Вы находитесь на странице: 1из 18
The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

The current issue and full text archive of this journal is available at

www.emeraldinsight.com/0307-4358.htm

The impact of passive investing on corporate valuations

Eric Belasco

Department of Agricultural Economics and Economics, Montana State University, Bozeman, Montana, USA

Michael Finke

Department of Personal Financial Planning, Texas Tech University, Lubbock, Texas, USA, and

David Nanigian

The American College, Bryn Mawr, Pennsylvania, USA

Abstract

Purpose – The purpose of this paper is to explore the impact of S&P 500 index fund money flow on the valuations of companies that are constituents of the index and those that are not. Design/methodology/approach – To examine the impact of passive investing on corporate valuations, the authors run panel regressions of price-to-earnings ratio on aggregate money flow into S&P 500 index funds and control for various accounting variables that impact price-to-earnings ratio. These regressions involve two samples of stocks. The first sample consists of S&P 500 constituents. The second consists of large-cap stocks that are not constituents of the S&P 500. The authors also run a set of separate regressions with price-to-book ratio rather than price-to-earnings ratio as the dependent variable. Findings – It is found that the valuations of S&P 500 constituents increased by 139 to 167 basis points relative to nonconstituents, depending on valuation metric, due to S&P 500 index fund money flow when evaluated at mean values of money flow and valuation metrics. The valuations of firms within the S&P 500 index respond positively to changes in S&P 500 index fund money flow while the valuations of firms outside the index do not. Additionally, the impact of money flow on valuations persists the month after the flow occurs, suggesting that the impact does not dissipate over time. Practical implications – Mispricings among individual stocks arising from index fund investing may reduce the allocative efficiency of the stock market and distort investors’ performance evaluations of actively managed funds. Originality/value – The paper is the first to explore the long-run relationship between S&P 500 index fund money flow and corporate valuations.

Keywords Investments, Indexing, Fund management, Active management, Passive management, Index fund, Index premium, Demand curves for stocks, S&P 500

Paper type Research paper

JEL classification – G11 (investment decisions), G12 (asset pricing), G23 (private financial institutions) The authors appreciated the feedback given by Ozzy Akay, David Blitzer, Dale Domian, Harold Evensky, Scott Hein, Jason Hsu, Vitali Kalesnik, Masha Rahnama, John Salter, William Waller, and Walter Woerheide, and the excellent research assistance provided by Melissa Cenneno. The authors also thank seminar participants at Texas Tech University (hosted by the Division of Personal Financial Planning) and The American College, as well as participants at the Academy of Financial Services Annual Meeting, the Financial Management Association International Annual Meeting, and the Journal of Indexes Editorial Board Meeting for their comments and suggestions.

The impact

of passive

investing

1067

Received August 2012 Revised January 2012, March 2012 Accepted May 2012

2012 Revised January 2012, March 2012 Accepted May 2012 Managerial Finance Vol. 38 No. 11, 2012

Managerial Finance Vol. 38 No. 11, 2012 pp. 1067-1084 q Emerald Group Publishing Limited

0307-4358

DOI 10.1108/03074351211266793

MF

38,11

1068

I. Introduction Neoclassical asset pricing theory assumes that the price of a stock changes as a result of new information about a corporation’s value to its shareholders. Event studies on changes to the composition of an index are frequently used to examine whether a non-valuation motivated change in the quantity demanded of a given security results in a change in its price. While these studies suggest an increase in valuation resulting from

inclusion in the index, it is possible that the increase accurately reflects greater intrinsic value of the stock from increased liquidity. This study explores the impact of S&P 500 index fund money flow on the valuations of companies that are constituents of the index and those that are not. We find that money flow increases the price-to-earnings and price-to-book ratios of companies that are constituents relative to those that are not. This provides evidence of downward-sloping demand curves for stocks. In a broad review of studies investigating indexation price effects, Brealey (2000) identifies a trend over the period 1966 to 1995 toward larger positive abnormal returns following the addition of a stock to the S&P 500 and larger negative abnormal returns following a deletion of a stock from the index. Petajisto (2011) finds that the average abnormal returns to additions are 8.8 percent in the period 1990 through 2005, and the average abnormal returns to deletions are 2 15.1 percent. The percentage of US equity mutual fund assets invested in index funds increased from 1.0 percent in 1984 to 12.4 percent in 2002 (French, 2008), and more than half of all US equity index fund assets are invested in S&P 500 index funds (Investment Company Institute, 2011). This increase in passive equity investment among S&P 500 firms suggests that the growth of the event effects may be largely due to the shift toward passive investing. The slope of a demand curve for a company’s stock is gauged by the (absolute)

magnitude of its price elasticity of demand, E D . In the stock market setting:

P

ð1Þ

where Q denotes the quantity of shares of a company’s stock demanded by investors and P denotes the price per share of the company’s stock. This concept is shown in Figure 1. Shleifer (1986) and Petajisto (2011) estimate that E D is near unity. This runs contrary to

the predictions of Fama’s (1970) efficient market hypothesis, which predicts that arbitrage trading will correct mispricings among stocks[1]. Arbitrageurs have not flattened the demand curve. Petajisto (2011) also discovers that in the period 1990 through 2005 the average abnormal returns to additions to the Russell 2000 was 8.0 percent and the average abnormal returns to deletions was 2 13.4 percent. Price effects in the purely market cap-based Russell 2,000 that are similar to those observed in the S&P 500 provide further evidence of downward-sloping demand curves for stocks. In a related strand of research, Baberdis et al. (2005) theorize that, in the presence of limited arbitrage, money flowing into a segment of the market impacts the correlation of returns between stocks in that segment. Consistent with the authors’ habitat theory of return comovement, they find that a stock’s correlation with other stocks in the S&P 500 increases when it is added to the index and, commensurate with the trend towards S&P 500 index fund investing, the correlations have increased in recent years. Relatively higher stock ownership by active funds reduces the magnitude of the correlation between stocks and the index; however, index additions result in increased

E D ¼

P

DQ

=

Q

D P = P

P

;

S D 0 D 1 P 1 P 0 Price (per share of stock)
S
D 0
D 1
P
1
P
0
Price (per share of stock)

Quantity (Volume of stock)

Note: This figure shows how the price of a stock responds to a change in the quantity of the stock demanded by market participants when the demand curve is downward-sloping

The impact

of passive

investing

1069

Figure 1. Supply and demand

passive ownership share and a decrease in firm-specific price changes (Ye, 2011). Wermers and Yao (2011) find stronger pricing anomalies, such as those related to momentum and accruals, among stocks with high levels of passive fund ownership, which provides additional evidence that index fund investing impedes the efficiency of capital markets. Increasing correlation among funds within the S&P 500 and pricing anomalies associated with increased passive investment may counteract alternative explanations for the gains from S&P 500 inclusion, including reduced agency costs from increased monitoring by analysts (Yu, 2008). Event studies only partially address price effects associated with index fund investing. This is because investor cash is continuously flowing in and out of index funds, yet changes to the composition of an index only occur a few times a year. Changes in the quantity of a stock demanded by index fund managers as a result of fund flows may also impact the price of stocks if stocks within the index exhibit downward-sloping demand curves. Goetzmann and Massa (2003) examine the relationship between index fund flows and returns on the index and find that a strong same-day relationship exists. However, it is difficult to determine whether the flows are driving the returns or vice versa. To disentangle demand effects from potential feedback effects, the authors perform a Geweke-Messe-Dent (1982) (GMD) test. The GMD test confirms that the direction of causality is from flows to returns. However, the GMD test is weak if the causality is at a higher frequency than the data. This motivated the authors to conduct a series of tests using higher frequency data, which arrive at the same results as the GMD test. Hence, strong evidence suggests that flows drive returns. However, the question of whether or not S&P 500 index fund money flow has inflated the fundamental value of companies that are constituents of the index relative to nonconstituents has not yet been determined.

MF

38,11

1070

We hypothesize that flows into S&P 500 index funds positively impact the PE ratio of companies that are in the index relative to those that are not. In an unpublished manuscript, Morck and Yang (2002) find evidence that money flow into index funds alters the Tobin’s Q ratios. Our empirical work is an extension of Morck and Yang and is different in several ways. First, we perform regressions that involve net cash flows (NCF) rather than time-series plots. Second, we take a comprehensive view of index fund assets by looking at all S&P 500 index funds, including exchange traded funds (ETF). Third, we focus on ratios that are more commonly used by practitioners to gauge valuation. Fourth, we use more recent data. This paper examines the extent to which price multiple measures of valuation are impacted by the trend toward passive investing. To conduct this analysis, we run panel regressions of price-to-earnings ratio (PE) on aggregate NCF into S&P 500 index funds controlling for various accounting variables that impact PE. Additionally, regressions are run with price-to-book ratio (PB), an alternative valuation measure, as the dependent variable[2]. The results from regression models that control for aggregate US equity fund NCF demonstrate that the valuations of constituents increased by 139 to 167 basis points relative to nonconstituents, depending on valuation metric, due to S&P 500 index fund NCF when evaluated at mean value of NCF and valuation metrics. Valuations of firms within the S&P 500 index respond positively to changes in S&P 500 NCF while the values of firms outside the index do not increase during the same time period. Additionally, flow induced price-pressure effects do not appear to dissipate over time.

II. Data To maximize coverage of index fund assets, we incorporate both mutual fund and ETF data into our analysis. Mutual fund and ETF data come from Morningstar Direct. Index and individual stock data come from Compustat. Because Morningstar’s coverage of fund flows begins in February 1993, we control for lagged one-month fund flows, and because Compustat’s price-dividends-earnings file ends in February 2007, the sample period spans March 1993 to February 2007.

A. Description of data on fund flows

Monthly NCFs are gathered from Morningstar Direct’s fund flows database on all US-domiciled open-end mutual funds and ETFs excluding money market funds and funds of funds[3]. We gather two time-series of NCFs. The first consists of all S&P 500 index funds which, based on conversations with Morningstar, is defined as all funds with an institutional category of “S&P 500 Tracking.” The second consists of all US equity funds and is used to control for market-wide fund NCF in our regression analysis; it consists of all funds with a US broad asset class of “US Stock.”

B. Description of data on S&P 500 index valuation

For each month in the sample, information on month-end closing prices, earnings per share, book value per share, and a variable indicating historical S&P 500 index constituency status is gathered from Compustat on all publicly-traded large-cap companies. Following Fama and French (1996), we define large-cap as greater than median NYSE market capitalization and gather monthly market capitalization breakpoints from Kenneth French’s web site. We restrict our analysis to only large-cap companies in order to prevent the “size premia” from biasing our results. We exclude

non-operating establishments, financial services companies (due to non-reporting of working capital data), and companies headquartered outside of the USA (ADRs, ADSs, AMs, and GDRs)[4] from our analysis. We define earnings per share as 12 month moving average quarterly as reported basic earnings per share. We define book value per share as the total value of common equity excluding intangible assets based on fiscal year end data divided by common shares outstanding. Additionally, for each month in the sample, data on various accounting line items are gathered from Compustat to control for return on invested capital (ROIC), free cash flow (FCF) growth, and the internal growth rate of dividends, which are commonly regarded to impact price multiple ratios[5]. To control for seasonality, all income statement and statement of cash flow items are defined as the cumulative value of the associated variables over the prior four-quarters. Balance sheet items are defined as the associated value as of the most recent quarter. Companies lacking four-quarters of continuous data in a given month are excluded from analysis for that period. Line item values that are coded by Compustat as “not meaningful” or “insignificant” are imputed as zero. Following Damodaran (2007), ROIC is calculated as after-tax operating income divided by the book value of lagged four-quarter invested capital, where after-tax operating income is calculated as net income plus after-tax interest expense minus after-tax non-operating income. The applicable tax rate is assumed to be the maximum marginal tax rate in the associated time period. Based on information provided by the Urban Institute and the Brookings Institution (2010), the effective tax rate was 34 percent prior to 1993 and 35 percent from 1993 through 2007. Following Brealey et al. (2006), FCF is calculated as net income plus depreciation and amortization expense minus the change in working capital minus capital expenditures. In the interest of preserving degrees of freedom, FCF Growth is defined

as the percentage change in FCFs between the current and lagged fourth quarter. Following Brealey et al. (2006), the internal growth rate of dividends is calculated as the product of the retention ratio (also known as the plowback ratio) and return on equity. The retention ratio is equal to one minus the indicated annual dollar value of dividends paid to common shareholders divided by income before extraordinary items (adjusted for common stock equivalents). The return on equity is calculated as income before extraordinary items divided by lagged four-quarter book value of equity. To control for the effect of outliers and data entry errors from the data feeds, all accounting variables are winsorized (i.e. truncated) at the 1 percent tails.

The impact

of passive

investing

1071

III. Methods

A series of least-squares regressions is used to analyze the impact of money flow into

S&P 500 index funds on the valuation of individual companies. Our choice of a panel study involving individual companies in contrast to time-series analysis involving portfolios of stocks was driven by two rationales. First, the panel study approach

allows us to better control for effects across the cross-section and time domains. Second, more information is utilized through the panel approach as our results are not limited to a single portfolio or a moment in time. We employ a fixed effects model with a one-way error term in our regression analysis

to account for company-specific individual effects. The use of this model is supported by

the results from the Hausman’s specification test for fixed versus random effects.

MF

38,11

1072

A. Description of regression model To test our hypothesis, we estimate regressions that are variants of the following form:

PE ð i ;t Þ ¼ a ð i ;t Þ þ b 1 £ ROIC ð i ; t Þ þ b 2 £ FCF

growth ð i ;t Þ þ

b 3

£ Expected growth rate of dividends ð i ; t Þ þ b 4 £ Aggregate

ð2Þ

where subscript i denotes the individual stock and subscript t denotes the month. We also run a set of regressions with an alternative dependent variable, PB.

IV. Results Table I presents summary statistics on the main variables of interest. NCFs are reported in billions of dollars. It is interesting to note that S&P 500 index fund NCFs are relatively more volatile than those of the aggregate market of US equity funds. For example, the standard deviation of S&P 500 index fund NCF is 2.25 times greater than its mean value, yet the standard deviation of aggregate US equity fund NCF is only 0.90 times the size of its mean value. The factors that affect aggregate US equity fund NCF, such as aggregate demand for liquidity among market participants, also affect S&P 500 index fund NCF. However, the Pearson correlation coefficient between the two money flow variables is only 0.55. The correlation coefficient is not indicative of multicollinearity, suggesting that the increased popularity of indexing over our sample period, in addition to market-wide factors, contribute to the high volatility of S&P 500

 

Mean

SD

25th percentile

Median

75th percentile

Panel A: net cash flow variables ( T ¼ 168 ) S&P 500 index fund net cash flow US equity fund net cash flow

0.97

2.18

0.14

0.88

2.06

8.22

7.43

4.07

8.44

12.20

Panel B: financial variables constituents sample PE regressions ( n ¼ 622 )

Price-to-earnings ratio Return on invested capital Free cash flow growth Expected growth rate of dividend stream

21.03

39.19

13.95

19.61

28.23

0.16

0.38

0.06

0.12

0.19

(1.23)

4.94

(2.07)

(1.23)

(0.41)

0.17

0.26

0.08

0.16

0.24

Panel C: financial variables nonconstituents sample PE regressions ( n ¼ 1,864 )

Price-to-earnings ratio Return on invested capital Free cash flow growth Expected growth rate of dividend stream

22.20

47.13

13.47

20.69

32.04

0.15

0.38

0.06

0.12

0.19

(1.23)

5.79

(2.34)

(1.40)

(0.38)

0.15

0.31

0.08

0.16

0.25

Panel D: financial variables constituents sample PB regressions ( n ¼ 619 )

Price-to-book ratio Return on invested capital Free cash flow growth Expected growth rate of dividend stream

4.25

4.30

2.03

3.07

4.91

0.16

0.39

0.06

0.12

0.19

(1.23)

4.97

(2.08)

(1.24)

(0.41)

0.16

0.24

0.07

0.16

0.24

Panel E: financial variables nonconstituents sample PB regressions ( n ¼ 1,851 )

Price-to-book ratio Return on invested capital Free cash flow growth Expected growth rate of dividend stream

4.73

5.15

2.14

3.30

5.47

0.15

0.38

0.06

0.12

0.19

(1.24)

5.80

(2.35)

(1.41)

(0.39)

0.15

0.29

0.07

0.16

0.24

Table I.

Notes: This table reports summary statistics of regression variables; net cash flow variables are

Summary statistics

reported in billions of dollars

index fund NCF relative to its mean. The heterogeneity is shown in the time plots of the two series in Figure 2. It is also interesting to note that the ROIC, FCF growth, and expected growth rate of dividends of S&P 500 constituents are similar to those of nonconstituents.

A. Description of results from the analysis of price-to-earnings ratio Table II presents the results from equations that relate PE ratio to the explanatory variables. The odd columns present results from regressions involving the sample of S&P 500 index constituents. The even columns present results from regressions involving the sample of large-cap stocks that are not constituents of the S&P 500 index. Column (5) shows that the PE ratios of constituents are positively and significantly (at the 1 percent level) impacted by S&P 500 index fund NCF. The regression coefficient is economically meaningful as well. The overall mean of PE was 21.03 and the time-series mean of S&P 500 index fund NCF was $0.97 billion. When evaluated at these average values, S&P 500 index fund NCF corresponds to a 1.7 percent increase in the value of constituents:

0:37 £ 0:97

21:03

¼ 1:7%:

The impact

of passive

investing

1073

The results from a regression of the same model on the sample of nonconstituents is conveyed in column (6). The results from the sample of nonconstituents are not significant, perhaps due to S&P 500 index fund NCF subsuming the effect of aggregate US equity fund NCF. Therefore, regressions which control for aggregate US equity fund NCF are run. The results from these regressions for constituents and nonconstituents are conveyed in

40,000,000,000

30,000,000,000

20,000,000,000

10,000,000,000

0

(10,000,000,000)

(20,000,000,000)

(30,000,000,000)

(40,000,000,000)

Aggregate US equity fund NCF S&P 500 index fund NCF Mar-93 Mar-94 Mar-95 Mar-96 Mar-97
Aggregate US equity fund NCF
S&P 500 index fund NCF
Mar-93
Mar-94
Mar-95
Mar-96
Mar-97
Mar-98
Mar-99
Mar-00
Mar-01
Mar-02
Mar-03
Mar-04
Mar-05
Mar-06

Notes: This figure plots the time-series of aggregate NCF into US equity funds and into the subset of such funds that Morningstar categorizes as having an objective of tracking the S&P 500 Index; the sample period spans March 1993-February 2007

Figure 2. Time plot of money flows

even-numbered columns report results from the sample of stocks outside of the S&P 500 index with greater than median market capitalizations; net cash

flow variables are reported in billions of dollars; coefficients are estimated using the fixed effects model; standard errors are given below the regression

Nonconstituents

Notes: Significant at: * p , 0.10, * * p , 0.05 and *** p , 0.01; odd-numbered columns report results from the sample of stocks in the S&P 500 index;

14.66 ***

0.10 ***

6.07 ***

0.06 * *

(0.15) *

31.81 *

(0.12)

(0.08)

(0.03)

(0.03)

(0.09)

(0.71)

(0.03)

(0.52)

(16.35)

0.30

0.01

(8)

1,864

Constituents

47.46 ***

0.12 ***

2.80 ***

0.25 ***

0.07 ***

14.70 ***

0.22 * *

(0.08)

(0.03)

(0.03)

(0.08)

(0.78)

(0.03)

(0.46)

(0.04)

(11.03)

(7)

0.14

622

Nonconstituents

6.01 ***

14.69 ***

31.95 *

(0.03)

(0.03)

(0.07)

(0.07)

(0.71)

(0.52)

(0.03)

(16.35)

0.30

0.01

(6)

1,864

Constituents

0.37 ***

48.17 ***

2.72 ***

0.36 ***

14.84 ***

(0.07)

(0.07)

(0.03)

(0.46)

(0.78)

(0.04)

(11.03)

(5)

0.14

622

Nonconstituents

6.10 ***

14.65 ***

0.08 ***

0.05 * *

31.78 *

(0.02)

(0.02)

(0.29)

(0.52)

(0.71)

(16.35)

1,864 0.30

0.01

(4)

Constituents

2.74 ***

14.68 ***

47.19 ***

***

***

0.09 (0.02)

0.14 (0.02)

(0.78)

(0.46)

(0.03)

(0.04)

(11.03)

(3)

622 0.14

Nonconstituents

6.02 ***

14.68 ***

31.94 *

(0.52)

(0.03)

(0.71)

(16.35)

1,864 0.30

0.01

(2)

Constituents

2.57 ***

47.97 ***

14.85 ***

(0.46)

(0.03)

(0.78)

(0.04)

(11.04)

(1)

622 0.14

index

Return on invested capital Free cash flow growth

Expected growth rate of dividend stream Aggregate net cash flow Lagged aggregate net cash flow

index fund

flow

cash 500

500 flow

Observations

R 2

net S&P

coefficients

Adjusted

Intercept

S&P cash

Lagged

fund

net

MF

38,11

1074

Table II. Price-to-earnings ratio regression results (March 1993-February 2007)

columns (7) and (8), respectively. After controlling for aggregate US equity fund NCF, the negative impact of S&P 500 index fund NCF on the PE ratio of nonconstituents becomes statistically significant at the 10 percent level. The difference in the regression coefficients between the constituent and nonconstituent samples are quantitatively unchanged relative to the models that do not control for aggregate US equity fund NCF. It is also interesting to note that fund flows do not explain much of the total variation in corporate valuations, as the introduction of fund flow variables into the regression models (columns (5) through (8)) does not improve adjusted R 2 values or substantially alter the value of the regression coefficients associated with the valuation control variables. The results from these regressions fail to show that NCF into S&P 500 index funds explains a substantial amount of the total variation of corporate valuations. However, they do show that money flow into S&P 500 index funds leads to considerable changes in the value of companies in the index relative to those outside of the index. This central finding is consistent with the hypothesis. A related question is whether or not price-pressure effects dissipate over time. To address this, we examine the coefficients on lagged S&P 500 index fund NCF. If market participants correct the flow induced price-pressure effects, then we would observe a reversal or at least a reduction in the difference in the coefficients on lagged S&P 500 index fund NCF between the constituent and nonconstituent samples relative to the difference in the coefficients on contemporaneous S&P 500 index fund NCF. However, this is not apparent in the data. The differences in the coefficients on the lagged and contemporaneous values in the regressions that do not control for aggregate US equity fund NCF ((5) and (6)) are 0.39 and 0.40, respectively. The difference in the coefficients on the lagged and contemporaneous values for S&P 500 index fund NCF in the regressions that do control for aggregate NCF ((7) and (8)) are both 0.37. The difference in the coefficients on lagged S&P 500 index fund NCF between the constituent and nonconstituent samples are similar to the difference in the coefficients on contemporaneous S&P 500 index fund NCF. Moreover, the standard errors of the coefficients on the lagged values are not greater than those associated with the contemporaneous values. The results from this analysis do not show that an adjustment process exists.

The impact

of passive

investing

1075

B. Description of results from the analysis of price-to-book ratio Table III presents the results from equations that relate PB ratio to the explanatory variables. The results pertaining to this alternative valuation metric are similar to those obtained through the use of price-to-earnings ratio. While S&P 500 index fund NCF does not explain much of the total variation in corporate valuations, it does positively impact on the valuation of constituents relative to nonconstituents. When evaluated at the overall average value of PB ratio (4.25), the PB ratio of constituents increased by a statistically significant (at the 1 percent level) 2.1 percent due to S&P 500 index fund NCF when evaluated at its time-series mean ($0.97 billion). The PB ratio of nonconstituents also increased but only by 0.6 percent when evaluated at mean levels. We attribute the slight increase in the value of nonconstituents to aggregate US equity fund NCF, which were omitted from this set of regressions. After controlling for aggregate US equity fund NCF, when evaluated at the overall average of PB ratio, the PB ratio of constituents increased by a statistically significant (at the 1 percent level) 1.6 percent while the value of nonconstituents was quantitatively unchanged.

even-numbered columns report results from the sample of stocks outside of the S&P 500 index with greater than median market capitalizations; net cash

flow variables are reported in billions of dollars; coefficients are estimated using the fixed effects model; standard errors are given below the regression

Nonconstituents

Notes: Significant at: * p , 0.10, * * p , 0.05 and *** p , 0.01; odd-numbered columns report results from the sample of stocks in the S&P 500 index;

0.01 ***

0.62 ***

0.02 ***

2.10 ***

0.01 ***

(0.02) * *

5.81 * *

(0.01)

(0.00)

(0.00)

(0.01)

(0.00)

(2.29)

(0.04)

(0.06)

0.62

0.01

(8)

1,851

Constituents

5.92 ***

0.02 ***

0.07 ***

9.74 ***

0.02 ***

0.12 ***

0.06 ***

(0.00)

(0.01)

(0.00)

(0.00)

(0.01)

(0.85)

(0.00)

(0.04)

(0.07)

(7)

0.57

619

Nonconstituents

6.06 ***

0.03 ***

2.11 ***

0.01 ***

0.61 ***

(0.01)

(0.01)

(0.06)

(0.04)

(2.29)

(0.00)

0.62

0.01

(6)

1,851

Constituents

5.95 ***

0.10 ***

0.09 ***

0.09 ***

9.87 ***

(0.01)

(0.01)

(0.00)

(0.00)

(0.04)

(0.07)

(0.86)

(5)

0.57

619

Nonconstituents

0.02 ***

0.62 ***

0.01 ***

2.10 ***

0.01 ***

5.83 * *

(0.00)

(0.00)

(0.00)

(0.06)

(0.04)

(2.29)

1,851 0.62

(4)

Constituents

9.66 ***

5.91 ***

***

***

0.10 ***

0.03 (0.00)

0.02 (0.00)

(0.00)

(0.00)

(0.07)

(0.04)

(0.85)

(3)

619 0.57

Nonconstituents

0.01 ***

6.01 ***

2.11 ***

0.60 ***

(0.06)

(0.04)

(0.00)

(2.29)

1,851 0.62

(2)

Constituents

9.82 ***

***

*

(0.86)

0.06 (0.03)

5.95 (0.07)

(0.00)

(0.00)

(1)

619 0.57

index

Return on invested capital Free cash flow growth

Expected growth rate of dividend stream Aggregate net cash flow Lagged aggregate net cash flow S&P 500 index fund net cash flow

flow

cash 500

Observations

R 2

net S&P

coefficients

Adjusted

Intercept

Lagged

fund

MF

38,11

1076

Table III. Price-to-book ratio regression results (March 1993-February 2007)

To examine the possibility that flow induced price-pressure effects dissipate over time, we turn to the coefficient on the lagged value of S&P 500 index fund NCF. Interestingly, the difference in the coefficient on lagged S&P 500 index fund NCF between the constituent and nonconstituents samples ((5) and (6)) is 0.08 yet the difference in the coefficient on the contemporaneous values between the two samples is 0.06. The same phenomena is observed among the regressions that control for aggregate US equity fund NCF ((7) and (8)). While the slightly greater impact of lagged S&P 500 index fund NCF than contemporaneous S&P 500 index fund NCF on valuations is somewhat perplexing, it clearly provides no support for the idea that the price-pressure effects dissipate over time. Lastly, we acknowledge the concern that our results may be partially an artifact of the price effects associated with changes to the S&P 500 index. Evidence of asymmetric price responses upon changes to index constituency status, detailed in Chen et al. (2004), and competing explanations for the price effect validate this concern. To examine how sensitive our results are to the event effects, we run an additional set of regressions in which we exclude companies from our samples if they were either added to or deleted from the S&P 500 index over the prior two months. The results obtained from these regressions were quantitatively unchanged.

The impact

of passive

investing

1077

C. Subperiod analysis

Tables IV and VI provide subperiod results for March 1993 through February 2000 for price-to-earnings ratio and price-to-book ratio, respectively. Tables V and VII provide subperiod results for March 2000 through February 2007 for price-to-earnings ratio and price-to-book ratio, respectively. Although the impact of NCF on valuation is statistically significant and positive among constituent firms in both sub-periods, the effect is more pronounced between 1993 and 2000. This may be due to heavier inflows into index funds during this period (the fraction of total US equity mutual fund assets tracking the S&P 500 increased by 151 percent between 1993 and 2000 versus 29 percent between 2000 and 2006). The more pronounced effect in the first subperiod may also be due to the increased popularity of day trading over that period (Tables VI and VII). As in the full period analysis, the subperiod analysis also does not provide much evidence to show that fund money flow explains much of the total variation in corporate valuations but the analysis also does not show that flow induced valuation effects dissipate over time.

V. Concluding remarks

A. Summary of research findings

This study is the first to explore the long-run relationship between S&P 500 index fund money flow and corporate valuations. Through a series of panel regressions, we examine whether money flow into S&P 500 index funds impacts the price-to-earnings ratio and price-to-book ratio of firms within and outside the index. The results are consistent with the hypothesis that money flow into S&P 500 index funds positively impacts the price multiples of companies that are in the index relative to those that are not. The impact on valuations persists the month after the flow occurs, suggesting that the impact does not dissipate over time. These results are consistent with the theory

Notes: Significant at: * p , 0.10, * * p , 0.05 and *** p , 0.01; odd-numbered columns report results from the sample of stocks in the S&P 500 index; even-numbered columns report results from the sample of stocks outside of the S&P 500 index with greater than median market capitalizations; net cash flow variables are reported in billions of dollars; coefficients are estimated using the fixed effects model; standard errors are given below the regression

Nonconstituents

(0.74) ***

0.18 ***

19.56 ***

0.80 ***

0.12 ***

0.18 ***

(51.41) *

(0.05)

(0.06)

(0.27)

(0.23)

(1.15)

(0.04)

(0.22)

(26.69)

0.04

0.40

(8)

1,335

Constituents

18.37 ***

0.22 ***

0.81 ***

0.41 *

(0.05)

(0.01)

(0.06)

(0.06)

(0.29)

(0.25)

(1.54)

(0.22)

(0.05)

(24.98)

(7)

0.18

9.92

0.21

493

Nonconstituents

0.12 ***

0.74 ***

19.61 ***

(46.47) *

(0.12)

(0.24)

(0.22)

(0.21)

(0.04)

(1.15)

(26.69)

0.40

0.15

(6)

1,335

Constituents

0.95 ***

18.54 ***

0.39 *

0.47 *

(0.02)

(0.25)

(0.24)

(0.05)

(0.22)

(1.54)

(24.97)

(5)

12.84

0.21

493

Nonconstituents

19.61 ***

0.83 ***

0.15 ***

0.12 ***

0.11 * *

(51.42) *

(0.05)

(0.05)

(0.04)

(0.21)

(1.15)

(26.69)

1,335 0.40

(4)

Constituents

18.28 ***

0.29 ***

0.36 *

(0.01)

(0.06)

(0.05)

(0.05)

(1.54)

(0.22)

(24.98)

(3)

0.04

9.04

493 0.21

Nonconstituents

0.12 ***

0.74 ***

19.61 ***

(46.79) *

(0.04)

(0.21)

(1.15)

(26.68)

1,335 0.40

(2)

Constituents

18.60 ***

(0.05)

(1.54)

0.20 (0.21)

(0.02)

(24.98)

(1)

15.15

493 0.21

index

Return on invested capital Free cash flow Growth

flow aggregate net

Expected growth rate of dividend stream Aggregate net cash flow

S&P 500 index fund

flow

cash 500

net cash flow

Observations

R 2

net S&P

coefficients

Adjusted

Intercept

Lagged

Lagged

fund

cash

MF

38,11

1078

Table IV. Price-to-earnings ratio regression results (March 1993-February 2000)

Notes: Significant at: * p , 0.10, * * p , 0.05 and *** p , 0.01; odd-numbered columns report results from the sample of stocks in the S&P 500 index; even-numbered columns report results from the sample of stocks outside of the S&P 500 index with greater than median market capitalizations; net cash flow variables are reported in billions of dollars; coefficients are estimated using the fixed effects model; standard errors are given below the regression

Nonconstituents

(0.25) ***

2.25 ***

12.23 ***

0.02 ***

0.02 ***

0.03 * *

(0.01)

(0.02)

(0.00)

(0.00)

(0.02)

(0.02)

(0.09)

(1.96)

(0.00)

0.00

0.73

(8)

1,324

Constituents

0.20 ***

0.23 ***

0.02 ***

3.24 ***

0.01 ***

0.07 ***

(0.02)

(0.00)

(0.00)

(0.02)

(0.10)

(1.58)

(0.00)

(0.01)

(0.00)

(7)

1.45

0.71

491

Nonconstituents

(0.18) ***

12.81 ***

2.26 ***

(0.02)

(0.02)

(0.02)

(0.00)

(0.09)

(1.97)

0.00

0.03

0.73

0.01

(6)

1,324

Constituents

0.25 ***

0.25 ***

3.26 ***

0.07 ***

(0.02)

(0.02)

(1.58)

(0.10)

(0.01)

(0.00)

(0.00)

(5)

1.84

0.71

491

Nonconstituents

2.29 ***

12.24 ***

0.04 ***

0.01 * *

(0.01)

(0.00)

(0.00)

(1.97)

(0.00)

(0.09)

(0.02)

0.00

1,324 0.73

(4)

Constituents

0.05 ***

0.04 ***

0.05 ***

3.20 ***

(0.00)

(0.00)

(0.10)

(0.01)

(0.00)

(1.60)

(0.00)

(3)

0.79

491 0.71

Nonconstituents

12.33 ***

2.28 ***

0.04 ***

(1.97)

(0.02)

(0.09)

(0.00)

0.00

1,324 0.73

(2)

Constituents

3.30 ***

(0.00)

0.00 (0.01)

(0.10)

(1.62)

(0.00)

(1)

2.44

491 0.70

index

Return on invested capital Free cash flow growth

flow aggregate net

Expected growth rate of dividend stream Aggregate net cash flow

index fund

flow

cash 500

500 flow

Observations

R 2

net S&P

coefficients

Adjusted

Intercept

S&P cash

Lagged

Lagged

fund

cash

net

The impact

of passive

investing

1079

Table V. Price-to-book ratio regression results (March 1993-February 2000)

even-numbered columns report results from the sample of stocks outside of the S&P 500 index with greater than median market capitalizations; net cash

flow variables are reported in billions of dollars; coefficients are estimated using the fixed effects model; standard errors are given below the regression

Nonconstituents

Notes: Significant at: * p , 0.10, * * p , 0.05 and *** p , 0.01; odd-numbered columns report results from the sample of stocks in the S&P 500 index;

0.08 ***

43.17 ***

(0.18) * *

33.73 * *

(0.03)

(0.03)

(0.09)

(0.09)

(0.07)

(1.00)

(0.04)

(2.44)

(0.01)

(15.87)

1.29

0.03

0.33

(8)

1,164

Constituents

39.82 ***

38.36 ***

0.07 ***

2.73 * *

0.06 * *

(0.03)

(0.03)

(0.09)

(0.09)

(1.19)

(0.05)

(2.64)

(0.06)

(10.76)

(7)

0.14

0.18

0.03

476

Nonconstituents

43.34 ***

33.85 * *

(0.09)

(0.05)

(0.08)

(0.08)

(1.00)

(2.44)

(0.04)

(0.01)

(15.87)

0.33

1.25

(6)

1,164

Constituents

38.72 ***

40.18 ***

0.20 * *

2.73 * *

(0.08)

(0.08)

(1.19)

(2.64)

(0.05)

(0.06)

(10.76)

(5)

0.18

0.13

476

Nonconstituents

43.36 ***

0.05 * *

33.72 * *

(0.03)

(0.03)

(1.00)

(0.04)

(2.44)

(0.01)

(15.87)

1.27

0.03

1,164 0.33

(4)

Constituents

38.31 ***

39.68 ***

***

***

2.74 * *

0.07 (0.02)

0.08 (0.02)

(1.19)

(0.05)

(2.64)

(0.06)

(10.76)

(3)

476 0.18

Nonconstituents

43.43 ***

33.83 * *

(1.00)

(2.44)

(0.04)

(0.01)

(15.87)

1,164 0.33

1.25

(2)

Constituents

38.70 ***

39.95 ***

2.75 * *

(1.19)

(0.05)

(2.64)

(0.06)

(10.77)

(1)

476 0.18

index

flow growth

of dividend stream Aggregate net cash flow Lagged aggregate net

rate

index fund

Return on invested capital

flow

Free cash growth

cash 500

500 flow

Observations

R 2

net S&P

coefficients

cash flow

Expected

Adjusted

Intercept

S&P cash

Lagged

fund

net

MF

38,11

1080

Table VI. Price-to-earnings ratio regression results (March 2000-February 2007)

even-numbered columns report results from the sample of stocks outside of the S&P 500 index with greater than median market capitalizations; net cash

flow variables are reported in billions of dollars, coefficients are estimated using the fixed effects model; standard errors are given below the regression

Nonconstituents

Notes: Significant at: * p , 0.10, * * p , 0.05 and *** p , 0.01; odd-numbered columns report results from the sample of stocks in the S&P 500 index;

(0.03) ***

0.02 ***

1.01 ***

5.83 ***

2.21 ***

0.01 ***

0.02 ***

(0.01)

(0.00)

(0.00)

(0.01)

(0.09)

(2.01)

(0.19)

(0.00)

0.67

0.01

(8)

1,173

Constituents

0.01 ***

0.03 ***

0.02 ***

4.41 ***

9.51 ***

2.33 ***

0.01 *

(0.00)

(0.01)

(0.00)

(0.00)

(0.01)

(0.23)

(0.00)

(0.12)

(0.84)

(7)

0.59

477

Nonconstituents

1.00 ***

2.26 ***

0.02 ***

6.00 ***

0.01 * *

0.01 * *

(0.01)

(0.01)

(0.19)

(2.02)

(0.09)

(0.00)

0.66

(6)

1,173

Constituents

0.04 ***

0.04 ***

9.59 ***

2.43 ***

4.40 ***

(0.01)

(0.01)

(0.00)

(0.84)

(0.23)

(0.00)

(0.12)

(5)

0.59

477

Nonconstituents

5.89 ***

0.02 ***

0.01 ***

2.23 ***

1.01 ***

0.01 ***

(0.00)

(0.00)

(0.19)

(0.00)

(0.09)

(2.01)

1,173 0.67

(4)

Constituents

***

***

9.50 ***

2.28 ***

4.42 ***

(0.00)

0.02 (0.00)

0.02 (0.00)

(0.00)

(0.23)

(0.12)

(0.84)

(3)

477 0.59

Nonconstituents

6.01 ***

1.00 ***

2.26 ***

0.02 ***

(0.00)

(0.19)

(2.02)

(0.09)

1,173 0.66

(2)

Constituents

9.59 ***

***

***

(0.85)

2.35 (0.23)

4.41 (0.12)

(0.00)

(0.00)

(1)

477 0.59

index

Return on invested capital Free cash flow growth

Expected growth rate of dividend stream Aggregate net cash flow Lagged aggregate net cash flow S&P 500 index fund net cash flow

flow

cash 500

Observations

R 2

net S&P

coefficients

Adjusted

Intercept

Lagged

fund

The impact

of passive

investing

1081

Table VII. Price-to-book ratio regression results (March 2000-February 2007)

MF

38,11

1082

of downward-sloping (i.e. not horizontal) demand curves for stocks and suggest that money flow from index funds can distort stock prices.

B. Practical implications of research findings Based on the results from our empirical study, it appears that the preference shift towards index fund investing is reducing the informational efficiency of stock prices. Informed investors may recognize the oversupply of capital allocated to stocks in indices and then place trades which counteract the effect (Grossman and Stiglitz, 1980)[6]. However, the speed of adjustment back to equilibrium valuations will be slow in the presence of inattentive investors (Duffie, 2010). By their nature, index fund investors are inattentive to asset valuations and, as described in De Long et al. (1990) and Shleifer and Vishny (1997), arbitragers (and perhaps most importantly those who provide them with capital) are rather impatient[7]. Prior theoretical work implies that, in the specific setting examined in this study, the preference shift towards index fund investing is an endogenous determinant of the speed of adjustment and as a result intertemporal arbitrage opportunities stemming from the shift will be unattractive. Moreover, until the preference shift abates, attempting to arbitrage the mispricing away may drown those informed traders swimming against the current of passive investment. This is because, as discussed in De Long et al. (1990), the impact of noise trader risk on asset prices is increasing in the proportion of noise traders in the market. This implies that the distortive effect of index fund investing on stock prices may be intensified by the perverse tendency of many active fund managers to “closet index” (Cremers and Petajisto, 2009). Mispricing among equities within the index may have adverse implications, including a reduction of the allocative efficiency of the stock market and investors’ performance evaluations of actively managed funds. The flow of non-valuation motivated passive funds may also reduce oversight among firms within the S&P 500, even if index inclusion increases the number of analysts providing information to investors about firm governance (Yu, 2008). These and other economic consequences are discussed in Wurgler (2010), Woolley and Bird (2003), Wermers and Yao (2010) and Ye (2011). This study adds to the literature on the importance of passive investing on valuations by linking flows into index funds with the valuations of companies that are constituents of the index relative to those that are not.

Notes

1. See, for example, Shleifer (2000) for a discussion of this hypothesis.

2. Some advantages of PB ratio are that it provides more meaningful valuation estimates in the event that earnings are negative and book values tend to be less volatile than earnings. However, it is important to bear in mind that one downside of book value is that it does not consider human capital as an operating factor. Additionally, it is difficult to compare companies in different industries based on PB ratios due to differing levels of hard asset intensity across industries.

3. Details on the methods employed by Morningstar to estimate net cash flows can be found at: http://corporate.morningstar.com/us/documents/Direct/INS_MDT_ EstimatedNetFlowsMethodology.pdf

a Standard Industrial Classification code ranging from 6,000 to 7,000. ADRs, ADSs, AMs, and GDRs are detected through the company name (i.e. – ADR in a company name indicates that the security is an ADR).

5. For a discussion of how these factors impact valuations, see Koller et al. (2005) and Damodaran (2007).

6. Alexander et al. (2007) find that valuation-motivated trades by mutual fund managers outperform non-valuation motivated trades, and that this outperformance was greater in the more recent period (1992-2003) than it had been between 1980-1991. The increased magnitude of the effect over a period characterized by a preference shift towards passive investing lends some credence to Grossman and Stiglitz’s (1980) theory.

7. De Long et al. (1990) suggest that this is because investors’ evaluations of professional money managers often occur over short time intervals. Gromb and Vayanos (2010) provide a detailed review of the literature on the limits of arbitrage.

References Alexander, G.J., Cici, G. and Gibson, S. (2007), “Does motivation matter when assessing trade performance? An analysis of mutual funds”, Review of Financial Studies , Vol. 20 No. 1, pp. 125-50.

Baberdis, N., Shleifer, A. and Wurgler, J. (2005), “Comovement”, Journal of Financial Economics , Vol. 75 No. 2, pp. 283-317. Brealey, R. (2000), “Stock prices, stock indexes and index funds”, Bank of England Quarterly Bulletin , February, pp. 61-8.

Brealey,

McGraw-Hill/Irwin, New York, NY. Chen, H.H., Noronha, G. and Singal, V. (2004), “The price response to S&P 500 index additions and deletions: evidence of asymmetry and a new explanation”, Journal of Finance , Vol. 59 No. 4, pp. 1901-29. Damodaran, A. (2007), “Return on capital (ROC), return on invested capital (ROIC) and return on equity (ROE): measurement and implications”, New York University Stern School of Business Working Paper, available at: http://ssrn.com/abstract¼1105499 De Long, J.B., Shleifer, A., Summers, L.H. and Waldmann, R. (1990), “Noise trader risk in financial markets”, Journal of Political Economy, Vol. 98 No. 4, pp. 703-38. Duffie, D. (2010), “Presidential address: asset price dynamics with slow-moving capital”, The Journal of Finance , Vol. 65 No. 4, pp. 1237-67. Fama, E. (1970), “Efficient capital markets: a review of theory and empirical work”, The Journal of Finance, Vol. 25 No. 1, pp. 383-417. French, K.R. (2008), “Presidential address: the cost of active investing”, Journal of Finance , Vol. 63 No. 4, pp. 1537-73. Geweke, J., Meese, R. and Dent, W. (1982), “Comparing alternative tests of causality in temporal systems: analytic results and experimental evidence”, Journal of Econometrics , Vol. 21 No. 2, pp. 161-94. Goetzmann, W.N. and Massa, M. (2003), “Index funds and stock market growth”, Journal of Business , Vol. 76 No. 1, pp. 1-28. Gromb, D. and Vayanos, D. (2010), “Limits of arbitrage”, in Lo, A.W. and Merton, R.C. (Eds), Annual Review of Financial Economics , Vol. 2, Annual Reviews, Palo Alto, CA, pp. 251-75.

R.,

Allen,

F.

and

Myers,

S.

(2006), Principles

of Corporate Finance, 8th ed.,

The impact

of passive

investing

1083

MF

38,11

1084

Grossman, S. and Stiglitz, J. (1980), “On the impossibility of informationally efficient markets”, The American Economic Review , Vol. 70 No. 3, pp. 393-408. Investment Company Institute (2011), 2011 Investment Company Fact Book , 51st ed., The Investment Company Institute, Washington, DC, pp. 1-236, available at: www.ici.org/pdf/

2011_factbook.pdf

Morck, R. and Yang, F. (2002), “The mysterious growing value of S&P 500 membership”, National Bureau of Economic Research Working Paper 8654. Petajisto, A. (2011), “The index premium and its hidden cost for index funds”, Journal of Empirical Finance, Vol. 18 No. 2, pp. 271-88. Petersen, M. and Rajan, R. (1997), “Trade credit: theories and evidence”, Review of Financial Studies, Vol. 10 No. 3, pp. 661-91. Shleifer, A. (1986), “Do demand curves for stocks slope down?”, Journal of Finance, Vol. 41 No. 3, pp. 579-90. Shleifer, A. (2000), Inefficient Markets: An Introduction to Behavioral Finance , Oxford University Press, New York, NY, p. 2. Shleifer, A. and Vishny, R.W. (1997), “The limits of arbitrage”, Journal of Finance , Vol. 52 No. 1, pp. 35-55. Urban Institute and the Brookings Institution (2010), Historical Top Bracket and Rate Tax , available at: www.taxpolicycenter.org/taxfacts/displayafact.cfm?Docid¼65& Topic2id¼70 (accessed January 15, 2010). Wermers, R. and Yao, T. (2010), “Active vs passive investing and the efficiency of individual stock prices”, working paper. Woolley, P. and Bird, R. (2003), “Economic implications of passive investing”, Journal of Asset Management, Vol. 3 No. 4, pp. 303-12. Wurgler, J. (2010), “On the economic consequences of index-linked investing”, in Allen, W.T., Khurana, R., Lorsch, J. and Rosenfeld, G. (Eds), Challenges to Business in the Twenty-first Century: The Way Forward , American Academy of Arts and Sciences, Cambridge. Ye, P. (2011), “The value of active investing: can active institutional investors remove excess comovement of stock returns?”, Journal of Financial and Quantitative Analysis , May 1. Yu, F. (2008), “Analyst coverage and earnings management”, Journal of Financial Economics , Vol. 88, pp. 245-71.

Corresponding author David Nanigian can be contacted at: david.nanigian@theamericancollege.edu

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com Or visit our web site for further details: www.emeraldinsight.com/reprints