Published and Working Papers:nb: For published papers, the final published versions can differ from the working paper versions posted below.
Created: 12/99; Updated: 7/01
Previous tests for liquidity constraints using consumption Euler equations have frequently split the sample on the basis of wealth, arguing that low-wealth consumers are more likely to be constrained. We propose alternative tests using different and more direct information on borrowing constraints obtained from the 1983 Survey of Consumer Finances. In a first stage we estimate probabilities of being constrained, which are then utilized in a second sample, the Panel Study of Income Dynamics, to estimate switching regression models of the Euler equation. Our estimates indicate stronger excess sensitivity associated with the possibility of liquidity constraints than the sample splitting approach. In particular, households without access to credit are less able to smooth their consumption past income fluctuations than households with access to credit. (published version (JStore))
Previous tests of the life-cycle/permanent-income theory on micro-data have usually had difficulty isolating the predictable component of income. This paper avoids this difficulty by focusing on income tax refunds, which are both predetermined and transitory. Using the Consumer Expenditure Survey, it finds significant evidence of excess sensitivity. This sensitivity is due in part to increased spending on nondurables at the time of refund-receipt by those likely to be liquidity constrained. However, there is also evidence of increases in spending, mostly on durables, by those unlikely to be constrained; as well as in spending well after the receipt of refunds, which cannot be due to liquidity constraints.
The large magnitudes of these responses to refunds, interpretable as marginal propensities to consume, evidence a greater impact of fiscal policy than found by most previous studies. The effect on the timing of durables purchases is not easily explained by standard models of durables. The responses also serve to evaluate some recent behavioral theories of saving, mental accounts and self-control/forced savings. (published version (JStore))
Despite the high cost of college, there has been little study of the adequacy of household savings and other resources available to fund college. To gauge their adequacy this paper examines households' standard of living as they pay for college. Using the Consumer Expenditure Survey, the main finding is that households appear to do a relatively good job smoothing their consumption into the academic year, despite large expenses. This is consistent with the Life-Cycle Theory of saving and consumption. There is some evidence of a delayed decline in consumption, and of a decline for households with children first beginning college, but the magnitudes of these declines are rather small. (Elsevier)
"An Empirical Analysis of Personal Bankruptcy and Delinquency" (with D. Gross), Review of Financial Studies, 15(1), Spring, 2002.
This paper uses a new panel data set of credit card accounts to analyze credit card delinquency, personal bankruptcy, and the stability of credit risk models. We estimate duration models for default and assess the relative importance of different variables in predicting default. We investigate how the propensity to default has changed over time, disentangling the two leading explanations for the recent increase in default rates -- a deterioration in the risk-composition of borrowers versus an increase in borrowers' willingness to default due to declines in default costs, including social, information, and legal costs. Even after controlling for risk-composition and other economic fundamentals, the propensity to default significantly increased between 1995 and 1997. Conversely, increases in credit limits and other changes in risk-composition explain only a small part of the change in default rates. Standard default models appear to have missed an important time-varying default factor, consistent with a decline in default costs. (Published version, (c)Society for Financial Studies)
"Consumer Response to the Reagan Tax Cuts," Journal of Public Economics, 85(1), 2002.
This paper uses micro data from the Consumer Expenditure Survey to estimate the response of household consumption to the second and third phases of the Reagan tax cuts. Since these phases were pre-announced, they allow for an unusually clean test of the canonical Life-Cycle /Permanent-Income model. Consumption is found to be excessively sensitive to the tax cuts, counter to the model. Liquidity constraints and other standard explanations do not appear to explain this excess sensitivity. The consumption response is larger than previously estimated for tax refunds and more concentrated in nondurables. These differences have important implications for the structure of tax changes, in particular for choosing between varying withholding rates versus varying 'lumpy' final tax payments and refunds. (Elsevier)[Home] [Teaching] [Research/Top]
"Do Liquidity Constraints and Interest Rates Matter for Consumer Behavior? Evidence from Credit Card Data" (with D. Gross), Quarterly Journal of Economics, 117(1), February 2002.
This paper utilizes a unique new dataset of credit card accounts to analyze how people respond to changes in credit supply. The data consist of a panel of thousands of individual credit card accounts from several different card issuers, with associated credit bureau data. We estimate both marginal propensities to consume (MPCs) out of liquidity and interest-rate elasticities. We also evaluate the ability of different models of consumption to rationalize our results, distinguishing the Permanent-Income Hypothesis (PIH), liquidity constraints, precautionary saving, and behavioral models.
We find that increases in credit limits generate a significant rise in debt, counter to the PIH. The average "MPC out of liquidity" (dDebt/dLimit) ranges between 10%-14%. The MPC is much larger for people starting near their limits, consistent with binding liquidity constraints. However, the MPC is significant even for people starting well below their limit. We show this response is consistent with buffer-stock models of precautionary saving. Nonetheless there are other results that conventional models cannot easily explain, for example, why so many people are borrowing on their credit cards, and simultaneously holding low yielding assets. Unlike most other studies, we also find strong effects from changes in account-specific interest rates. The long-run elasticity of debt to the interest rate is approximately -1.3. Less than half of this elasticity represents balance-shifting across cards; most reflects net changes in total borrowing. The elasticity is larger for decreases in interest rates than for increases, which can explain the widespread use of temporary promotional ("teaser") rates. (Working paper, published version)
In spite of the important role of income expectations in economics, little is known about how people actually form these expectations. We use a unique data set that contains explicit income expectations of almost 30 cohorts of medical students to examine how people form income expectations. We find that medical students significantly condition their income expectations on personal characteristics such as gender and ability. For instance, female students expect to earn substantially less than male students, even controlling for differences in the hours they expect to work. Income expectations also increase with the contemporaneous income of physicians currently practicing in the specialty a students plan to enter, which is consistent with learning models. Nonetheless, expectations are also significantly forward-looking: after students report relatively high income expectations for a given specialty, physicians practicing in that specialty subsequently tend to experience high income growth relative to physicians in other specialties. We also find that explicit income expectations are useful for predicting behavior. A specialty-choice model that uses the students' explicit income expectations has a better fit than a model that assumes income expectations are formed statically, and even a model that assumes that students have perfect foresight regarding their future income. (working paper)
Although income expectations play a central role in many economic decisions, little is known about the sources of income prediction errors and how people respond to income shocks. This paper uses a unique panel data set to examine the accuracy of physicians' income expectations, the sources of income prediction errors, and the effect of income prediction errors on physician behavior. The data set contains direct survey measures of income expectations for a generation of medical students, their corresponding income realizations, and a rich summary of the shocks hitting their medical practices. We find that income prediction errors were positive on average over the sample period, but varied significantly over time and cross-sectionally. We trace these results to aggregate and group-level shocks, especially persistent specialty-specific shocks such as the growth of HMOs and other changes in the health care market. Physicians who experienced negative income shocks were more likely to respond by increasing their hours worked. (working paper)
"Expectations, Heterogeneous Forecast Errors, and Consumption: Micro Evidence from the Michigan Consumer Sentiment Surveys" Journal of Money, Credit, and Banking, 36(1), February 2004. (Formerly titled "Consumer Sentiment: Its Rationality and Usefulness in Forecasting Expenditure: Evidence from the Michigan Micro Data".)
This paper provides one of the first comprehensive analyses of the household data underlying the Michigan Index of Consumer Sentiment. This data is used to test the rationality of consumer expectations and to assess their usefulness in forecasting expenditure. The results can also be interpreted as characterizing the shocks that have hit different types of households over time. Expectations are found to be biased, at least ex post, in that forecast errors do not average out even over a sample period lasting almost 20 years. People underestimated the disinflation of the early 1980's and in the 1990's, and generally appear to underestimate the amplitude of business cycles. Forecasts are also inefficient, in that people’s forecast errors are correlated with their demographic characteristics and/or aggregate shocks did not hit all people uniformly.
Further, sentiment is found to be useful in forecasting future consumption, even controlling for lagged consumption and macro variables like stock prices. This excess sensitivity is counter to the permanent income hypothesis [PIH]. Higher confidence is correlated with less saving, consistent with precautionary motives and increases in expected future resources. Some of the rejection of the PIH is found to be due to the systematic demographic components in forecast errors. But even after controlling for these components, some excess sensitivity persists. More broadly, these results suggest that empirical implementations of forward-looking models need to better account for systematic heterogeneity in forecast errors. (Published version)
This paper extends the empirical literature on portfolio choice in three ways. First, consistent with theoretical models of portfolio choice, it estimates tobit models of the ratio of risky securities to total wealth, proxied by consumption, not just financial wealth. Second, in response to evidence that transactions costs are important, (S,s)-type models of securities purchases (asset flows) are also estimated. Third, alternative measures of household risk, both hedging motives and background risks, are considered. In addition to previously studied income risk, whose drawbacks are discussed, the measures include "consumption risk". The Consumer Expenditure Survey is used to calculate the standard deviation of household consumption growth and the correlation of consumption growth with market returns, both instrumented to avoid endogeneity. Another set of measures is taken from the monthly Michigan consumer sentiment surveys, which have households themselves identify the risks they believe they will face in the future.
Both securities holdings and securities purchases are found to vary significantly with the alternative measures of household risk. Households with exogenously more volatile consumption, or a larger consumption-return covariance, hold and buy fewer securities. Households that are pessimistic about the future, expecting a deterioration in financial conditions or an increase in unemployment risk, also hold and buy fewer securities. By contrast, income risk is generally less significant. Securities purchases are also found to increase with excess market returns and decrease with the initial securities-to-wealth ratio. This latter result is consistent with the rebalancing motive generated by (S,s)-type dynamics. The marginal effects of the household risks are greater than the marginal effect of past returns. However, the sensitivity of securities purchases to returns has increased in recent years, even controlling for changes in the composition of investors. (working paper)
Conventional wisdom assumes that homeownership is risky because house prices are
volatile. But all households start life "short" housing services, and homeownership could be a less
risky way of obtaining those services than the alternative, renting. While a renter faces year-toyear
fluctuations in rent, a homeowner receives a guaranteed flow of housing services at a known
price, and so is hedged against rent risk. Although the homeowner is in turn exposed to asset price
risk when she sells her house, that risk can be relatively small since it arrives at the end of the stay
in the house and so is discounted, or it is deferred even later if the homeowner moves to a
correlated housing market. We show in a stylized model with endogenous house prices that rent
risk can indeed outweigh asset price risk. The net benefit of homeownership increases in the
owner's expected horizon in the home, as the number of rent risks avoided rises and the asset price
risk occurs later in time. This effect of horizon on the demand for owning should increase
multiplicatively with the magnitude of the volatility of rents. Another implication of our analysis is
that the aggregate wealth effect from fluctuations in house prices may be small since higher prices
are generally offset by equivalent increases in the expected cost of future housing services.
We test these implications using MSA-level data on house prices and rent volatility matched with CPS data on homeownership. Consistent with the model, the difference in the probability of homeownership between households with long and short expected horizons in their residences is 2.9 to 5.4 percentage points greater in high rent variance MSAs than in low rent variance MSAs. The sensitivity to rent risk is greatest for households that exogenously must devote a larger share of their budgets to housing. Similarly, the "younger" elderly who live in high rent variance MSAs are more likely to own their own homes on average, but their probability of homeownership falls faster as they approach the end of life and their horizon shortens. Finally, we find that the house price-to-rent ratio capitalizes not only expected future rents, but also the associated rent risk premia, consistent with asset pricing models. At the MSA level, a one standard deviation increase in rent variance increases the house price-to-rent ratio by 2 to 4 percent. (working paper, published version)
This paper analyzes securitization and more generally "special purpose vehicles" (SPVs), which are now pervasive in corporate finance. The first part of the paper provides an overview of the institutional features of SPVs and securitization. The second part provides a model to analyze the motivations for using SPVs and the conditions under which SPVs are sustainable. We argue that a key source of value to using SPVs is that they help reduce bankruptcy costs. Off-balance sheet financing involves transferring assets to SPVs, which reduces the amount of assets that are subject to bankruptcy costs, since SPVs are carefully designed to avoid bankruptcy. Off-balance sheet financing is most advantageous for sponsoring firms that are risky or face large bankruptcy costs. SPVs become sustainable in a repeated SPV game, because firms can implicitly "commit" to subsidize or "bail out" their SPVs when the SPV would otherwise not honor its debt commitments, despite legal and accounting restrictions to the contrary. The third part of the paper tests two key implications of the model using unique data on credit card securitizations. First, riskier firms should securitize more, ceteris paribus. Second, since investors know that SPV sponsors can bail out their SPVs if there is a need, in pricing the debt of the SPV investors will care about the risk of the sponsor defaulting, above and beyond the risk of the SPVs assets. We find evidence consistent with these implications. (working paper)
This paper takes a portfolio view of consumer credit. Default models (credit-risk scores) estimate the probability of default of individual loans. But to compute risk-adjusted returns, lenders also need to know the covariances of the returns on their loans with aggregate returns. Covariances are independently relevant for lenders who care directly about the volatility of their portfolios, e.g. because of Value-at-Risk considerations or the structure of the securitization market. Cross-sectional differences in these covariances also provide insight into the nature of the shocks hitting different types of consumers. We use a unique panel dataset of credit bureau records to measure the 'covariance risk' of individual consumers, i.e., the covariance of their default risk with aggregate consumer default rates, and more generally to analyze the cross-sectional distribution of credit, including the effects of credit scores. We obtain two key sets of results. First, there is significant systematic heterogeneity in covariance risk across consumers with different characteristics. Consumers with high covariance risk tend to also have low credit scores (high default probabilities). Second, the amount of credit obtained by consumers significantly increases with their credit scores, and significantly decreases with their covariance risk (especially revolving credit), though the effect of covariance risk is smaller in magnitude. It appears that some lenders take covariance risk into account, at least in part, in determining the amount of credit they provide. (working paper)
During 2001, most U.S. taxpayers were mailed a Federal tax rebate in a randomly assigned week between July and September. Using special questions added to the Consumer Expenditure Survey, we use this historically unique experiment to measure the change in consumption expenditures caused by receipt of the rebate and to test the Permanent Income Hypothesis and related models. Households spent about 20-40 percent of their rebates on non-durable goods during the three-month period in which they received their rebates, and roughly two-thirds of their rebates cumulatively during the quarter of receipt and subsequent three-month period. The implied effects on aggregate consumption demand are substantial. Responses are larger for households with low liquid wealth or low income, consistent with liquidity constraints. (working paper, published version; summary [NBER Digest], discussion [from Congressional Budget Office])
This paper documents the trends in the life-cycle profiles of net worth and housing equity between 1983 and 2004. The net worth of older households significantly increased during the housing boom of recent years. However, net worth grew by more than housing equity, in part because other assets also appreciated at the same time. Moreover, the younger elderly offset rising house prices by increasing their housing debt, and used some of the proceeds to invest in other assets. We also consider how much of their housing equity older households can actually tap, using reverse mortgages. This fraction is lower at younger ages, such that young retirees can consume less than half of their housing equity. These results imply that 'consumable' net worth is smaller than standard calculations of net worth. (working paper)
We use a new panel dataset of credit card accounts to analyze how consumers responded to the 2001 federal income tax rebates. We estimate the monthly response of credit card payments, spending, and debt, exploiting the unique, randomized timing of the rebate disbursement. We find that on average consumers initially saved some of the rebate, by increasing their credit card payments and thereby paying down debt. But soon afterwards spending increased, counter to the canonical Permanent-Income model. Consistent with liquidity constraints, spending rose most for consumers who were most likely to be initially constrained by their credit limits, whereas debt declined most (so saving rose most) for unconstrained consumers. More generally, the results suggest that there can be important dynamics in consumers' response to 'lumpy' increases in income like tax rebates, working in part through balance sheet (liquidity) mechanisms. (working paper; discussion [from Congressional Budget Office])
A number of studies have pointed to various mistakes that consumers might make in their consumption-saving and financial decisions. We utilize a unique market experiment conducted by a large U.S. bank to assess how systematic and costly such mistakes are in practice. The bank offered consumers a choice between two credit card contracts, one with an annual fee but a lower interest rate and one with no annual fee but a higher interest rate. To minimize their total interest costs net of the fee, consumers expecting to borrow a sufficiently large amount should choose the contract with the fee, and vice-versa. We find that on average consumers chose the contract that ex post minimized their net costs. A substantial fraction of consumers (about 40%) still chose the ex post sub-optimal contract, with some incurring hundreds of dollars of avoidable interest costs. Nonetheless, the probability of choosing the sub-optimal contract declines with the dollar magnitude of the potential error, and consumers with larger errors were more likely to subsequently switch to the optimal contract. Thus most of the errors appear not to have been very costly, with the exception that a small minority of consumers persists in holding substantially sub-optimal contracts without switching. (working paper)
In the summer of 2003, the US government mailed around $14 billion in child tax credit payments to millions of households. Using special questions added to the Consumer Expenditure Survey, we estimate the change in consumption expenditures caused by receipt of these payments, by comparing the spending of households that receive payments in a given period to the spending of those that do not. On average, households spent about a quarter of their payments on nondurable consumption goods during the three-month period in which the payments were received. There is also less precisely estimated evidence of an ongoing but smaller response in the subsequent three-month period, so that roughly one-third of the payment was spent cumulatively during the quarter of receipt and subsequent three-month period. These responses are larger for households with relatively low liquid wealth or low income, which is consistent with their facing binding liquidity constraints. (working paper)
This paper empirically examines the benefits of relationship banking to banks, in the context of the credit card market. Using a unique panel dataset that contains rich information about the relationships between a large bank and its credit card customers, we estimate the effects of relationship banking on the customers' default, attrition, and usage behavior. We find that relationship accounts exhibit lower probabilities of default and attrition, and have higher utilization rates, compared to non-relationship accounts. Such effects become more pronounced with increases in various measures of the strength of the relationships, such as relationship length, breath, depth, and proximity. Moreover, dynamic information about changes in the behavior of a customer's other accounts at the bank help predict and thus monitor the behavior of the credit card account over time. These results imply significant benefits of relationship banking to banks in the retail credit market. (working paper)