Dividend Volatility and Asset Prices: A Loss Aversion/Narrow Framing ...

Dividend Volatility and Asset Prices: A Loss Aversion/Narrow Framing Approach Yan Li and Liyan Yang∗

Abstract This paper documents that the aggregate dividend growth rate exhibits strong volatility clustering. We incorporate this feature into a theoretical model and study the role of dividend volatility in asset prices when investors are loss-averse over fluctuations in the value of their financial wealth. We find that our model explains many salient features of the stock market, including the high mean, excess volatility, and predictability of stock returns; the low correlation between consumption growth and stock returns; time-varying Sharpe ratios; the GARCH effect and the volatility feedback effect in stock returns; and the decline of equity premiums in the postwar period. JEL classification: G11, G12 Keywords: dividend volatility clustering, loss aversion, narrow framing, asset pricing

∗

Yan Li: Department of Finance, Fox School of Business, Temple University, Philadelphia, PA 19122; Email: [email protected]; Tel: 215-204-4148. Liyan Yang: Department of Finance, Joseph L. Rotman School of Management, University of Toronto, 105 St. George Street, Toronto, Ontario M5S 3E6; Email: [email protected]; Tel: 416-978-3930. We thank Nick Barberis, David Easley, Ming Huang, Maureen O’Hara and participants in 2009 China International Finance Conference and 2010 Northern Finance Association Meetings for helpful suggestions. All errors are our responsibility.

1

Dividend Volatility and Asset Prices: A Loss Aversion/Narrow Framing Approach

Abstract This paper documents that the aggregate dividend growth rate exhibits strong volatility clustering. We incorporate this feature into a theoretical model and study the role of dividend volatility in asset prices when investors are loss-averse over ‡uctuations in the value of their …nancial wealth. We …nd that our model explains many salient features of the stock market, including the high mean, excess volatility, and predictability of stock returns; the low correlation between consumption growth and stock returns; time-varying Sharpe ratios; the GARCH e¤ect and the volatility feedback e¤ect in stock returns; and the decline of equity premiums in the postwar period. JEL classi…cation: G11, G12 Keywords: dividend volatility clustering, loss aversion, narrow framing, asset pricing

1

1

Introduction

How does aggregate dividend volatility a¤ect asset prices?1 Despite the fact that the dividend is one of the most important …nancial variables, there has been little research on this subject largely because, in standard consumption-based models, the pricing kernel is entirely driven by consumption. Thus, it is the covariance between consumption and dividends, not the volatility of dividends, that determines stock prices.2 By contrast, the role of dividend volatility has been highlighted by the “loss aversion/narrow framing approach” advocated in the recent behavioral …nance literature.3 As can be seen from its name, the key elements of this approach are loss aversion and narrow framing, two of the most important …ndings from experimental literature on decisionmaking under uncertainty. In the context of …nancial markets, loss aversion means that investors are more sensitive to losses in their …nancial investments than to gains of the same magnitude. Narrow framing states that investors tend to isolate the stock market risk from their overall wealth risk (such as labor income risk and housing risk) and derive utility directly from changes in the value of their stock holdings. Since dividend volatility drives ‡uctuations in the value of their …nancial wealth, investors may perceive aggregate dividend volatility rather than consumption volatility, the commonly used measure in the 1

Throughout the paper, the term “dividend volatility”refers to the standard deviation of the growth rate of (not the level of) the aggregate dividends paid to all stocks. See equation (3) for a technical de…nition. 2 To the best of our knowledge, the only exception is Longsta¤ and Piazzesi (2004), who demonstrate that volatile and procyclical dividends can raise equity premiums in a representative agent model with power utility. However, they limit their discussion to the equity premium and do not explore whether dividend volatility can help explain other puzzling facts in the aggregate stock market, such as excess volatility and return predictability. 3 See Barberis and Huang (2007) for a comprehensive survey of models adopting the loss aversion/narrow framing approach.

2

literature, as a metric for representing risk. This immediately implies that dividend volatility has signi…cant implications for asset prices. Barberis and Huang (2007, p200) summarize by stating that models adopting the loss aversion/narrow framing approach “can generate a large equity premium and a low and stable risk-free rate, even when consumption growth is smooth and only weakly correlated with the stock market.” To date, studies adopting the loss aversion/narrow framing approach have assumed that the dividend growth rate process is independent and identically distributed (i.i.d.) over time to re‡ect the fact that dividend growth is di¢ cult to forecast at the mean level. But is the dividend growth rate truly an i.i.d. process? Is it possible that the dividend growth rate is forecastable at higher moments, such as volatility? If yes, what are the pricing implications of this forecastability in dividend volatility? In this paper, we empirically investigate the data and …nd strong evidence that dividend volatility indeed exhibits strong persistence (usually called volatility clustering) indicating the tendency of a large (small) change today to be followed by a large (small) change tomorrow. We then incorporate dividend volatility clustering into a standard model featuring loss aversion/narrow framing and …nd that this simple extension helps to explain many salient features of the aggregate market. Speci…cally, in addition to the low and stable risk-free rate, high equity premium, and low correlation between consumption and stock returns that can be generated from a standard model adopting the loss aversion/narrow framing approach with an assumed i.i.d. dividend growth rate,4 our extended model can explain the excess volatility and predictability of stock 4

For a review of a standard loss aversion/narrow framing model, see Barberis and Huang (2007). In fact, our model will be reduced to a standard model adopting the loss aversion/narrow framing approach if we remove the time varying feature in the dividend process, i.e., if we set u to be 0 in equation (4).

3

returns, the time-varying Sharpe ratios, the GARCH and volatility feedbacks e¤ect in stock returns, and the declining equity premium in the postwar period. The intuition of our new results is as follows. Suppose that the dividend volatility, which is the state variable in our model, is high this period. Given the persistence in dividend volatility, the investor forecasts highly volatile dividends in the next period; this implies highly volatile returns in the next period. Since she is loss-averse, she …nds this prospect scary and requires a higher equity premium. This pushes stock prices down, generating excess volatility. The same mechanism also produces predictability in stock returns: at moments when the investor forecasts especially high return volatility, she pushes current stock prices down, generating a low price-dividend ratio. But these are also moments when she requires a high equity premium; the price-dividend ratio therefore predicts returns. In addition, the conditional mean and conditional standard deviation of expected returns are driven di¤erently by dividend volatility; hence, the Sharpe ratio, as a measure of the price of risk, changes over time. Under the loss aversion/narrow framing approach, stock returns are ultimately driven by dividend news. Thus, they inherit its property of volatility clustering. This explains the well-documented volatility clustering e¤ect, also known as the GARCH e¤ect, in stock returns.5 Moreover, our model provides a simple preference-based explanation for the volatility feedback e¤ect (e.g., French, Schwert and Stambaugh, 1987; Campbell and Hentschel, 1992): an increase in future stock return volatility leads to a decrease in current stock return. In our 5

Volatility clustering is called the GARCH e¤ect because it is typically modeled using generalized autoregressive conditional heteroscedasticity proposed by Bollerslev (1986).

4

model, stock volatility is priced because the loss-averse investor does not like uncertainty; therefore, an anticipated increase in volatility would raise the required rate of return and lower current prices. Last but not least, the declining dividend volatility in the postwar period also helps to explain the decline in ex-ante equity premium, which has been documented in many studies (e.g., Blanchard, 1993; Fama and French, 2002; Buranavityawut, Freeman and Freeman, 2006). Our model generates the three results predicted by a standard loss aversion/narrow framing model through the same mechanism. Speci…cally, since the risk-free rate is determined by a smooth consumption process, it is stable and its level is also low for empirically plausible values of the intertemporal rate of substitution. Given that our model generates volatile stock prices, loss aversion is responsible for the high unconditional equity premium. In addition, unlike the consumption-based models, stock returns are ultimately driven by dividend news, which has a low correlation with consumption news; thus, the correlation between consumption and stock returns is low in our economy. Our paper is most closely related to the literature that relies on the loss aversion/narrow framing approach to explain aggregate stock market phenomena. Benartzi and Thaler (1995) are the …rst to incorporate loss aversion and narrow framing in a theoretical model to study the equity premium puzzle. Barberis, Huang, and Santos (2001) study the joint properties of stock returns and consumption growth by incorporating dynamic features of loss aversion, usually labeled the “house money e¤ect”, meaning that people are more (less) willing to bear risks when they have had prior gains (losses).6 Our study di¤ers from that of Barberis, 6

Unlike loss aversion itself, the idea that loss aversion changes over time is less well-supported by experi-

5

Huang, and Santos (2001) in three ways: (i) while their model makes two psychological assumptions, namely loss aversion and the “house money”e¤ect, we use loss aversion as the only psychological assumption; (ii) in terms of model mechanism, while Barberis, Huang, and Santos (2001) rely on stochastic risk aversion to explain the stylized facts in the stock market, our results are driven not by investors’changing risk aversion, but by salient properties of dividend volatility for which we provide strong econometric evidence; and (iii) our model can also match more dimensions of data, such as the GARCH and volatility feedback e¤ects in stock returns as well as the decline in equity premium in the postwar period. McQueen and Vorkink (2004) introduce a new feature, namely, state-dependent sensitivity to news, into the model of Barberis, Huang, and Santos (2001) to explain conditional volatility clustering in stock returns. Our study di¤ers from theirs in that our primary goal is to systematically investigate the role of dividend volatility in asset prices. Our work also contributes to the literature examining how low-frequency risk can provide a justi…cation for observed risk premiums (e.g., Bansal and Yaron, 2004; Hansen and Sargent, 2009). An important aspect of these models is that shocks to both the level and the volatility of consumption are persistent, as are shocks to the level and volatility of dividends. So far, this literature has been conducted within consumption-based models. Our work shows that under the loss aversion/narrow framing approach which has the advantage over consumptionbased models of matching the observed low correlation between consumption and stock returns, the presence of a low-frequency component in aggregate dividend volatility can also mental evidence. Moreover, Brunnermeier and Nagel’s (2008) recent work has shown that time-varying risk aversion …ts poorly with household-level data.

6

shed light on key observed …nancial market phenomena. Since we attempt to provide a comprehensive study of the role of aggregate dividend volatility in asset prices, our work also touches on several other areas of the literature, including studies that relate dividend volatility to asset prices (Longsta¤ and Piazzesi, 2004; Sheinkman and Xiong, 2003; Dumas, Kurshev and Uppal, 2009), studies that address the declining trend in equity premiums (Jagannathan, McGrattan and Scherbina, 2000; Pastor and Stambaugh, 2001; Kim, Morley and Nelson, 2004, 2005; Lettau, Ludvigson and Wachter, 2008), and studies that explore the GARCH and volatility feedback e¤ects in stock returns, to name a few, Engle (1982), Bollerslev (1986), French, Schwert and Stambaugh (1987), Campbell and Hentschel (1992) and Tauchen (2004). The rest of the paper is organized as follows. Section 2 summarizes the three key ingredients of our model. In particular, it provides empirical evidence that dividend volatility is persistent over time. Section 3 presents the model and characterizes the equilibrium asset prices. Section 4 calibrates the model, solves the price-dividend ratios, and analyzes model simulation results. It also provides a variety of sensitivity analyses. The …nal section concludes the paper.

2

Key Ingredients of the Model

In this section, we discuss the three key ingredients underlying our model: dividend volatility clustering, loss aversion and narrow framing. We …rst provide extensive empirical evidence that dividend volatility is highly persistent, which provides a rationale for incorporating this 7

feature of dividend volatility into our theoretical model in Section 2.2. We then review two of the most important ideas emerging from the experimental literature on decision-making under risk: loss aversion and narrow framing.

2.1

Dividend Volatility Clustering

Volatility clustering, which characterizes the persistence in volatility, has been documented as a standard feature of many …nancial series. For instance, Bollerslev, Engle and Wooldridge (1988) show that conditional variance of market returns ‡uctuates across time and is very persistent. For high-frequency return data, the ARCH literature …nds a very high coe¢ cient in the correlation of conditional standard deviations of returns. In this subsection, we document that the dividend growth rate also exihibits strong volatility clustering. Moreover, even though our data are at a low-frequency, the estimated coe¢ cient is very similar to those found in high-frequency data. Following Campbell (2000) and Bansal, Khatchatrian and Yaron (2005), we back out the aggregate quarterly dividend time series from CRSP stock return data from 1926.Q3– 2006.Q3.7 The detailed data construction is provided in Appendix A. This imputed dividend series has accounted for stock repurchases as an increasingly signi…cant component of dividends since the 1980s. We …rst ran two standard diagnostic tests, the Box-Pierce-Ljung test and the ARCH test, to determine whether the dividend growth rate can be predicted at the second moment. 7

In this section, we provide empirical evidence of dividend series at a quarterly frequency to match our benchmark model in section 4. In unreported results, we also examine dividend series at monthly and annual frequencies, which also exhibit strong volatility clustering.

8

The results are reported in Panel A of Table 1;

means that the estimates are signi…cantly

di¤erent from zero at the 1% level. The statistics from both tests signi…cantly reject the null hypothesis of no volatility clustering in the dividend growth rate, contradicting the standard i.i.d. assumptions in existing models. The results of the preliminary tests suggest that use of the exponential GARCH (EGARCH) model to identify the persistent component in dividend volatility is appropriate. We choose EGARCH instead of other types of GARCH models (such as (Fractionally) Integrated GARCH) because it matches best with our theoretical dividend volatility speci…cation in section 3. Moreover, the EGARCH speci…cation can capture potential asymmetric behaviors in volatility. Speci…cally, we consider the following regression:

gD;t+1 = log

2 t

Zt+1

=

0

+

+

1 gD;t

1

+

log

2 t 1

1;

+ A1 [jZt j

E jZt j] + L1 Zt ,

i:i:d: N (0; 1) ,

where gD;t+1 is the dividend growth rate, 0;

t Zt+1 ,

2 t

is the conditional variance of gD;t+1 , and

; G1 ; A1 and L1 are coe¢ cients. In what follows, we report results based on this

AR(1)-EGARCH(1,1) speci…cation. We also applied AR(2)-EGARCH(1,1) and other speci…cations, and the main results remained unchanged. In addition to the EGARCH speci…cation, we tried the speci…cations described in Bansal, Khatchatrian and Yaron (2005) and obtained similar results not reported here.

9

Panel B of Table 1 reports the estimation result. The standard errors of the estimated parameters are reported in parentheses;

and

mean that the estimates are signi…cantly

di¤erent from zero at the 1% and 5% levels, respectively. The coe¢ cient for dividend volatility is ^ 1 = 0:968, indicating that dividend volatility is highly persistent, which is consistent with the standard …ndings in the ARCH literature. This estimated persistence parameter ^ 1 represents an important input parameter for our model.8 Note that the coe¢ cient that measures the persistence in the dividend growth rate per se is much smaller ( ^ 1 = 0:451). In the long-run risk literature (e.g., Bansal and Yaron, 2004; Bansal, Kiku and Yaron, 2007), it is crucial to have persistence in both the mean and the volatility of the dividend growth rate to explain the high equity premium; in other words, both ^ 1 and ^ 1 are assumed to be close to one. In our model, we require the persistent component only in the volatility of dividend growth rate and not in the level of dividend growth rate; that is, we only require that ^ 1 is su¢ ciently high, which is supported by the data. To further con…rm that the persistence in dividend volatility is indeed very high, we resorted to the augmented Dicky-Fuller unit root test by running the following regression:

log (^ t ) =

0

+

1

log (^ t 1 ) +

2

log (^ t 1 ) + et ,

where ^ t is conditional dividend volatility obtained from the EGARCH estimation, and 0;

1

and

2

are coe¢ cients. Panel C of Table 1 reports the test statistics together with

the critical values at 1%, 5%, and 10% levels. We can hardly reject the null hypothesis 8

That is, in subsection 4.3, we use 0:968 for the value of

10

in our model (see equation [4]).

of

1

= 1 at the 10% critical level, which implies that dividend volatility is indeed very

persistent. For comparison, we also ran the unit root test in the dividend growth rate; the unreported result strongly rejects the unit root hypothesis at any critical level, which is not surprising given that ^ 1 is only 0:451 in Panel B of Table 1. Given the strong econometric evidence, we believe that dividend volatility clustering is an important feature of the actual dividend data. Our theoretical model incorporates this feature when we specify the dividend growth rate process.

2.2

Loss Aversion and Narrow Framing

Loss aversion is a central feature of Kahneman and Tversky’s (1979) prospect theory, a prominent descriptive theory of how people evaluate risk.9 According to this theory, people derive utility not from absolute wealth levels but rather from gains and losses measured relative to a reference point. Loss aversion is the …nding that people have a greater sensitivity to losses than to gains of the same magnitude; this is represented by a kink in the utility function. The most basic evidence for loss aversion is that people tend to reject the following 9

To improve upon Kahneman and Tversky (1979), Tversky and Kahneman (1992) propose a generalization of prospect theory, the cumulative prospect theory. Loss aversion remains a central feature of the cumulative prospect theory. In addition to loss aversion, this theory has two other main features: diminishing sensitivity and probability weighting. In the present paper, we study the implication of loss aversion alone. We would not expect our results to be very di¤erent even if it were possible to solve a dynamic asset pricing model that did incorporate all three features of prospect theory. The asset pricing implications of other features of prospect theory have been studied, for example, by Gomes (2005), Barberis and Huang (2008), Barberis and Xiong (2009), Hens and Vlcek (2009), and Li and Yang (2009).

11

gamble: (110; 50%; 100; 50%) ,

which says that there is a 50% probability of winning 110 dollars and a 50% probability of losing 100 dollars. It is very di¢ cult to reconcile this evidence with di¤erentiable utility functions because the very high local risk aversion required to reject this gamble will also lead people to reject large-scale gambles (Epstein and Zin, 1990; Rabin, 2000; Barberis, Huang and Thaler, 2006). Under traditional utility functions de…ned over consumption or total wealth, investors evaluate a new gamble by …rst merging it with the other risks they already face and then checking whether the combination is attractive. Experimental evidence, however, has shown that when o¤ered a new gamble, people often engage in narrow framing in the sense that they tend to evaluate the new gamble in isolation. In other words, they act as if they obtain direct utility from the outcome of this new gamble rather than indirect utility via its contribution to their total wealth as predicted by traditional utility functions. Several explanations have been proposed to explain why investors might frame stock market risk narrowly. One view is that narrow framing stems from non-consumption utility, such as regret. Regret is the pain that we feel when we realize that we could be better o¤ today had we taken a di¤erent action in the past. Investing in the stock market exposes the investor to possible future regret: if the stock market performs badly, she might regret having allocated some fraction of her wealth to the stock market. Another interpretation of narrow framing is proposed by Kahneman (2003), who argues that narrow framing arises 12

due to accessibility: the return distribution of a new gamble itself is much more accessible than the distribution of the investor’s overall wealth once the new gamble has been merged with other risks; as a result, it is easier for the investor to evaluate the stock market risk in isolation than it is for her to evaluate the wealth risk as a whole.10 Although we describe loss aversion and narrow framing separately, recent studies have shown that these two attitudes may form a natural pair (e.g., Kahneman, 2003; Barberis, Huang and Thaler, 2006). Kahneman (2003) provides an explanation for this, arguing that since both prospect theory and narrow framing capture the way people act when they make decisions intuitively, it is natural for loss aversion and narrow framing to appear in combination. In our context, this combination manifests as follows: even though stock market risk is just one of the many risks investors face, they still obtain direct utility from stock market ‡uctuations (narrow framing) and are more sensitive to gains than to losses (loss aversion).

3 3.1

The Model Setup

Consider an economy populated by a continuum of identical, in…nitely-lived, investors who exhibit narrow framing and loss aversion. Two assets are available to trade: a risk-free asset in zero net supply paying a gross interest rate Rf;t , and one unit of a risky asset paying a gross return Rt+1 between time t and t + 1. 10

People can frame the stock market risk at di¤erent horizons, see Thaler, Tversky, Kahneman, and Schwartz (1997). For more discussions on narrow framing, please refer to Barberis and Huang (2007, 2009).

13

The loss-averse investor chooses consumption Ct and risky asset holdings St to maximize the utility function E

"

1 X t=0

1 t Ct 1

+ b0 Ct

t+1

v (Xt+1 )

#

;

(1)

subject to the standard budget constraint, and

Xt+1 = St (Rt+1 Rf;t ) ; 8 > > < Xt+1 ; if Xt+1 0; v (Xt+1 ) = : > > : Xt+1 ; if Xt+1 < 0: The …rst term in the objective function is the standard utility over consumption, where 2 (0; 1) is the time discount factor and

> 0 measures the curvature of the investor’s

utility over consumption.11 For ease of expression, we call this term consumption utility. The second term deserves more attention than the …rst because it captures the direct utility that the investor derives from ‡uctuations in the value of her …nancial wealth. For simplicity, we call this additional utility the prospect theory utility; this description captures the idea that the investor exhibits narrow framing and loss aversion. Ct is the aggregate per capita consumption at time t, which is exogenous to the investor. The exogenous scalar Ct is introduced to ensure that consumption utility and prospect theory utility are of the same order as aggregate wealth increases over time.12 Depending on the return of the risky asset, the investor’s total portfolio excess return For = 1, we replace Ct1 = (1 ) with log (Ct ). Another tractable preference speci…cation that incorporates loss aversion and narrow framing but does not rely on a scaling to ensure stationarity can be found in Barberis and Huang (2007, 2009). 11

12

14

Xt+1 can be either positive or negative, with a positive value indicating a …nancial gain and a negative one indicating a …nancial loss. The function v (Xt+1 ) describes how she feels about her investment performance. Since she is loss-averse, the pain she receives from …nancial losses outweighs the happiness she derives from …nancial gains of the same magnitude. Therefore, v (Xt+1 ) takes di¤erent functional forms with respect to the values of Xt+1 : when Xt+1 is positive, showing that she makes money, v (Xt+1 ) is linear in Xt+1 with slope one; by contrast, when Xt+1 is negative, meaning that she loses money, v (Xt+1 ) ampli…es her utility loss by a magnitude of ; with

being greater than one. Figure 1 plots the function

v (Xt+1 ). Parameter b0 determines how much the second utility factors into the investor’s total utility. If b0 = 0, narrow framing and loss aversion do not play a role in the overall utility and the model is reduced to a traditional asset pricing setting with time-varying dividend volatilities. In this case, as we will show in subsection 4.2, higher dividend volatility leads to a higher dividend growth rate, resulting in a higher price-dividend ratio and a lower equity premium. However, as the value of b0 increases, the e¤ect of prospect theory utility increases such that the investor su¤ers more disutility from her …nancial loss and demands a higher risk premium in holding stocks. Given that we are considering a pure exchange economy, the technology is fully charac-

15

terized by the consumption and dividend growth processes, which are speci…ed as follows:

log (

t+1 )

gC;t+1 = log Ct+1 =Ct = gC +

C t+1 ;

(2)

gD;t+1 = log (Dt+1 =Dt ) = gD +

t "t+1 ;

(3)

u ut+1 ;

(4)

log ( ) =

[log ( t )

log ( )] +

where gC;t+1 is the growth rate of aggregate consumption Ct , gC and

C

are the mean and

standard deviation of the consumption growth rate, and Dt is the dividend, the growth rate of which is denoted as gD;t+1 ; with mean gD and standard deviation

t.

Equation (3)

is a standard assumption in asset pricing literature that is used to capture the idea that the mean level of dividend is di¢ cult to predict. Similar to other models that adopt the loss aversion/narrow framing approach, we model consumption and dividend as two distinct processes: besides the dividend stream, the agent also receives a stream of non-…nancial income, such as labor income, to …nance her aggregate consumption. We draw special attention to equation (4), which characterizes the evolution of dividend volatility. At this point, we depart from existing models such as that of Barberis, Huang and Santos (2001). That is, we incorporate the realistic volatility-clustering feature of the dividend growth rate documented in Section 2.1. To ensure that log (

t+1 )

instead of

t

t

is positive, we model

as an AR(1) process. In this sense, the dividend volatility equa-

tion (4) is very similar to the EGARCH speci…cation in subsection 2.1. magnitude of the innovation to the conditional volatility

16

t:

a larger

u

u

captures the

will increase divi-

dend volatility. A particularly interesting parameter is the coe¢ cient , which controls the strength of dependence on past volatilities. A larger

implies that the impact of a shock to

dividend volatility is more persistent. As has been shown in subsection 2.1, this persistence parameter

is very high in actual dividend data.

The innovations to the consumption growth

t,

the dividend growth "t , and the dividend

volatility ut are jointly normally distributed as 0

1

20

1 0

6B 0 C B 1 ! 0 6B C B 6B C B B C B i.i.d. N 6 6B 0 C ; B ! 1 6B C B 4@ A @ 0 0 1

B t C B C B C B " C B t C B C @ A ut

13

C7 C7 C7 C7 . C7 C7 A5

(5)

Parameter ! represents the correlation between consumption shocks and dividend shocks. Note that when there is persistence in dividend volatility, the unconditional correlation between gC;t+1 and gD;t+1 is !e

0:5

2= u

(1

2

) < !. For simplicity, we assume that consumption

growth shocks are independent of dividend volatility shocks, i.e., Cov ( t ; ut ) = 0. However, we do allow for interaction between shocks to the dividend growth rate "t and shocks to the dividend volatility ut ; the interaction of these two shocks is denoted by . As will be shown later,

also plays a role in generating certain model results, including the excess volatility

and negative autocorrelation of stock returns. Under equations (2)-(5), the consumption growth rate is, in our setting, an i.i.d. process, while the dividend growth rate is a white noise process featuring volatility clustering.

17

3.2

Equilibrium Prices

This subsection derives the equilibrium asset prices. We …rst construct a one-factor Markov equilibrium in which the risk-free rate is a constant and the state variable

t

(dividend volatil-

ity) determines the distribution of future stock returns. Assume that the price-dividend ratio is a function of

t:

ft

Pt =Dt = f ( t ).

We will verify that there is indeed an equilibrium satisfying this assumption. Given the one-factor assumption, the stock returns Rt+1 can be determined as

Rt+1 =

1 + Pt+1 =Dt+1 Dt+1 1 + f ( t+1 ) gD + Pt+1 + Dt+1 = = e Pt Pt =Dt Dt f ( t)

t "t+1

.

(6)

Intuitively, the change in stock returns can be attributed to either the news about dividend growth "t+1 ; the …nancial market uncertainty

t;

or changes in the price-dividend ratio f:

Given that the dividend process is exogenous, the key to solving Rt+1 is to solve the pricedividend ratio f: In equilibrium, the dynamics of the economy are fully captured by the Euler equations:13

1 = Rf;t Et

h

Ct+1 =Ct

i

(7)

13 The Euler equations are both necessary and su¢ cient to characterize the equilibrium. See Barberis, Huang and Santos (2001) for a proof.

18

h

i

1 = Et Rt+1 Ct+1 =Ct

where v^ (Rt+1 ) =

8 > > < > > :

(8)

+ b0 Et [^ v (Rt+1 )]

if Rt+1

Rt+1

Rf;t ;

Rf;t ;

(Rt+1

Rf;t ) ; if Rt+1 < Rf;t :

Equation (7) and the i.i.d. assumption regarding the consumption growth together imply a constant risk-free rate, Rf;t =

1

e

gC

2 2 =2 C

(9)

:

After substituting in the respective consumption and dividend processes, equation (8) becomes

1 = Et

1 + f ( t+1 ) gD + e f ( t)

t "t+1

e (gC +

C t+1

) + b E v^ 1 + f ( t+1 ) egD + 0 t f ( t)

t "t+1

, 8 t. (10)

In equilibrium, the function f must evolve according to equation (10), which also veri…es the conjectured one-factor Markov equilibrium price function.

Methodology of Numerical Computation We solve f numerically on a grid search of the state variable

t.

We begin by guessing a solution to (10), say, f (0) . According to

equation (4), the distribution of

t+1

is completely determined by

19

t

and ut+1 . We then

obtain a new candidate solution f (1) by the following recursion:

1 =

1 + f (i) ( t+1 ) ( t ! e f (i+1) ( t ) 1 + f (i) ( t+1 ) gD + t "t+1 , 8 t. e +b0 Et v^ f (i+1) ( t ) egD

gC + 12

2 2 C

(1

!2 )

Et

C )"t+1

We continue this process until f (i) ! f .

4 4.1

Model Results Benchmark Calibration of Parameter Values

There has been no consensus regarding the decision interval of loss-averse investors. Kahneman (2003) provides evidence that the frequency at which information is most salient or “available” is likely to be the frequency the investor uses to evaluate performance. Since much …nancial information is provided at an annual frequency, previous studies, such as that of Barberis, Huang and Santos (2001), assume an annual evaluation of performance by investors. By contrast, in this paper, we interpret one period in our model as one quarter; thus, the investor re-evaluates her investment performance each quarter. We choose to calibrate the model at a quarterly frequency because much information is also provided on a quarterly basis. In particular, …rms routinely report their earnings/dividends on a quarterly basis. Since the dividend is the main focus of the present paper, quarterly calibration

20

might be more relevant.14 For a robustness check, we also calibrate the model at an annual frequency and report the results in subsection 4.4. Table 2 summarizes our choice of parameter values for the model calibrated at a quarterly frequency. We choose similar values to those of Barberis, Huang and Santos (2001) for the consumption growth parameters and the preference parameters. More speci…cally, for gC and C,

the mean and standard deviation of log consumption growth, we follow Cecchetti, Lam

and Mark (1990) and set gC = 0:46% and

C

= 1:90%; this corresponds to an annual growth

rate of 1:84% with a volatility of 3:79%. The curvature time discount factor

of utility over consumption and the

are set as 1:0 and 0:995, respectively, bringing the net annual risk-free

rate close to 3:86 percent by equation (9) and the values of gC and parameter

C.

The loss aversion

is chosen to be 2:25, because many independent experimental studies have

estimated it as being around this level. Because b0 does not have an empirical counterpart, following Barberis, Huang and Santos (2001), we present results for a range of values of b0 . Using NYSE/AMEX data and Fama risk-free rate data from 1926.Q3 to 2006.Q4 from CRSP, we calibrate the unconditional mean of the quarterly dividend growth rate as its empirical mean, gD = 0:39%. By matching the …rst moment of Equation (3), i.e., E [log (jgD;t+1 log ( ) + E [log (j"t+1 j)], we calibrate log ( ) as

3:91. The parameter , which governs the

persistence in dividend volatility, takes the value of 0:968, the estimated value of ^ 1 in the 14

In reality, since each …rm chooses to issue its dividend/earning report on a di¤erent date, we could have modeled the aggregate dividend process as a continuous process with time-varying arrival rates. However, we believe our speci…cation is a good approximation. First, as shown in Lamont and Frazzini (2007), most …rms concentrate their reports in the …rst month of the quarter. Second, the decision interval of agents can be di¤erent from the interval at which information arrives. Thus, investors might aggregate all the information within the quarter to evaluate their investment performance.

21

gD j)] =

EGARCH model in subsection 2.1. Another important parameter is

u,

which measures the magnitude of dividend volatility.

We calibrate this parameter as 0:25, such that the model-implied annual dividend growth rate has a volatility equal to that of its empirical counterpart. Compared to Bansal and Yaron (2004), the value of

u

appears large. However, this is an artifact of our speci…cation

of the volatility process in equation (4), where the logarithm of dividend volatility, rather than its square as in Bansal and Yaron (2004), follows an AR(1) process. Indeed, given 1 and taking a …rst-order approximation of (4), we have w

= 2E ( 2t )

u

= 1:5

2 t+1

2 t

+

w ut+1 ,

where

10 3 , which is closer to the value found in Bansal and Yaron’s

(2004) speci…cation. Two more model parameters remain to be calibrated: , which captures the interaction between innovations in dividend growth rate and in dividend volatility, and !, the correlation between consumption and dividend. By equations (3) and (4), we calibrate

at

0:38; the

calibration details are given in Appendix B. Following Campbell (2000), we set ! = 0:16, which implies an unconditional correlation of 0:1 between consumption and dividend growth processes.

4.2

Price-dividend Ratio Function f

Figure 2 plots the price-dividend ratio f as a function of log ( t ) for b0 = 0, b0 = 0:7, and b0 = 100. The choices of b0 = 0 and b0 = 100 in Panels (a) and (b) correspond to two extreme cases: when b0 = 0, the investor only cares about consumption utility and does not

22

exhibit loss aversion and narrow framing in her preference; when b0 = 100, her preference is mostly determined by loss aversion and narrow framing. The choice of b0 = 0:7 in Panel (c) corresponds to the situation in which both consumption utility and prospect theory utility are important. We also try a variety of other intermediate values for b0 , for example, b0 = 2, b0 = 6, b0 = 20, etc. The essential pattern, however, is fully depicted by Figure 2. When b0 = 0, the investor’s preference corresponds to the standard consumption utility. In this case, the price-dividend ratio f ( t ) increases with the dividend volatility

t.

Now

that the model is reduced to a standard consumption-based model, the standard consumption growth rate becomes the stochastic discount factor (SDF). Because the dividend process is only weakly correlated with SDF, the investor does not care much about the stock risk and, thus, behaves as if she is risk-neutral toward the stock. A higher dividend volatility implies higher expected cash ‡ows from holding stocks and makes the stock more attractive to the risk-neutral alike investor.15 Therefore, when b0 = 0, stock prices rise with dividend volatility, resulting in a positive relationship between f ( t ) and

t.

When b0 = 100, the investor derives utility primarily from ‡uctuations in the value of her stock holdings (loss aversion and narrow framing) and derives very little consumption utility. This is similar to the situation studied in Benartzi and Thaler (1995). In this scenario, the overall relationship between

t

and the price-dividend ratio f ( t ) is negative. The intuition

is as follows. A higher value of

t

corresponds to a more volatile dividend process and,

hence, a more volatile return process, which implies that the investor is more likely to su¤er 15

In mathematical terms, by equation (3), the conditional expected dividend at date t + 1 is Et (Dt+1 ) = 2 Dt Et (egD;t+1 ) = Dt egD + t =2 , which is increasing in the state variable t .

23

…nancial losses. This causes her great pain and makes stocks less desirable. As a result, she requires more compensation when faced with more volatile dividend processes, which, in turn, results in lower stock prices or higher equity premiums. In fact, this negative slope of f function is consistent with Bansal and Yaron (2004) and Bansal, Khatchatrian and Yaron (2005), who document that asset prices drop as economic uncertainty rises, although their measure of economic uncertainty is conditional consumption volatility rather than conditional dividend volatility. It is rather di¢ cult to justify this negative relationship within the standard power utility framework, where, as we have shown above, a higher dividend volatility is associated with higher expected dividend growth and, thus, price-dividend ratios always vary positively with dividend volatility. However, it can be easily understood with the introduction of loss aversion/narrow framing preferences. When b0 = 0:7, both consumption utility and prospect theory utility will matter. Now the investor cares not only about consumption, the standard expected power utility term in (1), but also about ‡uctuations in the value of her investments, the additional prospect theory utility term in (1). As the standard consumption utility contributes to a positive relationship between

t

and the price-dividend ratio f ( t ), and because prospect theory

utility contributes to a negative relationship between

t

and the price-dividend ratio f ( t ),

the overall shape of the function f is U -shaped. For low values of

t,

the impact of loss

aversion and narrow framing is dominant; hence, the function f is downward-sloping. As

t

becomes larger, the impact of consumption utility catches up, and the function f eventually becomes upward-sloping.

24

Examining Panels (a) through (c) of Figure 2, we …nd that as b0 increases, the function f will move downward. This is because the larger the value of b0 , the more the investor cares about her wealth ‡uctuations, which, in turn, means that she requires a higher risk premium to hold stocks. What does f look like in the data? We performed an AR(1)-EGARCH(1,1) estimation on the quarterly dividend growth for 1926.Q3-2006.Q4; the resulting price-dividend ratios are plotted against the estimated conditional dividend volatilities ^ t in Figure 3. The data display a U-shaped pattern, as predicted by Panel (c) of Figure 2. On the one hand, 80 percent of the observed data display a negative relationship between the empirical price-dividend ratios and the estimated dividend volatility. On the other hand, for certain extremely high realizations of ^ t (for instance, when the logarithm of dividend volatility is larger than 0:05) price-dividend ratios begin to rise with dividend volatility.

4.3

Simulation Results

In this subsection, we generate arti…cial data under the parameter con…guration in Table 2 and show that the model-simulated data replicate the interesting patterns found in the stock market. In our benchmark results here, we report the results for three values of b0 : b0 = 0:7, b0 = 2, b0 = 6. In subsection 4.4, we will investigate how robust our results are with respect to other parameter con…gurations. To facilitate a comparison with historical data, we simulate the model at a quarterly frequency and time-aggregate results to obtain annual data. We do 10,000 simulations each

25

with 320 quarterly observations. We then calculate the statistics of interest and report their sample moments. Given that the simulation number is large enough, the sample moments should serve as good approximations to population moments.

4.3.1

Stock Returns and Stock Volatility

Two important features of the aggregate market include the historically signi…cantly higher average return that stocks have earned compared to T-bills (Mehra and Prescott, 1985) and the greater volatility of stock returns compared to the underlying dividends (Shiller, 1981). We examine whether our model is able to match these two dimensions of the data. Table 3 reports a variety of statistics calculated from model-simulated data and the corresponding statistics from historical data. It is noteworthy that the model can match the mean and standard deviations of excess stock returns fairly well. When b0 = 6 ; the model generates a sizable premium of 5:59% per annum, which is slightly lower than the empirical value 5:90%; the model also generates a standard deviation of 17:49%, which is close to the corresponding value of 19:17% in the data. We also report the mean and standard deviation of the simulated annual price-dividend ratios, E (Pta =Dta ) and The empirical value

(Pta =Dta ), where the superscript “a”indicates annualized variables.

(Pta =Dta ) = 12:43% is relatively high compared to that found in other

papers (Campbell and Cochrane, 1999; Barberis, Huang and Santos, 2001; Bansal and Yaron, 2004). This is due to the relatively high price-dividend ratios from 1996 to 2006, a period that includes the high-tech bubble.

26

We are able to match the volatility of stock returns even though the volatility of pricedividend ratios is lower than its empirical counterparts, which is a common problem with one-factor models. The reason behind the achievement of excess volatility in stock returns is the positive relationship between price-dividend ratios and dividend innovations. To see this more clearly, consider the following approximate relationship (Campbell, Lo and MacKinlay, 1997): rt+1

A + log [f (

t+1 ) =f

( t )] +

t "t+1 ;

where A is a constant. The excess volatility of market return relative to that of the dividend growth (or the “fundamental”), V ar (rt+1 )

V ar ( t "t+1 ), comes from two sources: the

volatility of price-dividend ratios, V ar (log (ft+1 =ft )), and the covariance between the pricedividend ratios and the news regarding the dividends, Cov (log (ft+1 =ft ) ; data, because

t "t+1 ).

In actual

< 0 in (5), good dividend news (positive "t+1 ) tends to be associated with

negative dividend volatility shock (negative ut+1 ). This implies that price-dividend ratios will increase next period when b0 takes the values in Table 3 (see Panel (c) of Figure 2). Therefore, the covariance term Cov (log (ft+1 =ft ) ;

t "t+1 )

is positive.

We notice that as b0 grows, both the means and standard deviations of model-simulated excess returns increase. When b0 increases, loss aversion becomes a more important feature of the investor’s preference. Thus, she becomes more fearful of the risky asset and requires a higher premium to hold it. A higher b0 also implies a larger covariance term Cov (log (ft+1 =ft ) ;

t "t+1 ),

resulting in a larger excess volatility of stock returns.

The model is also able to generate the low correlation between stock returns and con27

a a sumption growth, Corr rt+1 ; gC;t+1 = 0:1. This occurs because the variation in stock re-

turns is completely driven by the innovations in the dividend process, which are only weakly correlated with the consumption process.

4.3.2

Autocorrelation of Returns and Price-Dividend Ratios

Table 4 presents autocorrelation in returns and price-dividend ratios. Our model predicts a negative autocorrelation in stock returns, as documented by Poterba and Summers (1988) and Fama and French (1988a). This negative correlation comes from the negative correlation between shocks to the dividend growth rate "t and shocks to the dividend volatility ut . Suppose the stock return Rt is high in period t. This corresponds to good dividend news. That is, "t is high, which, in turn, implies a low realized dividend volatility,

t,

as "t and ut

are negatively correlated. Recall that, when b0 takes the values in Table 4, the price-dividend ratios f ( t ) decrease with the dividend volatility a high Rt leads to a high f ( t ) through the low low expected future stock return, Et (Rt+1 ) =

t

(see Panel (c) of Figure 2). Therefore,

t,

and the high f ( t ), in turn, leads to a

1 E f ( t) t

[(f (

t+1 )

+ 1) egD;t+1 ]. Moreover, our

model closely matches the highly positively correlated price-dividend ratios in the data. This is because the price-dividend ratios depend solely on the persistent AR(1) dividend volatility process.

28

4.3.3

Return Predictability

To analyze the predictability pattern of returns, we run the following regression on both simulated and historical data:

a a a rt+1 + rt+2 + ::: + rt+j =

j

+

j

(Dta =Pta ) +

j;t ,

a refers to the annual cumulative log returns from year t + j where rt+j

1 to t + j . Table 5

presents the regression result for di¤erent values of b0 . This estimation result from modelsimulated data resembles the classic pattern documented by Campbell and Shiller (1988) and Fama and French (1988b). The coe¢ cients are signi…cant and negative, indicating that high prices tend to predict low expected returns. Moreover, the forecasting power increases with forecasting horizons, as re‡ected by the increasing coe¢ cients and R20 s. The pattern of return predictability generated by our model can be understood through 1 the volatility test in Cochrane (1992). Starting from the accounting identity 1 = Rt+1 Rt+1

with Rt+1 = (Pt+1 + Dt+1 )=Pt , the log-linearization around the average price-dividend ratios, P=D, implies that, in the absence of rational asset price bubbles,

V ar(pt

dt )

1 X

h Cov(pt j

dt ; gD;t+j )

j=1

1 X

hj Cov(pt

dt ; rt+j ),

j=1

where the lower case indicates log values and h = P=D=(1 + P=D). This suggests that the variation in the price-dividend ratio will forecast either the change in expected dividend growth rate, that in the discount rate, or both.

29

In our model, even though dividend volatility is time-varying, the dividend growth rate per se is still a white noise process, meaning that Cov(pt

dt ; gD;t+j ) = 0. Given that the

risk-free rate is maintained as a constant, the only thing remaining for the price-dividend ratio to predict is the excess return. A high price-dividend ratio is associated with a decline in dividend volatility, and, thus, the required expected return will be lower. Therefore, our model implies an extreme version of the volatility test results. As shown in Cochrane (2005), both the estimated coe¢ cients and R20 s increase with the persistence of the price-dividend ratio, which, in our model, depends on dividend volatility. Therefore, it is the persistence in dividend volatility that drives the predictability of returns.

4.3.4

Time-varying Sharpe Ratios

Empirical evidence suggests that estimates of both conditional means and conditional standard deviations of returns change through time but do not move one-for-one. Hence, the Sharpe ratios are time-varying. Figure 4 presents the conditional means and conditional standard deviations as functions of the state variable log( t ) for the case of b0 = 6. Those conditional moments are computed numerically from the price-dividend ratio function and equations (3)-(6) as follows:

1 2 1 + f ( t+1 ) t ut+1 2 e ; Et (Rt+1 ) = egD + 2 (1 ) t Et f ( t) v u 2 u 2gD +2(1 2 ) 2 1+f ( t+1 ) tE u e e2 t ut+1 t f ( t) u t (Rt+1 ) = u h i2 : t 2 2 1+f ( ) t+1 e2gD +(1 ) t Et e t ut+1 f ( t)

30

Comparing the conditional means, Et (Rt+1 ), and conditional standard deviations,

t

(Rt+1 ),

of stock returns, we see that they are di¤erent functions of dividend volatility log ( t ), although they are both increasing in log ( t ). Most noticeably, for values of log ( t ) smaller than log ( ) =

3:91, the conditional standard deviation is almost a constant, whereas the

conditional mean exhibits more variation and increases with log ( t ). Therefore, the Sharpe ratio of the conditional mean to the conditional standard deviation varies over time. Its variation is due to the di¤erence between Et (Rt+1 ) and

t

(Rt+1 ).

We use simulations to get the distribution of Sharpe ratios, [Et (Rt+1 )

Rf;t ] =

t

(Rt+1 ).

Speci…cally, we make 160,000 random draws of "t+1 and ut+1 , calculate the conditional mean and conditional standard deviation of expected returns by numerical integration, and then obtain the conditional Sharpe ratios as a function of log( t ) when b0 = 6. Figure 5 presents the histogram of the simulated conditional Sharpe ratios. The unconditional mean and standard deviation of the simulated Sharpe ratios are 0:13 and 0:05, which closely matches their empirical values.

4.3.5

GARCH and Volatility Feedback E¤ects

Empirical studies show that stock returns exhibit a GARCH e¤ect (e.g., Bollerslev, 1986) and a volatility feedback e¤ect (e.g., French, Schwert and Stambaugh, 1987; Campbell and Hentschel, 1992). The central property of GARCH in stock returns is volatility clustering, in which large (small) return changes today tend to be followed by large (small) return changes tomorrow. The volatility feedback e¤ect states that an increase in future market volatility

31

raises current required stock returns and thus lowers current stock prices. Both e¤ects can be conveniently captured by Nelson’s (1991) exponential GARCH(1,1) (EGARCH(1,1)) model:

rt+1 = q0 + q1 rt +

p

ht zt+1 ,

log (ht ) = c0 +cG log (ht 1 ) + cA [jzt j zt+1

E jzt j] + cV zt ,

(11)

i:i:d: N (0; 1) ,

where rt+1 is the quarterly stock return between times t and t + 1, ht is the conditional variance of rt+1 , and q0 , q1 , c0 , cG , cA and cV are constants. Volatility clustering in EGARCH models is represented by a positive coe¢ cient cG , while the volatility feedback e¤ect is captured by a negative coe¢ cient cV . To evaluate our model along these two dimensions, we estimate an EGARCH(1,1) model on the simulated stock returns for a variety of economies and report the results in Table 6. For each economy, we make 160,000 random draws of "t+1 and ut+1 , compute the stock return time series using the solved price-dividend ratio function f ( ), and then implement an EGARCH(1,1) estimation on the simulated stock returns. The economies di¤er in the magnitude of parameter b0 , which controls the relative importance of prospect theory utility in the investor’s preference. For the …rst economy, we deliberately set b0 to be 0 to remove the e¤ect of loss aversion/narrow framing and single out the role of dividend volatility clustering. For the remaining two economies, we allow b0 to take positive values in order to examine the joint e¤ect of dividend volatility clustering and loss aversion/narrow framing. For comparison purposes, Table 6 also reports estimates from the historical quarterly value-weighted CRSP

32

market returns from 1926 to 2006 and veri…es that the historical return data indeed exhibit a GARCH e¤ect and a volatility feedback e¤ect. Table 6 suggests that dividend volatility clustering can itself generate return volatility clustering because when b0 = 0, that is, when volatility clustering is the only active feature, the estimate of cG is positive and close to 1. Therefore, stock returns inherit the GARCH property of the dividend process. However, when b0 = 0, the model predicts a reversed volatility feedback e¤ect, as cV is positive. The underlying reason for this reversed volatility feedback e¤ect is the same as that for the monotonic increasing price-dividend ratio function: when b0 = 0, the investor is almost risk-neutral toward the stock risk because the SDF is only weakly correlated with dividends; a higher return volatility corresponds to a higher dividend volatility and, thus, to higher expected cash ‡ows (see footnote 15); therefore, the risk-neutral alike investor is more likely to buy the stock, which pushes up both the current stock prices and the realized returns. Once we turn on the prospect theory utility, that is, once we allow b0 to be positive, our model can successfully predict a volatility feedback e¤ect as cV < 0 for b0 = 2 and 6. This result is intuitive: a higher stock volatility makes our loss-averse investor more likely to su¤er losses, and, consequently, she will pay less to hold the stock, leading to a decline in the stock price. Thus, our preference-based model incorporating the actual features of dividend volatility provides an explanation for both the GARCH e¤ect and the volatility feedback e¤ect in stock returns.

33

4.3.6

Declining Equity Premiums after WWII

Quite a few papers have documented that ex-ante equity premiums have declined during the past few decades (e.g., Blanchard, 1993; Fama and French, 2002; Freeman, 2004; Buranavityawut, Freeman and Freeman, 2006). The literature has used consumption volatility or return volatility to explain this phenomenon (e.g., Pastor and Stambaugh, 2001; Kim, Morley and Nelson, 2004, 2005; Lettau, Ludvigson and Wachter, 2008). Since our model uses dividend volatility to stand for risk, it is interesting to examine how this risk relates to the declining equity premiums in the …nancial market. In Panel A of Table 7, we re-calibrate our model according to the dividend data for 1954-2006. We …nd that, compared to Table 2, the mean dividend growth rate does not change signi…cantly but that the standard deviation of log ( t ) decreases from 0:25 to 0:17, a decline of roughly 30%.16 According to our model, lower dividend volatility means that stocks are less likely to perform poorly; thus, the lossaverse investor is less worried about ‡uctuations in her …nancial wealth. As a result, she is more willing to hold risky stocks, which pushes up stock prices and drives down expected returns.17 Panel B of Table 7 reports the historical equity premiums for the period 1954-2006 as well as those calculated from model-simulated data. Consistent with our intuition, we indeed 16

To formally characterize the decline in dividend volatility, we also follow Hamilton (1989) to estimate a regime-switching model. The basic idea is to model the dividend growth rate as deriving from one of two regimes, either a regime with a high dividend volatility or one with a low dividend volatility. Our estimation results show that the probability that the dividend growth rate was in a high-volatility state is very high for prewar data but exhibits sharp declines after the 1950s. The details are available upon request. 17 One may argue that the declining trend in dividend volatility is partly due to corporate managers’ intentions to smooth dividends. Whatever the reason, however, the investor in our theoretical model takes the dividend process as exogenously given when making her investment decisions, which is a standard assumption in the asset pricing literature.

34

…nd that, under the new parameter con…guration, the model-implied equity premiums are much lower than those in Table 3, which are obtained through calibration using all data from 1926-2006.18 In addition, from Panel B of Table 7, we see that when b0 = 6, our model generates up to two-thirds of the historical equity premium. This suggests that the decline in equity premiums might be a direct result of declining macroeconomic risk, which is characterized by dividend volatility.

4.4

Sensitivity Analyses and Discussions

In this subsection, we analyze the sensitivity of our results to various parameters of interest and model assumptions. For each parameter that we vary, we …x other parameters at the values shown in Table 2. In the benchmark results in subsection 4.3, we use the estimated value of

1

obtained

in subsection 2.1 as the input for parameter , the parameter that governs the persistency in dividend volatility. Table 8 reports the e¤ect of varying we change , we must also adjust the value of matches its empirical value, i.e., V ar

P4

u

for the case of b0 = 6. When

such that the annual dividend volatility

j=1 gD;t+j

2 log +2

= 4e

2 u 1

2

= 10:20%. Table 8

demonstrates that persistent dividend volatility is crucial to generate high means and excess volatility of stock returns. Without persistent dividend volatility, for example, when = 0 or 0:45, the model can only generate an excess return of up to half its empirical value. Conversely, if dividend volatility is very persistent, when they face a high dividend 18

To make this comparative static analysis sensible, we assume that, back to year 1926, the investor did not anticipate the decline in the dividend volatility in the year 1954.

35

volatility today, investors can reliably forecast high dividend volatility in the next period, which makes investing in the stock market very scary. Persistent dividend volatility and loss aversion/narrow framing together generate high and volatile stock returns close to their empirical values. Throughout our analysis, we …x the loss aversion parameter

at 2.25 because many

independent studies have estimated it at around this level. Table 9 shows the sensitivity of stock returns to the value of

for the case of b0 = 6. An increase in

raises the equity

premium because investors become more loss-averse. The analysis in subsection 4.3 focuses on intermediate values of b0 , the parameter that controls how much the investor cares about …nancial wealth ‡uctuations. Table 10 reports the results for extreme values of b0 : (i) b0 = 0, when the investor only cares about consumption utility; and (ii) b0 = 100, when loss aversion/narrow framing becomes a dominant feature of the investor’s preferences. This table provides sharp insight into the di¤ering roles of consumption utility and prospect theory utility in generating our results and, thus, helps us understand why we need to distinguish between consumption and dividend. When b0 = 0, the equity premium is negative, because in this case, as we explained in subsection 4.2, the investor is almost risk-neutral toward the stock, and the expected gross stock return is, therefore, close to the risk-free rate. That is, E (Rt+1 )

Rf;t , which, through Jensen’s in-

equality, implies a negative continuous equity premium or E [log (Rt+1 )] < log (Rf;t ). Bansal and Yaron (2004, page 1493) make a similar point. When b0 = 100, the equity premium is close to the one in our benchmark analysis, which suggests that it is the loss aversion/narrow

36

framing component that is driving our results. So far, we have reported the results assuming that investors evaluate their gains or losses on a quarterly basis. Table 11 re-interprets the length of the evaluation period as one year and re-calibrates the parameter values as follows: gC = 1:84%, gD = 1:57%, u

C

= 3:79%, log ( ) =

= 1,

2:75, ! = 0:13,

= 0:98,

= 2:25,

= 0:879,

= 0 and

= 0:35. Compared to our benchmark result in Table 3, now that the investor evaluates

her investment performance less frequently, she is less likely to experience losses and requires a lower premium to hold the stock.

A Note on Aggregation and Framing As with all models adopting the loss aversion/narrow framing approach, we assume that the key features of these preferences survive aggregation. As long as loss aversion and narrow framing are present in the aggregate, our main results will persist. Intuitively, if all investors isolate the stock market risk from their background risk and are loss-averse over ‡uctuations in stock market wealth, it is hard to see why the e¤ects of these factors would not be present in the aggregate. Berkelaar and Kouwenberg (2009) formally study a model with heterogeneous, loss-averse investors; their analysis suggests that heterogeneity in reference points and initial wealth of the loss-averse investors does not change the salient features of the equilibrium price process obtained in a representative agent model. To explain the aggregate market behavior, we assume in our analysis that investors engage in narrow framing at the stock market level. To deal with …rm-level stock returns, Barberis and Huang (2001) consider cases in which investors engage in narrow framing at 37

the individual stock level and the portfolio level. We can also extend our model in these dimensions. In particular, it seems an interesting research direction to examine whether individual dividend growth rates also exhibit dividend volatility clustering and, if so, to extend our current model to a multi-stock setting to deal with the cross section of stock returns.

5

Conclusion

A large body of experimental and psychological evidence supports the ideas of loss aversion and narrow framing. In the behavioral …nance literature, the loss aversion/narrow framing approach has been adopted in a variety of models to explain observed empirical facts. In this paper, we depart from existing models by incorporating a persistent component, for which we provide extensive empirical evidence, into the volatility of the dividend growth rate. We …nd that our model-simulated data exhibit similar patterns to those observed in actual data: stock returns have a high mean, high volatility and a low correlation with consumption; they are predicted by price-dividend ratios; the Sharpe ratios are time-varying; returns are negatively correlated and price-dividend ratios are positively correlated; returns exhibit the GARCH and volatility feedback e¤ects; and there is a decline in equity premium based on calibration using the postwar data. In sum, we provide a systematic investigation of the implications of dividend volatility for asset prices when investors derive direct utility from their …nancial investments. Facing an uncertain investment environment captured by dividend volatility, they are fearful of holding 38

risky assets; if this uncertainty is persistent, their fears are stronger. This mechanism can generate important asset price behaviors in …nancial markets.

39

Appendix A. Constructing Dividend Time Series We follow Bansal, Khatchatrian and Yaron (2005) to impute the dividend time series from the CRSP database. This appendix describes the details. The data cover quarterly samples from 1926.Q3 up to 2006.Q3. In order to construct the quarterly dividend variable, the following series are used: Pindx : Monthly stock price index on NYSE/AMEX. The price index for month j is calculated as Pindx;j = (V W RET Xj + 1) Pindx;j 1 , where V W RET X is the valueweighted return on NYSE/AMEX excluding dividends, taken from CRSP. Dindx : Monthly dividend index on NYSE/AMEX. The dividend for month j is calculated as Dindx;j =

1+V W RET Dj 1+V W RET Xj

1

Pindx;j , where V W RET D and V W RET X are,

correspondingly, the value-weighted returns on NYSE/AMEX including and excluding dividends, taken from CRSP. Dindx : Quarterly dividend index on NYSE/AMEX. The dividend for a quarter is the sum of the monthly dividend indices for the three months comprising the quarter. Then, a four-period backward moving average is taken to remove seasonality. That is, P Dindx;t = 14 3j=0 Dindx;3(t j) 2 + Dindx;3(t j) 1 + Dindx;3(t j) , where t indexes quarters.

Inf lation: Quarterly in‡ation index. The in‡ation index for a quarter is the in‡ation index in the last month of the quarter, taken from CRSP. The resulting quarterly dividend series and dividend growth rate series are calculated as follows: Dt =

Dindx;t , gD;t = log Inf lationt

40

Dt Dt 1

:

B. Calibrating Parameter By equation (3), gD j) = log ( t ) + log (j"t+1 j) .

log (jgD;t+1

(12)

Recursing the above equation one period forward, we have gD j) = log (

log (jgD;t+2

Substracting equation (12) multiplied by

= [log (

+ log (j"t+2 j) .

(13)

from equation (13) yields

gD j)

log (jgD;t+2

t+1 )

log (jgD;t+1

gD j)

t+1 )

log ( t )] + [log (j"t+2 j)

log (j"t+1 j)] :

t+1 )

log ( t ) = (1

u ut+1 :

(14)

Equation (4) implies log (

) log +

Substituting the above equation into equation (14) and rearranging terms, we have u ut+1

gD j)

= log (jgD;t+2 (1

) log

log (jgD;t+1

gD j) + log (j"t+1 j)

log (j"t+2 j) :

(15)

By equation (3), t "t+1

= gD;t+1

(16)

gD :

Thus, Cov ( t "t+1 ; = Cov (gD;t+1

u ut+1 )

gD ; log (jgD;t+2

+ Cov ( t "t+1 ; log (j"t+1 j)) : 41

gD j)

log (jgD;t+1

gD j)) (17)

By equation (5), Cov ( t "t+1 ;

u ut+1 )

= E ( t "t+1

Cov ( t "t+1 ; log (j"t+1 j)) =

e

log( )+

u ut+1 ) 2 u

2( 1

2)

=

log( )+

ue

2 u 2( 1

2)

;

E ["t+1 log (j"t+1 j)] ;

which are, in turn, substituted into equation (17), implying 2 =

4

Cov (gD;t+1

gD j)

gD ; log (jgD;t+2 + e

log( )+

2 u 2( 1

2)

log (jgD;t+1

E ["t+1 log (j"t+1 j)]

log( )+

ue

We then use data to estimate Cov (gD;t+1

2 u

2( 1

2)

gD ; log (jgD;t+2

gD j)

gD j))

3 5

:

log (jgD;t+1

(18)

gD j))

and use the fact that "t+1 follows a standard normal distribution to compute E ["t+1 log (j"t+1 j)]. Once the calibrated values for log ( ),

u

and

equation (18).

42

are available, we can easily estimate

using

References [1] Bansal, R., Yaron, A., 2004. Risks for the long run: a potential resolution of asset pricing puzzles. Journal of Finance 59, 1481-1509.

[2] Bansal, R., Khatchatrian, V., Yaron, A., 2005. Interpretable asset markets? European Economic Review 49, 531-560.

[3] Bansal, R., Kiku, D., Yaron, A., 2007. Risks for the long run: estimation and inference. Unpublished Working Paper. Duke University.

[4] Barberis, N., Huang, M., 2001. Mental accounting, loss aversion, and individual stock returns. Journal of Finance 56, 1247-1292.

[5] Barberis, N., Huang, M., 2007. The loss aversion/narrow framing approach to the equity premium puzzle. In: Mehra, R. (Ed.), Handbook of Investments: The Equity Premium. North Holland, Amsterdam, pp. 199-236.

[6] Barberis, N., Huang, M., 2008. Stocks as lotteries: the implications of probability weighting for security prices. American Economic Review 98, 2066-2100.

[7] Barberis, N., Huang, M., 2009. Preferences with frames: a new utility speci…cation that allows for the framing of risks. Journal of Economic Dynamics and Control 33, 1555-1576.

[8] Barberis, N., Huang, M., Santos, T., 2001. Prospect theory and asset prices. Quarterly Journal of Economics 116, 1-53.

[9] Barberis, N., Huang, M., Thaler, R., 2006. Individual preferences, monetary gambles, and stock market participation: A Case for Narrow Framing. American Economic Review 96, 1069-1090.

[10] Barberis, N., Xiong, W., 2009. What drives the disposition e¤ect? an analysis of a longstanding preference-based explanation. Journal of Finance 64, 751-784.

[11] Benartzi, S., Thaler, R., 1995. Myopic loss aversion and the equity premium puzzle. Quarterly Journal of Economics 110, 73-92.

[12] Berkelaar, A., Kouwenberg, R., 2009, From Boom ‘Til Bust: How Loss Aversion A¤ects Asset Prices, Journal of Banking and Finance, Volume 33, Issue 6, 1005-1013.

[13] Blanchard, O., 1993. Movements in the equity premium. Brookings Papers on Economic Activity 2, 75-118.

[14] Bollerslev, T.,1986. Generalized Autoregressive Conditional Heteroskedasticity. Journal of Econometrics 31, 307–28.

[15] Bollerslev, T., Engle, R., Wooldridge, J., 1988. A capital asset pricing model with time-varying covariances. Journal of Political Economy 96, 116-131.

[16] Brunnermeier, M., Nagel, S., 2008. Do wealth ‡uctuations generate time-varying risk aversion? micro-evidence on individuals’asset allocation. American Economic Review 98, 713-736

[17] Buranavityawut, N., Freeman, M., Freeman, N., 2006. Has the equity premium been low for forty years? The North American Journal of Economics and Finance 17, 191-205.

43

[18] Campbell, J., 2000. Asset prices, consumption and the business cycle. In: Taylor, J., Woodford, M. (Ed.), Handbook of Macroeconomics. North Holland, Amsterdam, pp. 1231-1303.

[19] Campbell, J., Cochrane, J., 1999. By force of habit: a consumption-based explanation of aggregate stock market behavior. Journal of Political Economy 107, 205-251.

[20] Campbell, J. and Hentschel, L. 1992. No News is Good News: An Asymmetric Model of Changing Volatility in Stock Returns. Journal of Financial Economics 31, 281-331.

[21] Campbell, J., Lo, A., MacKinlay, A., 1997. The econometrics of …nancial markets. Princeton University Press, Princeton.

[22] Campbell, J., Shiller, R., 1988. Stock prices, earnings, and expected dividends. Journal of Finance, XLIII, 661–676.

[23] Cecchetti, S., Lam, P., Mark, N., 1990. Mean reversion in equilibrium asset prices. American Economic Review, LXXX, 398–418.

[24] Cochrane, J., 1992. Explaining the variance of price-dividend ratios. Review of Financial Studies 5, 243-280.

[25] Cochrane, J., 2005. Asset Pricing. Princeton University Press, Princeton. [26] Dumas, B., Kurshev, A. and Uppal, R., 2009. Equilibrium portfolio strategies in the presence of sentiment risk and excess volatility. The Journal of Finance, LXIV, 579-629. [27] Engle, R., 1982. Autoregressive conditional heteroskedasticity with estimates of the variance of U.K. in‡ation, Econometrica 50, 987-1008.

[28] Espstein, L., and Zin, S., 1990. First-order risk aversion and the equity premium puzzle. Journal of Monetary Economics 26, 387–407.

[29] Fama, E., French, K., 1988a. Permanent and temporary components of stock prices. Journal of Political Economy, XCVI, 246–273.

[30] Fama, E., French, K., 1988b. Dividend yields and expected stock returns. Journal of Financial Economics, XXII, 3–25.

[31] Fama, E., French, K., 2002. The equity premium. Journal of Finance 57, 637-659. [32] Freeman, M., 2004. An explanation of the declining ex-ante equity premium. Unpublished Working Paper. University of Exeter.

[33] French, M., Schwert, W. and Stambaugh, R.. 1987. Expected Stock Returns and Volatility. Journal of Financial Economics 19, 3-30.

[34] Gomes, F., 2005. Portfolio choice and trading volume with loss-averse investors. Journal of Business 78, 675-706.

[35] Hamilton, D., 1989. A new approach to the analysis of nonstationary returns and the business cycle. Econometrica 57, 357-384.

[36] Hansen, L., Sargent, T., 2009. Fragile beliefs and the price of model uncertainty. Unpublished Working paper. University of Chicago.

44

[37] Hens, T, and Vlcek, M, 2009. Does prospect theory explain the disposition e¤ect? http://www.nccr-…nrisk.uzh.ch/wps.php?action=query&id=247.

[38] Jagannathan, R., McGrattan, E., Scherbina, A., 2000. The declining U.S. Equity Premium. Federal Reserve Bank of Minneapolis Quarterly Review. Vol. 24, No. 4, 3-19.

[39] Kahneman, D., 2003. Maps of bounded rationality: psychology for behavioral economics. American Economic Review 93, 1449-75.

[40] Kahneman, D., Tversky, A., 1979. Prospect theory: an analysis of decision under risk. Econometrica 47, 263-291.

[41] Tversky, A., and D. Kahneman. Advances in prospect theory: Cumulative representation of uncertainty. Journal of Risk and Uncertainty 5 (1992): 297–323.

[42] Kim, C., Morley, J., Nelson, C., 2004. Is there a positive relationship between stock market volatility and the equity premium? Journal of Money, Credit and Banking 36, 339-360.

[43] Kim, C., Morley, J., Nelson, C., 2005. The structural break in the equity premium. Journal of Business and Economic Statistics 23, 181-191.

[44] Lamont, O., Frazzini, A., 2007. The earnings announcement premium and trading volume. NBER Working paper w13090.

[45] Lettau, M., Ludvigson, S., Wachter, J., 2008. The declining risk premium: what role does macroeconomic risk play? Review of Financial Studies 21, 1653-1687.

[46] Li, Y., Yang, L., 2009. Prospect theory, the disposition e¤ect and asset prices. Unpublished Working paper, University of Toronto.

[47] Longsta¤, F., Piazzesi, M., 2004. Corporate earnings and the equity premium. Journal of Financial Economics 74, 401-421.

[48] McQueen, G., Vorkink, K., 2005, Whence GARCH? a preference-based explanation for conditional volatility. Review of Financial Studies 17, 915-949.

[49] Mehra, R., Prescott, E., 1985. The equity premium: a puzzle. Journal of Monetary Economics 15, 145–161.

[50] Nelson, D. 1991. Conditional Heteroscedasticity in Asset Returns: A New Approach. Econometrica 59, 347–70.

[51] Pastor, L., Stambaugh, R., 2001. The equity premium and structural breaks. Journal of Finance 56, 1207–1239.

[52] Poterba, J., Summers, L., 1988. Mean-reversion in stock returns: evidence and implications. Journal of Financial Economics, XXII, 27–59.

[53] Rabin, M., 2000. Risk aversion and expected utility theory: a calibration theorem. Econometrica 68, 1281-1292.

[54] Scheinkman, J., Xiong, W., 2003. Overcon…dence and speculative bubbles. Journal of Political Economy 111, 1183-1219.

45

[55] Shiller, R., 1981. Do stock prices move too much to be justi…ed by subsequent changes in dividends? American Economic Review 71, 421–436.

[56] Tauchen, G., 2004. Recent developments in stochastic volatility: statistical modelling and general equilibrium analysis, Unpublished Working paper, Duke University.

[57] Thaler, R., Tversky, A., Kahneman, D., Schwartz, A., 1997. The e¤ect of myopia and loss aversion on risk taking: an experimental test. Quarterly Journal of Economics 112, 647–61.

[58] Tversky, A., Kahneman, D., 1992. Advances in prospect theory: cumulative representation of uncertainty. Journal of Risk and Uncertainty 5, 297-323.

46

Table 1 Dividend Volatility Estimates Panel A: Dividend Volatility Clustering Tests Lag

4

8

12

Box-Pierce-Ljung Test

32:57

93:29

101:37

ARCH Test

24:96

66:72

73:58

Panel B: AR(1)-EAGRCH(1,1) Estimation Parameters Values

^0

^1

0:002

(0:001)

^1

^

0:451

0:224

(0:0618)

^1 A 0:423

0:968

(0:014)

(0:109)

(0:077)

^1 L 0:037 (0:034)

Panel C: Augmented Dicky-Fuller Test Critical Values Test Statistic 11:8

1%

5%

20:3

14:0

10% 11:2

Panel A reports the test statistics for the Box-Pierce-Ljung test and the ARCH test for lag=4, 8 and 12 on the quarterly dividend growth rate from 1926.Q3 to 2006.Q3. Panel B models the dividend growth rate, gD;t+1 , as AR(1)-EAGRCH(1,1), gD;t+1 =

+ Zt+1

1

log

2 t 1

+ A1 [jZt j

E jZt j] + L1 Zt , where

2 t

0

+

1 gD;t

+

t Zt+1 ,

log

2 t

=

is the conditional variance of gD;t+1 , and

i:i:d:N (0; 1). Panel C reports an augmented Dicky-Fuller test on the log of the conditional

volatility series estimated by an AR(1)-EAGRCH(1,1). In Panel B, the standard errors of the estimated parameters are reported in parentheses. In Panels A and B,

and

mean that the

estimates are signi…cantly di¤erent from zero at the 1% and 5% levels, respectively.

47

Table 2 Calibrated Parameters Parameters

Calibration Values

Preference 1:0 0:995 range

b0

2:25 Technology gC

0:46%

gD

0:39%

C

1:90%

log ( )

3:91

!

0:16 0:968 0:38 0:25

u

This table reports the calibration values for the preference parameters and technology parameters in the theoretical model. factor,

is the curvature of utility over consumption,

is the time discount

is the loss aversion parameter, and b0 controls the importance of loss aversion/narrowing

framing relative to consumption in the utility function. gC and gD are the means of the consumption and dividend growth rate, respectively.

C

is the volatility of consumption growth. log ( )

is the mean of the log of the conditional volatility of dividend growth. of dividend volatility, while

u

measures the persistence

controls the variation in dividend volatility. ! is the correlation

between consumption news and dividend news, and

is the correlation between dividend level news

and volatility news. The calibration for the quarterly dividend parameters is based on the dividend sample 1926.Q3-2006.Q3 constructed from the value-weighted NYSE/AMEX returns from CRSP.

48

Table 3 Asset Prices and Annual Returns

Variables

Empirical Value

Model

(1926-2006)

b0 = 0:7 b0 = 2 b0 = 6

Annual Excess Stock Return a E rt+1

a rf;t

5:90

1:96

3:96

5:59

a rt+1

a rf;t

19:17

14:89

16:36

17:49

0:31

0:13

0:24

0:32

0:1

0:1

0:1

0:1

E (Pta =Dta )

29:08

22:31

15:40

12:47

(Pta =Dta )

12:43

2:07

2:02

1:93

a E rt+1

a rf;t =

a rt+1

a rf;t

a a Corr rt+1 ; gC;t+1

Annual Price-Dividend Ratio

This table provides information regarding stock returns for the simulated and historical data. The historical data correspond to the period 1926-2006. The entries for the model are based on 10,000 simulations each with 320 quarterly observations, which are time-aggregated to an annual frequency. The parameter con…guration in simulation follows that in Table 2. The expressions a E rt+1

a rf;t and

a rt+1

a rf;t are, respectively, the mean and volatility of the annualized con-

a a tinuously compounded returns. Corr rt+1 ; gC;t+1 is the correlation between the annual stock

return and the annual consumption growth rate. E (Pta =Dta ) and volatility of the annualized price-dividend ratios.

49

(Pta =Dta ) are the mean and

Table 4 Autocorrelations of Returns and Price-Dividend Ratios

Corr rta ; rta

Empirical Value

Model

(1926-2006)

b0 = 0:7 b0 = 2 b0 = 6

j

j=1

0:09

0:01

0:01

0:01

j=2

0:17

0:02

0:03

0:04

j=3

0:06

0:02

0:03

0:03

j=4

0:12

0:01

0:03

0:02

j=5

0:06

0:01

0:02

0:02

Corr Pta =Dta ; Pta j =Dta

j

j=1

0:90

0:48

0:67

0:72

j=2

0:81

0:40

0:56

0:60

j=3

0:75

0:34

0:47

0:50

j=4

0:68

0:28

0:39

0:42

j=5

0:60

0:23

0:32

0:34

This table reports the autocorrelations of annualized stock returns and price-dividend ratios for the simulated and historical data. The historical data correspond to the period 1926-2006. The entries for the model are based on 10,000 simulations each with 320 quarterly observations, which are time-aggregated to an annual frequency. The parameter con…guration in these simulations follows that used in Table 2. The expressions Corr rta ; rta

j

and Corr Pta =Dta ; Pta j =Dta

j

are,

respectively, the autocorrelations of the annualized compounded equity returns and P/D ratios.

50

Table 5 Return Predictability Regressions Empirical Value

Model

(1926-2006)

b0 = 0:7 b0 = 2 b0 = 6

1

2:55

2:33

2:16

2:14

2

5:99

4:91

4:45

4:20

3

8:28

7:22

6:47

5:95

4

11:26

9:33

8:17

7:39

R2 (1)

0:04

0:04

0:05

0:06

R2 (2)

0:09

0:05

0:08

0:11

R2 (3)

0:13

0:07

0:11

0:14

R2 (4)

0:18

0:08

0:13

0:17

This table provides evidence for the predictability of future excess returns by price-dividend a a + + rt+2 ratios. For both historical and model simulateddata, we conduct this regression: rt+1

(Dta =Pta ) +

a refers to the annual cumulative log returns from where rt+j

a = ::: + rt+j

j

year t + j

1 to t + j . The historical data correspond to the period 1926-2006. The model

+

j

j;t ,

generated data are based on 10,000 simulations each with 320 quarterly observations, which are then time-aggregated to an annual frequency. The parameter con…guration used in the simulation follows that in Table 2.

51

Table 6 GARCH and Volatility Feedback E¤ects Empirical Value Variables

Model

(1926-2006)

b0 = 0

b0 = 2

b0 = 6

GARCH (cG )

0:854

0:969

0:982

0:984

ARCH (cA )

0:346

0:586

0:264

0:248

Volatility Feedback (cV )

0:112

0:017

0:215

0:209

This table shows estimates of the GARCH and volatility feedback e¤ects obtained using equation (11). The empirical values are based on the historical quarterly value-weighted CRSP market returns from 1926 to 2006. For the model-simulated values, we make 160,000 random draws of

"t+1 and ut+1 for di¤erent values of b0 , compute the stock return time series using the solved price-dividend ratio function f ( ), and then implement the estimation of equation (11) based on the simulated stock returns.

Table 7 Results for Calibration Using Postwar Data Panel A: Postwar Dividend Parameter Calibration Parameters Values

gD

u

0:17

0:50%

log ( ) 4:19

0:15

Panel B: Equity Premiums Variables a E rt+1

a rf;t

Empirical Value

Model

(1954–2006)

b0 = 0:7 b0 = 2 b0 = 6

4:87

1:36

2:35

3:15

This table reports the mean and volatility of stock returns for the simulated data for the parameter con…guration obtained from calibration using the postwar dividend sample 1954.Q32006.Q3. The preference parameters and consumption parameters are the same as in Table 2.

52

Table 8

Sensitivity of Stock Returns to Empirical Value

Variables

(1926-2006)

Model =0

= 0:45

= 0:968

a E rt+1

a rf;t

5:90

3:25

3:26

5:59

a rt+1

a rf;t

19:17

12:80

13:12

17:49

0:31

0:26

0:25

0:32

a E rt+1

a rf;t =

a rt+1

a rf;t

controls the persistence of dividend volatility. The results are for the case of

The parameter

b0 = 6; other parameters are …xed at the values used in Table 2.

Table 9

Sensitivity of Stock Returns to Empirical Value

Variables

(1926-2006)

Model = 1:5

= 2:25

=3

a E rt+1

a rf;t

5:90

1:41

5:59

9:63

a rt+1

a rf;t

19:17

14:36

17:49

19:57

0:31

0:10

0:32

0:50

a E rt+1

The parameter

a rf;t =

a rt+1

a rf;t

controls the investor’s loss aversion. The results are for the case of b0 = 6;

other parameters are …xed at the values used in Table 2.

53

Sensitivity of Stock Returns to b0

Table 10

Empirical Value Variables

Model

(1926-2006)

b0 = 0

b0 = 100

a E rt+1

a rf;t

5:90

0:69

6:64

a rt+1

a rf;t

19:17

12:68

18:11

0:31

0:05

0:37

a rt+1

a = rf;t

a E rt+1

a rf;t

Parameter b0 controls the importance of loss aversion/narrow framing relative to consumption in the utility function. Other parameters are …xed at the values used in Table 2.

Table 11

Results for Annual Decision Interval

Variables

Empirical Value

Model

(1926-2006)

b0 = 0:7 b0 = 2 b0 = 6

a E rt+1

a rf;t

5:90

0:77

1:42

1:85

a rt+1

a rf;t

19:17

10:20

10:25

10:43

0:31

0:09

0:15

0:20

a E rt+1

a rf;t =

a rt+1

a rf;t

This table assumes that the investor evaluates gains or losses on an annual basis. We use the annual frequency data to re-calibrate the parameter values as follows:

gC = 1:84%, gD = 1:57%,

C=

3:79%, log ( ) = 2:75, ! = 0:13,

54

= 1,

= 0:98,

= 0:879, = 0 and

= 2:25, u=

0:35.

Figure 1 Gain and Loss Function 20

10

v(x)

0

-10

-20

-30

-40

-5

-10

-15

-20

0

Gains/Losses x

20

15

10

5

 X , if X t +1 > 0 Figure 1 plots the gain and loss function v( X t +1 ) =  t +1 for λ > 1 . λX t +1 , if X t +1 ≤ 0

Figure 2 Price-Dividend Functions f (a) b0=0

(b) b0=100

380

(c) b0=0.7

65

360

105

60

Price-dividend Ratio f(σt)

100 340 55

320

95 50

300 45

90

280 40 260

85

35 240 80 30

220

200 -7

-6

-5

-4

log(σt)

-3

-2

-1

25 -7

-6

-5

-4

log(σt)

-3

-2

-1

75 -7

-6

-5

-4

log(σt)

-3

Figure 2 plots the equilibrium price-dividend ratios against the log of the conditional dividend volatility, log( σ t ), for b0 = 0, 100 and 0.7. 55

-2

-1

Figure 3 Historical Price-Dividend Ratios vs. Conditional Dividend Volatility 70

Price-dividend Ratios Pt/Dt

60

50

40

30

20

10

0

σt Estimates from AR(1)-EGARCH(1,1)

0.1

0.09

0.08

0.07

0.06

0.05

0.04

0.03

0.02

0.01

0

Figure 3 plots the historical price-dividend ratios against the conditional dividend volatility estimated from an AR(1)-EGARCH(1,1) regression for period 1926.Q32006.Q3.

Figure 4 Conditional Moments of Stock Returns (a)

(b)

1.15

0.7

Conditional Standard Deviation σt(Rt+1)

Conditional Expected Return Et(Rt+1)

0.6

1.1

1.05

0.5

0.4

0.3

0.2

0.1

1 -6

-5.5

-5

-4.5

-4

-3.5

-3

log(σt)

-2.5

-2

-1.5

0 -6

-1

-5.5

-5

-4.5

-4

-3.5

-3

log(σt)

-2.5

-2

-1.5

-1

Panel (a) and (b) plot the conditional expected stock return Et (Rt +1 ) and the conditional volatility of return σt (Rt +1 ) when b0=6. 56

Figure 5 Distribution of the Conditional Sharpe Ratios 4

3.5

x 10

3

Frequency

2.5

2

1.5

1

0.5

0

0

0.05

0.1

0.15

0.2

Quarterly Conditional Sharpe Ratios

0.25

0.3

The distribution is based on a simulation for the case b0=6. Specifically, we make 160,000 random draws of ε t +1 and ut +1 , calculate the conditional mean and conditional standard deviation of expected returns by numerical integration, and then obtain the conditional Sharpe ratios as a function of log( σ t ) when b0=6.

57