tax havens: the Cayman Islands, the Bermuda, the Bahamas, the British Virgin Islands, and the U.S. Virgin Islands.
Trades from these tax havens amount to about 10% of the total trading volume of foreign investors in the KOSPI market. This suggests that trades from tax havens make up a substantial proportion of all foreign trading and that their impact on the KOSPI is likely to be significant.
Next, I investigate whether foreign investors from tax havens pursue positive feedback trading strategies. These strategies are considered to have a destabilizing effect because their sales lead to falls in the stock prices and their purchases increasing in price. This may contribute to an increase in the volatility of stock returns. Such trading can also destabilize capital flows as investors rush into booming stocks and flee from falling ones. Empirical results show that tax haven investors buy more Korean stocks on days after a market rises and sell more on days after it falls. They also buy more of the stocks that outperformed the market the previous day. These results mean that foreign investors from tax havens are positive feedback traders.
Finally, I examine whether foreign trades from tax havens have a significant impact on stock prices and increase the volatility of stock returns. In the first study, I select event days with large net purchases (sales) from tax haven investors, and compute abnormal returns for 11 days surrounding these events. The null hypothesis that tax haven investors do not have an effect on the Korean stock market can be rejected if significant abnormal returns follow large trades from tax havens in the same direction. The results show that both large purchases and sales from tax havens have a significant price movement within the event days and that the impacts are larger for purchases. In particular, the price impact of large purchase trades from tax havens persists one day after the event, and their cumulative abnormal returns are significantly positive for five days. These results mean that large trades by tax haven investors have a significant price impact on Korean stocks. In the second study, I use panel regression to examine the relation between stock volatility and trades from tax havens. The results show that stock volatility increases significantly as both net purchases and net sales from tax havens increase.
Overall, the results in this paper allow the conclusion that capital flows from tax havens have significant effects on the Korean stock market. It is thus necessary for us to know who the tax haven investors are and what they do. However, tax havens remain domains of secrecy. This paper supports the recent efforts of the OECD and G20 to increase the transparency of capital flows from tax havens.