Where are the Trends? International Trading and Hedge Funds in Foreign Exchange Markets

Where are the Trends? International Trading and Hedge Funds in Foreign Exchange Markets YuChang Huang, Ph.D. Candidate of Department of International Business, National Taiwan University ABSTRACT Hedge funds are, without doubt, one of the least transparent types of financial assets, and numerous studies have debated the persistence of their performance. Trend-following strategies are mainly used in managed futures funds. However, their performance seems entirely different before and after the 2008 financial crisis. Using daily, before-fee profit/loss and position data provided by the one of the top foreign exchange CTAs worldwide, we find that the profit source of this top-class hedge fund could well be its competitors (including export and import companies and financial institutions) because their trading strategies may not be able to profit-maximizing. We also examine the trading strategies of the hedge fund and provide some insight into why the company uses Over-The-Counter (OTC) commodities in their investment portfolios despite the higher transaction costs. Keywords: international trading, hedge funds, CTAs, foreign exchange markets INTRODUCTION Many articles support the position that most hedge funds, or at least the leading ones, provide positive alpha returns; for example: Ackermann, McEnally, and Ravenscraft (1999), Brown, Goetzmann, and Ibbotson (1999), Liang (2000) ; Agarwal and Naik (2000), Kosowski, Naik, and Teo (2007), Agarwal, Boyson, and Naik (2009), Gregoriou, Hübner and Kooli (2009). The survey by Stulz (2007) takes this position and states that the bottom line of hedge fund research is that, at the very least, hedge funds have a non-negative alpha net of fees on average. There are also numerous articles that express the opposite viewpoint, i.e., hedge funds do not usually confer any real benefits; for example, Asness, Krail, and Liew (2001), Kat and Palaro (2005), Bhardwaj, Gorton, Rouwenhorst (2008), Griffin and Xu (2009), and Dichev and Yu (2011). Managed futures funds are a category of hedge funds that take long and short positions on futures, as well as options and forward contracts. They are operated by Commodity Trading Advisors (CTAs), who are regulated by the Commodity Futures Trading Commision (CFTC) and the National Futures Association (NFA). In this paper, we use managed futures and CTAs interchangeably. Like general hedge funds, CTAs are not obliged to publish their profiles, such as their investment methodology, holding positions, performance and AUMs (Assets Under Management). However, for marketing reasons, they usually provide their profiles to some databases. In this research, we mainly use the profiles tracked by Hedge Fund Research Inc. The AUMs of the major CTAs are currently above US 1 billion. Since CTAs diversify their portfolios by investing in numerous futures markets, most of them use systematic trading. According to the strategic description of systematic diversified CTA index by Hedge Fund Research Inc, systematic 68

diversified strategies have investment processes typically as function of mathematical, algorithmic and technical models, with little or no influence of individuals over the portfolio positioning. Strategies which employ an investment process designed to identify opportunities in markets exhibiting trending or momentum characteristics across individual instruments or asset classes. Strategies typically employ quantitative process which focus on statistically robust or technical patterns in the return series of the asset, and typically focus on highly liquid instruments and maintain shorter holding periods than either discretionary or mean reverting strategies. Although some strategies seek to employ counter trend models, strategies benefit most from an environment characterized by persistent, discernible trending behavior. Interestingly, CTAs do not perform like typical hedge funds during financial crises. For example, Figure 1 shows that the S&P 500 Total Return Index (adjusted for dividends) declined more than 50% between October 2007 and January 2009, and the Dow Jones Credit Suisse (DJCS) Hedge Fund Index started to decline after September 2008. However, the managed futures category of the DJCS Hedge Fund Index gained during September 2008 and January 2009. This may have been because many hedge funds held assets with lower liquidity and were therefore unable to adjust their positions. In contrast, CTAs were able to avoid large losses because they held more liquid assets, and even achieved gains by making timely switches to short positions. Figure 1: Dow Jones Credit Suisse Hedge Fund Index, Managed Futures Index and S&P500 Index in 2008 Financial Crisis, origin: Dow Jones Credit Suisse, S&P Nevertheless, when global stock markets were booming in 2009 and the performance of hedge funds was improving overall, CTAs were obviously losing ground during the same period. Szakmary, Shen and Sharma (2010) found that the performance of trend-following trading strategies earn hugely significant positive returns in future markets. However, various managed futures and CTA indices showed a poor result after 2008 financial crisis (see Table I). 69

Table 1: The Performance (%) of Various Managed Futures and CTA indices during 2005 and 2012 Index 2005 2006 2007 2008 2009 2010 2011 2012 DJCS managed futures index -0.11 8.05 6.01 18.33-6.57 12.22-4.19-2.93 HFRX systematic diversified CTA index 6.81 9.29 12.86 31.55-9.04 6.02-1.79-7.4 Barclay CTA index 1.71 3.54 7.64 14.09-0.1 7.05-3.09-1.7 Newedge CTA index 3.43 5.75 8.05 13.07-4.3 9.26-4.45-2.86 This phenomenon raises two questions. 1. Why was there such a big difference in the performances before and after the 2008 financial crisis? 2. Was the performance difference due to fluctuations in the financial markets? Table II from BIS (2010) Triennial Central Bank Survey, the daily trade volume in foreign exchange markets was US$4.0 trillion, up by 20% from US$3.3 trillion in 2007. Foreign exchange is one of the most heavily traded financial assets. If we compare this to the AUM of the largest known CTA fund worldwide in 2012, which is a mere US$14.7 billion, it would obviously be impossible for a handful of hedge funds to impact the entire market. This fact was pointed out by Eichengreen et al. (1998) and Fung and Hsieh (2000). Table 2: Global foreign exchange market turnover (in billions of USD) Average daily turnover in April 1998 2001 2004 2007 2010 Foreign exchange instruments 1,527 1,239 1,934 3,324 3,981 Spot transactions 568 387 631 1,005 1,490 Outright forwards 128 131 209 362 475 Foreign exchange swaps 734 656 954 1,714 1,765 Currency swaps 10 7 21 31 43 Options and other products 87 60 119 212 207 Turnover at April 2010 exchange rates 1,705 1,505 2,040 3,370 3,981 1. Adjusted for local and cross-border double-counting. 2. Non-US dollar legs for foreign currency transactions were converted into the original currency amounts at average exchange rates for April of each survey year and then reconverted into US dollar amounts at the average April 2010 exchange rates. It is noteworthy that between 2004 and 2007 (see Table III), there was a huge discrepancy between the growth of trading among dealers (only grew 37% according to data released by reporting dealers) and that of transactions with non-reporting financial institutions and non-financial customers (which grew more than 100%). Table 3: Reported foreign exchange market turnover by counterparty Daily average turnover Apr.1998 Apr.2001 Apr.2004 Apr.2007 Apr.2010 in billions of USD volume % volume % volume % volume % volume % Total 1,527 100 1,239 100 1,934 100 3,324 100 3,981 100 with reporting dealers 961 63 719 58 1,018 53 1,392 42 1,548 39 with other financial institutions 299 20 346 28 634 33 1,339 40 1,900 48 with non-financial customers 266 17 174 14 276 14 593 18 533 13 In the summary of BIS (2010) Triennial Central Bank Survey, they mentioned that the higher global foreign exchange market' is associated with the increased trading activity of 'other financial institutions' a category that includes non-reporting banks, hedge funds, pension funds, mutual funds, 70

insurance companies and central banks, among others. For the first time, activity of reporting dealers with other financial institutions surpassed inter-dealer transactions. According to our survey, many hedge funds use NDF to maintain their foreign exchange positions. However, this finding raises the following question: Why do funds opt for NDF over DF or spot transactions if their profit source only captures the price trend, given that the transaction costs of NDF are greater than those of DF and spot? Before we answer this question, we must consider the fact that most multinational corporations (MNCs) and many other institutions also use NDF to hedge their foreign exchange risks. There are three key reasons for using this strategy: 1. The amount and settlement date of NDF can be determined by the MNCs themselves, whereas typical assets traded on the futures or options markets tend to have standardized contract specifications. 2. Unlike DF, the NDF trader is not obliged to hold a substantial cash reserve on the settlement date; and there are almost no margin calls for MNCs, since financial institutions generally provide enough lines of credit for them. 3. In countries that have capital controls, NDF provides a legal path to reduce the volume of foreign exchange and realize a profit. In addition, Brown (2001) studied an anonymous MNC's foreign exchange risk management program, and concluded that hedging seems to be motivated by the desire to smooth earnings (perhaps to reduce information asymmetry), as well as the facilitation of internal contracting and competitive pricing. Therefore, MNCs might not be profit-seeking in NDF trading. However, once financial institutions undertake NDF trades with MNCs, they face a number of regulatory hurdles. For example, the Basel Accords require that financial institutions must comply with the rules on a minimum capital requirement because of their risk-taking. Under the Basel Accords, foreign exchange risk is counted as a category of market risk. If MNCs need short positions in USD, after they trade with some banks(dealers), any interbank(interdealer) markets are unable to spread the risk since the overall banks would be burdened by the depreciation of USD. Therefore, some non-financial institutions must undertake part of the imbalance of supply and demand between various currencies. Hence, we posit that hedge funds, which are less-regulated compared to financial institutions and more profit-maximizing than MNCs, have room to make a profit. This opinion is based on the following factors. 1. The decisions of MNCs are not only profit-maximizing in forward contracts. Generally, MNCs focus on their main business. 2. The regulations on financial institutions force them to reduce their risk exposure in foreign exchange. DATA Our objective is to identify the variables that impact the performance of foreign currency CTA funds. The subject of our case study is a flagship foreign exchange managed futures fund (fund F), one of the top 10 currency CTAs in the world. Fund F, which only trades in currencies, maintains positions in more than 20 currencies. The company's trading is mostly systematic, and its trading philosophy is basically trend-following. We obtain the daily before-all-fees performance and positions of fund F. Our study period ran from January 2006 to September 2009. The summary statistic of fund F s daily performance is in Table IV, and the histogram of the daily performance is in Figure 2. We found the annualized volatility of fund F's profit/loss is slightly lower than 10%, and the daily profit/loss is during -2.5% and 2.5% during our study period. 71

Table 4: The summary statistic of fund F s daily performance Percentiles Smallest -0.0232 1% -0.0155 Obs. 942 5% -0.0090 Mean 0.0003 25% -0.0031 Std. Dev. 0.0057 50% 0.0003 75% 0.0035 Variance 0.0000327 95% 0.0100 Skewness -0.0488 99% 0.0144 Kurtosis 4.0208 Largest 0.0219 Figure 2: The Histogram of Fund F s Daily Performance Although subscriptions to and redemptions of the fund are on a monthly basis, fund F provides daily performance estimates to its clients. As of 2013, fund F had published its performance figures for more than ten years continuously. Consequently, we were able to acquire the funds daily profit/loss figures, along with the percentage allocated daily to each long or short position in its portfolio. Hedge funds usually hold numerous illiquid assets that are difficult to price instantly (mark-to-market), but fund F only holds foreign exchange positions; and it uses the WM/Reuters Rate for the daily closing spot prices and forward rates to evaluate its profit/loss on a daily basis. However, since international financial markets have different holidays, there are always 3 or 4 days each year where we cannot align all real-time price information. Thus, when compiling our estimates, if some markets are closed due to a holiday, we merge the variation with the figure for the next day that all financial markets are open again. 72

First, we performed an autocorrelation test on the daily performance data for fund F (see Table V), and found that there were no instances of autocorrelation. The result confirms Getmansky, Lo, and Makarov's (2004) finding that managed futures funds are more liquid and less likely to smooth returns. Table 5: Autocorrelation Test of Daily Performance Data LAG AC PAC Q Prob>Q LAG AC PAC Q Prob>Q 1-0.0023-0.0023 0.00491 0.9441 21-0.0229-0.0311 13.529 0.889 2 0.0319 0.0319 0.96504 0.6172 22-0.0586-0.0561 16.843 0.772 3-0.01-0.0098 1.059 0.787 23 0.0157 0.0134 17.082 0.8052 4 0.0286 0.0277 1.8347 0.7661 24 0.0174 0.0263 17.376 0.8322 5-0.0565-0.0567 4.8665 0.4324 25-0.0293-0.0323 18.207 0.8335 6 0.0099 0.0076 4.9591 0.5491 26 0.0215 0.0202 18.654 0.8507 7 0.0051 0.0092 4.9837 0.662 27-0.008-0.0158 18.716 0.8801 8 0.0514 0.0503 7.4993 0.4838 28 0.0261 0.0234 19.379 0.886 9 0.0387 0.0429 8.9244 0.4443 29-0.0238-0.0153 19.929 0.8951 10 0.0126 0.0064 9.0747 0.525 30 0.0163 0.0161 20.189 0.9115 11 0.0102 0.009 9.174 0.6058 31 0.0044 0.0143 20.208 0.9312 12-0.0197-0.0227 9.5465 0.6557 32-0.0627-0.0746 24.053 0.8425 13 0.0123 0.0154 9.6905 0.719 33 0.0507 0.062 26.572 0.7781 14-0.0016 0.0027 9.6928 0.7843 34-0.0082-0.0138 26.638 0.8117 15 0.0029 0.0012 9.701 0.8381 35-0.0303-0.0309 27.536 0.8115 16-0.0274-0.0295 10.424 0.8436 36 0.0382 0.0452 28.968 0.791 17 0.0388 0.0327 11.874 0.8077 37-0.0298-0.0453 29.842 0.7921 18-0.0161-0.0154 12.123 0.8408 38 0.0114 0.0177 29.97 0.8205 19 0.0305 0.0273 13.019 0.8376 39-0.029-0.0283 30.798 0.8228 20 0.0019 0.0071 13.023 0.8764 40 0.0128 0.0152 30.96 0.8468 Next, we consider a number of factors that may affect the performance of foreign exchange CTAs. 1. Imbalance in currency demand: To assess the imbalance, we used the monthly Exports, Imports, and Balances figures (measured in millions of USD and seasonally adjusted) from the U.S. International Trade in Goods and Services data published by the Bureau of Economic Analysis of the Department of Commerce. We assumed that the Balances data represented a proxy index of the imbalance in global currency demand. A lower value indicated a greater imbalance in currency demand in international trade, given that the data was consistently negative throughout our study period. We derive a weighted average for the number of business days each week over a two-month period. We supposed that when the imbalance in currency demand increases, the market risk of the overall banks(dealers) which defined by the regulations increases. This causes the banks are willing to transfer more profits to the funds for the financial regulations. 2. Fluctuations in the Financial Market: We assume that the daily rise and fall of major indices correspond to changes in market behavior. Specifically, we take the S&P 500 Index as an indicator of the stock market's performance; the Dollar Index as an indicator of the foreign exchange market's performance; and the CBOE VIX Index to derive general market predictions about variances in the near future. 3. Turnover Rates: We compute the daily turnover rate for fund F by summing the absolute percentage daily differences of every currency used for its long and short positions. Due to the transaction costs, we supposed that the turnover rates have a negative impact on the fund performance. 73

EMPIRICAL RESULTS Table 6: Regression Analysis of Daily Performance Data (a) Regression with US dollar Index Movement Var Daily Coefficient (t-ratio) -0.0287 (-1.98)* -0.5715 (-2.46)* -0.4148 (-3.21)** *, **, *** = statistically significant at the 95%, 99%, 99.9% level -0.4243 (-5.27)*** 0.1836 (0.06) 0.1716 (4.59)*** 0.6033 (7.56)*** (b) Regression without US Dollar Index Movement Var Daily Coefficient (t-ratio) -0.0286 (-2.00)* -0.5695 (-2.46)* -0.4150 (-3.21)** *, **, *** = statistically significant at the 95%, 99%, 99.9% level -0.4245 (-5.31)*** 0.1712 (4.60)*** 0.6039 (7.61)*** We assume that pl is the daily profit/loss ratio of fund F; SPXch is the daily movement ratio of the S&P 500 Index; VIXch is the daily movement ratio of the VIX index; USdef = US exports minus imports in a month (millions of USD, months seasonally adjusted); turnover is the notional daily turnover of fund F; and USDXch is the daily movement ratio of the US Dollar Index. From the results of regression analysis in Tables VI(a) and VI(b) we draw the following conclusions: 1. The daily movement of the US dollar index does not explain the profit/loss ratio of fund F, since the other coefficients in the tables only change a little. The regression analysis results show that variations in the US dollar were not the profit source of fund F. In the other words, fund F did not simply derive profit from the weakness of the US dollar during the study period. (The US dollar index fell 15% during our study.) 2. As the unfavorable U.S. trade imbalance became more serious, fund F's profit tended to increase. This implies an imbalance in currency demand affects the performance of this foreign exchange funds. 3. When the VIX index increased, fund F's performance deteriorated. This implies that market volatility reduces the profit expectations of managed futures funds. It might because when the financial markets are volatile, the bid-ask spread are usually wider, cutting the fund performance. 4. Fund F's daily performance data exhibits ARCH and GARCH effects. 5. Increased turnover has a negative impact on the performance of fund F. This confirms our intuition about trading. We estimate that if the turnover rate were to increase by 100%, the fund's performance would decline by 0.42% to 0.44%. 6. When the S&P 500 index goes up, fund F's performance deteriorates. Since most of the physical assets are positively related with the stock market, we consider that fund F has an incentive to perform in the opposite way to the stock market, because it tends to attract investors with diversified portfolios. This fund may use some automated trading modules which are able to reduce the correlation between its performance and the stock markets' performance. 74

In addition, to test the robustness of our regression analysis, we check if Lehman Brothers' Chapter 11 Bankruptcy filing in September 2008 affected the performance of fund F. We used the following variables for the robustness test: 0 2008/09/15 2008/09/15 and 0 2008/09/15 2008/09/15 The statistical results in Table VII show that under daily data analysis, the impact of fluctuations in the VIX index and S&P 500 on fund F's performance did not change significantly after Lehman Brothers filed for Chapter 11 bankruptcy. Table 7: Robustness Check of Daily Performance Data Var Daily Coefficient -0.0203-0.6063-0.4125-0.4330 0.5761 0.1020 0.1688 0.6078 (t-ratio) (-0.68) (-2.08)* (-3.18)** (-5.41)*** (0.70) (0.03) (4.57)*** (7.71)*** *, **, *** = statistically significant at the 95%, 99%, 99.9% level We then merge the data on a weekly basis. Under similar method with daily data, we observe that the ARCH and GARCH effects were not significant in the weekly performance of fund F. However, the impacts of variations in the S&P500, VIX, U.S. deficit and turnover were as significant as those shown by the daily results. In addition, the robustness testing result shows that the impact of fluctuations in the VIX index and the S&P 500 on fund F's performance did not change significantly after Lehman Brothers filed for Chapter 11 bankruptcy in weekly data. Figure 3: U.S. Balance of Trade from 2006 to 2010 (origin: U.S. Department of Commerce) Figure 3 shows the trend of the monthly U.S. deficit from 2006 to 2010. Therefore, to investigate the ARCH effect for monthly data, we assume that variable represents the period prior to October 2008, and variable represents the period after that date. pl is the monthly profit/loss ratio of fund F; 75

SPXch is the monthly movement ratio of the S&P 500 Index; VIXch is the monthly movement ratio of the VIX index; turnover is the notional monthly turnover of fund F; USDXch is the monthly movement ratio of the US Dollar Index; and and represent the US trade balance before and after 2008/10/1 (million USD, months seasonally adjusted). Table 8: Regression Analysis of Monthly Performance Data (a) ARCH Effect Test Var Monthly Coefficient -0.1589-0.0749 0.0008 1.0511-0.3215-0.1096-0.0268 (t-ratio) (-1.01) (-2.29)* (0.00) (0.63) (-1.33) (-0.64) (-0.11) *, **, *** = statistically significant at the 95%, 99%, 99.9% level (b) Direct Regression Var Monthly Coefficient (t-ratio) -0.1105 (-0.95) -0.0643 (-2.86)** -1.0091 (-3.41)** -0.3000 (-1.41) -0.0894 (-0.54) *, **, *** = statistically significant at the 95%, 99%, 99.9% level Tables VIII(a) and VIII(b) show that the arch effect is not significant, and there is little difference in the coefficients corresponding to the variables. However, we find that the turnover rate and S&P 500 performance, which have high explicative significance for the high-frequency (daily and weekly) data, lose their significance when we merge the data on a monthly basis. Therefore, we conclude that, if we want to find statistically significant events in hedge fund performance and trading data, and then make valid inferences, we should examine high-frequency data instead of monthly performance data, which may even be after-fees. Thus, the fund performance data used in previous works may be too coarsely grained for accurate evaluation of a hedge fund's performance. CONCLUSION AND DISCUSSION The empirical data confirms our assumption that the imbalance in global currency demand is the source of the fund's profits. Given this imbalance, corporations and banks are willing to transfer potential profit opportunities to the fund because of self-interest and the limitations imposed by government regulations. However, the profit does not result directly from the weakness of the US dollar. The fact that the fund chooses NDF assets as its foreign exchange trading instrument (with relatively high associated transaction costs) probably means that trading parties who do not maximize for profit in this market already have enough profit opportunities. The recent financial crisis affected the fund's performance for two reasons: 1. the demand for currency declined globally; hence, there was less imbalance in the foreign exchange market; and 2. the reduction in the bargaining power of hedge funds in their dealings with banks. 76

Similarly, the Dow Jones Credit Suisse Hedge Fund Managed Futures Index declined 6.57% in 2009, while most of the financial markets (including hedge funds) were yielding substantial profits. The performance trends of CTAs and hedge funds were quite different after the financial crisis began in 2008. Thus, it is hard to claim their profits were derived primarily from their trading strategies. We surmise that, during the financial crisis, hedge funds could not adjust their assets due to liquidity problems. Therefore, the market price of the assets declined during the crisis, leading to losses by the hedge funds. In contrast, CTAs could easily clear, or even reverse, their positions in a short period. This advantage enabled them to perform well when the market was in turmoil. However, with the cooling of the global economy and subsequent deleveraging of the financial markets, the imbalance in the foreign exchange and commodity markets declined, thereby squeezing the potential profits of CTAs. The daily and weekly data shows that the turnover rate had a strong negative impact on fund F's performance. If the turnover rate increased by 100%, the fund's performance declined by 0.42% to 0.44%. Since transactions incur costs, this result seems entirely reasonable. However, we cannot determine in a statistically meaningful way whether the result arises from unnecessary transaction costs or if it is an inevitable expenditure incurred by following trends. At the very least, the result shows that the profit derived from non-financial firms is sufficient to ensure that the fund survives. We found that previous studies used data that may have been too low-frequency to capture the profit sources and operating characteristics of funds. Thus, we suggest that, in future studies of hedge fund operating behavior and characteristics, higher-frequency (at least weekly) trading data should be used because it would yield more accurate results. Finally, since managed futures funds trade in assets that are more transparent and easier to mark-to-market than ordinary hedge funds, they may be better subjects of study when we seek to understand the trading behavior of investors and the performance of alternative investments. REFERENCES Ackermann, C., R. McEnally, and D. Ravenscraft, (1999). "The Performance of Hedge Funds: Risk, Return, and Incentives", Journal of Finance 54, 833-874 Agarwal, V., N. M. Boyson, and N. Y. Naik, (2009). "Hedge Funds for Retail Investors? An Examination of Hedged Mutual Funds", Journal of Financial and Quantitative Analysis 44, 273-305. Agarwal, V., and N. Y. Naik, (2000). "Multi-Period Performance Persistence Analysis of Hedge Funds", Journal of Financial and Quantitative Analysis 35, 327-342. Asness, C., R. Krail, and J. Liew, (2001). "Do Hedge Funds Hedge?: Be Cautious in Analyzing Monthly Returns", Journal of Portfolio Management 28, 6-19. Bank for International Settlements, (2010). Triennial Central Bank Survey of Foreign Exchange and Derivatives Market Activity in 2010, Basel. Bhardwaj, G., G. Gorton, and K. G. Rouwenhorst, (2008). "Fooling Some of the People All of the Time: The Inefficient Performance and Persistence of Commodity Trading Advisors", Working Paper, Yale University. Brown, G. W., (2001). "Managing foreign exchange risk with derivatives", Journal of Financial Economics 60, 401-448 Brown, S. J., W. N. Goetzmann, and R. G. Ibbotson, (1999). "Offshore Hedge Funds: Survival and Performance, 1989-95", Journal of Business 72, 91-117. Dichev, D., and G. Yu, (2011). "Higher risk, lower returns: What hedge fund investors really earn", Journal of Financial Economics 100, 248-263 77

Eichengreen, B., D. Mathieson, B. Chadha, A. Jansen, L. Kodres, and S. Sharma., (1998). "Hedge Fund and Financial Market Dynamics", Washington, DC: International Monetary Fund. Fung, W., and D. A. Hsieh, (2000). "Performance Characteristics of Hedge Funds and Commodity Funds: Natural vs. Spurious Biases", Journal of Financial and Quantitative Analysis 35, 291-307 Getmansky, M., A. W. Lo, and I. Makarov, (2004). "An Econometric Model of Serial Correlation and Illiquidity in Hedge Fund Returns", Journal of Financial Economics 74, 529-609. Gregoriou, G. N., G. Hübner and M. Kooli, (2009). "Performance and persistence of Commodity Trading Advisors: Further evidence", Journal of Futures Markets 30, 725-752. Griffin, J. M., and J. Xu, (2009). "How smart are the smart guys? A unique view from hedge fund stock holdings", Review of Financial Studies 22, 2531-2570. Liang, B., (2000). "Hedge Funds: The Living and the Dead", Journal of Financial and Quantitative Analysis 35, 309-326. Kat, H. M., and H. P. Palaro, (2005). "Who needs hedge funds? A copula-based approach to hedge fund return replication", Alternative Investment Research Centre Working paper. Kosowski, R., N. Y. Naik, and M. Teo, (2007). "Do Hedge Funds Deliver Alpha? A Bayesian and Bootstrap Analysis", Journal of Financial Economics 84, 229-264. Stulz, R. M., (2007). "Hedge Funds: Past, Present, and Future", Journal of Economic Perspectives 21, 175-194. Szakmary, A., Q. Shen, and S. Sharma, (2010). "Trend-following trading strategies in commodity futures: A re-examination", Journal of Banking and Finance 34, 409-426. 78