best interest market money rate

V8/Jan/13/06 Senior Loan Funds: Interest Rate Hedge and Money Market/Bond Alternative Kam C. Chan William J. Trainor Jr. Edward R. Wolfe Department of Finance Gordon Ford College of Business Western Kentucky University Bowling Green, KY 42101 Key Words: Bank-loan/Senior loan funds Please direct correspondence to: Johnny Chan or William J. Trainor Jr. Western Kentucky University Department of Finance 1906 College Heights Blvd. Bowling Green, KY 42101 Telephone: (270) 745-6337 E-mail: Johnny.chan@wku.edu or Bill.Trainor@wku.edu Senior Loan Funds: Interest Rate Hedge and Money Market/Bond Alternative Abstract Senior or bank-loan funds are becoming an increasingly popular alternative to money-market funds despite their increased risk. Senior-loans are generally short-term variable interest rate loans made to corporations with average to poor credit ratings. The attraction for investors is that senior-loans should not lose their value as interest rates rise and in fact should generate higher yields as rates increase. In addition, senior-loans usually hold collateral against the borrower's assets and thus, the default risk is lower than with typical high yield bonds. This study explores the characteristics of these funds to determine: 1) whether they do in fact act as an interest rate risk hedge, and 2) find out what role they should play as a money-market alternative. 1 Senior Loan Funds: Interest Rate Hedge and Money Market/Bond Alternative I. Introduction Senior-loan funds, also called bank or prime rate loans funds, have become increasingly popular over the last few years as investors have scrambled for increased yields, while trying to mitigate the risks of a rising interest rate environment. The enhanced yield and interest rate hedge comes from the fact that these mutual funds basically purchase short-term floating rate loans that banks and other financial institutions lend to corporations with average to poor credit ratings. Most of the loans held by these funds are typically tied to the London Interbank Offer Rate (LIBOR). Hence, at each rate adjustment period date, (usually every 6 months or shorter) the interest rate, and therefore the interest income, is adjusted to reflect prevailing short-term interest rates. As Shaalan (2000) points out, because of the nature of the short-term floating rate feature, the returns of these senior-loan funds tend to be correlated with the interest rate cycle, i.e., senior-loan fund returns are positively correlated with interest rate changes, allowing the fund to act as a hedge as interest rates rise. Shaalan also points out that because senior loans are first in line for repayment when companies default, even in a rising interest rate environment where the probability of default may increase, senior-loan funds are able to claim real assets and potentially salvage a sizable portion of their investment. Thus, unlike high-yield bonds, which on average only recover 40% in the event of a default (Covitz and Han (2004)), these senior loans are somewhat insulated from such an outcome. To date, the research on senior-loan funds is rather limited. Besides Shaalan’s overview article, Opdyke (2005) also provides a general perspective on senior-loan funds. However, both 2 Shaalan and Opdyke only highlight the characteristics of these funds and do not perform analysis of these funds’ investment performance or their risk characteristics. II. Overall view of senior-loan funds In 2004, almost $10 billion flowed into senior-funds, up from just $700 million the year before. According to Standard & Poors, the default rate on these funds has also dropped dramatically over the last three years from 4.45% in 2002 to just 1.15% in 2005. This had led to increasing yield gains over standard money market funds. Examining senior-loan (referred to as bank-loan) fund data from Morningstar Principia, it should be noted that the first senior-loan fund, Eaton Vance Prime Rate Fund, was only launched in 1989. As of February 2005, there are now 42 senior-loan funds.1 Of the 42 funds, only 31 of them have more than $200 million in net assets under management, and because of the short history of these funds, only six of the 42 have at least a ten-year history. Thus, the analysis in this study is limited to a relatively small sample set and limited time span. However, the financial and economic environment has seen some relatively major fluctuations over the last 15 years, and therefore the results should be of interest. The analysis in this study is based on the bank-loan fund category as provided by Morningstar. We also employ returns for the S&P 500, the Lehman Brother Bond Index, t-bills, and the CSFB High Yield Bond Index as benchmarks for their respective asset categories. In addition, we collect consumer price index information to calculate a monthly inflation rate and the University of Michigan consumer sentiment index to gauge the sentiment from October 1989 to February 2005. All the economic data are from the government Federal statistics website (www.fed-stat.gov). 1 We count different share classes of funds as one fund. 3 III. Analysis General Table 1 presents the summary statistics of senior-loan funds and other benchmarks during 1991-2005. The average annual return for the senior-loan funds is 6.34%, which exceeds the tbill average return of 4.35%. On average, senior-loan funds have exceeded returns on t-bills by 2% and in fact, the returns are higher than those for t-bills every year except for 2001 and 2002 over the last 15 years. In contrast, senior-loan funds have, on average, underperformed the Lehman Brothers Aggregate Bond Index by 1.5%. What has made these funds so attractive relative to t-bills and even a typical bond fund in recent years is that senior-loan funds have dramatically outperformed both in the last two years. Table 2 presents the summary statistics for the last 15 years. As one might expect, the monthly mean return in descending order is for stocks, high-yield bonds, bonds, senior-loan funds, and t-bills. The risks of these instruments in terms of standard deviation is in ascending order of the same categories, hence consistent with what would be expected. When examining the Sharpe ratio for these asset categories, it is interesting to note that the senior-loan funds deliver the best risk adjusted return performance with a Sharpe ratio of 0.4456. This is larger than stock, high yield bond, and the aggregate bond index by a wide margin. Although the average values over this time period are very complimentary to senior loan funds, there have been some dramatic changes. Figure 1 shows the 36-month rolling average monthly return and standard deviation for senior-loan funds along with t-bill returns for the last 15 years using September 1989 as the starting date. As one can easily discern from the graph, since 2001 there has been a major decrease in return for these funds with a corresponding increase in risk. In addition, although the returns have followed t-bills very closely up through 4 2002, the returns to senior loans began to increase in direct contrast to t-bills. This is the first time since these funds were established that they have not moved in tandem with t-bills. It is also of great interest to note how the relationship between senior-loan funds and other asset categories has changed dramatically over the last several years. Figure 2 shows senior-loan fund's 36-month rolling correlation with t-bills, the S&P 500, the CSFB High Yield Bond Index, and the Lehman Brother's Aggregate Bond index. The most dramatic changes have been the increasing correlation with high-yield bonds and the decreasing correlation with T-bills, which as figure 1 previously showed, now appears to be significantly negative. The increasing correlation between senior-loan funds and high yield bonds suggest that senior-loan funds may have stepped up their exposure with average or low quality bonds in recent years. The decreasing correlation between senior-loan funds with t-bills in recent years suggests that the extent of the floating rate return may also be weakened. However, the timing of this breakdown also begs another explanation. Starting in 2000 with the tech crash and the September 11 terrorists attacks, there was a flight to quality where the yield differential between high yield and investment grade debt increased from the 2% range up to 6% through 2001 and 2002. This yield differential fell over the next couple of years back to the 2% range. During this time, over half of the return to senior loan funds has been in the form of capital gains due to the reduction in the yield spread and risk-premium. Thus, despite falling yields, these fund's returns have increased dramatically, and the disconnect between t-bill returns and senior-loan funds is most likely due to the yield spread changes and not to interest rate changes. Figure 2 also shows the relationship with the aggregate bond market, which has never been strong and is basically zero. Senior-loan funds’ low correlation coefficients with stocks (S&P500) and bonds (Lehman Brother Aggregate Bond Index) are generally expected because 5 senior-loan funds do not own common stocks and have only short-term average or low quality floating rate fixed income instruments under management. Investment Style Table 3 shows results based on an examination of 14 funds from Feb. 2000 to Feb. 2005. The average annualized return for these funds is 4.98% as compared to the overall average return reported by Morningstar, which is 4.74% (not shown). Over this time period, the worst fund returned 3.65% while the best fund returned 7.53%. The standard deviations differed between the funds by only 1.17% on an annualized basis so the differences in volatility between the funds are not large. This minor difference between the funds is confirmed when analyzing the correlation between the funds and the overall index. The average correlation between the funds and the index was over 0.9, and even the lowest correlation was only 0.78. Thus, it does not appear that returns are remarkably different between funds and the results derived in this study apply to most if not all the funds in this category. An investment style analysis of senior-loan funds is undertaken to determine how well the funds actually return what is promised, that is increasing returns in rising interest rate environments. This should also help answering the question as to what role senior-loan funds should play in investor's portfolios, and how similar or different the various funds perform within this asset category. We first examine the 14 funds that have been in existence since February 2000, and then use a generalized autoregressive heteroskedasticity (GARCH) model to examine the investment style of the overall senior-loan funds category, a large senior-loan fund (Eaton Vance Prime Rate Fund), and a medium size senior-loan fund (Merrill Senior Floating Rate Fund).2 2 As of February, 2005, Eaton Vance Prime Rate Fund has $1,591.4 million and Merrill Senior Floating Rate Fund has $715.2 million of net asset values. 6 For our model, consider the senior-loan fund return series, yt, which follows a GARCH (p, q) process. The conditional distribution of the series Y for time t is written as: yt | Ψt-1 ∼ N (0, ht) where Ψt-1 denotes the information at t-1. The conditional variance ht is: q p ht = ω + Σ αi yt-i2 + Σ γj ht-j i=1 j=1 where p ≥ 0, q > 0, ω > 0, αi > 0 and γj > 0 Hence, the regression model becomes: yt = xt’ β + εt εt = (√ht) * et et ∼ IN(0, 1) where xt is the vector of explanatory variables, β is a vector of coefficients, and et is an independently distributed random error term. We use the S&P 500 return, high-yield bond return, aggregate bond return, t-bill yield, inflation rate, and consumer sentiment index as the explanatory variables. Typically, a GARCH (1, 1) model is appropriate to capture the time-varying autoregressive heteroskedasticity of a financial time series such as the senior-loan fund returns. For instance, Kao, Cheng, and Chan (1988) use a similar GARCH model to examine international mutual fund returns. Because of the flight to quality, we also add a dummy variable to capture this. Since the yield spread reached a high of 6.56% in 2000 and didn't start to drop to any significant degree until several months after the September 11 terrorist attacks, we use a September 2001 dummy variable to account for this change. 7 The results of GARCH (1, 1) model for the senior-loan funds are shown in Table 4. In addition to the senior-loan fund overall average, we also use a large (Eaton Vance Prime Rate fund) and a medium senior-loan fund (Merrill Senior Floating Rate fund) in the style analysis. As expected, the GARCH model parameters (ω, α, and γ) are all significant suggesting the fund returns exhibit time-varying autoregressive heteroskedasticity. In terms of hedging interest rate risk, the senior-loan funds do in fact perform very well as suggested by the positive and significant coefficients associated with T-bills in all three funds returns equations. The magnitudes of the coefficients range from 0.79 in the average senior-loan fund to 1.05 for the large Eaton Vance Prime Rate fund. This suggests that the senior-loan funds are able to increase their return from .79% to 1.05% for a 1% increase in T-bill yield. Table 4 also shows the weak relations of senior-loan funds with S&P500, high yield bond, and aggregate bond returns. The magnitudes of the coefficients are also very small. In terms of investment styles, senior-loan funds are closer to T-bills than high yield bonds. The small magnitudes of these coefficients are consistent with the low correlation coefficients between senior-loan fund returns and the returns of these benchmarks found earlier. Thus, even though there was an apparent increase in the correlation between senior loan funds and high yield bonds over the last couple of years and a decoupling of their relationship to t-bills, this appears to be the sole result of the flight to quality and the dramatic change in quality yield spreads over the last few years. In addition, the September 11 dummy and combined t-bill and September 11 dummy variables are both significant, which help the style equations to control for flight for quality phenomenon due to tech crash and terrorists attacks. Finally, we ask the question as to how do senior-loan funds fair in hedging interest rate risk comparing with other benchmarks? To answer this, we again use a GARCH (1, 1) model to 8 compare senior loan funds with S&P 500, high yield bond, and aggregate bond. The results are in Table 5. For the T-bill variable, only the senior loan funds and aggregate bond return show positive and significant coefficients. The results suggest that both senior loan funds and bonds are a good hedge against interest rate risk. Nevertheless, senior loan funds deliver better riskadjusted performance during the last 15 years. Together with the fact that senior-loan funds deliver returns that are approximately 200 basis points better than T-bills (in Table 1), senior loan funds appear to be viable asset for diversification among different asst classes. IV. Summary We present the general characteristics, risk/return profile, investment style analysis, and relative hedging performance of senior-loan funds in this study. While senior-loan funds are becoming increasingly popular in recent years, there are few studies examining their characteristics. We find senior-loan funds have a risk-return spectrum below the typical investment classes of stocks, bonds, and high yield bonds but above t-bills. In terms of riskadjusted returns using Sharpe’s ratio, senior-loan funds dominated all asset classes. The investment style analysis shows that senior-loan funds are highly correlated with tbills with a sizable premium as well. This suggests that they are good interest rate risk hedge. It is also apparent that senior-loan funds and other asset classes are not highly correlated. Coupling the two facts above suggests that senior-loan funds appear to be a good alternative to moneymarket funds despite their increased risk. However, it must be noted that the basic characteristics of senior loan funds did change dramatically in 2000-01 as their returns fell while their volatility increased. This coincided with the decoupling of the relationship between t-bills and senior loans, and the dramatic increase in the correlation between senior loans and high yield bonds. However, this is likely a transitory 9 phase as it appears to be the result of a flight to quality rather than some fundamental shift. The tech crash and the Sept. 11 terrorist attacks were a large part of this flight to quality. During this time, the yield spread increased from 2% to over 6% and subsequently fell back to 2% over the next three years. Even though interest rates were falling, so was the yield spread and the increasing returns to senior loans while t-bill yields were falling was mainly the result of capital gains. This hypothesis is confirmed based on our GARCH models and using a dummy variable to account for this event. In fact, for every 1% change in t-bill yields, senior-loan fund returns change from .8% to 1.05%, thus supporting the notion that they do move in tandem with shortterm rates. Thus, senior loans do appear to be a great interest rate hedge and should be considered as a major alternative to money market funds and t-bills when interest rates are rising. 10 References Covitz, Daniel and S. Han. "An Empirical Analysis of Bond Recovery Rates: Exploring a Structural View of Default," Division of Research and Statistics, The Federal Reserve Board, (2004) Kao, Wenchi, Louis T.W. Cheng, and Kam C. Chan, 1998. International mutual fund performance in up and down market conditions, Financial Review, 33, 127-144. Opdyke, Jeff D., 2005. A haven as interest rates climb; bank-loan funds provide income and relative safety as rates threaten bonds, Wall Street Journal, Eastern Edition, April 28, D2. Shaalan, Amy, 2000. Banking on debt: bank loan funds, though often overlooked and easily misunderstood, have merits over other comparable low-grade fixed-income securities, Financial Planning, November, 1. 11 Figure 1. Rolling 36 month average monthly return and standard deviation for senior-loan funds from Sept 1989 to Feb. 2005. 0.8 0.7 0.6 0.5 Percent Senior Mean 0.4 0.3 0.2 0.1 0 Oct-92 Oct-93 Oct-94 Oct-95 Oct-96 Oct-97 Oct-98 Oct-99 Oct-00 Oct-01 Oct-02 Oct-03 Oct-04 Senior SD t-bill Mean 12 Figure 2. Rolling 36-month correlation between senior-loan funds and 3-month T-bills, S&P 500, High Yield Index, and Lehman Brother's Bond Index. 1 0.8 0.6 0.4 TBILL 0.2 0 ct -0 2 ct -0 3 ct -9 2 ct -9 3 ct -9 4 ct -9 5 ct -9 6 ct -9 7 ct -9 8 ct -9 9 ct -0 0 SP500 HYINDEX LBINDEX ct -0 1 ct -0 4 O -0.2 O O O O O O O O O O -0.4 -0.6 -0.8 O O 13 Table 1. Geometric average annual returns of senior-loan funds and benchmarks The table shows that senior-loan funds have exceeded returns on T-bills by 2%, and have only underperformed the Lehman Brothers Aggregate Bond Index by 1.5%. In addition, senior-loan funds annual returns are higher than those for T-bills every year except for 2001 and 2002 over the last 15 years. In recent years is that senior-loan funds have dramatically outperformed both in the last two years. Senior-Loan Year funds 1991 9.60% 1992 8.03% 1993 6.21% 1994 5.70% 1995 6.60% 1996 7.80% 1997 6.76% 1998 7.22% 1999 6.37% 2000 6.34% 2001 6.17% 2002 2.29% 2003 0.67% 2004 10.31% 2005 5.01% 19912005 6.34% T-bill S&P 500 7.99% -3.12% 5.68% 30.48% 3.59% 7.62% 3.12% 10.06% 4.45% 1.31% 5.79% 37.53% 5.26% 22.95% 5.32% 33.35% 5.01% 28.58% 4.87% 21.04% 6.31% -9.10% 3.67% -11.88% 1.68% -22.10% 1.05% 28.67% 1.43% 10.88% 4.35% 12.42% Lehman Brothers CSFB High Yield Aggregate Bond index Bond Index -6.38% 8.96% 43.75% 16.00% 16.66% 7.40% 18.91% 9.75% -0.97% -2.92% 17.38% 18.48% 12.42% 3.63% 12.63% 9.65% 0.58% 8.69% 3.28% -0.82% -5.21% 11.63% 5.78% 8.44% 3.11% 10.26% 27.93% 4.10% 11.96% 4.34% 10.79% 7.84% 14 Table 2. Summary statistics (September, 1989 to February, 2005) The table presents the summary statistics for the last 15 years. As expected, the monthly mean return falls in descending order is for stocks, high-yield bonds, bonds, senior-loan funds, and t-bills. The risk of these instruments in terms of standard deviations increases in ascending order of the same categories. The senior-loan funds deliver the best risk adjusted return performance with a Sharpe ratio of 0.4456, which is larger than stock, high yield bond, and bond by a wide margin. Variable Senior-loan funds SP500 High yield bond index Lehman Brothers Aggregate bond index T-bill Inflation rate Consumer sentiment index Standard Monthly deviations mean return (%) (%) or level or level 0.5187 0.3583 0.9340 4.1761 0.7781 0.6291 0.3590 0.2325 92.06 2.0398 1.1227 0.1598 0.2367 10.63 Minimum Maximum Sharp ratio: (%)or level (%)or level (mean – t-bill mean)/ std dev -1.3170 -14.4400 -7.6200 -3.3620 0.0770 -0.3946 63.90 1.5490 11.4400 8.6200 3.8700 0.6740 1.0309 112.00 0.4456 0.1377 0.2055 0.2405 15 Table 3. Stylized Facts based on 14 funds from Feb. 2000 to Feb. 2005. Table 4 shows results based on an examination of 14 funds since 2005. Over this time period, the worst fund returned 3.65% while the best fund returned 7.53%. The standard deviations difference between the funds is only 1.17% on an annualized basis so the differences in volatility between the funds are not large. The minor difference between the funds is confirmed when analyzing the correlation between the funds and the overall index. The average correlation between the funds and the index was over 0.9 and even the lowest correlation was only 0.78. Thus, it does not appear that returns are remarkably different between funds and the results derived in this study apply to most if not all the funds in this category. Monthly Annualized Average fund return 0.42% 4.98% Worst Fund average return 0.30% 3.65% Best Fund average return 0.63% 7.53% St. Dev. Between funds 0.97% 1.17% Avg. St. Dev. For any given month between funds 0.38% 4.56% Average worst monthly return -0.25% -3.04% Average best monthly return 0.99% 11.91% Avg. difference between best and worst -0.74% -8.88% Avg. correlation between funds and Senior-loan funds index 0.91 Worst correlation between funds and Senior-loan funds index 0.78 Best correlation between funds and Senior-loan funds index 0.97 Standard deviation of correlations 0.05 Note: There are only 14 funds that have complete data during this period. 16 Table 4. Investment style analysis of senior-loan funds with GARCH (1, 1) model GARCH (1,1) results for senior-loan fund overall average, a large fund (Eaton Vance Prime Rate), and a medium senior-loan fund (Merrill Senior Floating Rate fund) in the style analysis. As expected, the GARCH model parameters (ω, α, and γ) are all significant suggesting the fund returns exhibit time-varying autoregressive heteroskedasticity. In terms of hedging interest rate risk, the senior-loan funds do in fact perform very well as suggested by the positive and significant coefficients associated with T-bills in all three funds returns equations. Senior-loan funds Coeff. 0.4674 -.0039 t-stat 8.27*** -1.60 Intercept SP500 return (%) High yield bond index return (%) LB bond index return (%) T-bill yield*dummy (after Sept 2001 =1; otherwise 0) T-bill yield (%) Dummy (after Sept 2001 =1; otherwise 0) Inflation (in %) Consumer sentiment ω α1 γ Log likelihood N Eaton Vance Prime Rate (large) Coeff. t-stat 0.1315 2.76*** -0.0029 -1.62 Merrill Lynch Senior Floating Rate (medium) Coeff. t-stat 0.5458 2.97*** -0.0061 -1.15 0.0065 1.24 0.0039 0.95 -0.0027 -0.19 .0100 1.11 0.0152 2.26** 0.0173 0.86 -8.0541 -8.72*** -6.8813 -8.59*** -7.8361 -4.15*** 0.7870 1.0765 11.34*** 8.63*** 1.0497 0.9406 13.71*** 9.89*** 0.9324 1.1551 3.83*** 3.23*** -0.1608 -0.0021 -4.48*** -3.31*** -0.0303 -0.0003 -0.81 -0.47 -0.2967 -0.0033 -2.77*** -1.72* 0.0013 0.5785 0.5613 68.12 186 3.06*** 3.95*** 10.68*** 0.0003 0.6940 0.5670 87.18 186 2.08** 4.56*** 10.75*** 0.0021 0.3004 0.7551 -42.91 186 2.57** 3.44*** 16.99*** 17 Table 5. Comparative analysis of hedging performance of senior-loan funds and other assets with GARCH (1, 1) model We use a GARCH (1, 1) model to compare senior loan fund with S&P 500, high yield bond, and aggregate bond. For the T-bill variable, only the senior loan funds and aggregate bond return show positive and significant coefficients. The results suggest that both senior loan funds and bonds are good hedge against interest rate risk. Intercept T-bill yield*dummy (after Sept 2001 =1; otherwise 0) T-bill yield (%) Inflation (in %) Consumer sentiment Dummy (after Sept 2001 =1; otherwise 0) ω α1 γ Log likelihood N Senior-loan funds Coeff. t-stat 0.48 12.97*** -8.04 0.82 -0.17 -0.00 1.09 0.001 0.59 0.56 66.05 186 -8.57*** 12.98*** -4.92*** -4.61*** 8.55*** 3.42*** 4.09*** 11.18*** S&P 500 Coeff. t-stat 3.08 1.32 -15.34 2.69 -3.58 -0.02 2.13 0.55 0.11 0.86 -517.36 186 -0.89 0.91 -2.79*** -1.03 0.77 0.90 1.40 9.71** High yield bond Coeff. t-stat 4.47 3.13*** -14.75 -0.58 -0.11 -0.03 1.79 0.55 0.16 0.69 -372.75 186 Aggregate bond Coeff. t-stat 1.18 1.45 0.16 1.94* -1.28 -1.38 0.28 23.18*** -0.00 0.00 -2.66*** 0.62 -0.42 -1.78* -2.47** 1.72* 2.36** 2.66*** 6.08*** 1.69 -0.45 -0.01 0.17 1.20 -0.00 0.00 -281.27 186 18