V. Methodology and Results
2. Fund Flow
Sirri and Tufano (1998) state that searching cost plays a dominant role in the variation of fund flows, the higher the searching cost is, the more reluctant investors would be to change their positions of portfolios. In general, green investors care not only about financial returns but also the issue of the ecosystem. They take green criteria and environmental regulation into account when making investment decisions.
It is reasonable to assume that the volatility of green funds is smaller than that of common investment owing to their concern about the environment.
In terms of the relation between fund age and fund flows, this section presume that fund flows are negatively related to fund age because young funds lack of sufficient information for investors to evaluate their investments. It means that young
9
mutual funds endure more fluctuant fund flows while mature mutual funds own much steadier ones.
3. Flow-Performance Relation
Hypothesis 1: The flow-performance relation and fund flow volatility of green funds is equal to that of conventional funds.
This hypothesis assumes that investor preferences can be exhibited by a utility function defined over the moments of the return distribution of a portfolio. But this statement is the basis of the finance paradigm. Take the Capital Asset Pricing Model of Sharpe (1964), Linter (1965), and Mossin (1966) for example, the utility function is exclusive explained by expected returns and variance in their research. Furthermore, Berk and Green (2004) show a model which is used for investors to update the information of managerial ability as revealed in expected returns. They obtain a positive relation between past performance and subsequent fund flows owing to a rational rearrangement of capital to better managers. In addition, fund flow volatility increases on the sensitivity of investors to the past performance.
The first hypothesis hints that investors evaluate green funds the same way that they evaluate conventional funds.
Hypothesis 2: The flow-performance relation of green funds is stronger than that of conventional funds.
The second hypothesis can be inspired by the assumption that preferences of green investors can be showed by a multi-attribute utility function which is defined over the moments of a portfolio’s return as well as a variable which stands for if the investment decision is green. This combined goal not only fits the environmental demand but also satisfies the requirement for financial performance for companies when promoting green funds. This assumption is consistent with Statman (1999), who concludes that behavioral finance sees the investment decision as a kind of product selection in spite of the standard paradigm. As a result, “value-expressive”
characteristic of an asset matters on its desirability.
This section assumes that green investors can obtain additional utility from devoting to the green investments. However, this situation only happens when the green investments would have been chosen on its financial advantages. In addition, I refer to this as a conditional utility function. If green investors are conditional, then the flow-performance relation of green funds would be stronger than that of conventional counterparts. Lagged positive returns may boost larger fund flows of green funds than that of conventional funds because green investors alter their expectations of fund performance, as would conventional investors, but green investors can add investment to their portfolios to consume the green attribute additionally.
11
Hypothesis 3: The flow-performance relation of green funds is weaker than that of conventional funds.
This hypothesis can be motivated by the assumption that preferences of green funds can be stated by a multi-attribute utility function defined over the moments of a portfolio’s return distribution as well as if the investment decision is green.
Additionally, the utility function is additive in the attributes. The same as the definition by Keeney and Raiffa (1993), which implies that preferences for one attribute is independent of the other attribute. Precisely, the assumption of an additive utility function means that the utility which is derived from the green attribute is separate from and substitutable for the utility which is derived from the risk and return of an investment.
4. Fund Variation in Consideration of Subprime Loan Crisis
This paper will discuss the influence of Subprime Loan Crisis on fund flow volatility of green and conventional funds. As Hamilton (1989) mentioned, the duration of cash flow is much shorter when the market is confronted with severe recessions. Therefore, we can rationally suppose the fund flow volatility will be greater for green and conventional funds during the financial crisis. However, unlike ordinary investors who purchase for superior performance and sell for the sake of averting from financial losses, green investors are eco-friendly people even though
they are also interested in the profit. It is reasonable to assume that green investors will be more reluctant to change their positions of portfolios during the financial crisis, which implies that the green fund flow volatility will be smaller than that of conventional funds in the period of recession.
13
IV. Data
The CRSP Survivor-Bias-Free US Mutual Fund Database is the main resource of mutual fund data. This database contains information about names, managers, the investment style, portfolios and asset allocation of mutual funds from 1962 to September 2009. In terms of SR mutual funds, there are 159 SR funds identified by Social Investment Forum (SIF)4 which offers information about screening standards, fund portfolio, and fund general profile. As for green funds, I collect data from SR fund samples only for those funds which are engaging in positive or restricted investment in the “environment” category. In addition, non-green funds are defined as those funds without screening or investments on the “environment” category. After deleting unsuitable funds which own abnormal returns or lack of first offering date, there are 138 green funds in the end.
The rest of mutual funds in the CRSP database are defined as conventional funds, and the number of conventional funds is 43194. This section includes all available data from CRSP even though some of them are dead funds. Daily returns range from September 1998 to September 2009 while monthly returns range from December 1961 to September 2009.
This paper conducts the fund characteristics analysis with monthly returns from CRSP. Furthermore, the data of size factor, value factor and momentum factor come
4 http://www.socialinvest.org, the Social Investment Forum (SIF) is the US membership organization for professional, companies, and institutions, which are engaging in socially responsible investment. It also provides research information for the public.
from Kenneth R French data library which offers objective and balanced data from CRSP for academic research. In addition, the risk-free rate is the 90-day U.S. Treasury Bill collected from Datastream with the period from March 1990 to March 2010, and Crude Oil price is from Yahoo Finance in the period of September 2006 to December 2009.
Figure 1 exhibits total quarterly financial investment ($ billions) in clean energy from 2004 to 2009. It displays that financial investment in clean power had had steady growth before 2007. Under the condition of severe economic environment, the investment had dropped at a staggering rate after the fourth season in 2007. In the other hand, we can observe there is huge increase in the second season in 2009. The main reason lies in substantial investment in wind energy in China and U.K. In addition, Spain also put large amount of money to solar thermal electricity generation plants at the same time.
Figure 2 shows the comparison of total net assets between green funds and conventional funds. Obviously, the tendency of green funds is quite similar with that of conventional funds even after the breakout of financial crisis. There is an obvious discovery that Subprime Loan Crisis has greatly hampered the growth of total net assets for both green and conventional funds during the period from 2007 to 2008.
Figure 3 depicts the average monthly returns of green funds and conventional funds. By observing this figure, green funds and conventional funds seem to have
15
similar tendency in fluctuation. The average monthly returns of the two groups had dropped dramatically because of the financial crisis in 2007. The return of green funds decreased deeply to -0.15% around 2008. However, the average monthly returns of the two groups have increased gradually after 2008 as the economy recovers in recent years.
V. Methodology and Results
This chapter will describe the methodology and explain empirical results with different perspective, including fund performance, fund flows, fund characteristics, fund flow volatility, OLS regression, and fund variation in consideration of Subprime Loan Crisis for green funds and conventional funds. The formulas mentioned later will follow the steps in Bollen (2007).
1. Fund Performance
This section average the daily returns for green funds and conventional funds from 2000 to 2009, Table 1 5reveals the average of equally-weighted percentage daily returns for ten years. Accordingly, there is no significant difference in performance between green funds and conventional funds. Mallett and Michelson (2009) assert the same viewpoints. After comparing the performance of green funds, SR funds, and index funds, they find there is no real difference between green funds and index funds
5 This paper also averages the daily returns for non-green funds and conventional funds from 2000 to 2009. Readers who are interested in this section can see Appendix 1
as well as green funds and SR funds with parametric and non-parametric tests.
2. Fund Flow
The formula of returns is defined as:
(1)
, where is the net asset value per share and is the distribution amounts per share during the period. As a result, fund flow can be computed by total net assets between time t and t-1, that is:
(2)
, where denote dollar fund flows. We can translate dollar fund flows to percentage through dividing by , which assumes the fund flows occurring at the end of the period. This method is consistent with the assumption in Del Guercio and Tkac(2002), Sirri and Tufano(1998), as well as Barber, Odean, and Zheng(2005). In the other hand, Sirri and Tufano(1998) also compute fund flows with the hypothesis that fund flow happens at the beginning of the period, that is:
(3)
, for simplification, this section chooses (2) formula to analyze. In other words, this paper will use the following equation as the formula for fund flows.
17
(4)
After removing abnormal outliers which lack of sufficient information such as returns or distribution amounts from CRSP, this section average the monthly fund flows computed by the above formula to compare the difference in fund flow volatility between green funds and conventional funds. Figure 4 depicts average fund flow of green funds and conventional funds. When the market is immature, two of the mutual funds are unstable at first. Especially, the fund flows of green funds fluctuate increasingly dramatically than that of conventional funds. However, the two group have had gradually stabilized after 1998 when the industry becomes more mature.
Therefore, whether the industry is mature or not is an important factor in the volatility of mutual funds. The highest average fund flow of green funds is approximately 8.2%
at the beginning while the lowest average fund flow of green funds drops to roughly 0.4% when the market is mature.