You are on page 1of 16

Financial Analysts Journal Volume 68 Number 6 2012 CFA Institute

The Liquidity Style of Mutual Funds


Thomas M. Idzorek, CFA, James X. Xiong, CFA, and Roger G. Ibbotson
Recent literature indicates that a liquidity investment stylethe process of investing in less liquid stocks has led to excess returns relative to size and value. The authors examined whether this style, previously documented at the security level, can be uncovered at the mutual fund level. Across a wide range of mutual fund categories, they found that, on average, mutual funds that held less liquid stocks signicantly outperformed those that held more liquid stocks.

t is relatively well known that less liquid investments tend to outperform more liquid investments. The same holds true within the relatively liquid universe of publicly traded stocks. The generally accepted rationale for a liquidity premium is that all else equal, investors prefer greater liquidity; thus, in order to induce investors to hold less liquid assets, they must have the expectation (but not the guarantee) of a return premium. Using todays nomenclature, one could think of less liquidity as a risk factor, an exotic beta, or a structural alpha related to its extra costs. Recent literature indicates that the liquidity investment stylethe process of investing in relatively less liquid stocks within the liquid universe of publicly traded stocksproduces riskadjusted returns that rival or exceed those of the three best-known market anomalies: small minus large, value minus growth, and high minus low momentum (see Carhart 1997). For example, Amihud and Mendelson (1986) used the quoted bid ask spread to measure liquidity and tested the relationship between stock returns and liquidity over 19611980. They found evidence consistent with the notion of a liquidity premium. Datar, Naik, and Radcliffe (1998) used the turnover rate (the number of shares traded as a fraction of the number of shares outstanding) as a proxy for liquidity and found that stock returns are strongly negatively related to their turnover rates, which conrms the notion that less liquid stocks provide higher aver-

Thomas M. Idzorek, CFA, is president and global chief investment ofcer and James X. Xiong, CFA, is senior research consultant at Morningstar Investment Management, Chicago. Roger G. Ibbotson is chairman and chief investment ofcer at Zebra Capital Management, LLC, and professor in practice at the Yale School of Management, New Haven, Connecticut.
38 www.cfapubs.org

age returns. Overall, their results support the relationship between less liquidity and higher stock returns. Pstor and Stambaugh (2003) demonstrated that marketwide liquidity appears to be a state variable that is important in pricing common stocks. They found that expected stock returns are cross-sectionally related to the sensitivity of stock returns to aggregate liquidity. According to their measure, smaller stocks are less liquid and thus highly sensitive to aggregate liquidity. In addition, research by Li, Mooradian, and Zhang (2007) supports the hypothesis that marketwide liquidity is an important risk factor and has a signicant effect on expected returns. Recently, Lou and Sadka (2011) documented the importance of distinguishing between liquidity level as measured by the illiquidity measure of Amihud (2002) and liquidity risk, which measures sensitivity to changes in marketwide liquidity. They found that liquidity risk is a better predictor of stock prices during a crisis than liquidity level. Although stock-level liquidity has been explored by academics as an important explanatory risk factor (even though, as we shall see, the return premium associated with less liquid investments can be characterized by less risk) and as an ongoing concern for portfolios that need immediate liquidity, only recently has it been explored as an investment style similar to a preference for funds with a small-cap or value bias. To that end and perhaps most importantly for our purposes using monthly data for the largest 3,500 U.S. stocks by capitalization starting in 1972, Ibbotson, Chen, Kim, and Hu (2012) sorted stocks into equally weighted quartiles based on liquidity. Their results clearly show that annually rebalanced composites of relatively less liquid stocks signicantly outperform composites of more liquid stocks after controlling for size, valuation, and momentum. Ibbotson et al. (2012) attempted to distinguish between risk
2012 CFA Institute

The Liquidity Style of Mutual Funds

factors and an investment style, ultimately characterizing liquidity as the missing style.1 Despite these powerful stock-level liquidity ndings, we are aware of almost no mutual fund managers that actively seek less liquid stocks. Might this emerging investment style and risk factor be present and economically signicant among mutual funds? If so, methods of knowingly or unknowingly constructing portfolios of less liquid stocks may be benecial not only for creating mutual funds but also for selecting mutual funds that are more likely to outperform their peers. If the liquidity style exists in mutual funds, our research might encourage fund managers to avoid trading very liquid (heavily traded) stocks and discourage unnecessary trading.

Data and Methodology


Investigating whether mutual funds that hold less liquid stocks tend to outperform those that hold more liquid stocks is a data-intensive exercise. First, we needed an individual stock database that would enable us to estimate the liquidity of each individual stock. Next, we needed to know the holdings of each individual mutual fund throughout time. By combining data from Morningstars individual stock database with data from Morningstars mutual fund holding database, we were able to build composites of mutual funds based on the weighted-average liquidity of the individual stocks held by the mutual funds. We began with Morningstars open-end equity mutual fund universe containing both live and dead funds. Our primary focus was on U.S. equity mutual funds, but we also included a sample of non-U.S. equity mutual funds. We organized our study around Morningstar categories and the large-versus-small and growthversus-value style box for two primary reasons. First, we wanted to control for the two most common equity stylessize and valuationat a granular level. The Morningstar categories in question are those of the nine sizevaluation squares that constitute the style box representing the U.S. equity universe, the three valuation-based columns from the style box (value, core, and growth), the three size-based rows from the style box (large, mid, and small), and the non-U.S. category. Morningstar categorizes equity funds on the basis of size and a 10-factor value-versus-growth model applied to the individual stocks held by the funds.2 As such, the funds in any specic category share similar size and valuation attributes, in which the valueversus-growth determination goes well beyond the book-to-market ratio.

Second, the primary goal of this study was to determine whether funds with less liquid individual stock holdings tend to outperform similar funds with relatively more liquid individual stock holdings. When analyzing fund performance, the de facto standard for academicians is Fama French-based regressions (see Fama and French 1995), whereas sophisticated institutional investors often use a variation of Sharpes return-based style analysis or perhaps one of the commercially available factor models (e.g., Barra, Northeld, FactSet). However, the most common and intuitive practitioner-oriented approach for evaluating performance is relative to a category or peer group average. For example, the compensation for many money managers is partially linked to their category or peer group ranking (e.g., Lipper or Morningstar category ranking). Turning to our dataset, Morningstar has either monthly or quarterly mutual fund holdings data starting in 1983. However, wide-scale holdings data for most funds were not available until January 1995 for the U.S. equity fund universe (and January 2000 for the non-U.S. equity fund universe). For the U.S. equity fund universe, holdings data from January 1995 were used to form the starting composites that we began tracking in February 1995. The constituents of the composites are based on the previous months holdings information. Thus, our dataset consists of 14 years and 11 months of U.S. performance history and 9 years and 11 months of non-U.S. performance history. Table 1 summarizes the number of live funds in the various universes/categories with the required data at the start of the study period (February 1995 for U.S. equity categories and February 2000 for the non-U.S. equity fund universe) and at the end of the study period. There are a number of potential measures of liquidity for an individual stock. For simplicity and consistency, we focused on the basic stock-level turnover measure used in Ibbotson et al. (2012): average daily shares traded over the last year divided by the number of shares outstanding. No attempt was made to adjust the number of shares outstanding for free oat. In one of our robustness checks, we reran our analysis using an alternative denition of liquidity (i.e., the Amihud measure). Bringing the two databases together enabled us to estimate each mutual funds weighted-average liquidity at each point in time. For a given mutual fund, if we did not have a liquidity turnover ratio for a holding, we ignored the position and rescaled the other holdings prior to calculating the mutual funds weighted-average liquidity.3

November/December 2012

www.cfapubs.org

39

Financial Analysts Journal

Table 1.

Number of Mutual Funds with Required Data


Start Date No. of Funds (February 1995) 42 73 123 45 84 131 212 322 262 238 260 796 299 479 516 1,294 634 start in February 2000. End Date No. of Funds (December 2009)a 238 369 494 229 314 527 719 1,260 1,048 1,101 1,070 3,027 1,186 1,943 2,069 5,198 815

Morningstar Category Small value Small core Small growth Mid value Mid core Mid growth Large value Large core Large growth Small Mid Large Value Core Growth All U.S. All non-U.S.b
aIncluding defunct funds. bNon-U.S. mutual fund data

Armed with each mutual funds weightedaverage stock-level liquidity within any given category, we rank ordered the mutual funds on the basis of their weighted-average liquidity and used this information to form monthly rebalanced, equally weighted composites (in our case, quintiles) of mutual funds with similar weighted-average stock-level liquidity scores.4 Funds with the lowest weighted-average liquidity were assigned to the L1 quintile, and funds with the highest weightedaverage liquidity were assigned to the L5 quintile. The constituent mutual funds in the composite evolved each month as the weighted-average stocklevel liquidity of the mutual funds evolved.

Results
Table 2 summarizes the striking results for our primary universe of U.S. equity funds. Related to total returns, the table displays the annual geometric return, annual arithmetic return, standard deviation, and Sharpe ratio. In addition to these total return statistics, we also report alphas and t-statistics from two different regressions: (1) a single-factor regression of the total returns of each composite against the total returns of the appropri-

ate category-average composite and (2) a multifactor regression of each composites excess returns (over T-bills) against the three traditional Fama French factorsthe excess market return (adjusted for average U.S. mutual fund expenses), small minus big (SMB), and high minus low (HML), which most practitioners know as value (high book-to-market ratio) minus growth (low bookto-market ratio). The annualized alphas from the monthly return regressions as well as their corresponding t-statistics are also displayed in Table 2.5 The category-average composite is simply a composite representing the equally weighted return of all the funds in a particular category through time. Again, the determination of each funds category is based on an 11-factor model applied to individual stock holdings and then rolled up to determine the category. The nal row of each categorys section in Table 2 shows the difference in performance statistics from the lowest-liquidity composite (L1) and the highest-liquidity composite (L5), the t-statistic obtained by regressing L1 minus L5 on the category average, and the annualized alpha and t-statistic obtained by regressing L1 minus L5 on the three FamaFrench factors.

40

www.cfapubs.org

2012 CFA Institute

The Liquidity Style of Mutual Funds

Table 2.

Mutual Fund Liquidity Quintiles: U.S. Equity Universe, February 1995December 2009
Annualized Alpha Relative to Category Average (%) 1.68 0.77 0.14 0.61 2.17 3.93 2.93 0.59 0.61 1.19 1.87 4.88 2.81 0.47 0.90 0.01 1.98 4.88 2.31 0.49 0.20 0.46 2.83 5.27 2.71 0.81 0.33 0.91 2.85 5.71 3.36 1.27 0.13 t-Statistic of Alpha Relative to Category Average 2.69 1.46 0.26 1.17 2.20 3.21 3.21 0.88 1.35 1.76 2.09 3.19 2.15 0.81 1.77 0.01 2.58 2.59 3.66 0.69 0.35 0.75 2.52 3.92 2.32 1.13 0.54 1.38 2.24 2.90 2.36 1.87 0.30 Annualized Alpha Relative to Fama French Factors (%) 2.50 2.10 1.58 0.84 0.86 1.26 3.39 2.87 1.00 0.18 0.11 0.73 0.67 3.62 0.82 0.40 1.21 0.02 1.76 0.50 2.62 3.74 2.55 2.23 1.87 0.54 2.00 4.30 3.22 2.17 2.14 1.03 0.51 1.60 3.74 2.22 1.52 0.85 t-Statistic of Alpha Relative to Fama French Factors 1.71 1.33 0.96 0.52 0.67 0.90 2.98 1.98 0.67 0.12 0.08 0.57 0.52 3.35 0.70 0.31 0.88 0.01 1.14 0.41 1.83 2.88 1.53 1.54 1.29 0.40 1.54 3.73 2.48 1.76 1.68 0.75 0.40 1.50 2.48 2.10 1.17 0.57
(continued)

Geometric Arithmetic Mean Mean (%) (%) Small value L1 Small value L2 Small value L3 Small value L4 Small value L5
Small value average

Standard Deviation (%) 17.61 19.55 19.63 19.81 20.75 19.27 3.15 17.74 19.09 20.43 22.07 21.98 20.02 4.24 20.44 23.91 25.93 27.84 29.30 25.22 8.86 15.52 17.13 17.23 18.92 20.39 17.56 4.87 15.87 18.29 19.86 20.49 21.86 18.87 5.99 17.94 21.82 24.38

Sharpe Ratio 0.50 0.46 0.43 0.39 0.31 0.42 0.18 0.51 0.40 0.35 0.34 0.30 0.38 0.21 0.37 0.29 0.24 0.29 0.22 0.28 0.15 0.56 0.45 0.44 0.42 0.30 0.43 0.25 0.52 0.44 0.43 0.37 0.28 0.41 0.24 0.43 0.35 0.31

10.86 10.79 10.17 9.48 8.09 9.91 2.77 11.25 9.48 8.81 8.73 7.94 9.29 3.32 9.26 7.88 6.87 8.13 6.26 7.77 3.00 11.06 9.95 9.76 9.82 7.81 9.73 3.25 10.66 10.06 10.24 9.22 7.47 9.61 3.19 9.82 9.14 8.38

12.26 12.50 11.91 11.26 10.05 11.59 2.21 12.67 11.14 10.71 10.93 10.13 11.11 2.54 11.15 10.43 9.87 11.50 10.04 10.60 1.10 12.15 11.28 11.11 11.45 9.72 11.14 2.42 11.81 11.58 12.02 11.12 9.65 11.23 2.16 11.27 11.26 11.01

L1 L5 Small core L1 Small core L2 Small core L3 Small core L4 Small core L5 Small core average L1 L5 Small growth L1 Small growth L2 Small growth L3 Small growth L4 Small growth L5 Small growth average L1 L5 Mid value L1 Mid value L2 Mid value L3 Mid value L4 Mid value L5 Mid value average L1 L5 Mid core L1 Mid core L2 Mid core L3 Mid core L4 Mid core L5 Mid core average L1 L5 Mid growth L1 Mid growth L2 Mid growth L3

November/December 2012

www.cfapubs.org

41

Financial Analysts Journal

Table 2.

Mutual Fund Liquidity Quintiles (continued)


Annualized Alpha Relative to Category Average (%) 1.47 2.54 6.04 1.61 0.46 0.03 0.55 1.71 3.37 1.70 0.16 0.33 0.59 1.01 2.74 1.99 0.72 0.61 0.16 1.70 3.74 4.17 1.45 0.53 1.67 2.90 7.27 3.90 2.06 0.25 1.97 3.23 t-Statistic of Alpha Relative to Category Average 2.32 2.40 2.72 3.67 1.59 0.15 2.29 2.98 3.76 2.88 0.70 1.21 2.66 1.39 2.55 2.01 1.24 2.01 0.44 1.38 2.23 2.45 1.25 1.42 1.68 2.48 2.84 2.13 1.91 0.54 2.63 2.24 Annualized Alpha Relative to Fama French Factors (%) 0.04 0.58 0.78 2.82 2.07 1.09 0.65 0.19 1.11 0.62 3.21 1.50 0.25 0.01 0.43 0.97 0.08 2.48 1.21 0.31 0.82 0.18 1.21 0.18 2.46 2.66 0.74 0.23 0.59 1.63 0.19 4.36 2.75 2.14 1.33 0.01 0.63 t-Statistic of Alpha Relative to Fama French Factors 0.02 0.29 0.58 1.42 2.02 1.12 0.71 0.20 1.33 0.69 4.58 2.39 0.61 0.02 1.05 2.06 0.24 3.18 1.86 0.47 1.10 0.20 0.83 0.25 1.50 1.87 0.54 0.19 0.47 1.24 0.18 2.51 2.30 2.04 1.10 0.01 0.37
(continued)

Geometric Arithmetic Mean Mean (%) (%) Mid growth L4 Mid growth L5 Mid growth average L1 L5 Large value L1 Large value L2 Large value L3 Large value L4 Large value L5 Large value average L1 L5 Large core L1 Large core L2 Large core L3 Large core L4 Large core L5 Large core average L1 L5 Large growth L1 Large growth L2 Large growth L3 Large growth L4 Large growth L5 Large growth average L1 L5 Small L1 Small L2 Small L3 Small L4 Small L5 Small average L1 L5 Mid L1 Mid L2 Mid L3 Mid L4 Mid L5 7.35 6.63 8.38 3.18 8.44 7.65 7.42 7.07 6.11 7.35 2.33 7.95 6.91 6.66 6.35 6.30 6.86 1.65 7.61 6.92 5.86 6.74 5.87 6.68 1.75 11.12 9.36 8.40 7.76 6.75 8.82 4.37 10.24 10.00 9.25 7.75 6.91 10.39 10.19 10.82 1.08 9.41 8.75 8.61 8.34 7.52 8.52 1.89 8.95 8.10 7.97 7.63 7.84 8.10 1.11 8.80 8.36 7.48 8.63 8.60 8.38 0.20 12.61 11.19 10.72 10.59 9.91 11.00 2.70 11.42 11.58 11.28 10.34 10.08

Standard Deviation (%) 26.15 28.46 23.39 10.52 14.49 15.38 16.04 16.53 17.32 15.88 2.83 14.69 15.98 16.76 16.56 18.12 16.31 3.42 16.04 17.64 18.62 20.30 24.52 19.15 8.48 18.19 20.06 22.74 25.28 26.62 22.01 8.42 16.08 18.70 21.29 24.07 26.70

Sharpe Ratio 0.26 0.23 0.31 0.20 0.41 0.34 0.32 0.29 0.23 0.31 0.18 0.37 0.29 0.27 0.25 0.24 0.28 0.13 0.33 0.27 0.21 0.25 0.21 0.25 0.12 0.50 0.38 0.32 0.28 0.24 0.34 0.26 0.49 0.43 0.36 0.28 0.25

42

www.cfapubs.org

2012 CFA Institute

The Liquidity Style of Mutual Funds

Table 2.

Mutual Fund Liquidity Quintiles (continued)


Annualized Alpha Relative to Category Average (%) 7.34 2.42 0.94 0.45 1.07 1.76 4.24 2.38 0.65 0.09 0.13 2.44 4.93 2.14 0.25 0.61 0.65 1.14 3.32 1.76 0.42 0.06 0.54 1.76 3.58 2.95 1.07 0.29 0.66 2.54 5.62 t-Statistic of Alpha Relative to Category Average 2.62 2.10 1.46 1.29 2.84 1.21 2.23 1.72 0.80 0.30 1.18 1.88 2.12 2.86 0.42 1.28 1.44 0.89 2.38 3.28 1.09 0.22 1.91 1.96 3.08 2.04 1.04 0.54 1.07 1.37 2.36 Annualized Alpha Relative to Fama French Factors (%) 1.05 3.40 1.90 2.10 1.58 0.84 0.86 1.26 3.39 1.26 0.04 0.13 0.19 1.68 0.14 2.99 2.14 0.79 0.06 0.21 0.72 0.40 2.88 2.49 1.40 1.06 0.53 0.99 0.92 3.52 2.18 0.89 0.18 0.19 1.43 0.16 3.65 t-Statistic of Alpha Relative to Fama French Factors 0.99 1.53 1.71 1.33 0.96 0.52 0.67 0.90 2.98 1.84 0.05 0.15 0.16 1.09 0.17 1.74 2.68 1.40 0.11 0.29 0.79 0.69 2.98 2.48 1.37 0.96 0.49 0.97 0.94 4.20 2.44 1.44 0.30 0.25 1.16 0.29 2.16

Geometric Arithmetic Mean Mean (%) (%) Mid average L1 L5 Large L1 Large L2 Large L3 Large L4 Large L5 Large average L1 L5 Growth L1 Growth L2 Growth L3 Growth L4 Growth L5 Growth average L1 L5 Core L1 Core L2 Core L3 Core L4 Core L5 Core average L1 L5 Value L1 Value L2 Value L3 Value L4 Value L5 Value average L1 L5 All L1 All L2 All L3 All L4 All L5 All average L1 L5 9.01 3.33 8.34 7.49 6.41 6.01 6.03 6.93 2.30 8.10 7.18 7.34 7.83 5.85 7.40 2.26 9.12 7.84 7.21 7.43 7.48 7.87 1.63 9.29 8.40 8.27 7.92 7.01 8.20 2.28 9.09 7.98 7.15 7.58 6.44 7.80 2.65 10.94 1.34 9.35 8.66 7.71 7.44 8.11 8.25 1.24 9.38 8.78 9.36 10.50 9.18 9.44 0.20 10.15 9.10 8.57 8.90 9.39 9.22 0.76 10.30 9.56 9.55 9.26 8.59 9.45 1.71 10.16 9.24 8.58 9.44 9.22 9.33 0.94

Standard Deviation (%) 20.69 10.62 14.81 15.92 16.72 17.46 21.23 16.83 6.42 16.67 18.61 21.02 24.46 27.34 21.17 10.67 15.04 16.54 17.09 17.81 20.37 17.08 5.33 14.86 15.87 16.68 17.03 18.39 16.43 3.53 15.25 16.56 17.58 20.16 24.83 18.20 9.58

Sharpe Ratio 0.36 0.25 0.39 0.32 0.25 0.22 0.22 0.28 0.18 0.35 0.28 0.28 0.28 0.21 0.28 0.14 0.44 0.34 0.29 0.30 0.29 0.33 0.15 0.46 0.38 0.36 0.34 0.27 0.36 0.18 0.43 0.35 0.29 0.29 0.23 0.32 0.21

Notes: This table shows annualized results from monthly rebalanced composites. L1 represents the lowest-liquidity quintiles, and L5 represents the highest-liquidity quintiles.

November/December 2012

www.cfapubs.org

43

Financial Analysts Journal

For each of the 16 groupings, the lowestliquidity composite had a superior annual geometric return, annual arithmetic return, standard deviation, Sharpe ratio, annualized alpha versus the categorys composite average, and annualized alpha versus the three FamaFrench factors. With the exception of the growth category, the t-statistic of the alpha versus the categorys composite average of the lowest-liquidity composite exceeded 2.0, indicating that the alpha was statistically signicant at the 95% condence level. In 9 of the 16 groupings, the t-statistic of the alpha versus the three FamaFrench factors for the lowest-liquidity composite exceeded 2.0. In 11 of the 16 groupings, the t-statistic of the alpha versus the three Fama French factors for the zero-dollar portfolio created by L1 minus L5 exceeded 2.0. Furthermore, for the vast majority of groupings across the ve quintiles, the results are monotonic in favor of the lowerliquidity composites. Of note, for all 16 groupings in Table 2, for the annualized alpha versus the category-average composite, the alpha differential between the lowest- and highest-liquidity composites exceeded the annual geometric return differential. The same is true for 13 of the 16 groupings for the annualized alpha versus the three FamaFrench factors. This nding is a direct result of the lower standard deviation and lower beta of the lowest-liquidity composites relative to the highest-liquidity composites. For the regressions versus the appropriate category-average composite, the largest annualized alpha difference between the L1 and L5 quintiles occurred in the small category (727 bps) and the smallest annualized alpha difference occurred in the large core category (274 bps). For the regressions versus the FamaFrench factors, the largest annualized alpha difference between the L1 and L5 quintiles was again in the small category (436 bps) and the smallest annualized alpha difference occurred in the large growth category (246 bps). These results, as well as the rest of the results reported in Table 2, are consistent with the stocklevel results in Ibbotson et al. (2012), despite the difference in sample period. We highlight the performance of the All composites, representing our entire universe of U.S. equity funds, at the bottom of Table 2. The annual geometric return for the All L1 composite was 2.65% higher than that of the All L5 composite, the standard deviation was much lower (15.25% versus 24.83%), and the Sharpe ratio was nearly twice as high (0.43 versus 0.23). The annualized alphas relative to the category-average composite and the three FamaFrench factors for the lowest-liquidity composite were quite large and statistically sig-

nicant: 2.95% and 2.18%, respectively. For the L1 minus L5 regressions, the alphas were even larger: 5.62% and 3.65%. Many of the key results from Table 2 are summarized in the nine style box squares in Figure 1, which enables comparison of the liquidity premium with the value and size premiums. Within each square in the style box, the top number is the annual geometric return for that categorys lowest-liquidity composite, the second number is the annual geometric return for that categorys average, the third number is the annual geometric return for that categorys highest-liquidity composite, and the nal, bold number is the difference between the categorys L1 and L5 composites. The bold numbers to the right of the style box show the value minus growth differences for the appropriate size categories, and the bold numbers below the style box show the small minus large differences for the appropriate valuation categories. The most interesting comparisons are between the bold numbers; in general, the L1 minus L5 differences (the bold numbers inside the style box) exceed the value minus growth and small minus large differences (the bold numbers outside the style box). By organizing our study around the style box, we have largely controlled for size and valuation. Furthermore, by analyzing each composite relative to the FamaFrench factors, we have completely controlled for size and valuation (almost twice). The relationship between liquidity and the next most signicant factor, momentum, is left to further research and is the subject of a follow-up study Idzorek, Xiong, and Ibbotson (2011). Observing the nearly 15-year history for the All L1 and L5 liquidity quintiles reveals an interesting result (see Figure 2). For the most part, the lowest-liquidity composite dominated; however, for a brief period corresponding with the height of the technology bubble, the highest-liquidity composite (the dotted line in Figure 2) temporarily dominated. During this irrational period, investors could not get enough of the most liquid stocks, which beneted the mutual funds that held these glamour stocks. Interestingly, the brief outperformance of the high-liquidity composites during the technology bubble was either not as prevalent or nonexistent in the value-oriented categories, as illustrated in Figure 3, which shows the growth of a dollar invested in the mutual fund composites constructed from the value-oriented fund categories. We suspect this noteworthy pattern was less prevalent among value managers because they were unlikely to hold technology stocks at that time. Table 3 reports various up-market and downmarket return capture statistics for the All composites. The superior overall performance of the

44

www.cfapubs.org

2012 CFA Institute

The Liquidity Style of Mutual Funds

Figure 1.

Style Box Liquidity Performance for the U.S. Equity Mutual Fund Universe, February 1995December 2009
Value Core Valuation Spectrum Growth Value minus Growth

8.44 Large 7.35 6.11

7.95 6.86 6.30

7.61 6.68 5.87

0.67

2.33
Size Spectrum

1.65

1.75

11.06 9.73 7.81

10.66 9.61 7.47

9.82 8.38 6.63

Mid

1.35 L5 Compound Return


L1 L5

L1 Compound Return Style Squares Compound Return

3.25

3.19

3.18

10.86 Small 9.91 8.09

11.25 9.29 7.94

9.26 7.77 6.26

2.14

2.77
Small minus Large

3.32

3.00

2.56

2.43

1.09

Notes: This gure summarizes some of the key results from Table 2. All data are annualized geometric means measured in percent.

Figure 2.

All Liquidity Quintile Performance Comparison: Growth of $1, February 1995December 2009

Performance ($) 5 4

1 Jan/95

Jan/97

Jan/99

Jan/01

Jan/03

Jan/05

Jan/07

Jan/09 Average

All L1 (Lowest Liquidity)

All L5 (Highest Liquidity)

November/December 2012

www.cfapubs.org

45

Financial Analysts Journal

low-liquidity quintiles has primarily come from superior performance in down markets, as indicated by the down-market capture. Lower downmarket capture means a lower average loss in down markets. In particular, the losses for L1 in the two crisis periods (the 2000 tech crash and the 200809 nancial crisis) were signicantly worse than the losses for L5. This result is consistent with Lou and Sadka (2011), who found that illiquid stocks outperformed liquid stocks during the 200809 nancial crisis because liquid stocks are more sensitive to liquidity shocks. We repeated the monthly up-market/downmarket capture analysis for the 15 subcategories. The results paint a similar picture: In each case, the lowest-liquidity composite had a superior upmarket/down-market capture ratio relative to the corresponding highest-liquidity composite. Many people nd these results puzzling because their intuition tells them that in down markets, less liquid stocks (and the funds that hold them) should suffer the steepest declines. We posit that one cause of the superior downside performance of the lowest-liquidity quintile relates to the type of strategies typically used by low-liquidity managers versus high-liquidity managers. We suspect that, on average, the funds in the lowestliquidity quintile have less holdings turnover than those in the highest-liquidity quintile, which reects a general preference for a longer-holdingperiod strategy. In contrast, high-liquidity managFigure 3.

ers likely have higher holdings turnover and, on average, use strategies that involve more frequent trading. Funds that trade frequently pay greater attention to trading costs and are more likely to use liquidity-based measures, such as the bidask spread, to screen out relatively less liquid stocks. Furthermore, during periods of turmoil, highliquidity managers may be more likely to trade; thus, the most liquid stocks may, in fact, suffer the steepest declines because there is a greater propensity for their owners to trade them. We conrmed that this is the case by analyzing the standard holdings turnover statistic for the mutual funds that make up the composites. For the entire U.S. mutual fund universe, the average annual holdings turnover over time was 59% for the mutual funds with the least liquid stocks (L1) and 124% for the mutual funds with the most liquid stocks (L5).6 Turning again to our liquidity measure, the average liquidity measures for the ve quintiles of U.S. equity funds over the sample period are shown in Table 4. Once again, we measured the liquidity of a stock as its average daily shares traded over the last year divided by the number of shares outstanding. We then calculated the liquidity of a fund as the weighted-average liquidity of the stocks it holds. Multiplying the daily gures by 250 (which is the approximate number of trading days per year) produced annualized gures. The small growth category had the largest liquidity difference between the L5 and L1 composites: 932% (= 1,095% 163%).

Value Liquidity Quintile Performance Comparison: Growth of $1, February 1995December 2009

Performance ($) 5 4

1 Jan/95

Jan/97

Jan/99

Jan/01

Jan/03

Jan/05

Jan/07

Jan/09 Average

Value L1 (Lowest Liquidity)

Value L5 (Highest Liquidity)

46

www.cfapubs.org

2012 CFA Institute

The Liquidity Style of Mutual Funds

Table 3.

Monthly Up-Market/Down-Market Capture Statistics: U.S. Equity Mutual Fund Universe, February 1995December 2009
Average DownMarket Return 3.09% 3.75 4.19 4.75 5.63 4.28 Up-Market/ DownMarket Capture Ratio 1.14 1.02 0.96 0.95 0.89 0.98

Up Periods All L1 All L2 All L3 All L4 All L5 All average 117 112 109 107 106 109

Down Periods 62 67 70 72 73 70

Average Up-Market Return 3.03% 3.3 3.47 3.89 4.37 3.61

UpMarket Capture 86.31 93.91 98.73 110.45 123.01 102.78

DownMarket Capture 75.93 91.88 102.57 116.32 138.71 104.85

Loss from Apr. 2000 to Dec. 2001 11.0% 5.9 17.4 24.5 39.6 17.7

Loss from Sep. 2008 to Feb. 2009 40.3% 42.3 43.5 43.5 45.1 42.8

Notes: Monthly up-market/down-market capture statistics are from Morningstar EnCorr. The Up Periods and Down Periods columns report the total number of months with positive and negative monthly returns from the sample of 179 months. The average up-market and down-market returns are based on the performance of the market, which in this case is dened as the Russell 3000 Index. The up-market (down-market) capture identies the percentage of the markets upward (downward) movements that are captured, where numbers greater than 100 indicate greater sensitivity than the Russell 3000. The up-market/down-market capture ratio is the up-market capture divided by the down-market capture.

Table 4.

Annual Average Stock Turnover within Fund Categories: U.S. Equity Mutual Fund Universe, February 1995December 2009
Annual Stock Turnover L1 108% 118 163 110 118 158 93 105 128 125 128 105 95 110 135 110 L2 143% 163 228 138 165 218 113 135 160 180 183 135 120 140 178 145 L3 168% 200 290 163 208 273 130 150 188 233 233 155 138 160 223 175 L4 208% 263 380 203 260 355 150 168 230 315 308 185 163 195 295 230 L5 545% 793 1,095 493 610 808 220 270 450 928 728 343 328 455 725 573

Category Small value Small core Small growth Mid value Mid core Mid growth Large value Large core Large growth Small Mid Large Value Core Growth All U.S.

November/December 2012

www.cfapubs.org

47

Financial Analysts Journal

For the L5 and L1 composites, every outstanding share of stock traded approximately 10.95 and 1.63 times per year, respectively. The large value category had the smallest liquidity difference, 127% (= 220% 93%).7 In general, small funds and growth funds have larger liquidity differences than large funds and value funds, respectively, which indicates that small categories and growth categories tend to hold relatively heavily traded stocks. The liquidity measure for the L1 mutual funds in the All U.S. sample was 110% per year. All but two quintiles in Table 4 (large value L1 and value L1) had stock holdings with average annual turnover rates exceeding 100% per year.

the implementation lag slightly enhanced performance. Notice that with no implementation delay (Table 2), the geometric mean return for the All L1 composite exceeded that of the All L5 composite by 2.65%. Surprisingly, with the three-month implementation delay (Table 5), the difference increased to 3.26%. Similarly, the alphas for the L1 minus L5 regressions in Table 5 are even larger than the corresponding alphas in Table 2. Annual Rebalancing. For our monthly rebalanced composites, on average, about 70% of the funds from a given composite remained in the same composite in the following month; thus, for all practical purposes, buying and selling numerous different mutual funds each month or each quarter in order to hold the mutual funds with the least liquid stock holdings is impractical. Although one would expect it to be a less pure way of gathering exposure to low-liquidity stocks, would simply buying, each year, an annually rebalanced basket of mutual funds with the lowest average weighted-average liquidity measure in the previous year produce similar results? To determine the answer, we calculated the performance of annually rebalanced composites of mutual funds. Table 6 contains the results. Like the monthly rebalanced results reported in Table 2, the annually rebalanced results in Table 6 are extremely positive. Again, for space considerations, we present the results for only the U.S. equity universe All composites; however, for each of the 16 groupings, the lowest-liquidity composite had a superior annual arithmetic return, annual geometric return, standard deviation, Sharpe ratio, annualized alpha versus the categorys composite average, and annualized alpha versus the three FamaFrench factors. In contrast to the results

Robustness Checks
To test the robustness of the results reported in Table 2, we carried out a number of tests. We used a quarterly implementation delay, annually (instead of monthly) rebalanced composites, an alternative denition of liquidity, and a non-U.S. equity fund universe. Quarterly Implementation Delay. To test the sensitivity of our results to an implementation delay due to the availability of timely holdings data or to account for a potential lag effect due to stale pricing (which should be a nonissue), we repeated the analysis under the assumption of a one-quarter implementation delay. The results were quantitatively and qualitatively similar to those in Table 2. For space considerations, we present the results for only the U.S. equity universe All composites (see Table 5). The quarterly implementation lag decreased the number of monthly observations from 179 to 176, but the key statistics in Table 5 are very similar to those for the corresponding All composites listed at the bottom of Table 2. If anything, Table 5.

Mutual Fund Liquidity Quintiles with a Quarterly Implementation Delay: U.S. Equity Mutual Fund Universe, April 1995December 2009
Annualized Alpha Relative to Category Average (%) 3.05 1.26 0.32 0.97 2.75 5.95 t-Statistic of Alpha Relative to Category Average 2.19 1.29 0.61 1.59 1.48 2.44* Annualized Alpha Relative to FamaFrench Factors (%) 2.39 1.16 0.18 0.48 1.72 0.18 4.17 t-Statistic of Alpha Relative to Fama French Factors 2.66 1.92 0.30 0.61 1.37 0.32 2.42

Geometric Mean (%) All L1 All L2 All L3 All L4 All L5 All average L1 L5 8.83 7.74 6.68 6.76 5.58 7.26 3.26

Arithmetic Mean (%) 9.89 9.00 8.14 8.66 8.34 8.80 1.55

Standard Deviation (%) 15.17 16.48 17.7 20.28 24.69 18.24 9.52

Sharpe Ratio 0.42 0.33 0.26 0.25 0.2 0.29 0.23

Note: This table shows annualized results from monthly rebalanced composites. *Indicates a t-statistic of the alpha from an L1 versus L5 regression.

48

www.cfapubs.org

2012 CFA Institute

The Liquidity Style of Mutual Funds

based on a quarterly implementation delay, the t-statistics of the alphas for the L1 minus L5 regressions no longer exceeded 2. An Alternative Denition of Liquidity. Thus far, our analyses have focused on one of the simplest liquidity level measuresturnover, which we dene as the average daily shares traded over the last year divided by the number of shares outstanding. The literature has demonstrated that other liquiditylevel measures, such as the one proposed in Amihud (2002), have a signicant impact on performance; thus, we repeated our initial analysis with the Amihud measure, which is probably the best-known liquidity measure. A funds Amihud measure is dened as the weighted average of the Amihud measure of each stock holding, where the Amihud measure of a given stock holding is computed as 1 D | Ri ,d | Amihud measure = ln D d =1 Pi ,dVoli ,d where D = the number of trading days during the month (t) Ri,d = the stocks return on day d Pi,d = the adjusted price on day d Voli,d = the trading volume on day d For a given stock to be included during the month, we required a minimum of 10 days of corresponding price and volume data. Note that the Amihud measure calculates the absolute average return associated with dollar trading volume and, therefore, should be thought of as a measure of illiquidity: Low averages indicate small absolute returns associated with high dollar volumes, and high averages indicate large absolute returns associated with Table 6. ,

low dollar volumes. Because of the denominator price multiplied by volumethe Amihud measure does not adjust for market capitalization, and therefore, a stocks size has a signicant impact on the Amihud measure. Thus, one might expect that the lowest-liquidity composite based on the Amihud measure would contain small-cap funds that primarily hold small-cap stocks and would, therefore, have a higher standard deviation. In contrast, the turnover measure used in the rest of this article adjusts for market capitalization to some degree because both the numerator (shares traded) and the denominator (shares outstanding) are more or less equally affected by a stocks size (market capitalization). To rerun our analysis, we recalculated the Amihud measure for each fund in each month throughout the study. Based on the Amihud measure of illiquidity, we once again formed monthly rebalanced liquidity composites, in which L1 represents the lowest liquidity (highest Amihud measure). The results for the U.S. equity universe All composites are reported in Table 7. The results from Table 7 should be compared with those for the All composites in the nal section of Table 2. In Table 7, as in Table 2, the geometric and arithmetic means are similar, with the lowestliquidity composite producing signicantly better returns than the highest-liquidity composite. But the similarities between the two sets of results end there. The superior return of the lowest-liquidity composite in Table 7 was accompanied by the highest standard deviation, which seems logical given that small-cap funds are likely to have high Amihud measures (indicating low liquidity). The alphas and t-statistics in Table 7 are much less compelling than those in Table 2. The results for the other 15

Mutual Fund Liquidity Quintiles with Annual Rebalancing: U.S. Equity Mutual Fund Universe, January 1996December 2009
Annualized Alpha Relative to Category Average (%) 2.39 0.62 0.10 0.66 1.79 4.25 Annualized Alpha Relative to Fama French Factors (%) 2.31 1.04 0.60 0.39 0.40 0.69 2.72 t-Statistic of Alpha Relative to Fama French Factors 2.77 1.85 0.93 0.45 0.32 1.18 1.68

Geometric Mean (%) All L1 All L2 All L3 All L4 All L5 All average L1 L5 7.86 6.61 6.39 6.53 5.81 6.77 2.05

Arithmetic Mean (%) 8.92 7.86 7.83 8.42 8.42 8.29 0.50

Standard Deviation (%) 15.17 16.33 17.57 20.28 24.03 18.09 8.87

Sharpe Ratio 0.36 0.27 0.25 0.25 0.21 0.27 0.16

t-Statistic of Alpha Relative to Category Average 1.84 0.64 0.18 1.07 0.97 1.96*

Note: This table shows annualized results from annually rebalanced composites. *Indicates a t-statistic of the alpha from an L1 versus L5 regression.

November/December 2012

www.cfapubs.org

49

Financial Analysts Journal

categories are not reported, but they are similar to those reported in Table 7: better returns for the lower-liquidity composites accompanied by higher standard deviations and insignicant alphas. The Amihud measure of liquidity seems to measure a dimension of liquidity that is distinct from the one captured by our basic turnover measure.8 In an attempt to reconcile the performance differences between the composites based on turnover and the composites based on the Amihud measure, we ran our Amihud measure analysis with an additional twist: We conducted a double sort to control for fund volatility. To control for the effect of volatility, we sorted our U.S. fund universe into starting quintiles based on trailing, rolling 36-month volatility. Then, within each volatility-based quintile, we sorted the funds into ve equally weighted composites based on the turnover liquidity measure and the Amihud liquidity measure. Thus, for each liquidity measure (turnover and Amihud), we have 25 composites (i.e., ve L1 composites, ve L2 composites, and so on). Those who are familiar with the traditional method of displaying the results of a double sort (a single statistic across both sorting criteria) will notice that we have adopted a slightly unorthodox display that enables us to compare multiple statistics with far fewer tables. Table 8 contains two panels. Panel A presents the average annualized geometric return, arithmetic return, standard deviation, and Sharpe ratio for the composites based on the turnover liquidity measure. More specically, the rst row of data shows the average statistic for the ve lowest L1 turnover composites associated with the ve volatility quintiles. Panel B presents results based on the Amihud measure of liquidity. The results are consistent with those Table 7.

observed previously after controlling for volatility. On average, the L1 minus L5 arithmetic mean and volatility for the turnover liquidity measure were lower than those for the Amihud measure, so the geometric mean was similar for both measures. Next, for each liquidity measure and for each volatility quintile, we created an excess return series by subtracting the L5 return from the L1 return. We then regressed each L1 minus L5 excess return series on the three FamaFrench factors. Table 9 shows the average intercept and beta coefcients associated with the two liquidity measures. In Table 9, all coefcients are signicant at the 1% condence level. The intercept for turnover is highly signicant, whereas the intercept for the Amihud measure is not. The market and HML coefcients have the same sign for each liquidity measure, but the SMB coefcients have different signs (negative for turnover, positive for Amihud). This result indicates that the lowest-liquidity quintile for the turnover measure holds, on average, lowbeta, large value stocks and the lowest-liquidity quintile for the Amihud measure holds, on average, low-beta, small value stocks. Because small stocks tend to have higher volatility than large stocks, the lowest-liquidity quintile for the Amihud measure tends to have higher volatility than that for the turnover measure. Non-U.S. Equity Fund Universe. Finally, going beyond the universe of U.S. equity funds, we repeated our analysis (without the implementation delay and with monthly rebalancing) using a universe of non-U.S. equity funds (see Table 10). Unfortunately, our sample size was much smaller because we lacked the required individual stock data or holdings data (or both) for a relatively large

Mutual Fund Liquidity Quintiles Based on the Amihud Measure: U.S. Equity Mutual Fund Universe, February 1995December 2009
Annualized Alpha Relative to Category Average (%) 0.82 0.80 0.14 0.65 0.92 1.75 t-Statistic of Alpha Relative to Category Average 0.54 1.00 0.25 0.87 0.73 1.19* Annualized Alpha Relative to FamaFrench Factors (%) 0.31 1.10 0.23 0.33 0.19 0.20 0.49 t-Statistic of Alpha Relative to Fama French Factors 0.27 1.03 0.31 0.89 0.34 0.34 0.34

Geometric Mean (%) All L1 All L2 All L3 All L4 All L5 All average L1 L5 9.03 9.07 7.49 6.52 6.35 7.78 2.67

Arithmetic Mean (%) 11.06 10.94 8.87 7.86 7.85 9.31 3.21

Standard Deviation (%) 21.20 20.28 17.26 16.94 17.95 18.21 3.25

Sharpe Ratio 0.35 0.37 0.31 0.26 0.24 0.32 0.11

Note: This table shows annualized results from monthly rebalanced composites. *Indicates a t-statistic of the alpha from an L1 versus L5 regression.

50

www.cfapubs.org

2012 CFA Institute

The Liquidity Style of Mutual Funds

Table 8.

Volatility Quintiles: U.S. Equity Mutual Fund Universe, February 1995December 2009
Geometric Mean (%) Arithmetic Mean (%) 10.77 9.83 9.11 8.97 8.39 2.38 Standard Deviation (%) 15.45 16.63 17.51 18.17 18.66 3.21

Sharpe Ratio 0.44 0.35 0.29 0.27 0.24 0.20

A. Turnover liquidity measure All L1 All L2 All L3 All L4 All L5 L1 L5 B. Amihud liquidity measure All L1 All L2 All L3 All L4 All L5 L1 L5 9.24 8.35 7.69 6.82 6.26 2.97 11.04 10.11 9.39 8.56 8.01 3.03 17.65 17.49 17.26 17.54 17.65 0 0.40 0.35 0.31 0.26 0.23 0.17 9.41 8.26 7.36 7.08 6.41 3.00

Note: This table shows average annualized results from monthly rebalanced composites.

Table 9.

FamaFrench Three-Factor Regression: U.S. Equity Mutual Fund Universe, February 1995December 2009
Annualized Alpha Market 0.118 6.87 SMB 0.294 12.48 HML 0.135 7.70

A. Turnover liquidity measure L1 L5 t-Statistic B. Amihud liquidity measure L1 L5 t-Statistic 1.21% 0.87 0.194 7.33 0.561 15.49 0.289 10.75 3.66% 3.73

Note: This table shows results from monthly rebalanced composites.

Table 10.

Mutual Fund Liquidity Quintiles: Non-U.S. Equity Mutual Fund Universe, February 2000December 2009
Annualized Alpha Relative to Category Average (%) 0.54 1.66 0.60 0.22 2.10 2.69 t-Statistic of Alpha Relative to Category Average 0.40 1.28 0.37 0.15 0.77 0.85* Annualized Alpha Relative to FamaFrench Factors (%) 0.13 1.00 1.15 1.12 3.95 0.88 3.69 t-Statistic of Alpha Relative to Fama French Factors 0.09 0.65 0.72 0.61 1.43 0.99 1.16

Geometric Mean (%) All L1 All L2 All L3 All L4 All L5 All average L1 L5 1.5 2.69 1.58 0.74 1.63 1.15 3.13

Arithmetic Mean (%) 3.28 4.07 2.97 2.37 1.13 2.76 2.15

Standard Deviation (%) 19.16 16.9 16.92 18.27 23.84 18.19 4.69

Sharpe Ratio 0.03 0.08 0.01 0.02 0.07 0 0.1

*Indicates a t-statistic of the alpha from an L1 versus L5 regression.

November/December 2012

www.cfapubs.org

51

Financial Analysts Journal

number of funds. Therefore, this small sample may not represent non-U.S. equity funds well. Because of the lack of data availability, we moved our start date from February 1995 to February 2000 and did not break the universe into subcategories. Overall, this nearly 10-year period was not particularly good for stocks. The results are less compelling than those for the U.S. mutual fund universe. The non-U.S. All L2 quintile had the highest Sharpe ratio. The geometric mean return of L1 again trumped that of L5, but in this case, it was mostly due to the dismal return of L5 rather than standout performance by L1. Although the non-U.S. equity fund results are consistent with the U.S. equity fund results, none of the alphas are signicant for this smaller and shorter non-U.S. sample. For this non-U.S. universe, we conducted the FamaFrench three-factor regression analysis using the non-U.S. developed market FamaFrench factors from Kenneth Frenchs website (the market factor was adjusted for average fees).9 Perhaps the most bizarre and inconsistent number in our study is the large, positive alpha for the non-U.S. All L5 composite. We do not have an explanation for it, and more than any other result in the study, it contradicts our overall nding that mutual funds that hold less liquid stocks earn an illiquidity premium.

Conclusion
In this study, we analyzed the presence, impact, and signicance of the liquidity investment style in mutual funds. We showed that mutual funds that hold relatively less liquid stocks from within the liquid universe of publicly traded stocks outperformed mutual funds that hold relatively more liquid stocks by 2.65% (annualized geometric mean over nearly the last 15 years). The results were conrmed by the monthly rebalanced mutual fund composites for our universe of U.S. equity mutual funds, as well as for each of the nine sizevaluation squares that form the style box, the three valuation-based columns from the style box (value, core, and growth), and the three size-based rows from the style box (large, mid, and small). More specically, for each of the 16 groupings, the lowest-liquidity composite (L1) had a superior annual geometric return, annual arithmetic return, standard deviation, Sharpe ratio,

annualized alpha versus the categorys composite average, and annualized alpha versus the three FamaFrench factors. Surprisingly, the outperformance of the mutual funds that hold less liquid stocks was primarily due to superior performance in down markets. One possibility is that during periods of turmoil, highliquidity managers may be more likely to trade; thus, the most liquid stocks may, in fact, suffer the steepest declines because there is a greater propensity for their owners to trade them. We found similar results in four separate robustness tests based on permutations in the construction of our liquidity-based composites. We reran our analysis with a one-quarter implementation delay, annual rebalancing instead of monthly rebalancing, an alternative denition of liquidity, and a non-U.S. equity fund universe; each test demonstrated the aggregate superiority of investing in funds that hold less liquid stocks. The results based on a one-quarter implementation delay and those based on annual rebalancing, taken together, show that the less liquid investment style or signal seems to last a relatively long time. Constructing composites based on the Amihud denition of liquidity produced positive but less signicant results, which suggests that the Amihud denition of liquidity measures a different dimension of liquidity than does turnover. After controlling for volatility, we found that both the turnover and Amihud measures favor funds that hold stocks with lower betas; however, a low-liquidity fund based on turnover tends to have a large value bias, whereas a low-liquidity fund based on the Amihud measure tends to have a small value bias and thus more volatility. Finally, the results for non-U.S. equity funds are less compelling than those for the U.S. mutual fund universe, although this result is less conclusive because of a lack of available data. Overall, the liquidity investment style is clearly present in mutual funds and leads to dramatic differences in performance. The authors thank Eric Hu and Xuefeng Yuan for excellent data analysis assistance and Alexa Auerbach and Maciej Kowara for helpful comments.
This article qualies for 1 CE credit.

Notes
1. The results of Ibbotson et al. (2012) as well as earlier versions of that paper are so compelling that they are documented and updated each year in the Ibbotson Stocks, Bonds, Bills, and Ination (SBBI) Classic Yearbook. In our opinion, the results of Ibbotson et al. coupled with the results reported in this article suggest the ubiquitous four-factor modelmarket, size, valuation, and momentumshould be expanded to include liquidity as a fth factor. We leave direct testing of this vefactor model for further research. 2. The factors include market capitalization (size) and 10 value-versus-growth measures: price-to-projected-earnings ratio, price-to-book ratio, price-to-sales ratio, price-to-cash-

52

www.cfapubs.org

2012 CFA Institute

The Liquidity Style of Mutual Funds


ow ratio, dividend yield, long-term projected earnings growth, book value growth, sales growth, cash ow growth, and historical earnings growth. For additional details, see Morningstar (2008). 3. In the cases in which we lacked liquidity turnover ratios for more than 40% of the holdings, we ignored the fund completely. For U.S. equity funds, we had stock-level liquidity turnover ratios for more than 95% of funds. For non-U.S. equity funds, only about 10% of funds had 60% or more stock-level liquidity turnover ratios. 4. When calculating the total assets under management (AUM) of the composites, we looked for systematic patterns, somewhat expecting that the L1 (lowest-liquidity) composite might systematically favor smaller mutual funds that can more readily invest in less liquid stocks without a signicant market impact. In contrast to our expectation, the L1 composites AUM, on average, was greater than that of the L5 (highest-liquidity) composite: L1 = $929.87, L2 = $990.92, L3 = $984.73, L4 = $779.65, and L5 = $510.54 (these numbers are in millions of dollars and are averaged across the composite and over the entire period). A related measure is the volatility of fund ows. If large funds have relatively smaller fund ows, then they can afford to hold more illiquid stocks because they can accommodate redemptions with the liquid portion of their portfolios. Indeed, our data show that the volatility of fund ows was lowest for the L1 composite and highest for the L5 composite. The volatilities of fund ows were L1 = 3.65%, L2 = 3.82%, L3 = 3.80%, L4 = 4.22%, and L5 = 4.6%, where the volatility of fund ows is measured as the average absolute net inow or outow as a percentage of fund size over all the funds in the respective composite over the 15-year period. 5. The annualized alpha was estimated using the following formula: (1 + Monthly alpha)12 1. 6. Holdings turnover is a measure of how much a mutual fund turns over its portfolio and should not be confused with our liquidity turnover measure, which measures the average liquidity of the individual stock holdings. We conrmed that the average holdings turnover ratio of the mutual funds in the L5 composite was signicantly higher than the average holdings turnover ratio of the mutual funds in the L1 composite by calculating the average holdings turnover of each composite at each point in time and then taking the average through time. 7. Somewhat curiously, even though the stocks in the large value category seemed to have the smallest liquidity difference in Table 2, the alpha of the L1 composite for the large value category had the highest t-statistic. 8. Although the results are not reported, we tested a third denition of liquidity: the liquidity beta risk factor from Pstor and Stambaugh (2003). Using monthly returns and rolling ve-year returns, we calculated the liquidity beta for each mutual fund by following Equation 2 in Lou and Sadka (2011). We found that only about 10% of the U.S. equity funds have signicant liquidity beta coefcients at the 5% level. Consistent with Lou and Sadka (2011), we found that the lowest-liquidity-beta quintile outperformed the highestliquidity-beta quintile by 5% cumulatively from September 2008 to February 2009. Somewhat surprisingly, however, the lowest-liquidity-beta quintile underperformed the highestliquidity-beta quintile by about 40% cumulatively from April 2000 to December 2001, a period that included the tech crash. 9. http://mba.tuck.dartmouth.edu/pages/faculty/ken. french/data_library.html.

References
Amihud, Yakov. 2002. Illiquidity and Stock Returns: CrossSection and Time-Series Effects. Journal of Financial Markets, vol. 5, no. 1 (January):3156. Amihud, Yakov, and Haim Mendelson. 1986. Asset Pricing and the BidAsk Spread. Journal of Financial Economics, vol. 17, no. 2 (December):223249. Carhart, Mark M. 1997. On Persistence in Mutual Fund Performance. Journal of Finance, vol. 52, no. 1 (March):5782. Datar, Vinay T., Narayan Y. Naik, and Robert Radcliffe. 1998. Liquidity and Stock Returns: An Alternative Test. Journal of Financial Markets, vol. 1, no. 2 (August):203219. Fama, Eugene F., and Kenneth R. French. 1995. Size and Bookto-Market Factors in Earnings and Returns. Journal of Finance, vol. 50, no. 1 (March):131155. Ibbotson, Roger G., Zhiwu Chen, Daniel Y.J. Kim, and Wendy Y. Hu. 2012. Liquidity as an Investment Style. Working paper, Zebra Capital Management and Yale School of Management (August). Idzorek, Thomas, James X. Xiong, and Roger G. Ibbotson. 2011. Combining Liquidity and Momentum to Pick Top-Performing Mutual Funds. Working paper, Morningstar Investment Management (January). Li, Jinliang, Robert Mooradian, and Wei David Zhang. 2007. Is Illiquidity a Risk Factor? A Critical Look at Commission Costs. Financial Analysts Journal, vol. 63, no. 4 (July/August):2839. Lou, Xiaoxia, and Ronnie Sadka. 2011. Liquidity Level or Liquidity Risk? Evidence from the Financial Crisis. Financial Analysts Journal, vol. 67, no. 3 (May/June):5162. Morningstar. 2008. Morningstar Style Box Methodology. Morningstar Methodology Paper (28 April): http://corporate. morningstar.com/us/documents/MethodologyDocuments/ MethodologyPapers/MorningstarStyleBox_Methodology.pdf. Pstor, Lubo, and Robert F. Stambaugh. 2003. Liquidity Risk and Expected Stock Returns. Journal of Political Economy, vol. 111, no. 3 (June):642685.

November/December 2012

www.cfapubs.org

53