The current issue and full text archive of this journal is available at www.emeraldinsight.com/1753-9269.

htm

JERER 2,2

The performance of Italian real estate mutual funds

Giacomo Morri
Accounting, Control, Corporate and Real Estate Finance Department, SDA Bocconi School of Management, Bocconi University, Milan, Italy, and

170
Received March 2009 Accepted May 2009

Stephen L. Lee
Faculty of Finance, Cass Business School, City of London University, London, UK
Abstract
Purpose Italian real estate mutual funds have shown enormous growth over the past few years, however, little is known about their performance. The purpose of this paper is to correct this oversight. Design/methodology/approach The risk-adjusted performance, as measured by the Sharpe ratio, of 17 Italian real estate funds is valued using a number of fund characteristics and monthly data over the period 2005-2008. Two models are constructed and ordinary least squares regressions are applied. Findings Active property management, the level of property-type diversication and the way the fund is initially setup (either by subscription or by contribution) are found to be signicant factors in differentiating the performance between Italian real estate funds. Research limitations/implications The relatively short period of time (three years) used in the empirical analysis might represent its major drawback, also considering that data cover only the upward trend of the market cycle: a further research should be addressed when a longer time series is available. Practical implications Results are of interest to investors and nancial planners alike in revealing which factors should be considered in selecting Italian real estate funds. Originality/value The main contribution of the paper is to study those characteristics of Italian real estate funds which have power in explaining their performance. The area of research is well known, the sample is new. Keywords Real estate, Investment funds, Financial performance, Property management, Italy Paper type Research paper

Journal of European Real Estate Research Vol. 2 No. 2, 2009 pp. 170-185 q Emerald Group Publishing Limited 1753-9269 DOI 10.1108/17539260910978472

1. Introduction The academic literature on the topic of real estate fund performance can be broken down into three general areas. The rst area of interest is whether fund managers possess any market-timing or stock-picking skills. Little evidence in the real estate market supports the notion that they exhibit such skills (see Lee and Stevenson, 2003 for a review). The second group of research tests the issue of persistence of performance. This literature generally concludes that individual real estate returns and real estate funds are persistent (see Devaney et al., 2007 for a review). The third area of research examines the characteristics of real estate investment trusts (REITs), real estate operating companies (REOCs) and real estate mutual funds (REMFs) which have power in explaining performance (Howe and Shilling, 1990; Redman and Manakyan, 1995; Allen et al., 2000; ONeal and Page, 2000; Gallo et al., 2000; Kalberg et al., 2000;

Lin and Yung, 2004; Mooradian and Yang, 2001; Delcoure and Dickinson, 2004; Ooi and Liow, 2004; Gullett and Redman, 2005; Philpot and Peterson, 2006). However, these studies are almost all based on US data with no study examining the characteristics of Italian real estate mutual funds (IREMFs) that inuence returns. This paper corrects this oversight and considers the performance of Italian real estate funds, a number of fund characteristics and monthly data over the period 2005-2008. In particular, we examine a number of real estate fund characteristics under the control of the portfolio manager, such as the extent of geographical and property-type diversication, which could be used by IREMF managers to achieve higher returns. If the factors under the control of the portfolio manager are signicant in affecting return, then the portfolio decisions of the fund managers can add value to the fund and benet the investors. Otherwise, active portfolio management will have limited benet for investors, and the inuential factors will be related to the nature of the fund (such as size and expenses). As such, our results will be of interest to investors and nancial planners, in revealing which factors should be considered in selecting IREMFs. The rest of the paper is structured as follows. Section 2 discusses the characteristics and the growth of IREMFs. Section 3 outlines the characteristics that drive performance of funds that have been identied in previous studies. The data and results are presented in Section 4. Section 5 concludes the study. 2. Italian real estate mutual funds IREMFs are tax exempt closed-ended funds that invest predominantly in real estate and equity interests in real estate companies, which must represent at least two-thirds of the asset value within 24 months from the setup. IREMFs are not legal entities, but rather pools of investments managed by an ` di Gestione del Risparmio SGR) authorized savings management company (Societa on behalf of and in the interest of the unit holders. The term of each fund must be consistent with the nature of its investment activity and the maximum term for IREMFs was set by the legislator up to 30 years, but a three-year period of grace can be requested in case there are problems in divesting. The IREMF legal regime was rst introduced in 1994, but it was not until the end of 2003 that the IREMF market gathered the current legislative, mainly new scal rules, framework thanks to substantial amendments to the existing tax regime with the purpose of making it more advantageous both for domestic and foreign investors. REMFs are regulated by provisions contained in several statutes and regulations issued by the Bank of Italy, the Ministry for Economic Affairs (Ministero dellEconomia e delle Finanze) and the Italian authority for nancial markets (Commissione Nazionale ` e la Borsa); the control of the operating activity, including the calculation per le Societa of the net asset value (NAV), is conducted by a custodian bank. IREMFs are setup by either subscriptions, so-called blind pool funds (BPFs), which raise cash at the time of their inception and later invest in assets or by contributions also called seeded funds (SFs) that are setup by contributing an existing portfolio of properties, either in whole or in part for cash. BPFs need time to invest the money raised and to reach the optimal asset allocation, so historically for some quarters they are mainly invested in liquid assets and have used little debt, due to the need to complete their real estate investments. In contrast, SFs are already completely invested and nanced and so should provide higher performance from the start of the fund.

Italian real estate mutual funds

171

JERER 2,2

172

Owing to the fear that some private company could use SFs for the disposal of bad portfolios, private contribution funds have only been allowed only since 2003; before this contribution funds were reserved to public entities for the purpose of the disposal of the huge public property portfolio. In order to avoid excessive risk due to concentration and to promote effective diversication strategies, IREMFs are not permitted to invest more than one-third of the assets in anyone property or more than 10 per cent of the assets in the shares issued by the same company; to engage directly in the building activity is also forbidden, but development activity using external building companies is allowed. The funds NAV has to be calculated at least every six months with all properties valued by external independent appraisers. IREMFs are allowed to use debt up to a maximum of 60 per cent of the value of the real estate assets, real estate rights and interests in real estate companies and up to 20 per cent of other assets. As at the end of 2008 leverage represented 72 per cent of the total allowed level of debt. The management rules (regolamento) are setup by the managing company in order to set forth the funds operational rules: they contain the rights and the obligations of the funds investors and of the managing company, the distribution of earnings, the procedures for the funds winding up and the subscription and redemption of the funds units and the investment strategy. Listed IREMFs reect an expected IRR from 5 to 8 per cent. The scal framework has changed several times since the establishment of IREMFs and they are subject to special tax rules, the main set of which was introduced in 2001, and to which substantial amendments were subsequently introduced in 2004: IREMFs are now tax exempt and, therefore, they are not subject to the Italian corporate tax on income and to the Italian regional tax on productive activities. Nonetheless, upon redemption of the fund units the managing company withholds 20 per cent in dividends and on capital gains so that an individual investor pays no additional tax and corporations a tax credit. Prots and capital gains are taxed according to the scal status of the unit holders: private investors (individuals) do not pay other taxes, other companies pay taxes on general income and foreign investors do not pay taxes in Italy. However, IREMFs are not the same as REITs, called SIIQs in Italy, which are companies and are internally managed; IREMFS are mutual funds and are externally managed. There are several typologies of IREMFs: speculative funds with no constraints on debt and on other issues, reserved funds, which are only available to qualied investors, such as investment rms, banks, stockbrokers, savings management companies, open-end investment companies, pension funds, insurance companies are allowed to concentrate risk and to operate in conicts of interests with the SGR, and retail funds, which are aimed at for individual investors and subject to more limitations on investment and are obliged to be listed on the exchange. For instance, in order to guarantee retail investors the opportunity to liquidate their investment, the law requires that if the minimum unit is worth less than e25,000, an application for the listing on the stock exchange must be led within 24 months after the closing of the initial offer. At the end of 2008, there are 135 REMFs with total assets under management (AUM) of about e20.2 billion (Figure 1). The 25 retail funds representing just a fth of the total number of funds had more than 31 per cent (e6.4 billion) of total AUM, while

25

160 140

Italian real estate mutual funds

20 AUM 15 Bn Number of funds 120 100 80 10 60 40 5 20 0 Dec-01 Dec-02 Source: Assogestioni Dec-03 Dec-04 Dec-05 Dec-06 Dec-07 Dec-08 0

173

Figure 1. Total assets value and number of funds

reserved funds represented 81.5 per cent (110) by number had only 69 per cent (e13.8 billion) of total AUM. Of the 25 retail funds, 22 were listed on the Italian stock exchange (Borsa Italiana) at year end 2008 in the division dedicated to the closed end funds (Mercato Telematico dei Fondi). These funds have traditionally been trading at a signicant discount from their net asset value (Table I). The number of listed funds has constantly increased over time: with three to four funds seeking listing in every year from 2001 to 2006 (Figure 2). The growth of the listed funds is conrmed also by the market capitalization which was almost e6 billion in July 2007. So far, all existing funds have invested e31 billions, 87.3 per cent in real estate assets, 2.3 per cent in real estate companies; 6.9 per cent in nancial instruments, bank deposits and cash and the remaining 3.5 per cent in other activities. In other words, almost 90 per cent of IREMFs investments are in direct and indirect real estate, which suggest that their performance will be predominantly driven by how successful fund managers are at picking and managing such assets. 3. Fund characteristics There is a large body of literature where academics claim that different fund characteristics are useful devices in selecting either the top-performing funds or eliminating the worst. These include: risk, size, expenses, age, turnover, cash-ow, fund manager tenure, management structure and style (Peterson et al., 2001). Owing to data limitations we only consider risk, expenses, size, age and portfolio diversication in this study. Risk It is impossible to avoid risk when investing in mutual funds and as one might expect, the reported relationship between returns and risk is positive. However, academics have not come to an agreement on how to measure risk. The two most common ways of

JERER 2,2

Fund Valore Immobiliare Globale Securfond Unicredito Immobiliare Uno Bnl Portfolio Immobiliare Polis Piramide Globale Alpha CAAM RE Italia CAAM RE Europa Portfolio Immobiliare Crescita Investietico IImmobilium 2001 Tecla-FondoUfci Estense-Grande Distribuzione Invest Real Security Caravaggio Obelisco Olinda-Fondo Shops Beta Atlantic 2-Berenice Europa Immobiliare 1 Atlantic 1

Market cap (e) 115,687,500 143,929,680 243,854,560 182,523,346 130,109,400 45,096,225 202,556,250 91,528,263 80,610,350 187,878,000 87,101,965 117,086,632 203,077,930 67,003,200 180,604,876 133,821,600 256,049,640 101,037,618 151,411,030 268,237,221 67,148,800 178,858,309

NAV (e) 182,294,798.00 194,199,041.00 565,225,345.00 370,563,234.00 315,886,013.00 82,260,729.00 401,285,014.00 219,854,147.00 222,033,646.00 262,755,214.00 183,364,972.00 146,988,683.00 244,976,162,.00 155,195,123.00 316,366,410.00 162,950,471.00 422,189,846.00 278,128,199.00 362,832,501.00 439,007,087.00 174,844,583.00 417,847,780.00

Adj. NAV (e) 182,294,798.00 194,199,041.00 565,225,345.00 370,563,234.00 315,886,013.00 82,260,729.00 395,654,989.00 219,854,147.00 217,536,437.00 262,755,214.00 183,364,972.00 146,988,683.00 244,976,162,.00 155,195,123.00 176,872,689.08 162,950,471.00 397,399,585.40 278,128,199.00 350,448,122.96 400,396,893.95 174,844,583.00 410,416,120.00

Discount to NAV (%) 2 36.54 2 25.89 2 56.86 2 50.74 2 58.81 2 45.18 2 48.80 2 58.37 2 62.94 2 28.50 2 52.50 2 20.34 2 17.10 2 56.83 2.11 2 17.88 2 35.57 2 63.67 2 56.80 2 33.01 2 61.60 2 56.42

174

Table I. Listed funds description

6 Average annual market cap 5 Number of listed funds 4

25

20

15 Bn 3 10 2 5

Figure 2. Listed funds: number and market capitalisation

0 2001 2002 2003 2004 2005 2006 2007 2008

measuring risk in a fund is to calculate its beta or standard deviation. Beta is a measure of the systematic risk of a portfolio compared to the market. The standard deviation of a fund measures the risk by measuring the degree to which the fund uctuates in relation to its mean return over a period of time and includes both systematic and unsystematic risk.

The most appropriate measure depends on the investment assumption. If the mutual fund represents the entire investment for an individual investor the standard deviation is a more complete measure. In contrast, if the investor invests in many different funds the beta measure is preferable. If an investor only invests in one mutual fund it can imply that he is not fully diversied and therefore is exposed to both systematic and unsystematic risk. In addition, the elimination of unsystematic risk within a real estate fund requires hundreds of properties (Investment Property Forum, 2007). For instance, Chen and Peiser (1999) nd that REIT portfolio average returns show no positive relationship with beta. As a consequence, Redman and Manakyan (1995) and Ooi and Liow (2004) suggest that real estate risk should be measured by standard deviation. Expenses There is a claim that fund managers charging higher fees are more skilled and recoup charges by providing higher investment returns (Droms and Walker, 1995; Golec, 1996). On the other hand, numerous studies show that there exists a signicant negative correlation between expense ratio and performance (Rahman et al., 1991; Grinblatt and Titman, 1993; Malkiel, 1995; Elton et al., 1996; Hooks, 1996; Carhart, 1997; Malhotra and McLeod, 1997; Walker, 1997; Phelps and Detzel, 1997; Dellva and Olson, 1998; Otten and Bams, 2002) in equity mutual funds and Blake et al. (1993) and Domian and Reichenstein (1997, 2002) for bond portfolios. Then again Ippolito (1989) and Dowen and Mann (2004) nd management fees to be unrelated to fund performance. Studies in the real estate market also show conicting results, using different denitions of expenses. ONeal and Page (2000) and Gullett and Redman (2005) nd that REMF performance is inversely related to expenses ratios, i.e. the higher the funds expenses the lower the performance. In contrast, Philpot and Peterson (2006) nd that REMFs who charge higher management fees tend to outperform lesser compensated managers. Then, again Myer et al. (1997) nd that performance rankings are not affected by asset management fees. While, Lin and Yung (2004) show that the fund performance is unrelated to the expense ratio. Additionally, the authors nd that the expense ratio is negatively correlated to fund net assets, which they argue implies that there are economies of scale in the REMF industry. In a similar vein, Philpot and Peterson (2006) nd that the systematic risk of REMFs is positively related to management fees. The authors arguing that this implies that funds that charge higher management fees have an incentive to generate higher returns. Size Large mutual funds have several advantages over small ones. First, big funds are able to spread xed overhead expenses over a larger asset base. Second, managers of big funds can gain positions in benecial investment opportunities not available to smaller market participants (Ciccotello and Grant, 1996). Additionally, large funds are able to accomplish trades at more favourable terms given their market positions and use more resources for research. All together, these institutional and cost advantages should lead to large funds outperforming small ones. However, being big also presents management challenges (Chen et al., 2004). As a big fund keeps on growing it has to continue to nd worthwhile investment opportunities. Thus, after initial growth, a fund that has grown too large may cause its

Italian real estate mutual funds

175

JERER 2,2

176

managers to deviate from its original objectives by investing in lower quality investments (which otherwise would not be considered when the fund is smaller) and increase administrative costs by additional coordination and hiring of staff to manage sub-funds, which suggests that fund performance increases at a declining rate as fund sizes increase (Indro et al., 1999; Edwards and Caglayan, 2001). Furthermore, even though small funds can move more quickly many small funds are less diversied, especially in the real estate market, implying that luck plays a greater role in the performance rankings. As a result, a small (undiversied) real estate fund could be at the top because the few properties they own just happened to do well, and those at the bottom are there because the few properties they own just happened to do badly. The literature on the relation between mutual fund size and performance produces conicting ndings. Clark (2003) provides a comprehensive study in the USA and concludes that no signicant return differences can be found between small and large mutual funds on a variety of holding periods from 1991 to 2001. Conclusions supported by Ciccotello and Grant (1996) and Droms and Walker (1994) in the USA; Bird et al. (1983), Gallagher (2003) and Gallagher and Martin (2003) in Australia and Dahlquist et al. (2000) in Sweden. In contrast, Malhotra and McLeod (1997), Indro et al. (1999), Otten and Bams (2002) and Sing (2007) found that larger funds, on average, provided investors with superior returns. While, Grinblatt and Titman (1994) nd that size is not related to fund performance. In the hedge fund industry, Gregoriou and Rouah (2003) nd no correlation between the fund size and performance. In contrast, Herzberg and Mozes (2003) nd that smaller hedge funds display a performance that is better compared with larger funds. Hedges (2003) shows that smaller funds outperform larger funds, but nds that mid-sized funds perform the worst. In the real estate literature, Howe and Shilling (1990) observed that rm size may partially explain REIT performance. Chen and Peiser (1999) nd that smaller REITs had higher returns than larger REITs but also a higher standard deviation. Lin and Yung (2004) nd that size is the most signicant variable in explaining the risk-adjusted performance of REMFs, with a positive relationship between log fund net assets and risk-adjusted returns. The authors suggest that, is implies that, larger funds are able to reap the benets of economies of scale or that large funds may have better research and are able to afford better portfolio managers. Kalberg et al. (2000) show a positive relationship between returns and REMF size. In contrast, Ambrose et al. (2000) nd no evidence that large REITs have higher net operating income growth rates. In a similar vein, Mooradian and Yang (2001) nd that asset size does not explain hotel REIT growth. Ooi and Liow (2004) and Gullett and Redman (2005) nd rm size did not have a statistically signicant impact on risk-adjusted returns of real estate rms and mutual funds. Additionally, Philpot and Peterson (2006) nd that systematic risk is unrelated to fund size. Fund age The age of a fund could play a role in deciding performance since younger funds may face signicant higher costs in their start up period. This is due to marketing costs but also that the initial cash ows will place a greater burden on the funds transaction costs. There is also evidence showing that returns of new mutual funds may be

affected by an investment learning period (Gregory et al., 1997). Furthermore, younger funds are usually small and so are undiversied and so exposed to higher market risk leading to poor performance (Bauer et al., 2002). Again the literature on the relation between age and performance produces conicting results. Gregory et al. (1997) shows that mature funds perform better than younger ones. In contrast, Otten and Bams (2002) showed that younger European mutual funds did better than mature ones. Finally, Peterson et al. (2001) found no relationship between mutual fund performance and fund age. In the real estate literature, ONeal and Page (2000) nd that performance is negatively related to age, i.e. newer funds delivered better performance than older funds over the sample period. Lin and Yung (2004) and Gullett and Redman (2005) show manager tenure, which can proxy for fund age, had no impact on the risk-adjusted performance of REMFs. Additionally, Lin and Yung (2004) nd that manager tenure was negatively correlated with expense ratio and turnover, which implies that more experienced managers trade less frequently and therefore lowers trading costs. Philpot and Peterson (2006) nd that the longer the manager tenure the greater the risk of the fund but that senior mangers do not provide high-abnormal returns, which implies that older funds have lower risk-adjusted performance than their younger counterparts. Regional and property-type diversication In the real estate market literature, additional variables have also been suggested which have an impact on performance including: development activity and regional and property-type concentration. For instance, Lin and Yung (2004) examined the performance of a sample of REMFs over the period 1993-2001 and nd that the performance is largely determined by the returns of the real estate sector rather than any nancial variables. Brounen et al. (2000) found that property-developing REITs outperformed the market benchmark, while non-developers did not signicantly outperform the index. However, after taking risk into account the authors discovered that property-developing REITs hardly outperform the benchmark index. In other words, property-developing REITs have greater systematic risk that results from developing property. A result conrmed by Delcoure and Dickinson (2004) who nd that REITs and REOCs with greater investment in undeveloped land had higher systematic risk. Ooi and Liow (2004) make a similar conclusion for Asian real estate rms. However, the authors also nd that exposure to real estate development, as measured by the percentage of vacant land and projects under constructions relatively to total assets of real estate rms, had no impact on the risk-adjusted returns. Diversication can limit a funds exposure to the uctuations of any particular sector or region but diversication comes with attendant costs which may outweigh any diversication benets gained from the strategy. For instance, Capozza and Seguin (1999) examine the impact that diversication has on REIT value. Their main conclusions were that diversication has a positive impact on property-level cash ows but an offsetting negative impact on general and administrative expenses. In the case of property-type diversication, Gyourko and Nelling (1996) found that the systematic risk of equity REITs vary by the type of real estate in which they invest, with beta being signicantly higher for retail-oriented REITs than for REITs owning

Italian real estate mutual funds

177

JERER 2,2

178

industrial and warehouse properties during the 1988-1992 sample period. Redman and Manakyan (1995) using data on REITs from 1986 through 1990 nd that ownership of healthcare properties positively affected the risk-adjusted return. Chen and Peiser (1999) nd that ofce and industrial REITs performed signicantly better than the other property sectors over the period from 1993 to 1997, which the authors attribute from bargain purchase opportunities in the early 1990s following the real estate market crash. Gallo et al. (2000) examined the asset allocation decisions of REMFs and nd that the allocation of fund assets across the property-types explains most of the abnormal performance. Delcoure and Dickinson (2004) using data from 1997 to 2002 nd that REITs with greater industrial property holdings and REOCs with greater residential investments have lower systematic risk. However, Anderson and Beneeld (2005) note that while property-type diversication has an impact on REIT performance the effect is sometimes positive and sometimes negative, i.e. the impact of property-type diversication on REIT returns varies over time. The impact of geographic concentration also impacts on risk-adjusted performance. Chen and Peiser (1999) observed that geographically concentrated REITs (with investments in only one state) showed signicantly higher returns but also signicantly higher standard deviations than geographically diversied REITs (with investments in four or more states). Howe and Shilling (1990) found that properties located in the northeast were positively associated with risk-adjusted performance ` -vis the West) primarily in the 1979-1987 period, coinciding with a period during (vis-a which house values rose at extraordinary rates. Redman and Manakyan (1995) nd that in the late 1980s REITs with properties in the western US had a greater risk-adjusted return, which they attributed to the expanding economies of California, Washington, and Oregon. Similarly, Delcoure and Dickinson (2004) nd that REITs with greater investments in the mid-east, west, north-central and pacic regions of the USA shows lower systematic risk while REITs and REOCs with greater holdings in the southeast had greater systematic risk. In contrast, Ambrose et al. (2000) found that diversication strategy by geographical regions has no signicant benet on REIT performance. While, Gyourko and Nelling (1996) nd that diversication by property-type and broad economic region is unrelated to stock market-based measures of diversication. Lastly, Boer et al. (2005) examine the impact of geographical and property-type specialisation on risk-adjusted returns for a sample of 275 listed property companies in the USA and Europe over the 17-year period from 1984 to 2002. Boer et al. (2005) nd that an increase in geographical or property-type focus leads to and increase in risk-adjusted returns. The authors also nd that unsystematic risk increases with both geographical and especially the property-type specialisation of the company. In contrast, geographical and property-type specialisation lowers property company systematic risk. In other words, geographical or property-type specialisation property companies tend to have the highest returns but the highest risk. In contrast, Gullett and Redman (2005) nd that the level of diversication in 119 REMFs, as measured by total number of holdings and the percentage of holdings in the top ten REITs, had not impact on risk-adjusted returns. Finally, Morri and Erbanni (2008) analysing a sample of 119 US reits found that on a risk-adjusted basis more concentrated REITs have better returns, especially when focusing on property-types rather than on geographic areas.

4. Data and results Closing prices of the individual units on a monthly basis were employed to derive the monthly returns of the individual real estate funds. Following ONeal and Page (2000), Philpot and Peterson (2006) and Gullett and Redman (2005), we used a three-year performance period (2005-2008), in order to limit the effects of survivorship bias. This sampling procedure yielded 17 Italian real estate funds. For each of the 17 funds we then calculated the funds Sharpe ratio, which is dened as the funds returns in excess of the risk-free rate[1], divided by the funds standard deviation. We also collected observations on a number of real estate fund characteristics which have been used in previous studies (Table II), namely: . LNNAV. Natural log of the net asset value. . EFFGES. Management efciency (calculated as fund expenses/assets). . EFFGI. Active property management (calculated as properties expenses/properties). . HFDGEO. Portfolio Herndahl index for property locations (ve areas, according to Assogestioni[2] classication). . HFDTIP. Portfolio Herndahl index for property typologies (eight typologies, according to Assogestioni classication). . AGE. Years since inception. . INCSGR. Management fees (at December end)/assets (calculated as the average of June end and December end assets). A measure of the costs of the SGR which manages the fund. . MATSGR. SGR group (0 bank, 1 real estate company). . BPSF. Fund setup typology (1 blind pool, 0 seed fund). Table III contains sample descriptive statistics and show a number of features of interest. First, the Italian real estate funds showed an average Sharpe ratio of 0.13, with only ve funds outperforming the market benchmark. Second, the mean logarithm of net assets (LNAV), management efciency (EEFGES), property management (EFFGI) and management fees (INCSGR) ratios are fairly consistent among the funds. Third, the geographical (HFDGEO) and property-type diversication (HFDTIP) variables suggest that the funds are generally not well-diversied. Fourth, the mean age (AGE) variable indicates the funds are quite young. Fifth, the setup dummy (BPSF) indicates that the majority of funds are BPFs. Lastly, on an individual basis only active property management (EEFGI) is signicantly correlated with the Sharpe ratio at the usual level of signicance. We use the following model to estimate the relationship between the Sharpe ratio and the fund characteristics by means of ordinary least squares (OLS): Sharpe a b1 LNAV b2 EFFGES b3 EFFGI b4 HFDGEO b5 HFDTIP b6 AGE b7 INCSGR b8 MATSGR b9 BPSF 1 where: a the regression intercept term; bs the regression coefcients; 1 the regression error term and the dependent and independent variables are dened above.

Italian real estate mutual funds

179

180

JERER 2,2

Fund 0.135 0.066 0.050 0.157 2 0.044 0.284 0.173 0.217 0.058 0.432 0.092 0.316 0.322 2 0.113 0.119 2 0.082 0.096 19.022 19.086 20.081 19.732 19.575 19.475 19.793 19.181 19.284 19.303 19.016 18.840 19.961 19.254 18.860 18.803 19.702 1.381 1.846 1.260 2.010 1.775 1.897 1.860 1.179 0.907 2.151 1.531 1.606 1.363 1.858 1.526 1.918 0.980 0.859 0.714 0.223 0.536 0.797 0.989 1.656 0.386 0.092 0.519 0.865 0.748 1.359 0.381 0.344 0.169 1.210 87.301 55.552 57.553 51.180 62.448 74.157 80.066 65.190 73.526 65.612 83.849 65.541 60.555 74.548 74.754 62.862 67.019 64.930 65.693 56.088 66.207 77.347 55.594 77.415 67.607 83.792 79.320 70.945 82.659 94.868 100.00 75.386 58.928 81.794 7.836 7.022 7.022 6.578 6.499 6.447 5.795 5.713 5.181 4.992 4.795 4.507 2.792 3.518 3.214 2.899 2.025

Valore Immobiliare Globale Securfondo Unicredito Immobiliare Uno Bnl Portfolio Immobiliare Polis Piramide Globale Fondo Alpha CAAM RE Italia CAAM RE Europa Portfolio Immobiliare Crescita Investietico Immobilium 2001 Tecla-Fondo Ufci Estense-Grande Distribuzione Invest Real Security Caravaggio Olinda-Fondo Shops

Table II. Funds statistics Sharpe ratio LNAV EFFGES EFFGI HFDGEO HFDTIP AGE INCSGR 1.284 1.458 1.154 1.713 1.668 1.793 1.486 1.107 0.834 1.686 1.336 1.350 1.267 1.563 1.372 1.722 0.800 MATSGR 0 1 0 0 0 0 0 0 0 0 0 1 1 0 1 1 1 BPSF 1 1 1 1 1 1 0 1 1 1 1 1 0 0 1 1 0

Variables Sharpe ratio LNAV EFFGES EFFGI HFDGEO HFDTIP AGE INCSGR MATSGR BPSF

Mean 0.13 19.35 1.59 0.70 68.34 74.03 5.11 1.39 0.35 0.76

Max. 0.43 20.08 2.15 1.66 87.30 100.00 7.84 1.79 1.00 1.00

Min. 2 0.11 18.80 0.91 0.09 51.18 55.59 2.02 0.80 0.00 0.00

Corr. 0.12 0.09 0.36 2 0.03 0.04 0.05 0.04 0.03 0.06

p 0.20 0.12 2.24 0.01 0.02 0.04 0.02 0.01 0.05

Italian real estate mutual funds

181
Table III. Summary statistics

However, with such a model, correlation among independent variables, or multicollinearity, is of primary concern. Multicollinearity may cause a severe distortion of independent variable coefcients and levels of signicance. There are several classical tests for detecting multicollinearity but the most common method is by means of variance ination factors (VIFs), which measure how much the variance of an estimated regression coefcient increases if the independent variables are correlated. As a rule of thumb, multicollinearity is a serious problem for an individual variable if its VIF is greater than 4. Model 1 presented in the left columns of Table IV, which includes all fund characteristics variables, has an average VIF of 6.9 with a high of 15.4, which indicates multicollinearity is a serious problem. The VIF for INCSGR, for example, tells us that the variance of the estimated coefcient of 2 0.47 is inated by a factor of 14.3 because INCSGR is highly correlated with at least one of the other predictors in the model. In addition, while the R 2 is 0.45 the adjusted R 2 is 0.00 indicating that although one or more of the independent variables is signicantly related to excess returns the other variables are detracting from the results. To overcome the problem of multicollinearity, we considered all possible regressions to eliminate those fund characteristics which are confounding the regression results.

Variable C LNAV EEFGES EFFGI HFDGEO HFDTIP AGE INCSGR MATSGR BPSF R2 Adjusted R 2

Coeff 2 2.547 0.133 0.379 0.284 0.000 0.004 2 0.039 2 0.467 2 0.035 0.243 0.45 0.00

Model 1 SE t-stat. 5.85 0.26 0.44 0.18 0.01 0.01 0.05 0.53 0.19 0.23 0.44 0.51 0.86 1.59 0.01 0.71 0.84 0.88 0.19 1.79

VIF 6.5 15.4 3.5 4.8 2.5 3.7 14.3 5.1 6.0

Coeff 2 0.290 0.243 0.004 0.243 0.31 0.15

Model 2 SE t-stat. 0.29 0.10 0.00 0.13 1.01 2.38 1.12 1.85

VIF 1.7 1.9 2.9

Table IV. Regression results Sharpe ratios vs fund characteristics

JERER 2,2

182

Using all possible regressions the nal columns of Table IV shows that three variables have a signicant positive impact on Sharpe performance of Italian real estate funds; namely property market activity (EFFGI), the Herndahl index for property typologies diversication (HFDTIP), and fund setup typology (BPSF). In addition, the VIFs indicate that multicollinearity is not a serious problem in Model 2. The positive relation between EFFGI and adjusted performance suggests that funds with higher expenses for managing properties show, somehow unexpectedly, a better results for investors: given that this variable can be considered at the same time as a proxy of active management on the properties and of efciency in the property management, a possible explanation is that the former prevails over the latter and that the more active managed properties are able to get more stable returns. A further explanation can be found in the strategy of the fund: for instance, a value-added fund may be able to achieve higher returns, even though such a strategy requires higher expenses for managing the properties. The HFDTIP shows that more concentrated portfolios, and so higher specialization on specic properties, increases the return more than the risk, suggesting that the lack of diversication is more than compensated by the better knowledge of the market for the specic destination of use. Finally, the funds setup by contribution, notwithstanding better returns, present higher volatility. At a rst glance, due to the higher absolute returns of SFs, this result can appear surprisingly, but the higher volatility of those returns is more than compensated. 5. Conclusions This study examines the risk-adjusted performance of 17 listed IREMFs and a number of fund characteristics using monthly data over the period 2005-2008. While this relatively short period of time on one hand allows to avoid the survivorship bias issue, on the other it might represent the major drawback of this analysis, also considering that in the period taken into account data could cover only the upward trend of the market cycle: further research should be address when a longer time series of data will be available. The results indicate that active property management (EFFGI), Herndahl index for property typology diversication (HFDTIP) and the way to setup the funds (BPSF) have a signicant inuence on the risk-adjusted performance. In other words, an investor willing to choose an Italian real estate fund that offers a greater risk-adjusted performance, should focus on funds where the portfolio is actively managed, more concentrated across the property-types and setup as blind pool fund.
Notes 1. Risk-free rate is calculated as the average return in the period on Italian treasury bill one year. 2. The Italian association of mutual funds. References Allen, M.T., Madura, J. and Springer, T. (2000), REIT characteristics and the sensitivity of REIT returns, Journal of Real Estate Finance and Economics, Vol. 21 No. 2, pp. 141-52.

Ambrose, B.W., Ehrlich, S.R., Hughes, W.T. and Wachter, S.M. (2000), REIT economies of scale: fact or ction?, Journal of Real Estate Finance and Economics, Vol. 20 No. 2, pp. 211-24. Anderson, R.I. and Beneeld, J.D. (2005), REIT property type diversication, worse than ve?, paper presented at the Financial Management Association Annual Meeting, Na Chicago, IL. Bauer, R., Koedijk, K. and Otten, R. (2002), International evidence on ethical mutual fund performance and investment style, working paper March 2002, Limburg Institute of Financial Economics, Maastricht University, Maastricht. Bird, R., Chin, H. and McCrae, M. (1983), The performance of Australian superannuation funds, Australian Journal of Management, Vol. 8, pp. 49-69. Blake, C.R., Elton, E.J. and Gruber, M.J. (1993), The performance of bond mutual funds, Journal of Business, Vol. 66 No. 3, pp. 371-403. Boer, D., Brounen, D. and Opt Veld, H. (2005), Corporate focus and stock performance: evidence from international listed property markets, Journal of Real Estate Finance and Economics, Vol. 31 No. 3, pp. 263-81. Brounen, D., Eichholtz, P.M.A. and Kanters, P.M. (2000), The effects of property development activities on the performance of REITs, Real Estate Finance, Vol. 16 No. 4, pp. 17-22. Capozza, R. and Seguin, P.J. (1999), Focus transparency and value: the REIT evidence, Real Estate Economics, Vol. 27, pp. 587-619. Carhart, M.M. (1997), On persistence in mutual fund performance, Journal of Finance, Vol. 52 No. 1, pp. 57-82. Chen, J. and Peiser, R. (1999), The risk and return characteristics of REITs: 1993-1997, Real Estate Finance, Vol. 16 No. 1, pp. 61-8. Chen, J., Hong, H., Huang, M. and Kubik, J. (2004), Does fund size erode mutual fund performance? The role of liquidity and organization, The American Economic Review, Vol. 94 No. 5, pp. 1276-302. Ciccotello, C.S. and Grant, T.C. (1996), Equity fund size and growth: implications for performance and selection, Financial Services Review, Vol. 5, pp. 1-12. Clark, A. (2003), Does fund size affect performance?, Lipper Research Study, Lipper, New York, NY. Dahlquist, M., Engstrom, S. and Soderlind, P. (2000), Performance and characteristics of Swedish mutual funds, The Journal of Financial and Quantitative Analysis, Vol. 35, pp. 409-23. Delcoure, N. and Dickinson, R. (2004), REIT and REOC systematic risk sensitivity, Journal of Real Estate Research, Vol. 26 No. 3, pp. 1-18. Dellva, W.L. and Olson, G.T. (1998), The relationship between mutual fund fees, and expenses and their effects on performance, The Financial Review, Vol. 33, pp. 85-104. Devaney, S., Lee, S.L. and Young, M. (2007), Serial persistence in individual real estate returns in the UK, Journal of Property Investment and Finance, Vol. 25 No. 3, pp. 241-73. Domian, D.L. and Reichenstein, W. (1997), Performance and persistence in money market fund returns, Financial Services Review, Vol. 6, pp. 169-83. Domian, D.L. and Reichenstein, W. (2002), Predicting municipal bond fund returns, Journal of Investing, Autumn, pp. 53-65. Dowen, R.J. and Mann, T. (2004), Mutual fund performance, management behavior and investor costs, Financial Services Review, Vol. 13, pp. 79-91. Droms, W.G. and Walker, D.A. (1994), Investment performance of international mutual funds, The Journal of Financial Research, Vol. 1 No. 17, pp. 1-14.

Italian real estate mutual funds

183

JERER 2,2

184

Droms, W.G. and Walker, D.A. (1995), Determinants of variation in mutual fund returns, Applied Financial Economics, Vol. 5, pp. 383-9. Edwards, F.R. and Caglayan, M. (2001), Hedge fund performance and manager skill, Journal of Futures Markets, Vol. 21 No. 11, pp. 1003-28. Elton, E.J., Gruber, M.J. and Blake, C.R. (1996), The persistence of risk-adjusted mutual fund performance, Journal of Business, Vol. 69, pp. 33-157. Gallagher, D.R. (2003), Investment manager characteristics, strategy, top management changes and fund performance, Accounting and Finance, Vol. 43 No. 3, pp. 283-309. Gallagher, D.R. and Martin, K.M. (2003), Size and investment performance: a research note, Abacus, Vol. 41, pp. 55-65. Gallo, J.G., Lockwood, L.J. and Rutherford, R.C. (2000), Asset allocation and the performance of real estate mutual funds, Real Estate Economics, Vol. 28 No. 1, pp. 165-84. Golec, J.H. (1996), The effects of mutual fund managers characteristics on their portfolio performance, risk and fees, Financial Services Review, Vol. 5, pp. 133-48. Gregoriou, G.N. and Rouah, F. (2003), Large versus small hedge funds: does size affect performance?, Journal of Alternative Investments, Vol. 5 No. 5, pp. 75-7. Gregory, A., Matatko, J. and Luther, R. (1997), Ethical unit trust nancial performance: small company effects and fund size effects, Journal of Business Finance & Accounting, Vol. 24, pp. 705-25. Grinblatt, M. and Titman, S. (1993), Performance measurement without benchmarks: an examination of mutual fund returns, Journal of Business, Vol. 66 No. 1, pp. 47-68. Grinblatt, M. and Titman, S. (1994), A study of monthly fund returns and performance evaluation techniques, Journal of Financial and Quantitative Analysis, Vol. 29 No. 3, pp. 419-44. Gullett, N.S. and Redman, A.L. (2005), Do real estate mutual funds enhance portfolio returns and reduce portfolio risk?, Briengs in Real Estate Finance, Vol. 5, pp. 51-66. Gyourko, J. and Nelling, E. (1996), Systematic risk and diversication in the equity REIT market, Real Estate Economics, Vol. 24 No. 4, pp. 493-515. Hedges, R.J. (2003), Size vs. performance in the hedge fund industry, Journal of Financial Transformation, Vol. 10, pp. 14-17. Herzberg, M.M. and Mozes, H.A. (2003), The persistence of hedge fund risk: evidence and implications for investors, Journal of Alternative Investments, Vol. 6 No. 2, pp. 22-42. Hooks, J.A. (1996), The effect of loads and expenses on open-end mutual fund returns, Journal of Business Research, Vol. 36, pp. 199-202. Howe, J.S. and Shilling, J.D. (1990), REIT advisor performance, AREUEA Journal, Vol. 18, pp. 479-500. Indro, D., Jiang, C., Hu, M. and Lee, W. (1999), Mutual fund performance: does fund size matter?, Financial Analysts Journal, Vol. 55, pp. 74-87. Investment Property Forum (2007), Risk Reduction and Diversication in Property Portfolios: Main Report, Investment Property Forum, London, May. Ippolito, R. (1989), Efciency with costly information: a study of mutual fund performance 1965-1984, The Quarterly Journal of Economics, Vol. 104, pp. 1-24. Kalberg, J.G., Liu, C.L. and Trzcinka, C. (2000), The value added from investment managers: an examination of funds of REITs, Journal of Financial and Quantitative Analysis, Vol. 35 No. 3, pp. 387-408. Lee, S.L. and Stevenson, S. (2003), Empirical evidence on the micro and macro forecasting ability of real estate funds, Journal of Property Research, Vol. 20 No. 3, pp. 207-34.

Lin, C.Y. and Yung, K. (2004), Real estate mutual funds: performance and persistence, Journal of Real Estate Research, Vol. 26 No. 1, pp. 69-94. Malhotra, D. and McLeod, R. (1997), An empirical analysis of mutual fund expenses, The Journal of Financial Research, Vol. 20 No. 2, pp. 175-90. Malkiel, B.G. (1995), Returns from investing in equity mutual funds 1971 to 1991, Journal of Finance, Vol. 50 No. 2, pp. 549-72. Mooradian, R.M. and Yang, S.X. (2001), Dividend policy and rm performance: hotel REITs vs non-REIT hotel companies, Journal of Real Estate Portfolio Management, Vol. 7 No. 1, pp. 79-87. Morri, G. and Erbanni, E. (2008), Diversicazione del rischio nella gestione di portafogli immobiliari: analisi della performance dei REITs americani, Finanza, Marketing e Produzione, Vol. 3, pp. 27-46. Myer, F.C.N., Webb, J.R. and He, L.T. (1997), Issues in measuring performance of commingled real estate funds, Journal of Real Estate Portfolio Management, Vol. 3 No. 2, pp. 79-86. ONeal, E.S. and Page, D.E. (2000), Real estate mutual funds: abnormal performance and fund characteristics, Journal of Real Estate Portfolio Management, Vol. 6 No. 3, pp. 239-47. Ooi, J.T.L. and Liow, K-H. (2004), Risk-adjusted performance of real estate stocks evidence from emerging markets in Asia, Journal of Real Estate Research, Vol. 26 No. 4, pp. 371-95. Otten, R. and Bams, D. (2002), European mutual fund performance, European Financial Management, Vol. 8 No. 1, pp. 75-101. Peterson, J.D., Pietranico, P.A., Riepe, M.W. and Xu, F. (2001), Explaining the performance of domestic equity mutual funds, Journal of Investing, Vol. 10, pp. 81-92. Phelps, S. and Detzel, L. (1997), The nonpersistence of mutual fund performance, Quarterly Journal of Business and Economics, Vol. 36, pp. 55-69. Philpot, J. and Peterson, G.A. (2006), Manager characteristics and real estate mutual fund returns, risk and fees, Managerial Finance, Vol. 3 No. 12, pp. 988-96. Rahman, S., Fabozzi, F.J. and Lee, C. (1991), Errors-in-variables, functional form, and mutual fund returns, Quarterly Review of Economics and Business, Vol. 31, pp. 24-35. Redman, A.L. and Manakyan, H. (1995), A multivariate analysis of REIT performance by nancial and real asset portfolio characteristics, Journal of Real Estate Finance and Economics, Vol. 10, pp. 169-75. Sing, T.C. (2007), Effects of expenditures and size on mutual fund performance, Singapore Management Review, Vol. 29 No. 1. Walker, G.A. (1997), Testing the predictability of mutual fund returns, Review of Business, Vol. 4 No. 18, pp. 10-15. Corresponding author Giacomo Morri can be contacted at: giacomo.morri@sdabocconi.it

Italian real estate mutual funds

185

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com Or visit our web site for further details: www.emeraldinsight.com/reprints

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.