when they hire or fire asset managers in the institutional market. ...... We design a second measure to model a preliminary screening step where a sponsor.
Journal of Business & Economic Studies, Vol. 13, No. 1, Spring 2007 Why Do Institutional Plan Sponsors Hire and Fire Their Investment Managers? Jeffrey Heisler, Gottex Fund Management Christopher R. Knittel, University of California at Davis and NBER John J. Neumann, St. John’s University Scott D. Stewart, Boston University Abstract ______________________________________________________________________________ This study draws inferences about investment decisions by plan sponsors who oversee over $6 trillion in institutional investment assets by examining asset and account flows for activelymanaged U.S. equity products. Analysis reveals an expected role for benchmarks – primarily the S&P500 – and a curious role for total returns over short and long-term horizons in asset flow allocations. Sponsors punish for one-year losses, and screen on consistency, but not magnitude, of positive or negative active return over time. Style benchmarks are more prominent in decisions to move accounts, which are found to involve more criteria, possibly reflecting a higher hurdle decision requirement. ______________________________________________________________________________ INTRODUCTION Institutional plan sponsors, who allocate taxable corporate or tax-exempt endowment, pension, or foundation assets, have received little attention in the academic literature compared to the numerous studies of retail (individual) mutual fund investors. The behavior of these institutional investors, however, is important both due to the size of the institutional market and its potential impact on asset prices as of December 2000, institutional equity and bond funds held total assets of $6,646 billion compared to $4,770 billion for retail equity and bond mutual funds – and to the differences in their investment knowledge and the environment in which they function. Plan sponsors are typically financial professionals who possess knowledge of advanced techniques for evaluating manager skill and performance, or have access to consultants with this knowledge. In many cases, the individual with ultimate responsibility for an institutional investment is the CEO, CFO or Treasurer of a corporation. Lakonishok, Shleifer, and Vishny (1992) describe the agency relationships within which plan sponsors make asset manager selections, and argue that an environment which has them being responsive to multiple monitors – the oversight committee or senior management, the investors in the plan, including company shareholders, and perhaps even the external portfolio managers – may steer them into hiring asset managers or pursuing investment strategies that are designed to reduce job risk. With these differences in skill sets and the environments in which investment decisions are made, it is not unreasonable to expect that a profile of plan sponsor decisions will differ from that which has evolved in the literature about individual investors and mutual funds. LITERATURE REVIEW Prior research on the retail market largely focuses on the relationship between mutual fund performance, both prior and contemporaneous, and the flow of assets between funds. These results suggest that while, on average, retail investors tend to direct money to mutual funds with
88
Journal of Business & Economic Studies, Vol. 13, No. 1, Spring 2007 positive short-term total returns1 and positive short-term excess returns,2 they are less likely to withdraw assets from funds with poor short-term performance.3 In addition, older funds are found to grow their asset bases proportionately slower than younger funds.4 5 Del Guercio and Tkac (2002) take an initial step in contrasting the retail and institutional asset management markets by examining asset flows using a mutual fund sample and a pension fund sample. Among the differences they find between the samples: simply outperforming the S&P 500 market index is more important to plan sponsors than to individuals, while the magnitude of outperformance matters more to individuals; tracking error is significant in explaining flows in the pension fund sample but not in the mutual fund sample; risk-adjusted returns in the form of Jensen’s are related to flows in both samples, although the authors quite rightly suspect that mutual fund investors are probably not looking at such numbers and demonstrate correlation between Morningstar ratings and Jensen’s measure.6 They also find that the quantitative performance factors explain more variation in flows among mutual funds than pension funds, that there is autocorrelation in mutual fund flows but not in pension fund flows, and that the asymmetric (convex) flow-performance relation in the mutual fund market documented by Sirri and Tufano (1998) and others is more linear and symmetric among pension fund managers. This last finding suggests that institutional investors may be more willing to fire a manager that delivers poor performance than has been found to be the case among mutual fund investors. While not reporting results, they state examining changes in the number of client totals leads to similar conclusions. This article seeks to build upon the work in Del Guercio and Tkac (2002) to further develop the body of knowledge about the decision-making process followed by plan sponsors when they hire or fire asset managers in the institutional market. There remain many outstanding questions regarding the level of sophistication of institutional investors, including their investment horizon, their appreciation of manager style and sensitivity to performance consistency. Do they chase performance differently from retail investors? Do they truly adjust for risk or simply prefer performance figures which look attractive in a table? We examine how measures of fund performance and fund attributes affect the allocation of new money (and clients) and the reallocation of existing money (and clients) among institutional investment products. In so doing, we make inferences about the extent to which plan sponsors and their consultants, assumed to possess greater financial and investment management savvy than the average retail investor, rely on past performance, and attempt to identify which specific past performance measures play the more prominent role in their decisions. We address issues of benchmark use, investment horizon, and consistency of performance, and examine them from two perspectives. First, following all prior studies, we use asset flows among investment products to make inferences about plan sponsor decisions. To do this, we introduce a new measure of flows. But we also recognize that asset flows are actually an imperfect indicator of a hire or fire decision, since a sponsor can re-allocate all of a plan’s assets away from one manager to another (a “full” firing), or only a portion of its assets away from one manager to another – perhaps a “partial firing” that could also be an adjustment to “fix” a deviation from the plan’s stated target allocation objectives. Outflows could also simply represent withdrawals to meet a plan’s operational cash needs. Further, asset flows can be skewed by a few sponsors moving large amounts of assets with only a handful of hire/fire decisions. When an asset manager loses an entire account, however, it is very likely that the sponsor has fired the manager. Thus, our tests using accounts data – the number of accounts invested in each investment product – can be
89
Journal of Business & Economic Studies, Vol. 13, No. 1, Spring 2007 used both to robustness check our asset flows results and to respond to the question of whether different criteria are in play when a sponsor decides to move an entire account versus the multiple decisions that are collected together in asset flows. The dearth of research in the institutional area invites these analyses for practical considerations. Since asset flows impact security prices, it is important for investors and economists to understand the process institutional investors follow in reallocating their assets. Considering the amount of assets under management in this market, and the fact that some institutional products have twin retail mutual funds or, at the very least, are selecting from the same pool of securities, an understanding of the extent to which plan sponsor and individual investor decisions reinforce or temper each other is enhanced with a more complete understanding of the institutional market. These questions are also of interest for the investment industry, as most fund manager and fund company compensation, in the form of management fees, is based on the asset base managed (although there are often client-specific fee arrangements in the institutional market). Finally, this article extends the earlier 1987-1994 sample period to include 1984-2000. With the unusual performance of the financial markets from 1995 – 2000 and such potentially important financial implications arising from this area of research, it is worthwhile to confirm earlier results as well as grow the literature in this area. Reassuringly, our results are consistent with those presented by Del Guercio and Tkac (2002) on high-level questions comparing pension fund and mutual fund investment decisions. We find evidence that plan sponsors screen for managers who beat benchmarks while not necessarily considering the magnitude by which they beat them. We find low r2’s in our asset flows tests, control variables (product attributes) that explain as much or more variation in asset flows than do the quantitative performance variables, and replicate specific findings on two of these control variables: marginally significant positive autocorrelation of asset flows, negative relation between fund (investment product) size and asset flows. Lastly, we also find no evidence of the asymmetric flow-performance relationship documented in the literature which describes how mutual fund (retail) investors are more reluctant to withdraw money from poor performing funds than they are to invest it with good performers. A synthesis of the retail mutual fund literature, limited prior research on institutional investors, and firsthand anecdotal experience in the institutional money management industry raise additional important issues regarding plan sponsor investment decisions that have yet to be studied. All suggest that both total and excess return performance is pertinent in explaining asset flows between investment products. Retail investors appear to focus on recent performance, being drawn to mutual funds that have posted strong recent returns or have received publicity via appearance on a “top ten” list.7 Conversely, prudent man rules, professionalism and the longterm nature of pension, endowment, and foundation investment horizons all argue for plan sponsors to resist the lure of short-term performance and incorporate a longer horizon track record into their screening process for selecting managers. Longer horizons should lead to lower turnover, and in turn, trading volume, influencing securities prices differently than retail investors. This best intention may be mitigated, however, by the agency relationships which have them reporting to investors and an oversight committee, one or both of whom may create pressure to respond to recent performance. To identify the relevant horizon(s) in plan sponsor decisions, we incorporate total and excess return factors over 3-year and 5-year horizons, along with the prior year’s total and excess returns, into our tests. While it is expected that plan
90
Journal of Business & Economic Studies, Vol. 13, No. 1, Spring 2007 sponsors know better than to look at total (raw) returns given their financial training, we include them in the tests to detect evidence of unexpected behavior or outside influence from a monitor who probably does not share their level of expertise. Another important area for study is plan sponsor recognition of management style. Excess return is typically calculated relative to broad market indexes such as the S&P 500 and CRSP value-weighted index in studies of retail investor behavior, but this would represent a very basic screen for institutional sponsors who in practice either purchase outright or hire consultant software databases to research investment managers. Most pension plans, endowments and foundations set broad asset target weights and hire specialist managers specifically for the investment style objective their product follows, creating a portfolio of managers and a portfolio of investment styles. The proper evaluation of a manager’s performance, then, should use a benchmark that reflects the original hiring criteria for the manager and/or the stated strategy of the investment product and, perhaps more finely, the extent or degree to which the product pursued that strategy. To examine this issue, we employ four benchmarks in our tests: the S&P 500 along with three style benchmarks. The first of these style benchmarks is based on the style reported by the investment manager offering the product. The second is a simple style indicator based on the product’s “actual style” as measured by its correlation with Russell 1000 style indices over the preceding 24 quarters. The Russell indexes are widely published and considered to be the most popular industry benchmark beyond the S&P 500 for institutional domestic equity managers. The third benchmark is calculated to reflect not only the product’s style, but the degree to which the style was pursued via its beta exposure to style indices. This measure reflects the extremeness of manager style and differentiates between a “core” style manager versus a “deep” style manager versus a “mild” style manager. While the research databases can generate these style-exposure measures and it is expected that sponsors may use them to select managers, the fact that they require industry expertise to understand would make them difficult to explain to the sponsor’s monitors and so may mitigate their use in ex-post manager evaluation. Accordingly, the S&P 500 index is included in the tests, much as raw returns are included, as the null alternative – as the naïve alternative industry experts would prefer not to use but may be steered into using due to the involvement and reporting requirements to monitors who do not possess their same subject matter expertise. Stewart (1998) shows that the consistency with which a manager generates active returns should be a prime criterion in the plan sponsor screening process. Consistency is defined as the frequency, over short assessment periods within the evaluation period, with which the manager generated positive excess returns. In general, institutional investors prefer more consistent performance because it increases the likelihood of good long-term results, provides lower noise levels versus plan targets, and makes it easier for sponsor professionals to report this performance to their superiors. This suggests that the evaluation of performance may be path dependent, where investors differentiate between products that post identical active performance based on how this performance was achieved and may prefer the product that generates lower annualized, but more stable, growth. We combine this idea of consistency with our interest in investment horizon to test the path to the five-year excess return result through its one and threeyear interim assessment points. This measure has not been incorporated into earlier studies and should also shed additional light on the relative importance of simply over or under performing the benchmark and the magnitude of that over or under performance.
91
Journal of Business & Economic Studies, Vol. 13, No. 1, Spring 2007 The results support the importance of return consistency as a determinant for manager selection. The consistency with which managers deliver positive or negative active returns over multiple interim measurement horizons, without regard to the magnitude of these returns, plays a key role in determining the flow of assets among investment products. Active return consistency is based largely on the S&P 500, with a supplemental role for style benchmarks that do not represent the degree of style exposure. This suggests that plan sponsors are construing gains above the S&P 500 as skill, when in fact these gains may be due to style choice. This also raises the possibility of gaming by the fund managers, who are aware of their ability to manage their level of style exposure and have flexibility to do so if they are not being held accountable for this risk when evaluated by plan sponsors.8 Magnitude of 1-year loss and 3 and 5-year total returns are incremental factors in plan sponsors’ asset flow allocation decisions, while 5-year benchmark-based returns are incremental factors in their account movement decisions.9 Thus, we find that while institutional investors consider long-term returns in addition to short-term results and generally employ benchmarks in their evaluations, as expected, they can be swayed by total returns. Most intriguing is their sensitivity to one-year total return losses. Unlike retail investors but consistent with the pension fund analysis in Del Guercio and Tkac (2002), institutional investors seem willing to withdraw assets from poorly performing managers. Aside from quantitative performance criteria, sponsors may rely on qualitative features when selecting their asset managers. For example, if they are more comfortable with, and find it easier to defend the selection of, a manager with a longer and more-established track record, we expect fund age to have a positive relationship with asset flows and accounts. This would be especially true with accounts, since moving an entire account is a more drastic action than merely reallocating a portion of the plan’s assets. In addition, if plan sponsors value customer service and a personal relationship with their manager, or are predisposed to believe that a manager has a better chance of performing well with fewer assets, product size should have a negative relationship with flows. Our results are consistent with these expectations.10 If the sponsors expect out-performance of one manager relative to others to persist, we should see a positive relationship between current and lagged asset and account flows. We do find a positive and significant relationship for asset flows but a significant negative relationship for account flows. One explanation for this negative relationship is regression to the mean -- an account once gained can only be lost, and a manager that gains (loses) above average accounts regresses to the mean in the next period, producing a negative flows coefficient. METHODOLOGY Data The dataset of institutional managers and their products comes from the PSN Investment Manager Database compiled by Effron Enterprises Inc. This database provides historical information on over 7000 investment products, including annual summary information about each product and quarterly assets under management and performance data. This information is self-reported by the product managers. It is used by investment product managers for performance comparison to their peers and by plan sponsors and consultants to identify candidate investment managers.
92
Journal of Business & Economic Studies, Vol. 13, No. 1, Spring 2007 This article focuses on active domestic equity funds.11 These products constitute approximately 60% of the entire universe. While product performance information is available starting in 1979, assets under management figures are first available in 1984. Therefore, new asset flows and products returns are calculated beginning in 1985, and the analysis of annual flow behavior begins in 1989 to allow for a five-year lagged return calculation. Our test samples using asset flows have a maximum of 8,515 product-year observations and 6,969 product-year observations using account flows. Model Plan sponsors hire and fire investment managers. A direct measure of this decision would require knowledge of plan sponsors’ holdings, information which was unavailable. We therefore proxy plan sponsors’ hire/fire decisions by the relative changes in assets under management by the investment products and the number of client accounts invested in the products. The model estimates the relationship of asset and account flows to the product’s return, return consistency, and attributes: (1)
Asset Flowsi,t = f(
Returni,t- ,
(2)
Account Flowsi,t = g(
Returni,t- ,
Return Consistencyi,t- , Attributesi,t-1) +
i,t
Return Consistencyi,t- , Attributesi,t-1) +
i,t
The model is estimated using fixed-effects regression. Though both the asset and account flows data are unbalanced panel sets, there still is the possibility of cross-sectional or serial correlation. The latter is the larger concern because if it exists, the multiple observations for an investment product over time are not independent. Such correlation could arise from the overlap in the longer three-year and five-year return horizons, or from static structural features of the investment product. The fixed-effects control for unobserved features of a particular fund that are largely constant over time; example, the level of customer service or some other manager quality. Flows Asset Flows Asset flows are typically expressed as the change in assets adjusted for the return over the period of change: (3)
Dollar Flowsi,t = Assetsi,t – Assetsi,t-1(1 + Ri,t)
or as the percentage change in assets relative to the product’s beginning of year assets:
93
Journal of Business & Economic Studies, Vol. 13, No. 1, Spring 2007
(4)
Percentage Flowsi, t
=
Assets i, t - Assets i, t -1 (1 R i, t ) Assets i, t -1
where Ri,t is the return of product i in year t.12 However, asset flows do not necessarily indicate a hire/fire decision. Assets flows arise from normal-course-of-business withdrawals or deposits, net flows into the product’s asset class or style strategy, or institutional plan adjustments when their asset allocations deviate too significantly from target allocations. While one assumption could be that large net asset flows, in proportion to a product’s asset base, represents a hiring or firing decision, this requires a potentially arbitrary decision as to what represents a large change in assets. Further, unless the model controls for size this measure will suggest a negative relationship between size and flows.13 To address these issues, an alternative measure of assets flows is developed. We measure a product’s flows as the change in assets in proportion to all funds “on the move” within the equity industry that year. Funds “on the move” is the sum of the absolute value of all products’ dollar flows in that year. For a specific product, this measures the percentage of aggregate equity flow activity captured (or lost) by that product in that year. Scaling flows in each year this way also removes the need to control for year-by-year differences in aggregate flows, eliminating the need for year-style interaction variables. If there is a relationship between performance and flows, products with relatively better performance over some time horizon should capture a larger portion of the money entering the market or being reallocated by plan sponsors. This measure of captured flows for a product i is:
(5)
Asset Flowsi,t = j
Assetsi, t - Assetsi, t - 1 (1 R i, t ) | Assets j, t - Assets j, t - 1 (1 R j, t ) |
Account Flows There remains a limitation with asset flows in that a few large plan sponsors can distort the results, moving large amounts of assets with a few hire/fire decisions possibly based on criteria different from the average sponsor. Further, asset flows could also indicate re-allocation decisions to move some, but not all, assets from one manager to another. By some sponsors, this decision could be a firing; by others, it could be allocation maintenance relative to targets. Since sponsors tend to hold only one account with each product, an alternative approach to measuring the hire/fire decision is to examine the change in the number of accounts held by each product. While the PSN database contains information on the number of accounts invested with each product, changes in the number of accounts does not control for accounts gained from or lost to other product types. A product can gain (lose) an account from (to) a product within the equity market or from (to) a product outside of the equity market (a fixed income product, for example). To address these issues, and to provide a perspective of scale relative to both the product and the equity industry similar to what captured flows does, we first determine the number of accounts for the average equity product in the PSN database. We then calculate the year-by-year changes
94
Journal of Business & Economic Studies, Vol. 13, No. 1, Spring 2007 in the number of accounts for this average equity product and for each individual equity product. A product’s account flows is the difference between the proportional change in the number of its client accounts and the proportional change in the number of client accounts for the average equity product: (6)
Account Flowsi, t
A i, t
Ai, t
A i, t
1
1
At
At At
1
1
where Ai,t is the number of accounts for product i at time t and At is the average number of accounts per equity product at time t. Product Return We measure product returns five ways; the product’s total return, ri,t, excess return relative to the S&P 500, ri,t – rSP,t, and excess returns relative to a style-adjusted benchmark based on either the product’s self-reported style, ri,t – rSR,t, a style indicator variable, ri,t – rSI,t, or a style exposure variable, ri,t – rSE,t. Style-adjusted returns are calculated using the Russell 1000 Value and Russell 1000 Growth indexes, based on their common use as largecap benchmarks within the industry. We include a test for an asymmetric reaction to positive and negative performance, modeled using an < 0 interaction dummy variable that takes the value 1 if the total return is negative or the return difference between the product and benchmark is negative. The product’s self-reported style is available in the PSN database only for the last year of our sample – 2000. Tests based on self-reported style, therefore, assume that products do not change investment styles over the sample.14 A self-reported value product is benchmarked to the Russell 1000 Value index, while a self-reported growth product is benchmarked to the Russell 1000 Growth index. We benchmark all other products to the Russell 1000 index. To control for changing investment styles, potential style drift and a product’s style exposure, we develop two style measures: a style indicator variable and a style exposure variable. These style variables are based on the product’s sensitivity to the Russell 1000 Value and Russell 1000 Growth indexes as estimated from regressions using quarterly returns over the preceding six years, starting in 1979. Products with fewer than 20 quarterly observations or with adjusted-r2 less than 0.50 are discarded: (7)
Ri,t-1,t-24 =
i
+
G,iRRus1000G,t-1,t-24
+
V,iRRus1000V,t-1,t-24
+ ei,t-1,t-24
Style Indicator The style indicator variable categorizes a product as simply growth, value or something in between. If only the growth index coefficient estimated from the regressions is significant, the product is designated as a growth product and assigned a style indicator of 0. If the value index coefficient alone is significant, the product is designated as a value product and assigned a style indicator of 1. Products where both coefficients are significant are assigned an indicator between 0 and 1 calculated as a weighted-average coefficient estimate:
95
Journal of Business & Economic Studies, Vol. 13, No. 1, Spring 2007
(8)
V ,i ,t
Style Indicatori,t = V,i, t
G,i, t
Table 1. Distribution of Style Indicator This table reports the distribution of style indicators over the sample period. Product i is designated as growth (value), and assigned an indicator of 0 (1), if its estimated sensitivity to the Russell 1000 Growth (Value) index, G,i,t ( V,i,t), using bivariate regressions for the 6-year period t is significant. Products where both coefficients are significant are assigned a style indicator calculated as: Style Indicatori,t =
V ,i,t V, i, t
G, i, t
Products with less than 20 quarterly return observations or an adjusted-r2
Period
Obs 0
1979-1984 1980-1985 1981-1986 1982-1987 1983-1988 1984-1989 1985-1990 1986-1991 1987-1992 1988-1993 1989-1994 1990-1995 1991-1996 1992-1997 1993-1998 1994-1999 1995-2000
316 366 424 537 606 709 865 1062 1198 1296 1473 1606 1438 1365 1976 2260 2248
0
