#### DMCA

## By Force of Habit: A Consumption-Based Explanation of Aggregate Stock Market Behavior (1999)

### Cached

### Download Links

Citations: | 1452 - 72 self |

### Citations

1978 | A theory of the term structure of interest rates - Cox, Ingersoll, et al. - 1985 |

1750 | The equity premium: a puzzle. - Mehra, Prescott - 1985 |

1418 |
Substitution, Risk Aversion, and the Temporal Behavior of Consumption and Asset Returns: A Theoretical Framework.
- Epstein, Zin
- 1989
(Show Context)
Citation Context ...995) produces a risk-free rate that varies over time as a function of the surplus consumption ratio, whereas consumption growth is i.i.d. Therefore, this model predicts no time-series relationship at all between interest rates and expected consumption growth rates, consistent with the great difficulty the empirical literature has found in documenting any such relation in the data. In order to remove the tension between equity premia as in (22) and risk-free rates as in (23), our model uses non-time-separable preferences to distinguish intertemporal substitution and risk aversion. Weil (1989), Epstein and Zin (1991), Kandel and Stambaugh (1991), and Campbell (1996) use non-state-separable preferences to the same effect but do not generate time-varying risk aversion. Our solution to the risk-free rate puzzle has one other important advantage. Abel (1999) highlights the danger of accounting for an equity premium by a term premium. If a model assigned a high premium to the interest rate exposure of stock cash flows and longterm bond cash flows alike, it would account for the equity premium of stocks over short-term bonds, but it would counterfactually predict high expected returns for long-term bonds as wel... |

1267 |
Conditional Heteroskedasticity in Asset Returns: A
- Nelson
- 1992
(Show Context)
Citation Context ... right sign: high prices forecast low returns. Since high prices forecast low returns for many years in the future, the forecastability of returns in228 JOURNAL OF POLITICAL ECONOMY creases with the horizon, as we show next. The correlations are slightly smaller for the dividend claim since its return is slightly noisier. The cross-correlations between the price/dividend ratio or returns and subsequent absolute returns show that a low price/consumption ratio or a big price decline signals high volatility for several years ahead. This is the "leverage effect" that Black (1976), Schwert (1989), Nelson (1991), and many others have found in the data. As with the univariate autocorrelation of absolute returns, the data seem to indicate a somewhat shorter-lasting change in conditional variance than is predicted by the model, at least as viewed by this simple statistic. Again, the dividend claim behaves much like the consumption claim, despite the very low .2 correlation of dividend growth with consumption growth. Long-Horizon Regressions Table 5 presents long-horizon regressions of log excess stock returns on the log price/dividend ratio in simulated and historical data. We use excess returns to emph... |

811 | Asset Prices under Habit Formation and Catching Up with the Joneses. American Economic Review papers and proceedings
- Abel
- 1990
(Show Context)
Citation Context ...umption (e.g., Ferson and Constantinides 1991). This feature produces slow mean reversion in the price/dividend ratio, long-horizon return forecastability, and persistent movements in volatility. Third, we specify that habit adapts nonlinearly to the history of consumption. The nonlinearity keeps habit always below consumption and keeps marginal utility always finite and positive even in an endowment economy. In many models, including those of Sundaresan (1989), Ferson and Constantinides (1991), Heaton (1995), and Chapman (1998), consumption can fall below habit with undesirable consequences. Abel (1990,1999) keeps marginal utility positive by changing utility from u(C - X) to u(C/ X) , but this specification eliminates changing risk aversion. Most important, the nonlinear habit specification is essential for us to capture time variation in the Sharpe ratio (mean to standard deviation of returns) and a constant risk-free rate. 11. The Model A. Preferences and Technology Identical agents maximize the utility function AGGREGATE STOCK MARKET BEHAVIOR 209 Here X, is the level of habit, and 6 is the subjective time discount factor. It is convenient to capture the relation between consumption and ... |

782 |
Understanding Consumption,
- Deaton
- 1992
(Show Context)
Citation Context ...shes that are larger than the booms. We feed the model actual consumption data, and we find that the price/dividend ratios and returns predicted by our model provide a surprisingly good account of fluctuations in stock prices and returns over the last century. All these interesting and seemingly unrelated phenomena are in fact reflections of the same phenomenon, which is at the core of the model: a slowly time-varying, countercyclical risk premium. A. Habit Formation Habit formation has a long history in the study of consumption. Deaton and Muellbauer (1980) survey early work in the area, and Deaton (1992) gives a more recent overview. Ryder and Heal (1973), 208 JOURNAL OF POLITICAL ECONOMY Sundaresan (1989), and Constantinides (1990) are major theoretical papers on the subject. Habit formation captures a fundamental feature of psychology: repetition of a stimulus diminishes the perception of the stimulus and responses to it. Habit formation can explain why consumers' reported sense of well-being often seems more related to recent changes in consumption than to the absolute level of consumption. In macroeconomics, habit persistence can explain why recessions are so feared even though their effe... |

766 | ARCH Modeling in Finance: A Review of the Theory and Empirical Evidence,” - Bollerslev, Chou, et al. - 1992 |

677 | Habit Formation: A Resolution of the Equity Premium Puzzle, " - Constantinides - 1990 |

662 |
Why does stock market volatility change over time?
- Schwert
- 1989
(Show Context)
Citation Context ...returns with the right sign: high prices forecast low returns. Since high prices forecast low returns for many years in the future, the forecastability of returns in228 JOURNAL OF POLITICAL ECONOMY creases with the horizon, as we show next. The correlations are slightly smaller for the dividend claim since its return is slightly noisier. The cross-correlations between the price/dividend ratio or returns and subsequent absolute returns show that a low price/consumption ratio or a big price decline signals high volatility for several years ahead. This is the "leverage effect" that Black (1976), Schwert (1989), Nelson (1991), and many others have found in the data. As with the univariate autocorrelation of absolute returns, the data seem to indicate a somewhat shorter-lasting change in conditional variance than is predicted by the model, at least as viewed by this simple statistic. Again, the dividend claim behaves much like the consumption claim, despite the very low .2 correlation of dividend growth with consumption growth. Long-Horizon Regressions Table 5 presents long-horizon regressions of log excess stock returns on the log price/dividend ratio in simulated and historical data. We use excess ... |

519 |
Implications of security market data for models of dynamic economies.
- Hansen, Jagannathan
- 1991
(Show Context)
Citation Context ...free interest rate. B. Marginal Utility Since habit is external, marginal utility is uc(C,, X,) = (C, - X,) = s;yc;y. The intertemporal marginal rate of substitution is then -- AGGREGATE STOCK MARKET BEHAVIOR 2 1 1 It is related to the state variable st and the log consumption innovation vttl by We can now calculate moments of the marginal rate of substitution and find asset prices. Slope of the Mean-Standard Deviation Frontier The slope of the conditional mean-standard deviation frontier can be found from the conditional moments of the marginal rate of substitution. Following Shiller (1982), Hansen and Jagannathan (1991) show that the first-order condition 0 = Et(Mt+,R;+I)for an excess return Reimplies that the Sharpe ratio of any asset return must obey where p,denotes a conditional correlation. In our model, Mis conditionally lognormal,' so we can find the largest possible Sharpe ratio by max Et(RL1) = {e7202['tk(1t)12 - 1]1,2 yo[l + acs,>1. (7) (allassets) (7 t (R:+ 1) This formula helps us to specify the model. To produce a timevarying Sharpe ratio, X(s) must vary with s. To produce risk prices that are higher in bad times, when s is low, 3L (s) and hence the volatility of s must increase as s declines. Ri... |

492 | Income, Saving, and the Theory of Consumer Behavior - Duesenberry - 1949 |

458 |
Understanding risk and return
- Campbell
- 1996
(Show Context)
Citation Context ...turn variances are somewhat less thansk times 1-year return variances, so the market Sharpe ratio grows, ifsanything, faster than the square root of the horizon (MaCurdy andsShoven 1992; Siege1 1994; =-=Campbell 1996-=-).sIn our model, the k-period stochastic discount factor issEquation (6) implies that the standard deviation of this discountsfactor must increase roughly with the square root of the horizon tosbe con... |

435 |
Asset pricing with heterogeneous consumers.
- Constantinides, Duffie
- 1996
(Show Context)
Citation Context ...it rather than the more common internal-habit specification. At heart, all three objections have to do with the potential application of the model to microeconomic data. This is not our chief concern in this paper. Our goal, ambitious enough, is to find representative-agent preferences that explain the joint behavior of aggregate consumption and stock returns. These representative-agent preferences could take the same form as the underlying preferences of individual agents, but they could also result from aggregation of heterogeneous consumers with quite different preferences. As one example, Constantinides and Duffie (1996), building on Mankiw (1986), show how to disaggregate any representative-agent marginal utility process, including ours, to individual agents with power utility and low risk aversion in incomplete markets by allowing the crosssectional variance of idiosyncratic income to vary with the posited marginal utility process. Nonetheless, we find external habit forma242 JOURNAL OF POLITICAL ECONOMY tion appealing as a description of individual preferences, and the representative-agent model is clearly more compelling if its preferences can result from aggregation of individuals with similar preference... |

358 |
The equity premium puzzle and the risk-free rate puzzle.
- Weil
- 1989
(Show Context)
Citation Context ... explain a (gross return) Sharpe ratio of 0.50 with o = 1.22 percent, the power utility model needs a risk aversion coefficient q r 41 by equation (22). This is Mehra and Prescott's (1985) "equity premium puzzle." One can object to q r 41 as an implausibly large value of risk aversion, and we discuss this interpretation below. More important, a high value of q makes the term q g in the riskfree rate equation very large. Thus q = 41 and g = 1.89 percent means that we need p = 1.90 to get a 1 percent risk-free rate. Imposing p 2 1, one predicts a risk-free rate of more than 90 percent per year! Weil (1989) emphasizes this "risk-free rate puzzle," and Cochrane and Hansen (1992) discuss the level and variability of riskfree interest rates in high-risk aversion models. Despite its intuitive implausibility, one might argue that setting p = 1.90 resolves the risk-free rate puzzle. However, with P = 1.90 and q = 41, equation (23) implies that the risk-free interest rate should be quite sensitive to the mean consumption growth rate, which is not the case. Real interest rates do not vary across time or countries by 40 times the variation in predicted or average consumption growth. (Equivalently, one mu... |

354 | Asset Pricing in Production Economies.
- Jermann
- 1998
(Show Context)
Citation Context ...nomy. In practice, the asset pricing predictions of many habit persistence economies are strongly affected by the specification of technology. In many endowment economies with habits and random walk consumption, riskfree rates vary a great deal, as the varying surplus consumption ratio gives rise to strong motives for intertemporal substitution. When production is added to these economies, consumers make strong use of production opportunities to smooth marginal utility over time. The interest rate variation is quieted down, but the equilibrium consumption process moves far from a random walk (Jermann 1998).However, we shall pick the functional forms and parameters of our model to generate a constant real risk-free rate. Therefore, we can also close the model with a linear technology: where K , and E, denote the capital stock and an exogenous endowment or additive technology shock, respectively. This specification results in exactly the same process for consumption and asset prices as the endowment specification does. This fact suggests that the model's consumption and asset pricing implications will not be much affected if the model is closed with any standard concave specification of technolog... |

320 | Mean Reversion in Stock Prices: Evidence and Implications," - Poterba, Summers - 1988 |

317 | http://links.jstor.org/sici?sici=0012-9682%28196107%2929%3A3%3C315%3AREATTO%3E2.0.CO%3B2-G The Output-Inflation Trade-off When Prices Are Costly to Change Michael Parkin The - URL - 1986 |

291 |
Economics and Consumer Behavior,
- Deaton, Muellbauer
- 1980
(Show Context)
Citation Context ... negatively skewed stock prices and returns, with occasional crashes that are larger than the booms. We feed the model actual consumption data, and we find that the price/dividend ratios and returns predicted by our model provide a surprisingly good account of fluctuations in stock prices and returns over the last century. All these interesting and seemingly unrelated phenomena are in fact reflections of the same phenomenon, which is at the core of the model: a slowly time-varying, countercyclical risk premium. A. Habit Formation Habit formation has a long history in the study of consumption. Deaton and Muellbauer (1980) survey early work in the area, and Deaton (1992) gives a more recent overview. Ryder and Heal (1973), 208 JOURNAL OF POLITICAL ECONOMY Sundaresan (1989), and Constantinides (1990) are major theoretical papers on the subject. Habit formation captures a fundamental feature of psychology: repetition of a stimulus diminishes the perception of the stimulus and responses to it. Habit formation can explain why consumers' reported sense of well-being often seems more related to recent changes in consumption than to the absolute level of consumption. In macroeconomics, habit persistence can explain wh... |

287 | Market Volatility. - Shiller - 1989 |

276 | http://links.jstor.org/sici?sici=0022-3808%28197910%2987%3A5%3C940%3AOTDOTP%3E2.0.CO%3B2-K Incumbent Behavior: Vote-Seeking, Tax-Setting, and Yardstick Competition Timothy Besley - URL - 1995 |

264 | The Equity Risk Premium: A Solution,” - Rietz - 1988 |

258 |
Time-varying conditional covariances in tests of asset pricing models,
- Harvey
- 1989
(Show Context)
Citation Context ... one for oneswith estimates of conditional mean returns. Hence the slope of thesconditional mean-variance frontier, a measure of the price of risk,schanges through time with a business cycle pattern (=-=Harvey 1989-=-;sChou, Engle, and Kane 1992).sAs yet, there is no accepted economic explanation for these obser-svations. In the language of finance, we lack a successful theory andsmeasurement procedure for the fun... |

216 |
Stock returns, expected returns, and real activity.
- Fama
- 1990
(Show Context)
Citation Context ... Stock Return (7.) - 4 ~ ~ ~ -60 -40 -20 0 20 40 60 8 0 Stock Return (7 . ) FIG. 7.-a, Simulated monthly consumption growth vs. monthly consumption claim returns. b, Growth in simulated annual consumption vs. annual consumption claim returns. TABLE 7 MONTHLYMODEL ANNUALMODEL DATA SAMPLE - Consumption Dividend Consumption Dividend Quarterly Annual Annual CORRELATION Claim Claim Claim Claim Postwar Postwar Long 233 AGGREGATE STOCK MARKET BEHAVIOR sumption growth. The highest correlations occur between returns and the next year's consumption growth: .37 in postwar data and .49 in long-term data. Fama (1990) interprets similar correlations of returns with output as evidence that returns move on news of future cash flows. In our model, the unconditional correlation between monthly consumption claim returns and monthly consumption growth is .93. This value corresponds to panel a of figure 7 . It is less than the 1.0 of the standard time-separable model but much greater than the correlations we see in the data. In addition, there is no correlation between returns and consumption growth at any lead or lag. When we time-aggregate the artificial data to annual frequencies, the contemporaneous correlati... |

168 |
Optimum Growth with Intertemporally Dependent Preferences.” Rev.
- Ryder, Heal
- 1973
(Show Context)
Citation Context ... the model actual consumption data, and we find that the price/dividend ratios and returns predicted by our model provide a surprisingly good account of fluctuations in stock prices and returns over the last century. All these interesting and seemingly unrelated phenomena are in fact reflections of the same phenomenon, which is at the core of the model: a slowly time-varying, countercyclical risk premium. A. Habit Formation Habit formation has a long history in the study of consumption. Deaton and Muellbauer (1980) survey early work in the area, and Deaton (1992) gives a more recent overview. Ryder and Heal (1973), 208 JOURNAL OF POLITICAL ECONOMY Sundaresan (1989), and Constantinides (1990) are major theoretical papers on the subject. Habit formation captures a fundamental feature of psychology: repetition of a stimulus diminishes the perception of the stimulus and responses to it. Habit formation can explain why consumers' reported sense of well-being often seems more related to recent changes in consumption than to the absolute level of consumption. In macroeconomics, habit persistence can explain why recessions are so feared even though their effects on output are small relative to a few years' gro... |

168 | Intertemporally Dependent Preferences and the Volatility of Consumption and Wealth,
- Sundaresan
- 1989
(Show Context)
Citation Context ... price/dividend ratios and returns predicted by our model provide a surprisingly good account of fluctuations in stock prices and returns over the last century. All these interesting and seemingly unrelated phenomena are in fact reflections of the same phenomenon, which is at the core of the model: a slowly time-varying, countercyclical risk premium. A. Habit Formation Habit formation has a long history in the study of consumption. Deaton and Muellbauer (1980) survey early work in the area, and Deaton (1992) gives a more recent overview. Ryder and Heal (1973), 208 JOURNAL OF POLITICAL ECONOMY Sundaresan (1989), and Constantinides (1990) are major theoretical papers on the subject. Habit formation captures a fundamental feature of psychology: repetition of a stimulus diminishes the perception of the stimulus and responses to it. Habit formation can explain why consumers' reported sense of well-being often seems more related to recent changes in consumption than to the absolute level of consumption. In macroeconomics, habit persistence can explain why recessions are so feared even though their effects on output are small relative to a few years' growth. Our habit specification has three distinctive f... |

166 |
Habit Persistence and Durability in Aggregate Consumption: Empirical Tests,"
- Ferson, Constantinides
- 1991
(Show Context)
Citation Context ...Second, we specify that habit moves slowly in response to con-ssumption, in contrast to empirical specifications in which each pe-sriod's habit is proportional to the last period's consumption (e.g.,s=-=Ferson and Constantinides 1991-=-). This feature produces slow meansreversion in the price/dividend ratio, long-horizon return fore-scastability, and persistent movements in volatility.sThird, we specify that habit adapts nonlinearly... |

166 |
An empirical investigation of asset pricing with temporally dependent preference specifications,
- Heaton
- 1995
(Show Context)
Citation Context ... to empirical specifications in which each period's habit is proportional to the last period's consumption (e.g., Ferson and Constantinides 1991). This feature produces slow mean reversion in the price/dividend ratio, long-horizon return forecastability, and persistent movements in volatility. Third, we specify that habit adapts nonlinearly to the history of consumption. The nonlinearity keeps habit always below consumption and keeps marginal utility always finite and positive even in an endowment economy. In many models, including those of Sundaresan (1989), Ferson and Constantinides (1991), Heaton (1995), and Chapman (1998), consumption can fall below habit with undesirable consequences. Abel (1990,1999) keeps marginal utility positive by changing utility from u(C - X) to u(C/ X) , but this specification eliminates changing risk aversion. Most important, the nonlinear habit specification is essential for us to capture time variation in the Sharpe ratio (mean to standard deviation of returns) and a constant risk-free rate. 11. The Model A. Preferences and Technology Identical agents maximize the utility function AGGREGATE STOCK MARKET BEHAVIOR 209 Here X, is the level of habit, and 6 is the su... |

166 | http://links.jstor.org/sici?sici=0022-166X%28198321%2918%3A2%3C268%3AANLATC%3E2.0.CO%3B2-8 Changes in the Labor Market for Black Americans, - URL - 1948 |

149 | The Equity Premium and the Concentration of Aggregate Shocks,”
- Mankiw
- 1986
(Show Context)
Citation Context ...specification. At heart, all three objections have to do with the potential application of the model to microeconomic data. This is not our chief concern in this paper. Our goal, ambitious enough, is to find representative-agent preferences that explain the joint behavior of aggregate consumption and stock returns. These representative-agent preferences could take the same form as the underlying preferences of individual agents, but they could also result from aggregation of heterogeneous consumers with quite different preferences. As one example, Constantinides and Duffie (1996), building on Mankiw (1986), show how to disaggregate any representative-agent marginal utility process, including ours, to individual agents with power utility and low risk aversion in incomplete markets by allowing the crosssectional variance of idiosyncratic income to vary with the posited marginal utility process. Nonetheless, we find external habit forma242 JOURNAL OF POLITICAL ECONOMY tion appealing as a description of individual preferences, and the representative-agent model is clearly more compelling if its preferences can result from aggregation of individuals with similar preferences. Therefore, we now briefl... |

114 |
Expectations and Volatility of Consumption and Asset Returns,"
- KANDEL, STAMBAUGH
- 1990
(Show Context)
Citation Context ...d identically distributed (i.i.d.) lognormal process, with the same mean and standard deviation as postwar consumption growth. Our model can accommodate more complex consumption processes, including processes with predictability, conditional heteroskedasticity, and nonnormality. But these features are not salient characteristics of consumption data. More important, we want to emphasize that the model generates interesting asset price behavior internally, not from exogenous variation in the probability distribution of consumption growth. In this respect, our approach is the opposite of that of Kandel and Stambaugh (1990, 1991), who use fairly standard preferAGGREGATE STOCK MARKET BEHAVIOR 20'7 ences but derive some of these phenomena from movement over time in the conditional moments of consumption growth. We choose our model's functional form and parameters so that the risk-free interest rate is constant. We do this for several reasons. First, there appears to be only limited variation in the real risk-free rate in historical U.S. data, and the variation that does exist is not closely related to the business cycle or to movements in stock prices. Second, we want to show how the model can explain stock marke... |

60 | Volatility tests and efficient markets: A review essay,
- Cochrane
- 1991
(Show Context)
Citation Context ...ement cannot be explained by variation in expectedsdividends or interest rates, indicating large countercyclical variationsin expected excess returns (Campbell and Shiller 1988a, 1988b;sShiller 1989; =-=Cochrane 1991-=-, 1992). Estimates of conditional vari-sances of returns also change through time (see Bollerslev, Chou,sand Kroner [I9921 for a survey), but they do not move one for oneswith estimates of conditional... |

52 |
Risk and return: consumption beta versus market beta,
- Mankiw, Shapiro
- 1986
(Show Context)
Citation Context ...umption Growth and Stock ReturnssThe static capital asset pricing model (CAPM) often does a bettersjob of accounting for risk premia than the consumption-based assetspricing model with power utility (=-=Mankiw and Shapiro 1986-=-). Itsturns out that this is true in our artificial data as well, even thoughsthe data are generated by a consumption-based model. Campbellsand Cochrane (1998b) present detailed calculations. We show ... |

45 | Measuring risk aversion from excess returns on a stock index, - Chou, Engle, et al. - 1992 |

42 | the Theory of Consumer Behavior - Duesenberry, Income - 1949 |

33 | http://links.jstor.org/sici?sici=0893-9454%28199423%297%3A3%3C475%3ARRATCB%3E2.0.CO%3B2-K Investment Bank Reputation, Information Production, - URL - 1994 |

23 |
Consumption, asset markets, and macroeconomic fluctuations,
- Shiller
- 1982
(Show Context)
Citation Context ...y constant risk-free interest rate. B. Marginal Utility Since habit is external, marginal utility is uc(C,, X,) = (C, - X,) = s;yc;y. The intertemporal marginal rate of substitution is then -- AGGREGATE STOCK MARKET BEHAVIOR 2 1 1 It is related to the state variable st and the log consumption innovation vttl by We can now calculate moments of the marginal rate of substitution and find asset prices. Slope of the Mean-Standard Deviation Frontier The slope of the conditional mean-standard deviation frontier can be found from the conditional moments of the marginal rate of substitution. Following Shiller (1982), Hansen and Jagannathan (1991) show that the first-order condition 0 = Et(Mt+,R;+I)for an excess return Reimplies that the Sharpe ratio of any asset return must obey where p,denotes a conditional correlation. In our model, Mis conditionally lognormal,' so we can find the largest possible Sharpe ratio by max Et(RL1) = {e7202['tk(1t)12 - 1]1,2 yo[l + acs,>1. (7) (allassets) (7 t (R:+ 1) This formula helps us to specify the model. To produce a timevarying Sharpe ratio, X(s) must vary with s. To produce risk prices that are higher in bad times, when s is low, 3L (s) and hence the volatility of s ... |

18 |
Non-stationarity and Stage-of-the-Business-Cycle Effects in Consumption-Based Asset Pricing Relations,
- Ferson, Merrick
- 1987
(Show Context)
Citation Context ... Lars Hansen for helpful comments. Ubumal ofPolitical Economy, 1999, vol. 107, no. 21 O 1999 by The University of Chicago. All rights reserved. 0022-3808/99/0702-0004$02.50 206 JOURNAL OF POLITICAL ECONOMY I. Introduction A number of empirical observations suggest tantalizing links between asset markets and macroeconomics. Most important, equity risk premia seem to be higher at business cycle troughs than they are at peaks. Excess returns on common stocks over Treasury bills are forecastable, and many of the variables that predict excess returns are correlated with or predict business cycles (Ferson and Merrick 1987; Fama and French 1989). The literature on volatility tests mirrors this conclusion: price/dividend ratios move procyclically, but this movement cannot be explained by variation in expected dividends or interest rates, indicating large countercyclical variation in expected excess returns (Campbell and Shiller 1988a, 1988b; Shiller 1989; Cochrane 1991, 1992). Estimates of conditional variances of returns also change through time (see Bollerslev, Chou, and Kroner [I9921 for a survey), but they do not move one for one with estimates of conditional mean returns. Hence the slope of the conditional ... |

16 | Premia and Term Premia in General Equilibrium, - Risk - 1999 |

15 |
Asset Pricing Lessons for Macroeconomics." In
- Cochrane, Hansen
- 1992
(Show Context)
Citation Context ...this fact. The Correlation of Consumption Growth with Stock Returns Equilibrium consumption-based models typically imply that consumption growth and stock returns are highly, if not perfectly, correlated. For example, with log utility, the return on the wealth portfolio equals consumption growth, ex post, data point for data point. This implication is the basis for many theoretical models in finance that substitute portfolio returns for consumption growth. However, this implication is seldom checked or used to test asset pricing models, for the obvious reason that it is dramatically false. As Cochrane and Hansen (1992) emphasize, the actual low correlation between stock returns and consumption growth lies at the heart of many empirical failures of the consumption-based model. In our model, consumption growth and consumption claim returns are conditionally perfectly correlated since consumption growth is the only source of uncertainty. But the relation between consumption growth and returns varies over time with the surplus consumption ratio. Hence the unconditional correlation between consumption growth and returns is not perfect. Panel a of figure 7 shows this effect of conditioning information by plotting... |

15 |
Recursive Models of Dynamic Linear Economies." Manuscript.
- Hansen, Sargent
- 1998
(Show Context)
Citation Context ...umption and asset pricing implications. With internal habits, consumption today raises future habits, lowering the overall marginal utility of consumption today. But asset prices are determined by ratios of marginal utilities. If internal habits simply lower marginal utilities at all dates by the same proportion, then a switch from external to internal habits has no effect on allocations and asset prices. For example, we show in the appendix (Campbell and Cochrane 1998a) that this occurs with power utility (C - X)'-Y, a constant interest rate, and linear habit accumulation X,= 0C,"=l@ j Ct-j. Hansen and Sargent (1998) provide a similar example. Our model adopts a nonlinear habit accumulation equation to generate an exact random walk in consumption along with a constant risk-free rate. (In the linear habit example above, consumption is close to but not exactly a random walk.) The nonlinearity in the habit accumulation process is thus the only reason there is any difference between the internal-habit and external-habit specification of our model. Still, it is interesting to know how big this difference is. When habit is internal, marginal utility at time t in our model has extra terms reflecting the effect o... |

6 | Asset pricing lessons for modeling business cycles. - Boldrin, Christiano, et al. - 1995 |

3 |
Accumulating Pension Wealth with Stocks and Bonds," working paper,
- MaCurdy, Shoven
- 1992
(Show Context)
Citation Context ...-term bonds, and the entire equity premium is a risk premium, not a term premium. 239 AGGREGATE STOCK MARKET BEHAVIOR The Long-Run Equity Premium The equity premium puzzle is a feature of long as well as short horizons. Consumption is roughly a random walk at any horizon, so the standard deviation of consumption growth grows roughly with the square root of the horizon. The negative autocorrelation of stock returns means that k-year return variances are somewhat less than k times 1-year return variances, so the market Sharpe ratio grows, if anything, faster than the square root of the horizon (MaCurdy and Shoven 1992; Siege1 1994; Campbell 1996). In our model, the k-period stochastic discount factor is Equation (6) implies that the standard deviation of this discount factor must increase roughly with the square root of the horizon to be consistent with the long-run equity premium, and even faster to generate the negative autocorrelation of stock returns. One can think of our model as a member of a large class that adds a new state variable to the discount factor. However, most extra state variables-such as recessions, labor, and instruments for timevarying expected returns ("shifts in the investment oppor... |

1 |
Dividend Ylelds and Expected Stock Returns
- Fama, French
- 1988
(Show Context)
Citation Context ...se givensthat the dividend claim and consumption claim price/dividend ratios are almost identical.sReturns display a series of small negative autocorrelations thatsgenerate univariate mean reversion (=-=Fama and French 1988-=-b; PoTABLE 3sAUTOCORRELATIONS AND HISTORICALOF IMULATED DATAsLAG (Years)sp - d :sConsumption claimsDividend claimsPostwar samplesLong samplesr - rJ:sConsumption claimsDividend claimsPostwar samplesLon... |

1 | Stocksfor the Long Run: A Guide to Selecting Markets for LongT m Growth. Burr - Siegel - 1994 |

1 | Statis. - American - 1976 |

1 | Asset Pricing Lessons for Modeling Business Cycles." Manuscript. - Boldrin, Christiano, et al. - 1996 |

1 | Prices, Consumption, and the Business Cycle." - Asset - 1999 |

1 | You have printed the following article: By Force of Habit: A Consumption-Based Explanation of Aggregate Stock Market Behavior - Campbell, John - 1999 |