Structural VAR Approach to Mutual Fund Cash Flows: Net Flows, Inflows, and Outflows

• Author: Bong Soo Lee,Miyoun Paek,Yeonjeong Ha,Kwangsoo Ko
• Published date: 01 February 2015
• Date: 01 February 2015
• DOI: http://doi.org/10.1111/ajfs.12081

Structural VAR Approach to Mutual Fund

Cash Flows: Net Flows, Inflows, and

Outflows

Yeonjeong Ha

Department of Economics, Pusan National University

Bong Soo Lee*

College of Business, Florida State University

Miyoun Paek

Department of Business Administration, Pusan National University

Kwangsoo Ko

Department of Business Administration, Pusan National University

Received 17 September 2014; Accepted 23 December 2014

Abstract

In a dynamic structural VAR framework, we investigate how mutual fund cash flows respond

to market volatilities, market returns, and fund returns, using the data of inflows and out-

flows obtained from Form N-SAR filings with the SEC in the EDGAR system. We find that

market volatility (market return) shocks have contemporaneous negative (positive) effects on

net flows. Fund return shocks have a significant effect on net flows for more than 64% of

sample funds in each fund style group. Logistic regression shows the importance of diverse

fund characteristics. Disposition effect of fund investors depends also on fund characteristics.

Keywords Mutual fund cash flows; Structural VAR approach; Volatility timing; Disposition

effect

JEL Classification: G10, G11

1. Introduction

A number of studies have investigated the relation between fund returns a nd net

flows. At an aggregate market level, Warther (1995) and Edelen and Warner (2001)

analyze this relation using monthly and daily frequency data, respectively. At an

individual fund level, Chevalier and Ellison (1997) and Sirri and Tufano (1998) find

*Corresponding author: Bong Soo Lee, College of Business, Florida State University, 311

Rovette Building, Tallahassee, FL 32306-1110, USA. Tel: +1-850-644-4713, Fax: +1-850-644-

4225, email: blee2@cob.fsu.edu.

Asia-Pacific Journal of Financial Studies (2015) 44, 59–87 doi:10.1111/ajfs.12081

©2015 Korean Securities Association 59

a positive but asymmetric (or nonlinear) relation between lagged fund returns and

net flows. Edelen (1999) shows a negative relation between abnormal returns and

investor flows. Lynch and Musto (2003) develop a model to explain the nonlinear

or convex relation between past returns and net flows for mutua l funds. Spiegel

and Zhang (2013) show that this nonlinear relation disappears when the market

share of equity fund is used as a flow measure.

Cao et al. (2008) study the dynamic relation among market return, market vola-

tility, and aggregate mutual fund flow using United States daily data, which is nota-

ble in that they stress the importance of market volatility when analyzing fund flows.

They show that market volatility is negatively related to concurrent and lagg ed fund

flows. Ederington and Golubeva (2011) also find that aggregate net equity fund flows

are strongly negatively correlated with changes in VIX and positively with returns,

which implies that buy decisions are driven mainly by returns and sale decisions by

risk perceptions. Recently, Ben-Rephael et al. (2012) investigated aggregate monthly

net exchanges of equity funds, which are positively concurrently correlated with

excess stock market returns and negatively with changes in VIX.

At an individual fund level, Rakowski and Wang (2009) examine the relation

between fund returns and flows by estimating a reduced-form vector auto-regres-

sion (henceforth, VAR) model based on daily frequency. However, they do not con-

sider the concurrent relation.

1

Cashman et al. (2012a,b) further analyze the relation

between mutual fund flows and performance by decomposing net flows into inflows

and outflows. Recently, Johnson (2010) showed that both new and old shareholders

buy funds during periods of good returns, however, their fund sales are not related

to fund returns. These studies suggest that there exist dynamic relations among

market returns, market volatilities, fund returns, and cash flows. In particular, cash

inflows and outflows have been rarely investigated in the same manner, which

implies that net flow analysis fails to understand the determinants of fund cash

flows. In this respect, however, few studies investigate such dynamic relations at an

individual fund level.

We investigate the dynamic relations between market volatilities, market returns,

fund returns, and cash flows by estimating a structural VAR model, using not only

net flows but also inflows and outflows. Our approach differs from previous studies

in that the structural VAR model considers concurrent relations as well as dynamic

relations. Since our structural VAR model is a just-identified model, the results are

not dependent on the ordering of variables as in the conventional VAR model that

uses the Choleski decomposition. We focus on identifying the determinants of

mutual fund cash flows within a dynamic structural VAR framework and provide

new empirical evidence using monthly mutual fund cash flows (i.e. net flows,

inflows, and outflows) from Form N-SAR filings with the SEC in the Electronic

Data Gathering, Analysis, and Retrieval (EDGAR) database.

1

In a subsequent study, Rakowski (2010) finds a negative concurrent relation between fund

returns and the volatility of flows.

B. S. Lee et al.

60 ©2015 Korean Securities Association