Cash flow timing skills of socially responsible mutual fund investors

Cash flow timing skills of socially responsible mutual fund investors

FINANA-01036; No of Pages 15 International Review of Financial Analysis xxx (2016) xxx–xxx Contents lists available at ScienceDirect International R...

890KB Sizes 4 Downloads 33 Views

FINANA-01036; No of Pages 15 International Review of Financial Analysis xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

International Review of Financial Analysis

Cash flow timing skills of socially responsible mutual fund investors Fernando Muñoz ⁎ Centro Universitario de la Defensa de Zaragoza, Spain

a r t i c l e

i n f o

Article history: Received 6 May 2016 Received in revised form 6 September 2016 Accepted 16 September 2016 Available online xxxx Keywords: Cash flow timing Investor skills Sophisticated investors Green funds Religious funds Environmental Social and governance (ESG) funds Profit-seeking Value-driven

a b s t r a c t This paper studies, for the first time, the cash flow timing skills of socially responsible (SR) mutual fund investors. Our findings show that SR investors neither worsen nor improve their returns according to their cash flow timing decisions, although they show good timing for net purchase and perverse timing for net withdrawal decisions. When controlling for fund characteristics, investors in larger, institutional, with longer mean manager tenure, lower expense ratio, no load, lower mean turnover ratio and a fee level below the average funds, show better timing results; in other words, sophisticated and better informed investors make better cash flow timing decisions. Controlling for SR strategy, green fund investors (our proxy for profit-seeking investors) had the worst results (similar to those obtained for conventional investors in the prior literature), and religious fund investors (our proxy for the values-driven profile) had results that were most different from conventional investors. © 2016 Elsevier Inc. All rights reserved.

1. Introduction The socially responsible (SR) mutual fund industry has experienced huge growth in recent years. According to the US SIF,1 the value of assets under the management of SR mutual funds in the United States rose from $641 billion to $1.93 trillion between 2012 and 2014. Given these figures, academic interest in the analysis of this segment of the collective investment industry is understandable. Many of the papers that analyse SR mutual funds have focused on the management perspective. The first articles on this topic analysed the financial performance achieved by this type of portfolio (Luther, Matatko, and Corner (1992) or Cummings (2000)). Some of these first articles made comparative analyses of conventional mutual funds (see Mallin, Saadouni, and Briston (1995), Gregory, Matatko, and Luther (1997), Bello (2005), Bauer, Koedijk, and Otten (2005) or Bauer, Otten, and Tourani (2006); Bauer, Derwall, and Otten (2007), among others). More recent articles have studied the financial performance of SR mutual funds controlling for the type of SR mutual fund strategy (see Renneboog, Ter Horst, and Zhang (2008a), Nofsinger and Varma (2014) or Capelle-Blancard and Monjon (2014), among others) or for the skills of SR mutual fund managers (stock-picking, market timing ⁎ Centro Universitario de la Defensa, Academia General Militar, Ctra. de Huesca s/n, 50.090 Zaragoza, Spain. E-mail address: [email protected] 1 US SIF, the Forum for Sustainable and Responsible Investment, is the US membership association for professionals, firms, institutions and organisations engaged in sustainable, responsible, and impactful investing. The US SIF website is http://www.ussif.org/sribasics.

and/or style timing). The skills of mutual fund managers could be an important source of good financial performance for investors. Stockpicking is the ability to choose stocks that outperform other securities with a similar level of non-diversifiable risk. Market timing supposes to change exposure to the market at the right moment. A good manager will maintain a higher beta in bull markets and a lower beta in bear markets. A good manager could also improve their performance by anticipating which investment style will behave better, and by raising the exposure of their portfolio to this style (style timing). These managerial skills have been broadly analysed for conventional mutual fund managers, but we can also find articles dealing with this question for SR mutual fund managers (see Schröeder (2004), Kreander, Gray, Power, and Sinclair (2005), Gregory and Whittaker (2007), Ferruz, Muñoz, and Vargas (2012), Muñoz, Vargas, and Marco (2014), Muñoz, Vargas, and Vicente (2014), Muñoz, Vicente, and Ferruz (2015) or Leite and Cortez (2015), among others)2,3. 2 Chegut, Schenk, and Scholtens (2011) carry out an exhaustive literature review of the SR mutual fund financial performance literature. 3 The relationship between financial and non-financial performance (see Derwall, Guenster, Bauer, and Koedijk (2005), Kempf and Osthoff (2007), Statman and Glushkov (2009) or Manescu (2011), among others), the motivations of SR mutual fund investors (see McLachlan and Gardner (2004), Beal, Goyen, and Philips (2005), Nilsson (2009) or Derwall et al. (2011), among others) or SR mutual funds and criteria decision analyses (see Ballestero, Bravo, Pérez-Gladish, Arenas-Parra, and Pla-Santamaria (2012), BilbaoTerol, Arenas-Parra, Cañal-Fernández, and Bilbao-Terol (2013), Pérez-Gladish, MéndezRodríguez, M'Zali, and Lang (2013), Utz, Wimmer, Hirschberger, and Steuer (2014) or Petrillo, De Felice, García-Melón, and Pérez-Gladish (2016), among others) are other topics for which works can be found in the SR literature.

http://dx.doi.org/10.1016/j.irfa.2016.09.011 1057-5219/© 2016 Elsevier Inc. All rights reserved.

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

2

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

Mutual fund investors can achieve good financial result if they are able to make good investment decisions. The investment decisions of conventional mutual fund investors have been broadly analysed. A large number of papers analyse how past financial return affects investor decisions about fund flow allocation (Sirri and Tufano (1998) is one of the most representative works in this field), regarding the skill of investors to select funds that will outperform in the future (smart-money effect or selection skills) (Zheng (1999) is a good example) and/or regarding the ability of investors to properly time their purchase and redemption fund share decisions (Friesen and Sapp (2007) is one of the most relevant works for this topic). There are fewer articles dealing with these matters as regards SR mutual fund investors, however. There are several articles involving the relationship between past financial performance and fund flow allocation decisions, but, we have only found two articles studying the smart money effect, and we did not find any article studying the cash flow timing skills of SR mutual fund investors. The analysis of these questions for SR mutual fund investors is interesting due to the special characteristics of this kind of investor (they take into account both financial attributes and other nonfinancial issues such as environmental, religious or governance when making their investment decisions). In this article we analyse for first time, as far as we know, the cash flow timing skills of SR mutual fund investors, covering this gap in the literature. In the next section, we explain why the analysis of this issue is relevant for SR fund investors, showing the most relevant works for the topics dealt with in this article and posing the research questions that we want to answer with this research. The rest of article is structured as follows: in the third section, we explain the methodology employed to carry out the analyses; Section Four describes the data used in the study; the fifth section is devoted to presenting and discussing the main results of the paper; and finally, the sixth section draws the main conclusions. 2. Literature review 2.1. Investment decisions of SR mutual fund investors When considering the investment decisions of fund investors, the question most widely analysed for SR investors has been their reaction to past returns, that is, the relationship between fund flows and financial performance. In the case of conventional mutual fund investors, this relationship is asymmetric, that is the mutual funds with the best financial performance in the past attract higher inflows in subsequent periods, but the worst performing mutual funds do not suffer proportional outflows (Sirri & Tufano, 1998). For SR mutual fund investors this relationship is different. Bollen (2007) finds that the volatility of SR mutual fund flows was lower than that of conventional mutual funds in the period 1991–2002. This author shows how SR investors are less influenced by past negative returns than conventional mutual fund investors when making their investment decisions. Benson and Humphrey (2008) find that SR fund flows are less sensitive to returns than conventional flows, thus showing that SR investors have difficulty finding an alternative SR fund that meets their non-financial concerns. Renneboog et al. (2008b: p.1739) highlights two results from the US SIF (2001 and, 2003 report that suggest different fund flow behaviour for SR investors: i) “During the stock market downturn over the first 9 months of 2001, there was a 94% drop in the money inflows into all the US mutual funds, whereas the fall in the net investments in socially screened funds amounted to merely 54%. ii) Typically, a social investor's assets are ‘stickier’ than those of investors concerned only with financial performance. That is, social investors have been less likely to move investments from one fund to another and more inclined to stay with funds than conventional investors”.

Renneboog, Ter Horst, and Zhang (2011) studied the SR fund flows for a broad international sample of SR mutual funds, and controlled for the different kinds of SR investment strategy implemented. The findings of this study are very interesting. They show that SR money flows are minimally related to past fund returns, but, at the same time, that the kind of SR investment strategy that is implemented plays an important role in this relationship. From this, we can conclude that SR fund flows are less sensitive to past negative returns than conventional fund flows, and that this is especially true for SR funds implementing negative or sin/ethical screens. Social attributes also weaken the relationship between inflows and positive past returns. The opposite is found for environmental criteria: the relationship between past positive returns and inflows for funds with environmental criteria is stronger than for conventional funds. Peifer (2011) studied the stability4 of religious mutual fund investors and compared it with the stability of other kinds of investors (other SR mutual fund investors and conventional fund investors). The author found that religious mutual fund investors are the most stable; that is, past financial return has little impact on their fund flow decisions. Religious mutual funds usually implement negative or exclusionary screens as their SR strategy. In this way, the results obtained by Peifer (2011) are consistent with those reached by Renneboog et al. (2011). The relevance of the SR strategy implemented in the analysis of SR investor investment decisions means that the literature divides the SR movement into segments. A salient article is that of Derwall, Koedijk, and Ter Horst (2011), who split SR investors into two segments, values-driven and profit-seeking investors. Investors for whom social and personal values are the main reasons for investment decisions comprise the first group, and they are willing to sacrifice financial performance in exchange for utility from non-financial attributes. Investors for whom the main aim of implementing SR criteria when making investment decisions is the achievement of a good financial result, comprise the second group (profit-seeking). Derwall et al. (2011) sort SR investors into these two groups according to the SR strategy implemented by the fund in which they invest. These authors argue that values-driven investors put their money in SR mutual funds that implement negative or exclusionary screens, that is, funds that exclude, from their eligible stock universe, stock issued by companies in certain sectors that are considered morally reprehensible (some examples are the tobacco, alcohol and gambling sectors). These sectors are termed ‘sin sectors’ in the literature, and religious mutual funds, for example, usually implement this type of strategy (see, for example, Appendix 1 in Ferruz et al. (2012)). SR mutual funds implementing positive screens would represent profit-seeking investors. A positive screen strategy selects stocks with good records on environmental, social and/or governance issues. Derwall et al. (2011) justify their reasoning on the basis of the prior literature about the relationship between past returns and fund flows. Muñoz, Vargas, and Vicente (2014) developed work that controlled the results for the two SR investor segments described above. They analysed whether the decisions made by investors in profit-seeking and values-driven funds have an impact on the timing ability of socially responsible mutual fund managers. The findings indicate that controlling for fund flows has only a weak effect for the positive screening sample. Controlling for fund flows slightly improves the market timing coefficients for these funds. In short, this review of the literature about the relationship between fund flow and financial performance for SR investors suggests two main conclusions: i) The relationship between fund flow and financial performance differs between SR and conventional mutual fund investors.

4 The author defines stability as the extent to which investors hold onto their fund shares regardless of the performance in terms of past returns.

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

ii) Among SR investors, controlling for the SR investment strategy implemented by the fund is relevant, since this relationship differs according to the fund strategy.

Both conclusions are relevant to this paper since, as we show in the next subsection, the cash flow timing skills of mutual fund investors are influenced by the relationship between fund flows and past return. Another topic analysed in the literature in relation to the investment decisions of SR mutual fund investors has been the smart money effect, although this has been studied to a lesser extent than the relationship between financial performance and fund flow (in fact, we have only found two articles dealing with this issue). The smart money effect is defined as the ability to select funds that will show superior financial performance in the future (investor selection skills). Renneboog et al. (2008b) and Renneboog et al. (2011) have studied this topic with regard to SR fund investors; it was first analysed for conventional fund investors by Gruber (1996) and Zheng (1999). Renneboog et al. (2008b), without controlling for SR strategy, obtained mixed evidence about the smart money effect, in that they showed that SR investors are unable to select funds that will have a superior financial performance in the future, but are able to avoid funds that have a poor performance in subsequent periods. Renneboog et al. (2011), controlling for SR strategy, did not obtain any evidence of a smart money effect, since funds with more inflows do not perform differently from their benchmarks or from their conventional counterparts in subsequent periods. 2.2. Literature on the cash flow timing skills of conventional mutual fund investors Friesen and Sapp (2007) point out that investors can enhance their returns by selecting funds that will outperform in the future (smart money) and/or by properly timing their cash flows to the fund. The smart money effect (selection skill) has been analysed for SR investors, but the second method, cash flow timing, it is a topic that has not been explored for this kind of investor, as far as we know (we have not found any article exploring this question for SR mutual fund investors). The aim of this article is to study the timing skills of SR investors, filling this gap in the existing literature and extending knowledge about SR investor investment decisions. Friesen and Sapp (2007) tried to answer the following question: Do equity fund investors put cash in and take cash out at the right time on average? They examined the cash flow timing decisions of conventional US domestic equity mutual fund investors and find that, on average, investor timing decisions reduce the average return for fund investors by 1.56% annually. To study investor timing skills, Friesen and Sapp (2007) compared the geometric monthly return with the dollar-weighted average return (we give more detail about this timing measure in the methodology section). Friesen and Sapp (2007) also analysed the relationship between investor timing skills and several characteristics of the mutual fund, finding that timing ability is worse with increased fund load fees, turnover ratio and length of return history. This indicates that investors in older and more expensive funds show especially bad cash flow timing skill. Fund volatility and good financial performance also have a positive relationship with poor investor timing skills. Studying the reasons for this perverse ability further, the authors examined timing ability but controlled the results for positive and negative net cash flows. In this way, it was possible to discover if the wrong timing decisions are purchase decisions, withdrawal decisions, or both. The results indicated that net withdrawal decisions are more salient than net purchase decisions in an investor's poor timing skills, but, at the same time, they show that both elements contribute to this poor ability. Friesen and Sapp (2007) found their results consistent with the hypothesis that investors show return-chasing behaviour, where investors

3

purchase fund shares with high recent returns and withdraw fund shares with low recent returns. In this framework, poor timing ability can reduce the average return obtained by investors, although they are able to select funds with superior financial performance. These authors explain that whether fund returns are not persistent, if investors continue to purchase funds with returns high above the mean and withdraw money from funds with returns well below the mean, they will, on average, lose because of the tendency for returns to cluster at the mean. In order to explain this hypothesis, the authors considered ‘heuristic representativeness’, which was examined by Kahneman and Tversky (1972). Kahneman and Tversky demonstrate how people usually overestimate the degree to which one single event is similar to the parent population. Investors suffering from this cognitive bias will overestimate the predictability of fund returns, and assume that a single large return is equivalent to a large mean return. In order to provide additional empirical evidence supporting this hypothesis, Friesen and Sapp (2007) conducted additional analysis with simulated data, and this strengthened their explanation of the results. Dichev and Yu (2011) also studied the cash flow timing decisions of investors, in this case for hedge funds. Using the same methodology as Friesen and Sapp (2007), they found that investors' timing decisions reduce the average return for hedge fund investors by roughly 3%–7% annually. The authors also found that the hedge funds with the most stringent restrictions on redemptions reduced negative effects on the returns achieved by investors; in other words, limiting capital outflow is in the best interests of investors. These authors discussed above consider two possible explanations for their results. Friesen and Sapp (2007) suggest that one possible explanation could be return-chasing behaviour, in which investors invest in funds with high past returns. Alternatively, mutual fund managers may be unable to manage new cash flows properly, leading to lower future returns (Chevalier & Ellison, 1997). Dichev and Yu (2011) tested these two hypotheses and concluded that, as proposed by Friesen and Sapp (2007),5 the origin of investors' perverse timing ability is returnchasing behaviour. Navone and Pagani (2015) also implemented the methodology used by Friesen and Sapp (2007) and Dichev and Yu (2011) in order to study how front loads have an effect on investor timing abilities. Again, they found that investors have poor timing skills in average terms. Their results demonstrated that load fund investors show worse timing skills than investors in funds with no load, with some exceptions applicable to small, young and cheap funds. The findings of the above studies provided the motivation for our research proposal. The poor timing ability of investors is due to returnchasing behaviour, however the empirical evidence in the literature for the relationship between financial performance and fund flows for SR investors seems to indicate that this return-chasing behaviour is not found among SR investors (or at least among some of them, the values-driven investors). For this reason, we could expect different results for SR investors, and, among SR investors, different results for values-driven and profit-seeking investor profiles. In this study we thus want to answer the following research questions: 1) Do SR investors show cash flow timing skills? Are the timing skills of SR investors different from those of conventional fund investors? 2) How do net purchase and withdrawal decisions contribute to the cash flow timing skills of SR fund investors? 3) Do values-driven and profit-seeking investors show different results? 4) How do fund characteristics influence the results?

5 Chieh-Tse Hou (2012) is another author who has analysed these issues using the same methodology, but for the Taiwanese mutual fund market. The results are similar to those obtained by Friesen and Sapp (2007) and by Dichev and Yu (2011).

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

4

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

The dollar-weighted average return for fund i is computed according to the following expression:

3. Methodology To answer the research questions, we will implement the methodology used previously by Friesen and Sapp (2007), Dichev and Yu (2011) and Navone and Pagani (2015) in their analyses of the cash flow timing skills of fund investors in conventional and hedge funds. We assess investor timing skills by computing the difference between the geometric average monthly return and the dollar-weighted average monthly return (this difference is named on literature performance gap). The first of these represents the return from the fund that is achieved by a passive investor who follows a buy-and-hold investment strategy. This performance would be obtained by a mutual fund manager and would represent the average return on one dollar invested during the entire sample period. The next expression shows the geometric average monthly return for fund i:  rg ¼

1=T T ∏ ð1 þ rit Þ −1

t¼1

ð1Þ

where rit is the monthly return obtained by fund i in month t. As pointed out Dichev and Yu (2011), however, this is a bad representation of the actual return obtained by most investors. It represents the return achieved by an investor who joins the fund at the inception date and maintains the same position throughout the life of the fund, however most investors join a fund after the inception date, and make different capital contributions. An investor's actual return depends not only on the return on their investment but also on the amount of capital invested. The capital invested evolves according to cash flow. In this way, the dollar-weighted average return would be a better representation of the return achieved by the investor.6 This measures the return, weighted by the amount of money invested at each moment. For fund i, it is defined as the internal rate of return at which the accumulated value of the initial total net assets (TNA), plus the accumulated value of the net cash flows, equals the actual TNA at the end of the sample period. Dichev and Yu (2011) provide a very intuitive view of this measure. They say that the mutual fund should be considered as an investment project for which the initial TNA and the subsequent capital contributions are the inflows, and the capital redemptions and the final TNA are the outflows. Computing the internal rate of return of this investment project, we obtain a better measure of the actual return achieved by the average investor. In this way, the dollar-weighted monthly return reflects “the actual experience of real-life investors, who consciously or unconsciously time their capital flows into and out of the funds, and thus, their actual realized return can differ substantially from that of the fund” (Dichev & Yu, 2011 p. 251). In order to compute the dollar-weighted average return we need data on the net cash flows for funds. Following the prior literature, we compute the net cash flows of fund i in period t as:   NCFi;t ¼ TNAi;t −TNAi;t−1  1 þ ri;t

ð2Þ

where TNAi,t is the total net assets of fund i in period t and ri,t is the monthly return of fund i in period t. 6 As Dichev and Yu (2011) note, the main drawback of the geometric mean return as a measure of investment performance is that it assumes the equal-weighting of capital over time. Dichev (2007 p. 388) explains the following “Note that for both buy-and-hold and dollar-weighted returns, capital flows during a certain period do not affect the computed return for that period. Capital flows also do not affect the compounding of buy-and-hold returns. In contrast, capital flows affect the computation of dollar- weighted returns across periods because capital flows affect the weighting of the returns for each period in computing the overall return”. Dichev and Yu (2011) point out that the main advantage of dollar-weighted returns is that they properly consider the effect of capital flows and changing capital exposure on investor returns.

T X  −T  −t TNAi;T  1 þ ri;dw ¼ TNAi;0 þ NCFi;t  1 þ ri;dw

ð3Þ

t¼1

In this way, if the difference between the geometric and dollarweighted average return is positive, the investor cash flow timing decisions reduce their returns, and if the opposite occurs, the investors enhance their returns due to their timing skills.7 Performance gap ¼ ri;g −ri;dw

ð4Þ

Friesen and Sapp (2007) studied the cash flow timing skills of investors in more depth by calculating separately the dollar-weighted returns on the net positive and the net negative cash flows for each fund. In this way, they studied the relationship between net purchasing (net positive cash flows) and net withdrawal (net negative cash flows) decisions and investor skills. An investor with good timing skills will show a dollar-weighted return on positive net cash flow that overcomes the geometric average return, since they will invest in funds before periods of high returns. An investor with good timing skills will show a dollar-weighted return on negative net cash flows below the geometric average return. A skilful investor should obtain high dollar-weighted returns on positive net cash flows and low dollar-weighted returns on negative net cash flows. In order to calculate these returns Friesen and Sapp (2007) proposed the following formulas: T X

 T   ðT−tÞ X T   dw;þ þ NCFþ  1 þ r ¼ NCF ∏ 1 þ r i;s i;t i;t i

t¼1 T X

t¼1

s¼tþ1

 T   ðT−tÞ X T   dw;− − NCF−  1 þ r ¼ NCF ∏ 1 þ r i;s i;t i;t i

t¼1

t¼1

s¼tþ1

ð5Þ

ð6Þ

where, NCF+i,t = max(NCFi,t, 0) and NCF−i,t = min(NCFi,t, 0). Using expression (5), we can calculate the dollar-weighted return on positive net cash flows, and expression (6) lets us find the dollar-weighted return on negative net cash flows. 4. Data We selected all the funds in the Morningstar database with the label “socially conscious funds” that were domiciled in the US market in the period from January 1991 to May 2015 and have a domestic equity investment vocation. We compute the geometric monthly return and the dollar-weighted monthly return for each of the SR mutual funds in our sample over the entire sample period. The information collected for each of the funds includes the monthly total net assets (TNAs), the monthly net return and several characteristics of the mutual fund, such as the inception date, the mean manager tenure, the total expense ratio and the mean turnover ratio. The Morningstar database identifies SR mutual funds in several categories: religious mutual funds, Sharia mutual funds, environmental mutual funds and environmental, social and governance (ESG) mutual funds. The vast majority of SR mutual funds implement a mixed strategy of positive and negative screens, however the empirical evidence shows that the prevalent SR strategy for religious mutual funds involves exclusionary or negative screens. Green funds have shown similar behaviour to conventional funds on issues such as the relationship between financial performance and fund flows. Thus, in order to control our results for profit-seeking and values-driven investors, we take the religious and Sharia mutual funds to be a group of funds representing value-driven investors, and the 7 As Friesen and Sapp (2007) explain, this timing measure judges the success of investor cash flows against a buy-and-hold strategy in the respective fund.

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

environmental mutual funds to be a group of funds representing profitseeking investors (it is necessary to remember here the empirical evidence provided by Peifer (2011) and Renneboog et al. (2011), among others). ESG funds implement both social/governance and environmental issues, so these funds probably cater for SR investors belonging to both segments. Table 1 shows the main descriptive statistics for our sample of funds. Our SR fund sample comprises 194 funds, of which 34 are religious funds, 132 are ESG funds and 28 are green funds. The average total net assets (TNA) for all the SR mutual funds is around $111 million. Controlling for the type of SR strategy, the religious funds have an average TNA of $95.59 million, the ESG funds $130.13 million and the green funds $43.64 million. The monthly average net cash flow over all the SR mutual funds is $1.17 million, and the averages are $0.64 million for religious mutual funds, $1.38 million for ESG funds and $0.78 million for green funds. The monthly mean returns are 0.818% for all the SR mutual funds, 0.78% for religious mutual funds, 0.88% for ESG mutual funds and 0.56% for green funds. The average mean manager tenures are 6.06 years for all the SR funds considered together, 7.36 years for the religious mutual funds, 5.63 years for the ESG funds and finally 6.52 years for the green funds. The mean turnover ratios are 47.19% for all the SR

Table 1 Descriptive statistics for sample of SR mutual funds. Mean Panel A: all SR funds (194 funds) Total Net Assets ($ millions) 111.59 Monthly net cash flows ($ 1.17 millions) Monthly return (%) 0.818 Manager tenure (years) 6.06 Turnover ratio (%) 47.19 Prospectus net expense 1.21 ratio (%) Panel A.1: religious funds (34 funds) Total net assets ($ millions) 95.59 Monthly net cash flows ($ 0.64 millions) Monthly return (%) 0.78 Manager tenure (years) 7.36 Turnover ratio (%) 46.48 Prospectus net expense 1.34 ratio (%) Panel A.2: green funds (28 funds) Total net assets ($ millions) 43.64 Monthly net cash flows ($ 0.78 millions) Monthly return (%) 0.56 Manager tenure (years) 6.52 Turnover ratio (%) 44.57 Prospectus net expense 1.50 ratio (%) Panel A.3: ESG funds (132 funds) Total net assets ($ millions) 130.13 Monthly net cash flows ($ 1.38 millions) Monthly return (%) 0.88 Manager tenure (years) 5.63 Turnover ratio (%) 47.93 Prospectus net expense 1.11 ratio (%)

Median 25th 75th Standard percentile percentile deviation 87.66 0.41

44.04 −0.41

161.79 1.96

89.40 6.05

1.265 5.75 36.00 1.17

−1.642 3.50 24.00 0.90

3.704 8.17 68.00 1.47

4.526 3.39 33.62 0.54

61.53 0.24

45.87 −0.28

152.15 1.08

72.27 4.31

1.26 6.44 37.00 1.27

−1.61 4.98 29.00 0.95

3.49 10.03 70.00 1.51

4.42 3.80 27.76 0.51

32.97 0.16

21.77 −0.25

59.65 1.21

32.03 4.10

1.18 7.25 30.00 1.38

−2.48 4.84 27.50 1.20

4.03 7.80 62.00 1.86

5.38 3.49 31.09 0.54

105.99 0.51

48.29 −0.49

185.95 2.35

105.99 6.91

1.29 5.29 36.00 1.01

−1.47 3.06 21.00 0.81

3.69 7.76 68.00 1.29

4.37 3.18 35.63 0.51

This table presents the descriptive statistics for the SR mutual fund sample obtained from Morningstar. Panel A shows the statistics for all the SR mutual funds in the sample considered together. Panel A.1. shows the information for the religious funds, Panel A.2. reports the statistics for the green funds and Panel A.3. provides the information for the ESG funds. The information provided includes the total net assets in $ millions, the monthly net cash flow in $ millions, the monthly return in %, the mean manager tenure in years, the mean turnover ratio in % and the prospectus net expense ratio in %. For all these measures, the mean, median, 25th percentile, 75th percentile and standard deviation for each of the samples considered are reported.

5

mutual funds, 46.48% for the religious mutual funds, 47.93% for the ESG funds and 44.57% for the green funds. Finally, the average prospectus net expense ratios are 1.21% for all the SR mutual funds, 1.34% for the religious mutual funds, 1.11% for the ESG mutual funds and 1.50% for the green funds. As well as the average values for these measurements, Table 1 provides information about the median, the 25th percentile, the 75th percentile and the standard deviation.

5. Empirical results 5.1. Results for cash flow timing skill 5.1.1. Results for cash flow timing ability at aggregate level Table 28 provides the results for the cash flow timing skills for all the SR mutual funds considered together and when controlling for the different SR approaches, that is, for religious, green and ESG funds. For each of the SR funds we calculate, the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (that is, the difference between the geometric and the dollar-weighted returns) over the entire sample period. For all these variables, we provide the mean, the median, the 25th percentile, the 75th percentile and the standard deviation. In order to determine whether the geometric and dollar-weighted monthly returns are significantly different, we calculate t-test and non-parametric Wilcoxon test statistics. If we analyse the results for all the SR funds considered together, we see that the average geometric monthly return is 0.719%, whereas the average dollar-weighted monthly return is 0.696%, which is 0.023% less. This result could indicate a slightly perverse cash flow timing skill of SR investors, but this performance gap is insignificant; the cash flow timing decisions of the SR investors neither significantly worsen nor significantly improve the return they obtain. This shows that they have better timing skill than conventional investors, who have been shown in the prior literature to have a perverse cash flow timing skill, however if we consider the results for each of the types of SR mutual funds, different conclusions are reached: religious, green and ESG fund investors show different behaviour. In the case of religious fund investors, although insignificant, the performance gap is negative. The average geometric monthly return (0.683%) is 0.048% lower than the average dollar-weighted monthly return (0.731%). For ESG fund investors, the performance gap is also insignificant, but the geometric monthly return is slightly higher than the dollar-weighted monthly return (0.019% higher). Finally, the green fund investors show the most interesting result. These investors show a significantly perverse cash flow timing skill, since the average dollar-weighted monthly return is 0.127% lower (0.304%) than the average geometric monthly return (0.431%). These findings allow us to answer two of the research questions posed in the previous section. The prior literature (Friesen and Sapp (2007)) has shown us that conventional investors have poor cash flow timing skills, making their returns worse than those obtained by managers. Friesen and Sapp (2007) argue that their results are consistent with return-chasing investor behaviour, as we explained in the previous section, however when we analyse the results for US domestic equity SR fund investors, who show a different cash flow and performance behaviour than that found in the prior literature, the results indicate that they do not make their returns worse with their cash flow timing decisions. This result is consistent with the assertion that the behaviour of conventional and SR investors is different. More interesting still are the results obtained when we study the behaviour of the different kinds of SR investors. The result for green investors is consistent with the result achieved by Friesen and Sapp (2007)

8

Fig. 1 in Appendix 1 is a graphical summary of these results.

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

6

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

Table 2 Timing ability of SR mutual fund investors. Mean

Median

25th percentile

75th percentile

Standard deviation

Panel A: all SR mutual funds (194 funds) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) t-Statistic/Wilcoxon test

0.827 0.719 0.696 0.023 0.925/0.826

0.765 0.647 0.730 0.025

0.626 0.502 0.430 −0.161

1.093 0.991 1.036 0.189

0.403 0.439 0.542 0.348

Panel A.1: religious funds (34 funds) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) t-Statistic/Wilcoxon test

0.784 0.683 0.731 −0.048 −0.787/−0.385

0.752 0.633 0.710 −0.010

0.608 0.486 0.558 −0.112

0.871 0.763 0.898 0.101

0.270 0.292 0.343 0.354

Panel A.2: green funds (28 funds) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) t-Statistic/Wilcoxon test

0.596 0.431 0.304 0.127 1.906⁎/1.435

0.674 0.538 0.415 0.016

0.303 0.195 0.091 −0.119

0.857 0.763 0.853 0.304

0.605 0.701 0.822 0.354

Panel A.3: ESG funds (132 funds) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) t-Statistic/Wilcoxon test

0.888 0.789 0.770 0.019 0.646/0.579

0.783 0.676 0.842 0.047

0.651 0.551 0.472 −0.195

1.169 1.081 1.073 0.186

0.360 0.373 0.475 0.343

This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the sample period. A positive value for the performance gap represents a perverse cash flow timing skill for investors, whereas a negative value suggests good cash flow timing skill. For each of these measures, we provide the mean, median, 25th percentile, 75th percentile and standard deviation for each of the samples considered. t-Test and Wilcoxon test statistics are provided to check the significance of the performance gap. Panel A shows the information for all the SR mutual funds considered together, Panel A.1. the information for religious funds, Panel A.2. the information for green funds and finally Panel A.3 the information for ESG funds. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

for conventional investors (a perverse cash flow timing skill). The reasons for this can be found in the literature. Renneboog et al. (2011) find that, for investors in a fund with environmental criteria, the relationship between past positive returns and inflows is stronger than it is for conventional investors. Green fund investors therefore show return-chasing behaviour that leads to their cash flow timing skills being perverse, as is the case with conventional investors. Religious fund investors show better cash flow timing results (good skill, although insignificant). Again, we can find the reasons for this result in the literature. Renneboog et al. (2011) find that negative screens weaken the relationship between past performance and fund flow, and Peifer (2011) finds that religious fund investors hold onto their fund shares, ignoring the past return performance. These results could suggest that religious fund investors do not show return-chasing behaviour, leading to them making better cash flow timing decisions. In conclusion the evidence is mixed among SR fund investors. Green fund investors (our proxy for investors with a profit-seeking profile) show timing skills similar to conventional fund investors, and religious fund investors (our proxy for values-driven investors) and ESG investors make better cash flow timing decisions than green fund investors. 5.1.2. Cash flow timing ability results: Net purchase and withdrawal decisions We deepen our earlier analyses by studying the dollar-weighted returns separately for net positive and net negative cash flows. We report these results in Table 3.9 9 It is important to remember from the methodological section that a skillful investor should obtain a higher dollar-weighted returns on positive net cash flows than the geometric monthly fund return (that is a negative performance gap on net purchase decisions) and lower dollar-weighted returns on negative net cash flows than the geometric monthly fund return (that is, a positive performance gap on net withdrawal decisions). Figs. 2 and 3 in Appendix 2 show the main results graphically.

Panel A shows the results for all the SR mutual fund investors considered together. We can see that the dollar-weighted monthly return on positive net cash flows is 0.787% for the average fund, whereas the dollar-weighted monthly return on negative net cash flows is 0.914%. If we compare these figures with the average geometric monthly return (0.719%), we conclude that investors make well-timed purchase decisions, since they earn 0.068% above the average geometric monthly return, however they demonstrate poorly-timed withdrawal decisions, which cost them 0.176%. These findings are very interesting, because if we compare them with those obtained by Friesen and Sapp (2007), we see that SR investors show good timing skill in their purchase decisions, but lower their returns with their perversely timed withdrawal decisions. Friesen and Sapp (2007) provide empirical evidence that both purchase and withdrawal decisions contribute in the same way to the perverse timing skills of conventional investors, however, although the underperformance caused by withdrawal decisions is more relevant than the underperformance caused by purchase decisions. The advantage of SR investors over their conventional counterparts therefore proceeds from their well-timed purchase decisions. When we study this in more detail, by considering each of the subsamples, all SR investors, whether they are investors in religious, green or ESG funds, make bad withdrawal decisions that lower their monthly returns. This gap is more salient in the case of green fund investors (−0.315%), and is less important for religious and ESG fund investors (−0.134% and −0.158% respectively). Not all the SR fund investors show good skill in purchase decisions. For example, religious and ESG fund investors are able to improve their returns by their purchase decisions (by 0.056% and 0.084% respectively, although this is only significant in the second case), however green fund investors neither improve nor worsen their returns by their purchase decisions (the dollar-weighted monthly returns on net positive cash flows and the average geometric monthly returns are practically the same, at around 0.43%).

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

7

Table 3 Timing skill shown in the net purchase and withdrawal decisions of SR mutual fund investors. Mean

Median

25th percentile

75th percentile

Standard deviation

Panel A: all SR mutual funds (194 funds) Geometric monthly return (%) DW on pos net cash flows (%) DW on neg net cash flows (%) Gap for net purchase decisions (%) t-Test/Wilcoxon test Gap for net withdrawal decisions (%) t-Test/Wilcoxon test

0.719 0.787 0.914 −0.068 −3.087⁎⁎⁎/−3.332⁎⁎⁎ −0.176 −5.245⁎⁎⁎/−6.454⁎⁎⁎

0.647 0.807 0.898 −0.047

0.502 0.567 0.694 −0.257

0.991 1.073 1.158 0.102

0.439 0.443 0.429 0.306

−0.202

−0.512

0.019

0.468

Panel A.1: religious funds (34 funds) Geometric monthly return (%) DW on pos net cash flows (%) DW on neg net cash flows (%) Gap for net purchase decisions (%) t-Test/Wilcoxon test Gap for net withdrawal decisions (%) t-Test/Wilcoxon test

0.683 0.739 0.817 −0.056 −1.303/−1.599 −0.134 −2.37⁎⁎/−2.57⁎⁎

0.633 0.686 0.834 −0.046

0.486 0.601 0.644 −0.192

0.763 0.913 0.945 0.039

0.292 0.221 0.303 0.249

−0.157

−0.316

0.048

0.330

Panel A.2: green funds (28 funds) Geometric monthly return (%) DW on pos net cash flows (%) DW on neg net cash flows (%) Gap for net purchase decisions (%) t-Test/Wilcoxon test Gap for net withdrawal decisions (%) t-Test/Wilcoxon test

0.431 0.438 0.746 −0.007 −0.095/−0.751 −0.315 −3.108⁎⁎⁎/−3.37⁎⁎⁎

0.538 0.554 0.802 −0.036

0.195 0.175 0.557 −0.285

0.763 0.876 1.102 0.165

0.701 0.687 0.492 0.365

−0.396

−0.564

−0.074

0.536

Panel A.3: ESG funds (132 funds) Geometric monthly return (%) DW on pos net cash flows (%) DW on neg net cash flows (%) Gap for net purchase decisions (%) t-Test/Wilcoxon test Gap for net withdrawal decisions (%) t-Test/Wilcoxon test

0.789 0.873 0.946 −0.084 −3.149⁎⁎⁎/−2.887⁎⁎⁎ −0.158 −3.768⁎⁎⁎/−4.881⁎⁎⁎

0.676 0.891 0.927 −0.047

0.551 0.595 0.707 −0.278

1.081 1.118 1.203 0.104

0.373 0.381 0.456 0.307

−0.195

−0.528

0.022

0.481

This table provides information about the average geometric monthly return, the dollar-weighted monthly return (DW) on positive net cash flows, the dollar-weighted monthly return (DW) on negative net cash flows, the difference between the geometric monthly returns and the DW on positive net cash flows (gap for positive net purchase decisions), and the difference between the geometric monthly returns and the DW on negative net cash flows (gap for net withdrawal decisions). For all these measures the mean, median, 25th percentile, 75th percentile and standard deviation are provided. t-Test and Wilcoxon test statistics are reported in order to check the significance of the gaps for the net purchase and withdrawal decisions. It is important to remember from the methodological section that a skilful investor should obtain higher dollar-weighted returns on positive net cash flows than the geometric monthly fund return (that is a negative performance gap on net purchase decisions) and lower dollar-weighted returns on negative net cash flows than the geometric monthly fund return (that is, a positive performance gap on net withdrawal decisions). Panel A shows the results for all the SR mutual funds considered together, Panel A.1. shows the results for the religious funds, Panel A.2. the results for the green funds and Panel A.3. the results for the ESG funds. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

Friesen and Sapp (2007), when explaining the particularly bad results for withdrawal decisions, find them consistent with the limitsof-arbitrage theory of Shleifer and Vishny (1997). Investors may withdraw their money after experiencing negative returns, leading them to sell assets that are in fact undervalued. In this way, an explanation can be found in the prior literature about SR investor behaviour for the differences in these results for religious, green and ESG fund investors. The three groups of investors make bad withdrawal decisions, although the withdrawal decisions of green fund investors are worse (− 0.315%) than those of ESG and religious fund investors (− 0.158% and − 0.134% respectively). Renneboog et al. (2011) show how SR fund flows are less sensitive to past negative returns than conventional fund flows, especially if the SR fund is implementing negative or sin/ethical screens. Peifer (2011) found that religious mutual fund investors are the most stable, in that past financial return has little impact on their fund flow decisions. Investors in religious funds (which implement negative screens) and in ESG funds (which consider social and ethical criteria together with environmental issues) could thus react less strongly to negative past returns than green investors, for whom the literature predicts behaviour that is more similar to that of conventional investors with regards to the relationship between past financial performance and flows.

5.2. Controls for mutual fund characteristics The drivers of cash flow timing skills for conventional investors have been analysed and reported in the literature. Investors with more information and investors with a higher level of sophistication may show better cash flow timing skills than naïve investors. Following this line of argument, we controlled the analyses for various characteristics of the mutual fund, and report the results in this section. The characteristics of mutual funds are related to investor profiles. For example, investors in larger or older funds, or those with a longer mean manager tenure, may have more information when they make their cash flow decisions. The literature shows how cheaper funds and those with lower turnover ratios might cater for investors who are more sophisticated. 5.2.1. Effects of size, age and manager tenure: available information about funds Navone and Pagani (2015) demonstrate that cash flow timing skill is better in older and larger mutual funds. These authors consider size and age as proxies of available information, and conclude that when information about a fund is more readily available, investors make better cash flow timing decisions. Although these authors do not control for manager tenure, we believe that this mutual fund attribute could also be a proxy for the available information about the fund. For example, Gruber (1996), Khorana (1996) and Jain and Wu (2000) indicate that

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

8

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

advertised funds are, on average, somewhat larger and older, and have managers with longer tenures, than the other funds reported by Morningstar. In Table 4, we control for size in our results, and thus implement our analyses for funds with a mean TNA above, and those with a mean TNA below, the median mean TNA. Panel A shows the results for all the SR mutual funds considered together. We note that investors in small funds show worse cash flow timing skill than investors in large funds. For the first group the average geometric monthly return is 0.727%, whereas the dollar-weighted monthly return is 0.663%; that is, their cash flow timing decisions worsen their return by 0.063%. The dollar-weighted return is higher, however, for investors in large funds than the geometric monthly return, although this difference is insignificant. Whereas investors in small funds show a significant perverse timing skill, investors in large funds thus do not worsen their financial returns with their cash flow timing decisions. When we consider the results for each type of SR fund in the sample, ESG funds show a more striking result. Investors in small funds show significant perverse cash flow timing skills. In fact, their cash flow

Table 5 Timing ability of SR mutual fund investors (age effect).

Table 4 Timing ability of SR mutual fund investors (size effect). Below the median TNA (small funds) Panel A: all SR mutual funds (size effect) Arithmetic monthly return (%) 0.830 Geometric monthly return (%) 0.727 Dollar-weighted monthly return 0.663 (%) Performance gap (%) 0.063 Mean TNA ($ millions) 10.61 t-Statistic/Wilcoxon test 1.558/1.707⁎ Panel A.1: religious funds (size effect) Arithmetic monthly return (%) 0.767 Geometric monthly return (%) 0.665 Dollar-weighted monthly return 0.770 (%) Performance gap (%) −0.105 Mean TNA ($ millions) 10.83 t-Statistic/Wilcoxon test −0.896/−0.686 Panel A.2: green funds (size effect) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Mean TNA ($ millions) t-Statistic/Wilcoxon test Panel A.3: ESG funds (size effect) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Mean TNA ($ millions) t-Statistic/Wilcoxon test

timing decisions worsen their monthly returns by 0.078%. Investors in large funds show good but non-significant cash flow timing skill (0.039%). Table 5 provides the results when we control for the age of the fund. We measure the age of the fund from its inception date. We split the fund sample into two groups, one containing funds with an age above the median and the other containing funds with an age below the median. The results show that the effect of age on the cash flow timing skill results is insignificant. At the aggregate level, the performance gap is very similar for the two groups of funds (around 0.02% and insignificant). No significant results were obtained when we considered the results for each type of SR analysed. Finally, we studied the results when we control for the mean tenure of the managers. We again split the sample into two groups of funds, one containing all the funds with mean manager tenure above the median and the other the funds with mean manager tenure below the median. Focusing on the results for all the SR funds considered together, we can see in Table 6 how investors in funds with a mean manager tenure

Above the median TNA (large funds) 0.825 0.711 0.728 −0.017 212.57 −0.5798/−0.714 0.801 0.701 0.692 0.009 180.35 0.276/0.402

0.485 0.313 0.184

0.707 0.550 0.424

0.128 12.195 1.118/0.785

0.126 79.92 1.744/1.412

0.904 0.816 0.738

0.871 0.762 0.801

0.078 34.44 1.763⁎/1.952⁎

−0.039 96.13 −1.002/−1.211

This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the entire sample period. We control for size, for which we use the mean TNA. Two groups of funds are built for each of the samples, one containing the funds with mean TNA below the median and the other the funds with mean TNA above the median. The mean TNA in $ millions is provided, and the t-test and the Wilcoxon test statistics are given, to check the significance of the performance gap. Panel A gives information for all the SR mutual funds considered together, Panel A.1. the information for the religious funds, Panel A.2. for the green funds and finally Panel A.3 for the ESG funds. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

Below the median age (young funds) Panel A: all SR mutual funds (age effect) Arithmetic monthly return (%) 0.981 Geometric monthly return (%) 0.881 Dollar-weighted monthly return 0.854 (%) Performance gap (%) 0.027 Mean age (years) 5.09 t-Statistic/Wilcoxon test 0.800/0.599 Panel A.1: religious funds (age effect) Arithmetic monthly return (%) 0.937 Geometric monthly return (%) 0.850 Dollar-weighted monthly return 0.826 (%) Performance gap (%) 0.023 Mean age (years) 6 t-Statistic/Wilcoxon test 0.268/0.157 Panel A.2: green funds (age effect) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Mean age (years) t-Statistic/Wilcoxon test Panel A.3: ESG funds (age effect) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Mean age (years) t-Statistic/Wilcoxon test

Above the median age (old funds) 0.674 0.556 0.537 0.019 10.2 0.521/0.581

0.677 0.566 0.664 −0.097 18.68 −1.165/−0.448

0.936 0.832 0.671

0.301 0.084 −0.014

0.161 4.45 1.323/0.874

0.098 11.87 1.395/1.022

1.081 0.997 1.009

0.688 0.574 0.523

−0.012 5.11 −0.328/0.069

0.052 15.52 1.102/0.801

This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the entire sample period. For each of the samples of SR funds considered, this information is provided controlling for age (measured by the inception date of the fund). Two groups of funds are built for each of the samples, one containing the funds whose age is below the median and the other the funds whose age is above the median. We also provide the mean age, and the t-test and the Wilcoxon test statistics to check the significance of performance gap. Panel A shows the information for all the SR mutual funds considered together, Panel A.1. the information for the religious funds, Panel A.2. the information for the green funds and finally Panel A.3 the information for the ESG funds. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

below the median show a significantly perverse cash flow timing skill (the geometric monthly return is 0.778% whereas the dollar-weighted monthly return is 0.712%; these investors lower their monthly returns with their cash flow timing decisions by 0.065%). Investors in the group of funds with mean manager tenure above the median show good but insignificant cash flow timing ability. If we analyse the results for each of the kinds of SR funds, the more noticeable result appears to be for green funds, where investors in funds with a mean manager tenure below the median show strong evidence of perverse cash flow timing skill (their cash flow timing decisions lower their monthly returns by 0.29%). Considering all the results together, we find empirical evidence that size and mean manager tenure are the fund characteristics that are relevant for cash flow timing analyses. The empirical evidence obtained (except for the effect of age) is consistent with the results achieved by Navone and Pagani (2015) for conventional funds, suggesting that more information leads to better cash flow timing decisions. 5.2.2. Sophistication of investors There are analyses of the relationship between the characteristics of funds and the level of sophistication of the investors catered for by the funds reported in the literature. From works such as Houge and Wellman (2007), Zhao (2008), Bergstresser, Chalmers, and Tufano (2009) and Navone and Pagani (2015), among others, we know that funds with lower expenses cater for more sophisticated investors than

funds with higher expenses. Other authors such as Goetzmann and Peles (1997) and Bailey, Kumar, and Ng (2011) suggest that retail investors are prone to a number behavioural biases, and Keswani and Stolin (2008), for example, indicate that institutional investors should benefit from both better information and more sophisticated evaluation techniques. Another proxy for the sophistication of investors is the turnover ratio. Chalmers, Kaul, and Phillips (2013) indicate that funds with a lower turnover ratio cater for investors who are more sophisticated. In this section we therefore implement our analyses and control for the level of sophistication of investors. To split our samples between sophisticated and unsophisticated investors we use information about fees, expense ratios, whether a fund is institutional or non-institutional, and turnover ratios. Table 7 shows the results of our analyses when we split our SR fund sample according to the expense ratio of the funds. We create two groups of funds, those whose expense ratio is below the median and those with an expense ratio above the median. The results are consistent with the idea that lower expenses attract more sophisticated investors. In Panel A, we can see the results for all the SR funds considered together. These indicate that investors in funds with higher expense ratios show perverse cash flow timing skills (a 0.083% performance gap), however investors in the funds with lower expense ratios show good, although insignificant, cash flow timing ability (a performance gap of −0.036%).

Table 7 Timing ability of SR mutual fund investors (expense ratio).

Table 6 Timing ability of SR mutual fund investors (mean manager tenure). Below the median

9

Below the median expense ratio

Above the median

Above the median expense ratio

Panel A: all SR mutual funds (manager tenure) Arithmetic monthly return (%) 0.875 Geometric monthly return (%) 0.778 Dollar-weighted monthly return (%) 0.712 Performance gap (%) 0.065 Mean manager tenure (years) 3 t-Statistic/Wilcoxon test 1.768⁎/1.550

0.775 0.655 0.677 −0.023 8.9 −0.690/−0.446

Panel A: all SR mutual funds (expense ratio) Arithmetic monthly return (%) 0.866 Geometric monthly return (%) 0.767 Dollar-weighted monthly return (%) 0.803 Performance gap (%) −0.036 Mean expense ratio (%) 0.847 t-Statistic/Wilcoxon test −1.309/−1.042

0.789 0.670 0.588 0.083 1.863 2.015⁎⁎/1.984⁎⁎

Panel A.1: religious funds (manager tenure) Arithmetic monthly return (%) 0.784 Geometric monthly return (%) 0.683 Dollar-weighted monthly return (%) 0.750 Performance gap (%) −0.067 Mean manager tenure (years) 4.7 t-Statistic/Wilcoxon test −0.669/0

0.785 0.683 0.703 −0.020 11.14 −0.512/−0.785

Panel A.1: religious funds (expense ratio) Arithmetic monthly return (%) 0.851 Geometric monthly return (%) 0.755 Dollar-weighted monthly return (%) 0.727 Performance gap (%) 0.028 Mean expense ratio (%) 0.97 t-Statistic/Wilcoxon test 0.444/0.308

0.717 0.611 0.735 −0.124 1.80 −1.217/−0.781

Panel A.2: green funds (manager tenure) Arithmetic monthly return (%) 0.993 Geometric monthly return (%) 0.892 Dollar-weighted monthly return (%) 0.604 Performance gap (%) 0.288 Mean manager tenure (years) 3.54 t-Statistic/Wilcoxon test 2.327⁎⁎/2.045⁎⁎

0.338 0.134 0.110 0.023 8.45 0.348/−0.166

Panel A.2: green funds (expense ratio) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Mean expense ratio (%) t-Statistic/Wilcoxon test

0.947 0.835 0.757 0.078 1.124 0.785/0.094

0.244 0.028 −0.149 0.177 2.409 1.941⁎/1.664⁎

Panel A.3: ESG funds (manager tenure) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Mean manager tenure (years) t-Statistic/Wilcoxon test

0.950 0.845 0.866 −0.021 8.35 −0.564/−0.428

Panel A.3: ESG funds (expense ratio) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Mean expense ratio (%) t-Statistic/Wilcoxon test

0.870 0.771 0.830 −0.060 0.775 −1.618/−1.492

0.905 0.807 0.709 0.098 1.745 2.164⁎⁎/2.143⁎⁎

0.828 0.736 0.679 0.057 3.07 1.241/1.228

This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the entire sample period. For each of the samples of SR funds considered, this information is provided controlling for mean manager tenure. Two groups of funds are built for each of the samples, one containing the funds with mean manager tenure below the median and the other the funds with mean manager tenure above the median. We also provide the mean manager tenure, and the t-test and the Wilcoxon test statistics to check the significance of the performance gap. Panel A shows the information for all the SR mutual funds considered together, Panel A.1. the information for the religious funds, Panel A.2. the information for the green funds and finally Panel A.3 the information for the ESG funds. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the entire sample period. For each of the samples of SR funds considered, this information is provided controlling for expense ratio. Two groups of funds are built for each of the samples, with one containing the funds with expense ratio below the median and the other the funds with expense ratio above the median. We also provide the mean expense ratio, and the t-test and the Wilcoxon test statistics to check the significance of the performance gap. Panel A shows the information for all the SR mutual funds considered together, Panel A.1. the information for the religious funds, Panel A.2. the information for the green funds and finally Panel A.3 the information for the ESG funds. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

10

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

When we analyse the results for each of the types of SR mutual funds, the most interesting results are obtained for green and ESG funds. In both cases, investors in funds with a higher expense ratio show significantly poorer cash flow timing skills (0.18% and 0.10% respectively). Morningstar labels funds according to their fee levels. Using this label, we sorted the funds into three categories according to their fee level (below average, average and above average). Table 8 shows the results when controlling in this way. The results achieved when controlling for the level of fund fees also provide evidence of better cash flow timing skills of investors in funds with lower fees than investors in funds in other categories. Thus, investors in funds with fees below the average level show high cash flow timing skill for both the entire sample of SR funds considered together and for each of the categories of SR funds considered (religious, green and ESG), although these results are only significant for the sample of all SR funds considered together (performance gap −0.065%). Investors in funds with an average fee level show poor cash flow timing abilities in all cases, although the results are only significant when all the SR funds are considered together (0.094% performance gap). Finally, investors in funds in the category with above average fees only show significant results (of perverse skills) in the case of the ESG fund sample (a performance gap of 0.117%) (Table 9). When we control for the sophistication of investors according to mean turnover ratio, the results are less significant. Significant results are only obtained for investors in green funds with a mean turnover ratio above the median. Such investors show perverse cash flow timing skill (a performance gap of 0.214%). This result is also consistent with the idea that sophisticated investors show better cash flow timing skills. Finally, Table 10 shows the results when we implement the analyses for institutional and non-institutional funds. We can see how the investors in non-institutional funds show worse cash flow timing skills than investors in institutional funds. A significant perverse timing ability is thus demonstrated for the sample of all the SR funds considered together, and for green and ESG funds (performance gaps of 0.038%, 0.147% and 0.057% respectively). Finally, it is interesting to note that in the case of religious funds, the sign of cash flow timing skill is not consistent with the expected result when we are controlling for sophistication. In the case of investors in religious funds, the geometric monthly mean return is lower than the dollar-weighted monthly return for the funds with mean turnover ratio above the median, for non-institutional funds, and for funds with an expense ratio above the median or with an above average level of fees (although in all the cases these results are non-significant), however these results suggest that investors in religious funds show different behaviour. At the same time, the results for ESG and green funds are similar to those obtained previously in the literature for conventional investors. 5.2.3. Additional controls In this section, we provide additional controls, but only at aggregate level, considering all the SR funds together. For some of the characteristics that we used to control the results there were not enough funds in some of the categories of SR funds, and this led us to conduct these analyses only for all the SR funds considered together. Typically, researchers have used whether a fund is load or no load as a proxy for investor sophistication (see Bergstresser et al. (2009), among others). Morningstar provides information about the share class type of the fund. Each fund in our sample has one of the following labels: A, Adv, B, C, D, Inst, Inv, N, No Load, Other, Retirement and S. Each of these categories is characterized by a particular regime of load fees. Thus, share classes A and B charge high front load or deferred load fees (typically between 4% and 5.75%). Funds in the other categories either do not charge load fees or, as in the case of share types C and D, only charge them occasionally and with a limit of 1%. Finally, for those funds labelled ‘Other’, whether or not load fees are charged varies among

Table 8 Timing ability of SR mutual fund investors (fee level). Below average Panel A: all SR mutual funds Arithmetic monthly 0.900 return (%) Geometric monthly 0.802 return (%) Dollar-weighted monthly 0.867 return (%) Performance gap (%) −0.065 Number of funds 75 t-Statistic/Wilcoxon test −2.001⁎⁎/−1.859⁎ Panel A.1: religious funds Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Number of funds t-Statistic/Wilcoxon test Panel A.2: green funds Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Number of funds t-Statistic/Wilcoxon test Panel A.3: ESG funds Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Number of funds t-Statistic/Wilcoxon test

Average

Above average

0.824

0.763

0.725

0.636

0.631

0.566

0.094 0.070 47 71 2.003⁎/2.021⁎⁎ 1.438/1.587

0.803

0.855

0.744

0.700

0.764

0.639

0.778

0.654

0.735

−0.077 10 −1.042/−1.172

0.111 7 1.380/1.352

−0.095 17 −0.888/−0.355

0.928

0.616

0.464

0.823

0.464

0.272

0.867

0.181

0.139

−0.045 6 −0.579/−0.524

0.283 6 1.6914/0.943

0.133 16 1.449/1.293

0.914

0.855

0.897

0.817

0.763

0.788

0.882

0.706

0.671

−0.065 59 −1.671/−1.479

0.057 34 1.04/1.342

0.117 38 1.799⁎/1.661⁎

This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the entire sample period. For each of the samples of SR funds considered, this information is provided controlling for fee level group. Three groups of funds are built for each of the samples, the first containing funds with a fee level below the average, the second, funds with a fee level of around the average and the third, funds with a fee level above the average. The number of funds in each group is also shown, and the t-test and the Wilcoxon test statistics to check the significance of performance gap are provided. Panel A shows the information for all the SR mutual funds considered together, Panel A.1. the information for the religious funds, Panel A.2. the information for the green funds and finally Panel A.3 the information for the ESG funds (the information about fee level group fails for one fund). ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

funds. Using this information, we implement our analyses controlling for the type of share class. We provide these results in Table 11. We divided the sample of funds into three groups. The first of these includes the funds with share types A and B, representing load funds; the second group is formed of funds with share classes Adv, C, D, Inst, Inv, N, No Load, Retirement and S, representing the no load funds; and finally, the third contains the funds of ‘other’ share types. The results provide additional empirical evidence about the relevance of controlling for the level of sophistication of investors, and thus, in the case of load funds the results indicate a significantly poor cash flow timing skill (performance gap of 0.112%); in other words, unsophisticated investors make wrong cash flow timing decisions that lessen the monthly returns they achieve. Investors in no load funds show good but insignificant cash flow timing ability (performance gap of − 0.013%). The investors

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx Table 9 Timing ability of SR mutual fund investors (mean turnover ratio). Below the median

11

Table 10 Timing ability of SR mutual fund investors (institutional vs retail). Institutional

Non-institutional

0.782 0.662 0.645 0.017 92.45 0.444/0.300

Panel A: all SR mutual funds Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Number of funds t-Statistic/Wilcoxon test

0.895 0.792 0.802 −0.011 60 −0.233/−0.810

0.797 0.686 0.648 0.038 134 1.297/1.771⁎

Panel A.1: religious funds (turnover ratio) Arithmetic monthly return (%) 0.855 Geometric monthly return (%) 0.759 Dollar-weighted monthly return (%) 0.731 Performance gap (%) 0.028 Mean turnover ratio (%) 33.54 t-Statistic/Wilcoxon test 0.650/0.738

0.714 0.607 0.730 −0.124 94.03 −1.103/−0.686

Panel A.1: religious funds Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Number of funds t-Statistic/Wilcoxon test

1.084 1.006 0.854 0.152 9 1.531/1.362

0.676 0.567 0.686 −0.120 25 −1.712/−1.197

Panel A.2: green funds (turnover ratio) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Mean turnover ratio (%) t-Statistic/Wilcoxon test

0.794 0.650 0.588 0.062 34.99 0.721/−0.155

0.331 0.140 −0.074 0.214 85.48 2.064⁎/1.961⁎⁎

Panel A.2: green funds Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Number of funds t-Statistic/Wilcoxon test

0.619 0.475 0.406 0.068 7 0.444/−0.169

0.588 0.417 0.270 0.147 21 1.968⁎/1.686⁎

Panel A.3: ESG funds (turnover ratio) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Mean turnover ratio (%) t-Statistic/Wilcoxon test

0.878 0.791 0.789 0.002 34.44 0.045/0.489

0.900 0.789 0.745 0.044 96.13 1.006/0.548

Panel A.3: ESG funds Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Number of funds t-Statistic/Wilcoxon test

0.900 0.798 0.855 −0.057 44 −1.048/−1.517

0.881 0.784 0.727 0.057 88 1.627/1.964⁎⁎

Panel A: all SR mutual funds (turnover ratio) Arithmetic monthly return (%) 0.872 Geometric monthly return (%) 0.776 Dollar-weighted monthly return (%) 0.746 Performance gap (%) 0.029 Mean turnover ratio (%) 33.08 t-Statistic/Wilcoxon test 0.912/0.970

Above the median

This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the entire sample period. For each of the samples of SR funds considered, this information is provided controlling for mean turnover ratio. Two groups are built for each of the samples, the first containing the funds with mean turnover ratio below the median and the second the funds with mean turnover ratio above the median. We also provide the mean turnover ratio in each group and the t-test and the Wilcoxon test statistics to check the significance of the performance gap. Panel A shows the information for all the SR mutual funds considered together, Panel A.1. the information for the religious funds, Panel A.2. the information for the green funds and finally Panel A.3 the information for the ESG funds. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the entire sample period. For each of the samples of SR funds considered, this information is provided controlling for institutional/non-institutional character. Two groups are built for each of the samples, the first containing the institutional funds and the second the non-institutional funds. The number of funds in each group is also shown and the t-test and the Wilcoxon test statistics to check the significance of the performance gap are provided. Panel A shows the information for all the SR mutual funds considered together, Panel A.1. the information for the religious funds, Panel A.2. the information for the green funds and finally Panel A.3 the information for the ESG funds. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

in the ‘other’ category show perverse but insignificant timing skill (0.091%). Finally, several authors have provided empirical evidence about the relationship between mutual fund flows and the rating of the funds provided by specialised companies. One of the rating systems most commonly analysed in the literature has been Morningstar's star-rating system, which allocates a rating of one, two, three, four or five stars according to the financial performance of the fund (see for example, Nanda, Wang, and Zheng (2004), Del Guercio and Tkac (2008) or Khorana and Servaes (2012)). Ratings are a source of information for investors who seek to maximise their returns. We obtained information about the overall Morningstar star rating10 for each of the funds in our sample, managing to find this data for 186 out 194 funds (we failed to find information for only eight funds). Nine funds have a rating of 1 star, 50 funds have a rating of 2 stars, 80 funds have a rating of 3 stars, 27 funds have a rating of 4 stars and 20 funds have a rating of 5 stars. In Table 12, we show the results of controlling for each of the rating categories. The results obtained indicate that the cash flow timing skills of SR fund investors improve as the rating of the fund increases. Funds with

a Morningstar rating of 5 stars thus cater for investors with significantly good cash flow timing skills (their fund flow decisions improve their monthly returns by 0.091%). For the rest of the categories, it seems that the lower the rating of the fund, the worse the timing ability of its investors (however, the results for all these categories are nonsignificant).

10 Detailed information about the rating information can be found on the Morningstar Ratings for Funds FactSheet: http://corporate.morningstar.com/es/documents/MethodologyDocuments/FactSheets/ MorningstarRatingForFunds_FactSheet.pdf.

6. Conclusions Friesen and Sapp (2007) note that investors have two possible methods of enhancing their returns: first, by selecting funds that will outperform in the future (smart money); and second, by making welltimed cash flow decisions. Both issues have been analysed for conventional investors, or those who only consider financial attributes in their investment decisions, however these topics have been neglected for SR mutual fund investors. Two works have previously analysed only the smart money effect (selection skill). Cash flow timing is an unexplored topic for this kind of investor, as far as we know (we have not found any article dealing with this topic). The aim of this article is to study the timing skills of SR investors, filling this gap in the literature and extending knowledge about SR investor investment decisions. This is important for several reasons. This industry is of growing importance worldwide; and the literature has shown different results for conventional and SR mutual fund investors in areas such as their reaction to past financial performance. We also analyse this question by controlling

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

12

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

Table 11 Timing ability of SR mutual fund investors controlling for share type (load/no load). Panel A: share type A, B (load funds) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) Number of funds t-Statistic/Wilcoxon test

Table 12 Timing ability of SR mutual fund investors controlling for star rating. Panel A: SR funds with rating of 1 star (9 funds)

0.767 0.653 0.541 0.112 37 1.775⁎/1.939⁎

Panel B: share type Adv, C, D, Inst, Inv, N, retirement, S, No load (no load funds) Arithmetic monthly return (%) 0.814 Geometric monthly return (%) 0.704 Dollar-weighted monthly return (%) 0.717 Performance gap (%) −0.013 Number of funds 134 t-Statistic/Wilcoxon test −0.433/−0.474 Panel C: share type other (existence of load depends on the fund) Arithmetic monthly return (%) 1.004 Geometric monthly return (%) 0.910 Dollar-weighted monthly return (%) 0.819 Performance gap (%) 0.091 Number of funds 23 t-Statistic/Wilcoxon test 1.523/1.277 This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the entire sample period. For each of the samples of SR funds considered, this information is provided controlling for the load/no load character of the fund. Whether a fund is a load fund or not is decided from the share class information. Three samples of funds are built, the first containing the load funds, the second the no load funds and the third those funds for which it is not possible determine their load/no load character from the share type information. The number of funds in each group is also shown, and the t-test and the Wilcoxon test statistics to check the significance of the performance gap are provided. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

for the type of SR strategy implemented. This is important, since the literature that has studied the investment decisions of SR mutual fund investors provides empirical evidence suggesting that SR mutual fund investors should be split into two segments, profit-seeking and values-driven investors. As we have explained in previous sections, the first segment shows an investment decision behaviour that is more similar to that of conventional fund investors, whereas investors in the second segment would behave differently from conventional investors. We also control for two important issues in cash flow timing decisions: i) the available information about the fund at the time the cash flow decision is made; and ii) the level of sophistication of the investor. Both of these have been identified in the literature as determinants of the cash flow timing skills of conventional investors. The results obtained meet our expectations. When we study all the SR mutual funds together, we see evidence that investors have neither good nor bad cash flow timing skills. The literature has provided evidence of perverse cash flow timing skills for conventional mutual fund investors. When we analyse the results for the different types of SR funds in the study, we find significantly poor timing skill for investors in green funds, insignificantly poor timing skill for investors in ESG funds and insignificantly good timing skill for investors in religious funds. From these results we extract two conclusions: i) SR mutual fund investors and conventional fund investors show different cash flow timing skills; and ii) among SR mutual fund investors, investor timing skills vary according to the type of strategy implemented by the SR fund. Green fund investors (our proxy for investors with a profit-seeking profile) show worse timing skills (they behave in a similar way to conventional investors), while values-driven investors (our proxy for this is religious fund investors) make better cash flow timing decisions. We deepen our analysis of cash flow timing skills by studying dollarweighted returns separately for net positive and net negative cash flows. When we study all the SR funds considered together, we see that investors time their purchase decisions well, since they earn

Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) t-Statistic/Wilcoxon test

0.753 0.659 0.542 0.116 0.827/0.770

Panel B: SR funds with rating of 2 stars (50 funds) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) t-Statistic/Wilcoxon test

0.762 0.651 0.592 0.059 1.034/1.511

Panel C: SR funds with rating of 3 stars (80 funds) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) t-Statistic/Wilcoxon test

0.771 0.657 0.640 0.017 0.489/0.758

Panel D: SR funds with rating of 4 stars (27 funds) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) t-Statistic/Wilcoxon test

0.845 0.737 0.754 −0.017 −0.237/−0.577

Panel E: SR funds with rating of 5 stars (20 funds) Arithmetic monthly return (%) Geometric monthly return (%) Dollar-weighted monthly return (%) Performance gap (%) t-Statistic/Wilcoxon test

1.082 0.967 1.058 −0.091 −1.882⁎/−1.643

This table provides information about the arithmetic monthly return, the geometric monthly return, the dollar-weighted monthly return and the performance gap (computed as the difference between the geometric monthly return and the dollar-weighted monthly return). We compute all these measures for each fund over the entire sample period. This information is provided controlling for the Morningstar star rating of the fund. Five samples of funds are built, one for each of the star ratings that Morningstar allocates to a fund. The number of funds in each group is provided, and the t-test and the Wilcoxon test statistics to check the significance of the performance gap are given. ⁎⁎⁎Significant at 1%, ⁎⁎significant at 5%, ⁎significant at 10%.

0.068% over the average geometric monthly return, however they make poorly timed withdrawal decisions, which cost them 0.176%. When we analyse the results for each of the categories of SR funds in the study, we see that, in all cases, religious, green and ESG fund investors make wrong withdrawal decisions which lower their monthly returns, although this underperformance is more important for green fund investors (− 0.315%) than for religious and ESG fund investors (−0.134% and −0.158% respectively). When we look at purchase decisions, we see that not all the SR fund investors show good skills. From this, we conclude that religious and ESG fund investors are able to improve their returns through their purchase decisions (by 0.056% and 0.084% respectively, although this is only significant in the second case). Green fund investors neither improve nor worsen their returns through their purchases. All these results are evidence that green fund investors (our proxy for the profit-seeking profile) show cash-flow timing skills evidence that are more similar to those of conventional investors. When we control for available information about a fund, the results indicate that for those funds for which there is more information (larger funds, and those with a higher mean manager tenure), the timing results are better. Another interesting result is that funds with the highest Morningstar rating (5 stars) cater for investors with significantly good cash flow timing skills. The sophistication of the investor is approached by controlling our results for expense ratio, fee level group, mean turnover ratio, institutional/non-institutional character and load/no load regime. The results demonstrate that sophisticated investors have better results

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

than unsophisticated ones (where sophisticated investors are those who invest in funds with lower expense ratios, fee levels below the average, no load funds, institutional funds and funds with lower mean turnover ratios). If we look in more detail at the results of controlling for all these characteristics, we find, interestingly, that the conclusions reached are mainly driven by the results for green and ESG funds, whereas the results for religious funds show the opposite behaviour. Taking into account that the results of applying the controls are consistent with those obtained in the conventional investor literature, this suggests that green (our proxy for profit-seeking investors) and ESG fund investors behave in a similar way to conventional investors, while religious

13

mutual fund investors (our proxy for values-driven investors) show different behaviour.

Acknowledgements The author acknowledges financial support from Centro Universitario de la Defensa and Universidad de Zaragoza for the funding from the research project UZCUD2015-SOC-03. The author also would like to express his thanks to the Aragon government for funding received as part of the Public and Official Research Group (S14/2 CIBER Análisis Económico-Financiero de la Empresa y los Mercados).

Appendix 1. Appendix

Fig. 1 summarises the main results shown in Table 2. The mean geometric monthly return, the mean DW monthly return, and as the difference between the two, and the performance gap are shown for each of the SR mutual fund subsamples analysed. It is important to remember from the methodological section that a skilful investor should obtain higher dollar-weighted returns than the geometric monthly fund return (that is, a negative value of the performance gap).

Fig. 1. Cash flow timing skills of SR mutual fund investors.

Appendix 2. Appendix

Fig. 2 provides information about the geometric monthly return, the dollar-weighted monthly return (DW) on positive net cash flows and the difference between the two (gap for net purchase decisions). The mean value of these magnitudes is reported for each of the subsamples of SR funds analysed in the article. It is important to remember from the methodological section that a skilful investor should obtain higher dollarweighted returns on positive net cash flows than the geometric monthly fund return (that is, a negative value of the performance gap on net purchase decisions).

Fig. 2. Net purchase SR mutual fund investors timing skills.

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

14

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx

Fig. 3 provides information about the geometric monthly return, the dollar-weighted monthly return (DW) on negative net cash flows and the difference between the two (gap for net withdrawal decisions). The mean value of these magnitudes is reported for each of the subsamples of SR funds analysed in the article. It is important to remember from the methodological section that a skilful investor should obtain lower dollarweighted returns on negative net cash flows than the geometric monthly fund return (that is, a positive value of the performance gap on net withdrawal decisions).

Fig. 3. Net withdrawal SR mutual fund investors timing skills.

References Bailey, W., Kumar, A., & Ng, D. (2011). Behavioral biases of mutual fund investors. Journal of Financial Economics, 102(1), 1–27. Ballestero, E., Bravo, M., Pérez-Gladish, B., Arenas-Parra, M., & Pla-Santamaria, D. (2012). Socially responsible investment: A multicriteria approach to portfolio selection combining ethical and financial objectives. European Journal of Operational Research, 216(2), 487–494. Bauer, R., Derwall, J., & Otten, R. (2007). The ethical mutual fund performance: New evidence from Canada. Journal of Business Ethics, 70(2), 111–124. Bauer, R., Koedijk, K., & Otten, R. (2005). International evidence on ethical mutual fund performance and investment style. Journal of Banking & Finance, 29(7), 1751–1767. Bauer, R., Otten, R., & Tourani, A. R. (2006). Ethical investing in Australia: Is there a financial penalty? Pacific-Basin Finance Journal, 14(1), 33–48. Beal, D. J., Goyen, M., & Philips, P. (2005). Why do we invest ethically? The Journal of Investing, 14(3), 66–78. Bello, Z. Y. (2005). Socially responsible investing and portfolio diversification. Journal of Financial Research, 28, 41–57. Benson, K. L., & Humphrey, J. E. (2008). Socially responsible investment funds: Investor reaction to current and past returns. Journal of Banking & Finance, 32(9), 1850–1859. Bergstresser, D., Chalmers, J. M., & Tufano, P. (2009). Assessing the costs and benefits of brokers in the mutual fund industry. Review of Financial Studies, 22, 925–957. Bilbao-Terol, A., Arenas-Parra, M., Cañal-Fernández, V., & Bilbao-Terol, C. (2013). Selection of socially responsible portfolios using hedonic prices. Journal of Business Ethics, 115(3), 515–529. Bollen, N. (2007). Mutual fund attributes and investor behavior. Journal of Financial and Quantitative Analysis, 42, 683–708. Capelle-Blancard, G., & Monjon, S. (2014). The performance of socially responsible mutual funds: Does the screening process matters. European Financial Management, 20(3), 494–520. Chalmers, J., Kaul, A., & Phillips, B. (2013). The wisdom of crowds: Mutual fund investors' aggregate asset allocation decisions. Journal of Banking & Finance, 37, 3318–3333. Chegut, A., Schenk, H., & Scholtens, B. (2011). Assessing SRI fund performance research: Best practices in empirical analysis. Sustainable Development, 19, 77–94. Chevalier, J. A., & Ellison, G. (1997). Risk taking by mutual funds as a response to incentives. Journal of Political Economy, 105, 1167–1200. Chieh-Tse Hou, T. (2012). Return persistence and investment timing decisions in Taiwanese domestic equity mutual funds. Managerial Finance, 38(9), 873–891. Cummings, L. S. (2000). The financial performance of ethical investment trusts: An Australian perspective. Journal of Business Ethics, 25(1), 79–92. Del Guercio, D., & Tkac, P. A. (2008). Star power: The effect of Morningstar ratings on mutual fund flow. Journal of Financial and Quantitative Analysis, 43(04), 907–936. Derwall, J., Guenster, N., Bauer, R., & Koedijk, K. (2005). The eco-efficiency premium puzzle. Financial Analysts Journal, 61(2), 51–63. Derwall, J., Koedijk, K., & Ter Horst, J. (2011). A tale of values-driven and profit-seeking social investors. Journal of Banking & Finance, 61(2), 2137–2147. Dichev, I. D. (2007). What are stock investors' actual historical returns? Evidence from dollar-weighted returns. The American Economic Review, 97(1), 386–401. Dichev, I. D., & Yu, G. (2011). Higher risk, lower returns: What hedge fund investors really earn. Journal of Financial Economics, 100, 248–263. Ferruz, L., Muñoz, F., & Vargas, M. (2012). Managerial abilities: Evidence from religious mutual fund managers. Journal of Business Ethics, 105(4), 503–517.

Friesen, G. C., & Sapp, T. R. (2007). Mutual fund flows and investor returns: An empirical examination of fund investor timing ability. Journal of Banking & Finance, 31(9), 2796–2816. Goetzmann, W. N., & Peles, N. (1997). Cognitive dissonance and mutual fund investors. Journal of Financial Research, 20(2), 145–158. Gregory, A., & Whittaker, J. (2007). Performance and performance persistence of ethical unit trusts in the UK. Journal of Business, Finance and Accounting, 34(7–8), 1327–1344. Gregory, A., Matatko, J., & Luther, R. (1997). Ethical unit trust financial performance: Small company effects and fund size effects. Journal of Business, Finance and Accounting, 24(5), 705–725. Gruber, M. J. (1996). Another puzzle: The growth in actively managed mutual funds. The Journal of Finance, 5-7, 783–810. Houge, T., & Wellman, J. (2007). The use and abuse of mutual fund expenses. Journal of Business Ethics, 70(1), 23–32. Jain, P. C., & Wu, J. S. (2000). Truth in mutual fund advertising: Evidence on future performance and fund flows. The Journal of Finance, 55(2), 937–958. Kahneman, D., & Tversky, A. (1972). Subjective probability: A judgment of representativeness. Cognitive Psychology, 3, 430–454. Kempf, A., & Osthoff, P. (2007). The effect of socially responsible investing on portfolio performance. European Financial Management, 13(5), 908–922. Keswani, A., & Stolin, D. (2008). Which money is smart? Mutual fund buys and sells of individual and institutional investors. The Journal of Finance, 63(1), 85–118. Khorana, A. (1996). Top management turnover an empirical investigation of mutual fund managers. Journal of Financial Economics, 40(3), 403–427. Khorana, A., & Servaes, H. (2012). What drives market share in the mutual fund industry. Review of Finance, 16, 81–113. Kreander, N., Gray, R. H., Power, D. M., & Sinclair, C. D. (2005). Evaluating the performance of ethical and non-ethical funds: A matched pair analysis. Journal of Business, Finance and Accounting, 32(7–8), 1465–1493. Leite, P., & Cortez, M. C. (2015). Performance of European socially responsible funds during market crises: Evidence from France. International Review of Financial Analysis, 40, 132–141. Luther, R. G., Matatko, J., & Corner, D. C. (1992). The investment performance of UK ethical unit trusts. Accounting, Auditing & Accountability Journal, 5(4), 57–70. Mallin, C. A., Saadouni, B., & Briston, R. J. (1995). The financial performance of ethical investment funds. Journal of Business, Finance and Accounting, 22(4), 483–496. Manescu, C. (2011). Stock return in relation to environmental, social and governance performance: Mispricing or compensation for risk? Sustainable Development, 19, 95–118. McLachlan, J., & Gardner, J. (2004). A comparison of socially responsible and conventional investors. Journal of Business Ethics, 52, 11–25. Muñoz, F., Vargas, M., & Marco, I. (2014). Environmental mutual funds: Financial performance and managerial abilities. Journal of Business Ethics, 124(4), 551–569. Muñoz, F., Vargas, M., & Vicente, R. (2014). Fund flow bias in market timing skill. Evidence of the clientele effect. International Review of Economics and Finance, 33, 257–269. Muñoz, F., Vicente, R., & Ferruz, L. (2015). Stock-picking and style-timing abilities: A comparative analysis of conventional and socially responsible mutual funds in the US market. Quantitative Finance, 15(2), 345–358. Nanda, V., Wang, Z. J., & Zheng, L. (2004). Family values and the star phenomenon: Strategies of mutual fund families. Review of Financial Studies, 17, 667–698. Navone, M., & Pagani, M. (2015). Brothers from different mothers how distribution fees change investment behavior. Journal of Banking & Finance, 51, 12–25. Nilsson, J. (2009). Segmenting socially responsible mutual fund investors: The influence of financial return and social responsibility. The International Journal of Bank Marketing, 27(1), 5–31.

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011

F. Muñoz / International Review of Financial Analysis xxx (2016) xxx–xxx Nofsinger, J., & Varma, A. (2014). Socially responsible funds and market crises. Journal of Banking & Finance, 48, 180–193. Peifer, J. L. (2011). Morality in the financial market? A look at religiously affiliated mutual funds in the USA. Socio-Economic Review, 9, 235–259. Pérez-Gladish, B., Méndez-Rodríguez, P., M'Zali, B., & Lang, P. (2013). Mutual funds efficiency measurement taking into account financial and socially responsible criteria. Journal of Multi-Criteria Decision Analysis, 20(3–4), 109–125. Petrillo, A., De Felice, F., García-Melón, M., & Pérez-Gladish, B. (2016). Investing in socially responsible mutual funds: Proposal of non-financial ranking in Italian market. Research in International Business and Finance, 37, 541–555. Renneboog, L., Ter Horst, J., & Zhang, C. (2008a). The price of ethics and stakeholder governance: The performance of socially responsible mutual funds. Journal of Corporate Finance, 14, 302–328. Renneboog, L., Ter Horst, J., & Zhang, C. (2008b). Socially responsible investments: Institutional aspects, performance and investor behavior. Journal of Banking & Finance, 32, 1723–1742. Renneboog, L., Ter Horst, J., & Zhang, C. (2011). Is ethical money financially smart? Nonfinancial attributes and money flows of socially responsible investment funds. Journal of Financial Intermediation, 20(4), 562–588.

15

Schröeder, M. (2004). The performance of socially responsible investments: Investment funds and indices. Financial Markets and Portfolio Management, 18(2), 122–142. Shleifer, A., & Vishny, R. (1997). The limits of arbitrage. The Journal of Finance, 52, 35–55. Sirri, E. R., & Tufano, P. (1998). Costly search and mutual fund flows. The Journal of Finance, 53(5), 1589–1622. Social Investment Forum (SIF) (2001). Report on responsible investing trends in the US. Statman, M., & Glushkov, D. (2009). The wage of social responsibility. Financial Analysts Journal, 65, 33–46. Utz, S., Wimmer, M., Hirschberger, M., & Steuer, R. (2014). Tri-criterion inverse portfolio optimization with application to socially responsible mutual funds. European Journal of Operational Research, 234, 491–498. Zhao, X. (2008). Conflict of interests between load fund investors and brokers and financial advisors. (Working paper) China Europe International Business School. Zheng, L. (1999). Is money smart? A study of mutual fund investors' fund selection ability. The Journal of Finance, 54(3), 901–933.

Please cite this article as: Muñoz, F., Cash flow timing skills of socially responsible mutual fund investors, International Review of Financial Analysis (2016), http://dx.doi.org/10.1016/j.irfa.2016.09.011