Mutual Fund Intermediation, Equity Issues, and the Real Economy Bill Y. Zu University of Melbourne Australian Centre for Financial Studies, Funds Management RRG October 2014
Motivation Rapid expansion of the mutual fund industry Mixed evidence of ability to beat benchmarks (Elton and Gruber 2013) Does the real economy benefit from institutional investors? (Greenwood and Scharfstein 2013) There can be two possible channels: Secondary market: improve price efficiency and capital allocation (Boehmer and Kelley 2009, Piacentino 2013) This paper: Primary market as financiers of equity capital Emphasis on mutual funds: Largest group of institutional investor, delegated by retail investors Retail flows can easily move in and out of the fund sector
Mutual funds as informed financiers Existing literature on mutual funds: Superior performance in the primary market (IPO & SEO) (Gibson, Safieddine and Sonti 2004; Chemmanur, He and Hu 2009) Screening ability alleviates information costs (Greenwood and Jovanovic 1990; Holmstrom and Tirole 1997) Simple theoretical framework: Firms issue equity to finance real investments, produce output Mutual funds have informational advantages over households In equilibrium, fund investment conveys information about productivity and predicts output
Empirical findings Aggregate and industry-level: Mutual fund investment in new equity issues predicts real output New issues in low-participation times predict lower output, while issues in high-participation times predict higher output Lower fund flows leads to higher participation-output sensitivity flight-to-quality Effect on cyclical component of output Results are driven by active funds rather than passive funds At the firm level: Issues with higher mutual fund participation exhibit higher productivity growth
A description of the model Production: Firms with different productivity issue equity to invest Representative mutual fund: Informed: can identify firm type at a cost (diminishing returns) Invests some of their fund flows in primary market Representative household: Uninformed: cannot differentiate between firms; has own beliefs about productivity Allocates between MF and direct primary markets In equilibrium: Mutual fund investment in new issues positively predict output Output is decreasing in the interaction of fund investment and fund flows Decreasing fund flows increases the effect of fund investment on output
How do we measure fund investment in new issues? Do not have direct allocations use holdings Investment in new issues = Total issues participation rate from mutual funds Participation rate is proxied by change in holdings Data source: Thomson Reuters s12 quarterly mutual fund holding filings Participation j is the change in holdings of the equity-issuing firm j by all mutual funds, scaled by new shares issued Average over all issues in each quarter/year PART i,t dummy variable for above/below median participation
Other data Investment Company Institute (ICI): aggregate fund flows, 1984-2011 SDC: IPOs and SEOs in the U.S. CRSP/Compustat Bureau of Economic Analysis: real GDP, real industry value-added, deflators, non-financial business assets (proxy for W) Federal Reserve: credit issuance, Survey of Professional Forecasters Hodrick-Prescott (1997) filter for cyclical component
Regression results Aggregate Output Industry Value-added GDP Growth Cyclical VA Growth VA Cyclical Flow t 1 2.8991 0.2999** 1.1770 0.0026 (1.11) (2.53) (1.13) (0.07) Issue t 1-8.4938-0.8798*** 1.2864-1.1039*** (-1.49) (-2.64) (1.17) (-3.98) PART t 1-0.0042* -0.0004*** -0.0120-0.0014** (-1.73) (-2.96) (-1.07) (-2.15) Issue PART (f t 1) 6.6818 0.5377** 1.3525* 1.7154*** (1.41) (2.47) (1.74) (3.14) ft 1 2-0.0044-0.0002 0.4318-2.7251** (-1.03) (-1.26) (0.03) (-2.61) Flow t 1 f t 1 0.0749 0.0058 0.1107 0.0160 (0.71) (1.30) (0.38) (0.84) Flow t 1 PART t 1-2.1787-0.2025** 0.5053-0.1588** (-0.95) (-2.39) (0.40) (-2.14) Flow t 1 Issue t 1-0.0000-0.0000-0.0000* 0.0000 (-0.26) (-0.83) (-1.83) (1.44)
Firm productivity Can mutual funds screen out the productive firms? Dependent variable: Post-SEO five-year productivity growth TFP i,t+5 RevGrth i,t+5 AvgROA i,t+5 Issue i,t 0.0364 0.3448*** -0.0641 (0.48) (2.63) (-1.59) PART i,t -0.1066*** -0.0411 0.0271** (-3.04) (-1.05) (2.14) Issue i,t PART i,t 0.2543** 0.3434*** -0.0532* (2.29) (2.81) (-1.69) + Controls
Robustness: reverse causality? Potential alternative explanation Households invest more in MF when perceived productivity is high (Jank, 2012) If households are correct on average, MF investment is correlated to output, even when MFs do not screen Low correlation between participation rate and fund flows (0.16 in aggregate, 0.002 at industry level) Active funds v.s. passive funds Expect effects to come from active funds only Identify passive funds by name Results are driven by active funds passive funds do not have any effect
Summary and implications First paper to link mutual fund investment to the real economy Screening ability allows funds to pick productive equity issues Higher mutual fund investment predicts higher real cyclical output Lower fund flows hamper funds ability to intermediate Heterogeneity in mutual funds more effect from active funds Positive role of mutual funds in the real economy Through efficient capital allocation Without altering the investment opportunity set
Potential future work... Real output v.s. stock returns Aggregate issues predict lower stock returns (Pontiff and Woodgate 2008) Long-run underperformance is more pronounced when institutions increase their holdings around the issue (Edelen, Ince and Kadlec 2013) Maybe due to change in risk following new equity issues (Eckbo, Masulis and Norli 2000) Financial development and economic growth Cross-country growth regressions (a la King and Levine 1993) Use banking sector as proxies for financial development, and stock market liquidity and size as proxies for stock market development This paper: size of the market matters less than who is in the market Strategic decisions for firms to issue equity (Chiu and Kini 2013)