Content uploaded by Crystal Lin
Author content
All content in this area was uploaded by Crystal Lin on Nov 27, 2021
Content may be subject to copyright.
Original article
Real estate mutual funds: a style analysis
Crystal Yan Lin, PhD,
a,
* Kenneth Yung, PhD
b
a
Department of Finance, Eastern Illinois University, School of Business, Charleston, IL 61920, USA
b
Department of Finance, Old Dominion University, College of Business & Public Administration,
Norfolk, VA 23529, USA
Abstract
We find that the characteristics of real estate related securities are different from those of the
general common equities. To help investors understand better the products offered by real estate
mutual funds, we develop style descriptors that are specifically created for real estate related securities.
Among the universe of real estate securities, we find real estate funds tilt toward large stocks and favor
growth moderately over value. Growth managers outperform value mangers in this sector by 1.51%
to 2.30% per year. However, there is evidence of shifts in the investment style among the funds. Our
results help investors in evaluating real estate fund performance and making better asset allocation
decisions. © 2007 Academy of Financial Services. All rights reserved.
Jel classifications: G11; G12
Keywords: Mutual fund performance; Style investment; Real estate mutual fund
1. Introduction
Over $8.1 trillion are currently managed by the U.S. mutual fund industry.
1
A significant
portion of this amount is actively managed by professional investment managers who
presumably rely on superior stock selection skills to outperform passive strategies. The
bewildering variety of approaches followed by investment managers very often makes it
difficult for investors to choose funds that are suitable. The institutional investment com-
munity has responded to the proliferation of investment methods by scrutinizing more
closely an investment manager’s investment style. The attention to investment style has
* Corresponding author. Tel.: ⫹1-217-581-2227; fax: ⫹1-217-581-6247.
E-mail address: cylin@eiu.edu (C.Y. Lin).
Financial Services Review 16 (2007) 261–273
1057-0810/07/$ – see front matter © 2007 Academy of Financial Services. All rights reserved.
several benefits. Among them, accounting for style helps performance evaluation by giving
a clearer picture of a manager’s stock selection skill. For example, the manager of a portfolio
of large stocks may appear disappointed relative to a broad market index, but performance
may be outstanding relative to a large stock benchmark. In addition, style-investing appeals
to investors as it gives them a convenient framework with which to organize their investment
strategies. Essentially, in style investing, investors group assets into different asset classes
referred to as styles and move money into and out of these styles. According to Jeremy
Siegel, style investing refers to “rotate between small and large and value and growth stocks’
(Siegel, 1998).
In addition to the extensive studies on mutual fund performance, in recent years financial
economists have also examined mutual fund investment styles. Brown and Goetzmann
(1997) and Carhart (1997) find that size and value help explain the differences in fund
performance. Chan, Chen and Lakonishok (2002) use the Fama-French factors as style
indices and find mutual funds adopt investment styles that tend to cluster around a broad
market benchmark, and the few funds that deviate from the index are more likely to favor
growth stocks and past winners. Barberis and Shleifer (2003) show how funds’ pursuit of
styles can account for observed patterns in stock returns. On the profitability of style
momentum strategies, Moskowitz and Grinblatt (1999) and Asness, Liew and Stevens (1997)
successfully apply momentum strategies to industry portfolios and country portfolios, re-
spectively. Lewellen (2002) reports that momentum strategies based on size and book-to-
market portfolios are at least as profitable as individual stock momentum. Chen and De
Bondt (2004) find evidence of style momentum within the S&P 500 index.
Extant studies on mutual funds have typically focused on general equity funds. To our
knowledge, there are very few published articles on real estate mutual funds and none has
examined specifically the issue of real estate mutual fund investment styles. O’Neal and Page
(2000) study the performance of 28 real estate mutual funds over a three-year period from
1996 to 1998. Their results show that real estate mutual funds do not offer positive abnormal
performance relative to several broader equity market indices. Lin and Yung (2004), using
a larger sample and a longer sample period, report that real estate mutual fund performance
is largely tied to that of the real estate industry. They also conclude that factors such as size,
book-to-market, and momentum are immaterial after accounting for the real estate market
factor (NAREIT index). Though Gallo, Lockwood and Rutherford (2000) consider invest-
ment styles of real estate mutual funds, they define investment style according to the types
of investment properties held.
In this study, we add to the literature by specifically examining the investment styles of
real estate mutual funds using style descriptors created for real estate related securities.
Damodaran and Liu (1993) and Kallberg, Liu and Trzcinka (2000) have suggested that
money managers investing in the real estate sector could produce positive abnormal returns
because of their specific appraisal skills and information. An investigation of the investment
styles of real estate funds hence will give us a clearer picture of a manager’s selection skill
in the sector. In addition, such an understanding would benefit investors who are more
interested in indirect real estate investments than direct real estate ownerships. In this study,
we use style descriptors that are similar to the Fama-French factors (SMB and HML) for
evaluating the styles of real estate funds. An advantage of this approach is that it is consistent
262 C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
with the ‘large or small’ and ‘value or growth’ rotations in Wall Street as described by
Jeremy Siegel. Moreover, Chan et al. (2002) have shown that the Fama-French factors
perform very well as style descriptors relative to other style classification schemes. A
significant difference between our study and the others is that we create the style descriptors
using only real estate related common stocks. We believe this is more appropriate given that
real estate mutual funds invest primarily in real estate related assets. When a real estate fund
manager ponders the question ‘large versus small’ or ‘value versus growth,’ the reference is
more likely the universe of real estate related assets instead of the entire population of
common stocks. In addition, researchers have found that the risk characteristics of real estate
securities are different from those of general common equities (Reilly & Brown 2000). Thus,
using the conventional Fama-French factors directly will give a biased analysis.
Our results show that the style descriptors specifically created for real estate related
securities perform well in understanding the investment styles of real estate mutual funds.
Among the universe of real estate securities, we find real estate funds tilt toward large stocks
on average. In addition, there is a moderate tendency to prefer growth to value stocks. On
average, growth managers outperform value mangers in this sector by 1.51% to 2.30% per
year. We also find evidence that fund managers, especially those among the losers, shift their
investment styles. Taken together there is little evidence that real estate fund managers are
able to time the style factors. Our investigations help us understand better the products
offered by real estate funds and the performance evaluation of fund managers. It helps
particularly in asset allocation decisions as shifts in investment style of real estate funds
represent disruptions to investor’s overall portfolio characteristics. Our concern regarding a
better understanding of the investment style of real estate mutual funds echoes that of Detzel
(2006) in that he also finds it necessary for mutual fund investors to be able to readily identify
each fund’s equity class in their investment decisions. In addition, Kadiyala (2004) points out
the importance of understanding the determinants of mutual fund performance in making
asset allocation decisions.
2. Data
Our sample period starts from January 1, 1997 and ends on December 31, 2004. The
sample covers all real estate mutual funds with at least 24 months of daily return data. Daily
data are obtained from Morningstar, Inc. Monthly returns are used in the analysis, which is
calculated from compounding daily returns. Fund characteristics data such as net assets,
expense ratio, and turnover are obtained from the respective fund prospectus. Table 1
provides selected descriptive statistics of the funds. The 1990s represents one of the fastest
growing periods for the mutual fund industry, with the size of managed assets achieving an
annual growth rate in excess of 19%. During this period, real estate mutual funds grew faster
than the fund industry as a whole, achieving a 44% growth per annum. The share of real
estate mutual funds in the industry grew from 0.05% in 1993 to about 0.30% in 2004 (source:
Investment Companies Yearbook). The mean expense ratio (1.50% to 1.98%) over the study
period appears high relative to the industry (1.17% to 1.20%).
2
This is consistent with the
implications of Damodaran and Liu (1993) and Kallberg, Liu and Trzcinka (2000) that
263C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
investment managers in this sector require specific appraisal skills and information. A quick
comparison between the annual returns of the real estate funds and the S&P 500 shows a
pattern that is consistent with the general observation that real estate investments have low
or negative correlations with the stock market (Goetzmann & Ibbotson 1990). In fact, we
have found a correlation coefficient of ⫺0.04, and that supports the general argument that
real estate investment could be a good defensive play for portfolio risk diversification.
Table 1 Sample descriptive statistics
Panel A: Net assets (millions)
1997 1998 1999 2000 2001 2002 2003 2004
Mean 183.70 159.19 93.08 98.51 109.05 128.76 131.77 108.07
Median 53.31 32.26 19.80 22.21 21.50 25.88 36.24 52.06
Maximum 3433.00 2480.00 1465.00 1309.00 1387.10 1270.00 1681.30 550.18
Minimum 2.11 0.00 0.00 0.00 0.00 0.95 0.92 1.26
SD 592.06 426.17 234.05 215.72 238.68 270.71 267.78 135.80
Panel B: Expense ratio (%)
1997 1998 1999 2000 2001 2002 2003 2004
Mean 1.52 1.50 1.74 1.78 1.78 1.86 1.98 1.79
Median 1.25 1.36 1.66 1.71 1.59 1.71 1.74 1.76
Maximum 3.49 2.60 4.18 4.14 5.24 4.57 4.85 2.92
Minimum 0.48 0.24 0.26 0.33 0.33 0.28 0.69 0.52
SD 0.66 0.58 0.77 0.76 0.91 0.88 0.91 0.62
Panel C: Turnover (%)
1997 1998 1999 2000 2001 2002 2003 2004
Mean 71.18 53.72 42.26 65.70 66.93 72.19 58.97 42.00
Median 57.00 42.58 38.00 39.00 42.75 47.00 45.55 27.34
Maximum 205.00 196.00 198.00 482.00 274.00 327.00 213.45 158.00
Minimum 8.40 2.00 2.52 7.00 5.00 6.00 13.11 13.00
SD 56.56 45.95 30.05 89.46 61.44 65.47 51.76 34.99
Panel D: Annual fund return (%)
1997 1998 1999 2000 2001 2002 2003 2004
Mean 4.86 ⫺25.52 ⫺12.34 13.85 ⫺0.15 ⫺3.86 28.13 14.68
Median 4.48 ⫺25.63 ⫺12.97 15.17 ⫺1.34 ⫺4.23 27.17 15.32
Maximum 22.99 ⫺3.41 20.12 21.59 18.95 14.70 77.10 23.14
Minimum ⫺11.43 ⫺39.59 ⫺22.20 ⫺7.81 ⫺13.97 ⫺20.11 ⫺38.85 ⫺1.60
SD 7.60 5.86 5.27 5.49 5.52 4.73 9.80 4.13
Observations 24 49 69 87 103 126 141 141
S&P500 33.36 28.58 21.04 ⫺9.1 ⫺11.89 ⫺22.1 28.69 10.88
Fund characteristics data (net assets, expense ratio, and turnover) is obtained from the respective fund
prospectus. Fund return data is from Morningstar, Inc.
264 C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
3. The Style Descriptors
We follow the method of Fama and French (1993) in constructing our style descriptors.
At the end of each year, all real estate related stocks with SIC codes 15, 16, 17, 65, and 6798
are ranked on size (price time shares). The median size is then used to split the stocks into
two groups, small and large (S and B). We also break the stocks into three book-to-market
equity groups based on the breakpoints for the bottom 30% (Low), middle 40% (Medium),
and top 30% (High) of the ranked ME/BE values. Similar to Fama and French, the decision
to sort firms into three groups on BE/ME and only two on ME follows the evidence in Fama
and French (1992) that book-to-market ratio has a stronger role in average stock returns than
size. We construct six portfolios (S/L, S/M, S/H, B/L, B/M, B/H) from the intersections of
the two ME groups and the three BE/ME groups.
3
The portfolios are reformed every year.
Our first style descriptor RESMB (small minus big among real estate related securities only)
is meant to mimic the risk factor related to size. It is computed as the difference, each month,
between the simple average of the returns on the three small-stock portfolios (S/L, S/M, and
S/H) and the simple average of the returns on the three big-stock portfolios (B/L, B/M, and
B/H). Thus, RESMB is the difference between the returns on small- and big-stock portfolios
with about the same weighted-average book-to-market equity. This difference is therefore
considerably free of the influence of BE/ME. Our second style descriptor REHML (high
minus low among real estate related securities only) is meant to mimic the risk factor in
returns related to book-to-market equity. REHML is the difference, each month, between the
simple average of the returns on the two high-BE/ME portfolios (S/H and B/H) and the
simple average of the returns on the two low-BE/ME portfolios (S/L and B/L). Thus,
REHML is the difference between the returns on high- and low-BE/ME portfolios with about
the same weighted-average size. This difference is, therefore, largely free of the influence of
the size factor in returns. In addition, our proxy for the market factor in stock returns is the
excess market return; RERM-RF. RERM is the return on the value-weighted portfolio of all
the real estate related stocks in the six size-BE/ME portfolios.
Table 2 reports selected descriptive statistics for the six real estate related stock portfolios.
The portfolio S/H has the most stocks. The large number of small stocks with high BE/ME
is consistent with the findings of Wang, Erickson, Gau and Chan (1995) that real estate
related securities are relatively less well researched. Throughout the study period, portfolio
B/H consistently has the least number of stocks. It appears likely that large real estate related
stocks are more followed and hence more efficiently priced. The range of mean BE/ME ratio
(1.03 to 2.07) is considerably higher than the 29-year average (0.30 to 1.80) of all the
NYSE-AMEX-NASDAQ stocks reported in Fama and French (1993). This again could
imply real estate related securities are priced differently in the market. The median size of
real estate firms over the study period ranges from US$ 211 million to US$588 million. On
average, the size of real estate firms is much smaller than that of all the NYSE-AMEX-
NASDAQ firms. Table 2 clearly points out that characteristics of real estate related securities
are very different from those of the NYSE-AMEX-NASDAQ universe. We consider this a
strong support for using our specifically created real estate related style descriptors, RESMB,
REHML, and RERM.
In Table 3, we report descriptive statistics of our real estate securities style descriptors
265C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
versus those of the conventional Fama-French factors (RMRF, SMB, and HML). Over the
study period, our real estate related size descriptor RESMB has a mean (median) that is much
higher than the Fama-French SMB. The median of RESMB is 0.703 while that of the
Table 2 Descriptive statistics of stocks used to construct real estate factors
Panel A: Number of stocks in the portfolios
Year-end Number of stocks in portfolio
S/L S/M S/H B/L B/M B/H Total
1996 24 21 48 38 40 14 185
1997 34 28 54 43 49 23 231
1998 35 29 68 53 58 20 263
1999 31 30 66 54 54 19 254
2000 19 38 71 66 47 14 255
2001 21 33 65 58 47 14 238
2002 24 29 62 52 49 14 230
2003 23 35 61 56 45 18 238
Panel B: Descriptive statistics for stocks in factor portfolios
Year-end Size ($ millions) BE/ME
Mean Median SD Mean Median SD
1996 402.6 211.6 504.5 2.07 0.64 8.06
1997 577.2 343.7 734.4 1.53 0.67 5.16
1998 581.4 297.1 795.7 1.77 0.83 5.41
1999 572.4 254.0 848.3 1.58 1.00 3.48
2000 703.3 238.2 1,300.1 1.49 0.88 3.51
2001 810.2 331.1 1,285.9 1.65 0.76 4.05
2002 801.1 369.2 1,247.8 1.52 0.80 4.05
2003 1,206.4 588.3 1,687.9 1.03 0.61 2.78
Note: At the end of each year, we sort stocks with SIC codes 15, 16, 17, 65, and 6798 by their market
capitalization and book-to-market ratio. Six portfolios are constructed: S/L, small-growth; S/M, small-neutral;
S/H, small-value; B/L, big-growth; B/M, big-neutral; B/H, big-value. The monthly returns of these six portfolios
in the following year are used to construct size and book-to-market factors.
Table 3 Descriptive statistics for real estate factors
RERMRF RESMB REHML RMRF SMB HML
Mean 0.721 0.488 0.324 0.430 0.386 0.428
Median 0.653 0.703 ⫺0.106 1.415 0.310 0.625
Maximum 10.005 6.658 13.847 8.180 22.090 13.740
Minimum ⫺11.735 ⫺8.500 ⫺13.564 ⫺16.200 ⫺16.780 ⫺13.200
SD 3.769 2.706 4.030 5.306 4.958 4.587
Skewness ⫺0.406 ⫺0.554 ⫺0.234 ⫺0.609 0.691 ⫺0.009
Kurtosis 3.676 4.076 6.035 2.881 7.502 4.000
Note: Descriptive statistics for market, size, and book-to-market factors. RMRF, SMB, and HML are Fama-
French excess market return, small minus big, and high minus low return series. RERMRF, RESMB, and
REHML are excess real estate market return, real estate small minus big, and real estate high minus low return
series. Sample period is from January 1997 through December 2004.
266 C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
conventional SMB is 0.310. The considerably larger small-stock premium among real estate
related securities implies a heightened market inefficiency in the sector because of firm size.
The source of the inefficiency could be related to the difficulties in analyzing the assets
and/or lack of analyst coverage, a recurring observation among published research in real
estate. The real estate securities value descriptor REHML, on the other hand, is smaller than
that of the Fama-French HML. The mean (median) of REHML is 0.324 (⫺0.106) whereas
the mean (median) of the conventional HML is 0.428 (0.625). The REHML has a higher
degree of negative skewness, causing the median negative. The large differences between
real estate descriptors (RESMB, REHML) and the Fama-French factors (SMB, HML)
confirm our earlier conjecture that investment styles of real estate mutual funds cannot be
appropriately described by conventional factors. Risk characteristics are quite different in the
real state sector.
4. Real Estate Mutual Fund Factor Exposures
At the end of each year over the study period the following model is estimated for each
real estate mutual fund that has a complete history of returns over the prior 24 months:
r
it
⫺r
ft
⫽
␣
i
⫹

1i
关r
remt
⫺r
ft
兴⫹

2i
RESMB
t
⫹

3i
REHML
t
⫹
it
(1)
r
it
is the return in month tfor fund i,r
ft
is the return on a one-month Treasury Bill, r
remt
is
the return on the value-weighted portfolio of real estate stocks with SIC code 15, 16, 17, 65,
and 6798, and RESMB
t
REHML
t
are the returns on zero-investment factor-mimicking
portfolios for size and book-to-market. The estimates of

2i
and

3i
measure fund i’s
orientation toward firm size and book-to-market. A positive (negative)

2i
means the fund is
oriented toward small (large) real estate securities. A positive (negative)

3i
means the fund
is tilted toward value (growth) real estate stocks.
We report the regression results in Table 4. Panel A gives the distribution of funds’ factor
loadings. Funds are assigned to quintile portfolios based on the estimated coefficients from
the model. Within each quintile, the equal-weighted average of the coefficients is calculated.
Then the weighted averages over all the years of the coefficients are reported in Panel A. For
comparison purpose, Panel B reports the loadings on the NAREIT index return and four of
our earlier created factor portfolios S/L (small-growth), S/H (small-value), B/L (big-growth),
and B/H (big-value). Regarding size orientation, the results show that most funds are tilted
(4 out of 5 quintiles have negative

2i
) toward large real estate stocks. A likely reason is that
real estate firms are on average relatively small (see Table 2) and institutional funds have
minimum size requirements regarding their investment targets. The NAREIT index has an
average sensitivity to the size factor of ⫺0.199. Only two quintiles of real estate funds have
sensitivities above this value.
With respect to the value-growth orientation, it appears that real estate funds on average
moderately favor growth over value real estate securities. Three quintiles have negative
coefficients of

3
.On the other hand, the NAREIT index has an average sensitivity to the
REHML factor of 0.080. The tilt towards growth real estate stocks among mutual funds has
been documented frequently in academic research. Some argue it could be because of
267C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
investment strategies, others suggest it is influenced by personal interests of the fund
manager. In summary, we find real estate funds tilt toward large real estate stocks and
moderately favor growth over value real estate securities. In results not detailed here, we find
real estate funds are oriented toward small-value if we applied the conventional Fama-French
factors in the regression. This proves that conventional Fama-French factors could lead to a
biased analysis for real estate mutual funds.
5. Fund Style and Fund Performance
A major focus in the study of mutual funds is that of the fund performance. Starting with
Jensen (1968), many studies claim that the net return provided by the average actively
managed mutual fund is inferior to that of a comparable passive benchmark. Conflicting
results, however, have been reported in the eighties and nineties. For example, Coggin and
Trzcinka (1997) and Davis (2001) find growth-oriented funds are associated with higher
alphas. Chen, Jegadeesh and Wermers (2000) also report that growth-oriented funds exhibit
better stock selection skills than income-oriented funds. Grinblatt and Titman (1989) find
that the risk-adjusted gross returns of growth and aggressive growth funds are significantly
positive.
Table 5 provides estimates of alphas and loadings for portfolios of real estate mutual funds
that are sorted by size and book-to-market characteristics. At the end of each year over the
Table 4 Distribution of estimated factor loadings for mutual funds
Panel A: Distribution of factor loadings for real estate mutual funds
Loading on 1 (Low) 2345(high)
Market 0.639 0.752 0.792 0.838 1.020
Size ⫺0.398 ⫺0.308 ⫺0.238 ⫺0.149 0.235
Book-to-market ⫺0.160 ⫺0.057 ⫺0.016 0.039 0.161
Panel B: Distribution of factor loadings for benchmark portfolios
Loading on S/L S/H B/L B/H NAREIT
Market 0.975 1.017 1.035 0.992 0.871
Size 1.198 0.972 ⫺0.019 0.207 ⫺0.199
Book-to-market ⫺0.643 0.697 ⫺0.165 0.495 0.080
Note: At the end of each year over the study period the following model is estimated for funds with a complete
history of monthly returns over the prior 24 months: r
it
⫺r
ft
⫽
␣
i
⫹

1i
关r
remt
⫺r
ft
兴⫹

2i
RESMB
t
⫹

3i
REHML
t
⫹
it
.r
it
is the return in month tfor fund i,r
ft
is the return on a one-month Treasury Bill, r
remt
is
the return on the value-weighted portfolio of stocks with SIC codes 15, 16, 17, 65, and 6798, and RESMB
t
REHML
t
are the returns on zero-investment factor-mimicking portfolios for real estate size and book-to-market,
respectively. Funds are assigned to quintile portfolios based on the estimated coefficients from the model and the
equal-weighted average coefficient across funds within a quintile is calculated. The numbers reported in Panel A
are the weighted average across years, where the weights are the number of fund observations available in that
year. Panel B reports the loadings on the NAREIT index return and our six factor portfolio returns: S/L,
small-growth; S/H, small-value; B/L, big-growth; B/H, big-value. Sample period is from January 1997 through
December 2004.
268 C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
study period all real estate funds are sorted by their value-weighted average size rank and
book-to-market characteristics and assigned to one of four portfolios. For each of the
resulting portfolios, equally weighted returns are calculated over the subsequent 12 months,
and the process is repeated. At the end of the sample period, our regression model is applied
to the complete history of returns on each portfolio. Results in Table 5 shows that growth-
oriented real estate fund managers perform better than value-oriented managers on a style-
adjusted basis. The difference between the alphas of growth and value managers for large
caps is 0.194% per month (2.33% per year). The difference between the alphas of growth and
value managers for small caps is 0.131% per months (1.57% per year). This finding is
consistent with those of Coggin and Trzcinka (1997) and Chen et al. (2000). It is also
consistent with Damodaran and Liu (1993) and Kallberg et al. (2000) that money managers
investing in the real estate sector could produce positive abnormal returns because of their
specific appraisal skills and information. Value managers of real estate funds have on average
either a negative alpha or an alpha that is near zero.
6. Style Shifts, Past Performance, and Market Timing
It is understandable that a fund manager may shift his investment style if past performance
has been less than satisfactory. Peer pressure and remuneration concerns frequently provide
Table 5 Real estate mutual fund performance (percent per month) and loadings from three-factor models,
classified by style
Rank on size Loading on Rank on book-to-market
Value Growth
Large cap Constant ⫺0.212 ⫺0.018
Market 0.865*** 0.813***
Size ⫺0.133** ⫺0.175***
Book-to-market 0.081* 0.016
Adjusted
R-squared 0.862 0.858
Small cap Constant 0.016 0.147
Market 0.775*** 0.782***
Size ⫺0.311*** ⫺0.322***
Book-to-market 0.013 0.014
Adjusted
R-squared 0.814 0.819
Note: At the end of each year the following model is estimated for funds with a complete history of monthly
returns over the prior 24 months: r
it
⫺r
ft
⫽
␣
i
⫹

1i
关r
remt
⫺r
ft
兴⫹

2i
RESMB
t
⫹

3i
REHML
t
⫹
it
.r
it
is the
return in month tfor fund i,r
ft
is the return on a one-month Treasury Bill, r
remt
is the return on the value-weighted
portfolio of stocks with SIC codes 15, 16, 17, 65 and 6798, and RESMB
t
REHML
t
are the returns on
zero-investment factor-mimicking portfolios for real estate size and book-to-market, respectively. Funds are
assigned to one of four portfolios by their factor loading’s rank on size and book-to-market. The median is used
to classify large and small, value and growth portfolios. Sample period is from January 1997 through December
2004.
***Significant at the 1% level; **significant at the 5% level; *significant at the 10% level.
269C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
the needed impetus. To investigate the occurrence of shifts in investment style among real
estate funds, we sort funds into portfolios based on a two-way within-group classification.
The first sort is by a fund’s past performance (the compounded return on the fund over the
past two years), and the second sort (in two iterations) by fund size and book to market value.
In the classification by past fund return, we classify the top 25% of funds with the highest
past return as winners; and the bottom 25% as losers. Regarding size, the large caps include
the top third of funds and the small caps are those among the bottom third. Similarly, value
(growth) funds are those in the top (bottom) third regarding book-to-market. The average
across all portfolio formation years is reported in Table 6. We then compare each group’s
current style with its future style in the subsequent year.
In Table 6, with the real estate mutual funds classified according to size (large cap vs.
small caps), results in Panel A show that the mean absolute difference between past and
future RESMB ranks across all categories are quite small and comparable. Only the winners
among small caps have a mean absolute difference larger than 0.1. In other words, there is
no clear pattern regarding the shift between large and small stocks among the real estate
funds. On the other hand, the mean absolute difference with respect to book-to-market
(REHML) is large across all the categories of funds. The mean absolute difference for style
ranks with respect to book-to-market is 0.433 (0.250) for large-cap (small-cap) funds with
good past performance, compared with 0.366 (0.426) for large-cap (small-cap) funds with
poor past performance. The losers, as a whole group, have higher mean absolute difference
Table 6 Style shifts and past returns
Panel A: Fund style shifts and past performance: size classified
Past 24
month
return
Past
SMB
rank
Future
SMB
rank
Mean
absolute
difference
Past
HML
rank
Future
HML
rank
Mean
absolute
difference
Winners Large cap 3.283 0.126 0.143 0.072 0.250 0.471 0.443
Small cap 3.054 0.424 0.281 0.148 0.511 0.589 0.250
Losers Large cap ⫺15.654 0.125 0.209 0.084 0.330 0.541 0.366
Small cap ⫺9.822 0.293 0.258 0.045 0.281 0.586 0.426
Panel B: Fund style shifts and past performance: book-to-market classified
Past 24
month
return
Past
SMB
rank
Future
SMB
rank
Mean
absolute
difference
Past
HML
rank
Future
HML
rank
Mean
absolute
difference
Winners Value 3.963 0.359 0.255 0.110 0.509 0.612 0.258
Growth 1.841 0.171 0.193 0.080 0.199 0.467 0.356
Losers Value ⫺11.402 0.273 0.250 0.072 0.417 0.609 0.359
Growth ⫺13.115 0.211 0.218 0.082 0.189 0.506 0.438
Note: At the end of each year every fund with available data is sorted by past 24-month return. The top 25%
performance funds are classified as winners, and the bottom 25% funds are classified as losers. These funds are
classified as large cap versus small cap in Panel A (or value vs. growth in Panel B) at the same time by their
loadings on the size factor and book-to-market factor. For each of these two by three portfolios, the simple
average of the loadings on size or book-to-market factor is calculated. Data reported in Panel A and B is the
average across all portfolio formation years that are weighted by the number of funds in each year. Sample period
is from January 1997 through December 2004.
270 C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
with respect to book-to-market than the winners. That is, losers are more likely to shift their
investment style. Another observation is that small-cap funds with poor past performance
have more pronounced shifts in investment style regarding book-to-market.
When funds are classified according to book-to-market (value vs. growth), Panel B shows
again notable shifts in investment style among the losers. Specifically, the losers (both the
value and growth funds) have higher mean absolute difference for style ranks with respect
to book-to-market than the winners. The mean absolute difference for style ranks with
respect to book-to-market is 0.359 (0.438) for value (growth) funds with poor past perfor-
mance, compared with 0.258 (0.356) for value (growth) funds with good past performance.
The shift in style ranks with respect to size (RESMB), however, is small and comparable
between losers and winners. In summary, the results in Table 6 show that when performance
has been poor, the fund manager is likely to make a change regarding the investment
strategy. The shift could have been a temporary attempt to cover earlier losses, or it could
be a rotation simply because of the cyclical nature of the real estate industry. However, from
an investor’s perspective, such style changes by poorly performing funds represent disrup-
tions to the investor’s overall portfolio structure.
A change in investment style may represent a manager’s attempt to take advantage of
short-term market movements. Extant literature, however, reports little evidence of market
timing by mutual funds (e.g., Connor & Korajczyk, 1991; Ferson & Schadt, 1996; Chan et
al., 2002).
To investigate market timing by real estate funds, we follow Henriksson and Merton
(1981) in using the following regression: (2)
r
pt
⫺r
ft
⫽
␣
p
⫹

1p
关r
remt
⫺r
ft
兴⫹

2p
RESMB
t
⫹

3p
REHML
t
⫹

4p
max共0,r
remt
⫺r
ft
兲
⫹

5p
max共0,RESMB
t
兲⫹

6p
max共0,REHML
t
兲⫹
pt
(2)
Coefficients

4
,

5
, and

6
estimate a fund’s market timing effects. Based on results in Table
7, we can say that real estate funds do not have market-timing ability given that all the
Table 7 Style timing
Independent variables
Constant 关r
remt
⫺r
ft
兴RESMB REHML 关r
remt
⫺r
ft
兴⫹ RESMB⫹REHML⫹Adjusted R
2
⫺1.001 *** 0.802 *** ⫺0.219 0.075 0.123 0.137 ⫺0.132 0.840
(⫺3.023) (9.532) (⫺1.632) (0.934) (0.886) (0.592) (⫺0.940)
Note: We formed an equal-weighted portfolio for all real estate mutual funds with available data in each month
over the study period. The following regression is estimated:
r
pt
⫺r
ft
⫽
␣
p
⫹

1p
关r
remt
⫺r
ft
兴⫹

2p
RESMB
t
⫹

3p
REHML
t
⫹

4p
max(0,r
remt
⫺r
ft
)⫹

5p
max(0,RESMB
t
)⫹

6p
max(0,REHML
t
)⫹
pt
.
r
pt
is the return in month tfor portfolio p,r
ft
is the return on a one-month Treasury Bill, r
remt
is the return on
the value-weighted portfolio of stocks with SIC codes 15, 16, 17, 65, and 6798, and RESMB
t
REHML
t
are the
returns on zero-investment factor-mimicking portfolios for real estate size and book-to-market, respectively.
关r
remt
⫺r
ft
兴⫹, RESMB⫹, and REHML⫹are defined to be max (0, 关r
remt
⫺r
ft
兴), max (0, RESMB), and max (0,
REHML), respectively. t-statistics are reported in parentheses. Sample period is from January 1997 through
December 2004.
271C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
coefficients of

4
,

5
, and

6
are insignificant. This finding is consistent with results in
existing literature.
7. Conclusions
One of the most important developments in active equity management in the last decade
has been the creation of portfolio strategies based on value- and growth-oriented investment
styles. It is now common for money mangers to describe themselves as “value stock
mangers” or “growth stock managers” when selling their services to clients. Academic study
of style investing, however, has been at its incipient stages only. In this study, we investigate
the investment styles of real estate mutual funds. We create two real estate related style
descriptors, RESMB (size) and REHML (book-to-market), to measure a fund’s inclination
toward large versus small and value versus growth stocks. We create these style descriptors
from real estate related stocks because we find the risk characteristics of real estate securities
different from those of the NYSE-AMEX-NASDAQ universe. Existing literature reports that
most mutual funds adopt styles that bunch around an overall market index, with few funds
taking extreme positions away from the index. Relative to the real estate market portfolio
(NAREIT index), we find real estate funds on average tilt toward large real estate stocks and
moderately favor growth over value real estate securities. We also find growth stock
managers outperform value stock managers by 1.51% to 2.30% a year among real estate
mutual funds. Growth stocks generally have a favorable history of past returns and hence
may appear to be safer choices as far as managers’ personal career risks are concerned. We
also find evidence of shifts in investment style, especially among the losers. Taken together
real estate funds do not appear to be able to time the style factors. Our analysis of real estate
mutual fund investment styles provides insights regarding the kind of product offered. Our
findings are useful for evaluating real estate mutual fund performance and also for control-
ling the risk of the investor’s overall portfolio.
Notes
1. The 2004 year-end figure (source: Investment Company Institute, Washington, DC).
2. Source: Investment Company Institute, Washington, DC.
3. We have also performed analyses based on a 2X2 portfolio classification. Results are
similar in general.
References
Asness, C., Liew, J., & Stevens, R.(1997). Parallels between the cross-sectional predictability of stock and
country returns. Journal of Portfolio Management,23, 79 – 87.
Barberis, N., & Shleifer, A.(2003). Style investing. Journal of Financial Economics,68, 161–199.
Brown, J., & Goetzmann, W.(1997). Mutual fund styles. Journal of Financial Economics,43, 373–399.
Carhart, M.(1997). On persistence in mutual fund performance. Journal of Finance,52, 57– 82.
272 C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
Chan, L., Chen, H., & Lakonishok, J.(2002). On mutual fund investment styles. Review of Financial Studies,15,
1407–1437.
Chen, H., & De Bondt, W.(2004). Style momentum within the S&P-500 index. Journal of Empirical Finance,11,
483–507.
Chen, H., Jegadeesh, N., & Wermers, R.(2000). The value of active mutual fund management: an examination
of the stockholdings and trades of fund managers. Journal of Financial and Quantitative Analysis,35,
343–368.
Coggin, T., & Trzcinka, C. (1997). Analyzing the performance of equity managers: a note on value versus growth.
In T.D. Coggins, FJ. Fabozzi, and R.D. Arnotts (Eds.), The handbook of equity style management (pp.
167–170). New Hope, PA: John Wiley.
Connor, G., & Korajczyk, R.(1991). The attributes, behavior and performance of U.S. mutual funds. Review of
Quantitative Finance and Accounting,1, 5–26.
Damodaran, A., & Liu, C.(1993). Insider trading as a signal of private information. Review of Financial Studies,
6, 79 –119.
Davis, J.(2001). (2001). Mutual fund performance and manager style. Financial Analysts Journal,57, 19 –27.
Detzel, L.(2006). Determining a mutual fund’s equity class. Financial Services Review,15, 199 –212.
Fama, E., & French, K.(1992). The cross-section of expected stock returns. Journal of Finance, 47, 427– 465.
Fama, E., & French, K.(1993). Common risk factors in the returns on stock and bonds. Journal of Financial
Economics,33, 3–56.
Ferson, W., & Schadt, R.(1996). Measuring fund strategy and performance in changing economic conditions.
Journal of Finance,51, 425– 461.
Gallo, J., Lockwood, L., & Rutherford, R.(2000). Asset allocation and the performance of real estate mutual
funds. Real Estate Economics,28, 165–184.
Goetzmann, W., & Ibbotson, R.(1990). The performance of real estate as an asset class. Journal of Applied
Corporate Finance,3, 65–76.
Grinblatt, M., & Titman, S.(1989). Mutual fund performance: an analysis of quarterly portfolio holdings. Journal
of Business,62, 393– 416.
Henriksson, R., & Merton, R.(1981). On market timing and investment performance. Journal of Business,54,
513–533.
Investment Companies Yearbook, various issues. Rockville, MD: CDA/Wiesenberger, Inc.
Jensen, M.(1968). The performance of mutual funds in the period 1945–1964. Journal of Finance,23, 389 – 416.
Kadiyala, P.(2004). Asset allocation decisions of mutual fund investors. Financial Services Review,13, 285–302.
Kallberg, J., Liu, C., & Trzcinka, C.(2000). The value added from investment managers: an examination of funds
of REITs. Journal of Financial and Quantitative Analysis,35, 387– 408.
Lewellen, J.(2002). Momentum and autocorrelation in stock returns, Review of Financial Studies,15, 533–563.
Lin, C., & Yung, K.(2004). Real estate mutual funds: performance and persistence. Journal of Real Estate
Research,26, 69 –93.
Moskowitz, T., & Grinblatt, M. (1999). Do industries explain momentum? Journal of Finance,54, 1249 –1290.
O’Neal, E., & Page, D.(2000). Real estate mutual funds: abnormal performance and fund characteristics. Journal
of Real Estate Portfolio Management,6, 239 –247.
Reilly, F., & Brown, B. (2000). Investment analysis and portfolio management. Mason, OH: Thomson South-
Western.
Siegel, J. (1998). Stocks for the long run. New York: McGraw-Hill.
Wang, K., Erickson, J., Gau, G., & Chan, S.(1995). Market microstructure and real estate returns. Real Estate
Economics,23, 85–100.
273C.Y. Lin, K. Yung / Financial Services Review 16 (2007) 261–273
