This action might not be possible to undo. Are you sure you want to continue?
The Dissertation Committee for Tansel Yilmazer certiﬁes that this is the approved version of the following dissertation:
Household Saving Behavior, Portfolio Choice and Children: Evidence from the Survey of Consumer Finances
Daniel T. Slesnick, Supervisor Don Fullerton Maxwell B. Stinchcombe Peter J. Wilcoxen Jacqueline Angel
Household Saving Behavior, Portfolio Choice and Children: Evidence from the Survey of Consumer Finances
by Tansel Yilmazer, B.S., M.A.
DISSERTATION Presented to the Faculty of the Graduate School of The University of Texas at Austin in Partial Fulﬁllment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY
THE UNIVERSITY OF TEXAS AT AUSTIN December 2002
MI 48106-1346 .UMI Number: 3110711 ________________________________________________________ UMI Microform 3110711 Copyright 2004 by ProQuest Information and Learning Company. United States Code. This microform edition is protected against unauthorized copying under Title 17. ____________________________________________________________ ProQuest Information and Learning Company 300 North Zeeb Road PO Box 1346 Ann Arbor. All rights reserved.
G¨rkem Celik. for his support.Acknowledgments I am grateful to many people who shared the best and worst moments of ‘my dissertation years. Maxwell Stinchcombe. patience. Peter Wilcoxen and Jacqueline Angel for their valuable feedback and comments. Anne Golla. iv . in spite of the thousands of miles between us. Anne Gorney. Mala Velamuri. Angela Lyons. o ¸ and Vivian Goldman-Leﬄer for their stimulating conversations and friendship. I would like to thank my advisor. Adam Winship. I would also like to thank my committee members Don Fullerton. I wish to thank Fikret for always being there for me. Steve Trejo. Daniel Slesnick. I am indebted to my family for their love and believing in me over these years. guidance and encouragement. Finally. Special thanks go to Asli Kes. Matias Fontenla.’ First.
Portfolio Choice and Children: Evidence from the Survey of Consumer Finances Publication No. v . 2002 Supervisor: Daniel T. Tansel Yilmazer. As a result of the portfolio constraint. this dissertation examines the relationship between having children and the motives of saving: (i) to hold assets because of the return they provide. such as owner-occupied housing.D.Household Saving Behavior.’ and (iii) to accumulate for anticipated future needs. The University of Texas at Austin. The ﬁrst chapter examines how the number of children living in the household aﬀects the way households allocate their wealth across diﬀerent assets. risky assets and interest-bearing accounts. Slesnick Using the Survey of Consumer Finances (SCF). Ph. The results show that the number of children increases the housing consumption of homeowners and the share of the portfolio allocated to owner-occupied housing. such as educational expenses. (ii) to build up reserves as a precaution for a ‘rainy day. The portfolio allocation of homeowners is compared to that of renters by taking into account the portfolio constraint imposed by the consumption demand for housing.
homeowners decrease the share of the portfolio invested in retirement assets as the number of children increases. income uncertainty has little eﬀect on household savings. and after controlling for family size. vi . Further. Also. this chapter extends the empirical work on precautionary savings. the second chapter investigates the relationship between household saving and fertility decisions. The third chapter examines the eﬀect of ﬁnancing children’s college education on household savings. The results show that parents save for college expenses of their children. By examining the implications of income uncertainty on the demand for children. having an additional child reduces savings of households with young heads and increases savings of those with older heads. Using a life-cycle model that incorporates precautionary motives for saving. savings for college increase with the age of the household head. The results show that households with higher income uncertainty are less likely to have a child. Using the actual college expenditures reported in the 1983-86 SCF. The results are consistent with the predictions the lifecycle theory of saving that households save in advance for expected expenses to smooth their consumption. the empirical model estimates the expected expenditures on children’s college education and investigates the eﬀect of expected college expenses on household savings.
.5 Conclusion . . . . . . . . . . Conclusion . . . . . . . . . . . . 2. .4 3. . . . . . . . . . 2. . on Household . Introduction iv v ix xi 1 6 6 12 12 15 17 24 30 44 44 48 51 58 63 Chapter 2. . . . . . 2. .3 Data . . . . . . The Relationship between Fertility and Saving . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 3. . . . . . . . . .2 Empirical Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. . 3. . . . . . . . . . . . . . . . . . . . . . . Data . . . . . . . . . . . . . . . . . . .Table of Contents Acknowledgments Abstract List of Tables List of Figures Chapter 1. . .2 3. . . . .1 Introduction . . . . . .2. . . .3 3. . . . . . . .2 The Model . vii . . . . . . . .5 The Eﬀect of Precautionary Motives Saving and Fertility Introduction . . . . . . . .1 Theory . . . . . . . . . . .2. . . . Chapter 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Estimation and Results . . . . . . . . . . . . . . . . . . . . .4 Estimation and Results . . . . . . . . . . . . . . . . . . 2. . . . 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Do Children Aﬀect Household Portfolio Allocation? 2.
. . . . . . . . . . . . . 4. . Appendix for Chapter A. 4. .3 Estimating Permanent Income . . . . . . . .Chapter 4. . . .2 A Model of Saving for College 4. . 73 73 80 83 86 88 93 103 Appendix A.1 Introduction . . . .2 Deﬁnition of Variables . . . A. . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Estimating Marginal Tax Rates . . . . . . . . . Appendices College . . . . . . . . . . . . 111 Bibliography Vita 113 121 viii . . . . .3 Data . . . . . . . . . . . . . . . .1 Deﬁnition of Variables . . . . . . . . . . . . . . . . . . 4. . 104 . Appendix for Chapter 4 111 C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Saving for Children’s 4. . . 106 Appendix B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Education . . 105 .6 Conclusion . . . . . . Appendix for Chapter 3 109 B. . . . . . . . . . . . . . .5 Estimation and Results . . . 109 Appendix C. . . . . A. . . . 2 104 . . . . . . . . . . .1 Deﬁnition of Variables . .4 Empirical Speciﬁcation . . . . . . . . . . . . . . . 4.
. . . . . . . . . . .6 3. . . . . . . . . . . . . . . .7 2. Descriptive Summary of Variables . . . . . . . . . . . . . . . 1998 . . . . . . . . . . The Eﬀect of a Change in the Fertility Decision on SAVE1 . . . . . . . . . . . . . 1998 . . . . . . . . . . Savings and College Expenses by the Number of College .1 3. .5 3. . . . . . . . . . . Results: Asset Shares and Housing Expenditure of Renters . .2 4. . . . . . . . . . .8 4. . . . . . . Regressions of SAVE1 on Income Uncertainty with Endogenous Fertility Decision . . . . . . . . . . . . . . . . . .3 2. Savings. . . . . . . Results from Probit Estimation . . . . . . .2 2. . .1 4. 1983 . Expenditure on Housing. . . Portfolio Shares for Assets by the Number of Children and Age Saving Motives by Age Groups. . . . 1998: Continued . . .2 3. . Saving Motives By the Number of Children . . . Renters: Continued . . . Tobit Estimates of College Expenditure Equation ix . . . . . . . . . . . . . . . . . . . .4 4. . . . . Income and Income Uncertainty by Age and Fertility Probit: Fertility Decision of Fecund Households . . . .4 2. 33 34 35 36 37 38 39 40 41 42 43 65 66 67 68 69 70 71 72 95 96 97 98 99 . . Children in . . . .1 2. . .8 2. . . .9 2. . . . Descriptive Statistics by Household Fertility Decision . . . . .5 Descriptive Statistics by Year . . . . . . . . .11 3. . . . . . . . . . Mean Asset Shares. .5 2. . . . . . . . Poisson Regression: Number of Children . . . . . . . .10 2. . Regressions of SAVE2 on Income Uncertainty with Endogenous Fertility Decision . . Mean Asset Shares by Year . . . . . . . . . . . . . . . . . . . . . . . . Mean Asset Shares. . . . .List of Tables 2. . . . . . . . . . . . . . . . . . . . . Results: Asset Shares and Housing Expenditure of Homeowners Homeowners: Continued . . . . . . . .3 3. Mean Income Uncertainty by Household Demographics . . . . . . . . . .6 2.4 3. . . . . . . . . .3 4. .7 3.
. . . . Eﬀect of Anticipated College Expenses on Savings . . . . . . . . . .6 4. . . . . . . . . . . . . . . . . . . .4.7 College Expenditures and Savings by the Number of Children in College . . . 100 101 x .
List of Figures
The Importance of Educational Expenses on Savings . . . . .
Chapter 1 Introduction
Raising children is costly with their housing, educational and other expenses. To meet the costs of raising their children, parents use both current income and intertemporal transfers. Children living in the household, therefore, are likely to aﬀect the level of household savings, portfolio composition and the life-cycle proﬁle of savings. Using data from the Survey of Consumer Finances (SCF), this dissertation examines the relationship between children and the motives of saving: (i) to hold assets because of the return they provide, (ii) to build up reserves as a precaution for a ‘rainy day,’ and (iii) to accumulate for anticipated future needs, such as educational expenses. Most U.S. households hold a large portion of their wealth in the form of owner-occupied housing. According to the 1995 SCF, 65 percent of households are homeowners, and the value of an average homeowner’s property is 60 percent of its total assets. Owner-occupied housing diﬀers from other types of wealth in its dual role as both a consumption good and an investment good. Since households cannot separate the level of consumption of housing services from investment in housing as an asset, the optimal level of owner-occupied housing may be higher than the optimal level for households only interested
in long run returns. The demand for housing services is likely to increase with the number of children living in the household. Therefore, the consumption constraint can be even more binding for households with children. Chapter 2 uses the 1989, 1992, 1995 and 1998 SCF to investigate how the number of children living in the household aﬀect the portfolio choice between housing and other assets. The portfolio allocation of homeowners is compared to that of renters by taking into account the portfolio constraint imposed by the consumption demand for housing. The empirical model also examines the eﬀect of children on the demand for housing services and homeownership decision. The results show that the number of children increases the housing consumption of homeowners as well as the share of the portfolio allocated to owner-occupied housing. As a result of the portfolio constraint, homeowners decrease the portfolio share of retirement assets as the number of children increases. Low levels of retirement savings of U.S. households have generated signiﬁcant concern in the last twenty years. The ﬁndings of Chapter 2 show that households with children decrease the portfolio share for retirement savings considerably while they increase the portfolio share for housing. If the return on housing is less than the return on retirement accounts, there is a hidden cost of children. Explaining the size of the portfolio eﬀect allows a better understanding of the cost of children. Also, changes in housing programs or tax deduction rules for mortgage interest payments inﬂuence the portfolio allocation of households with children considerably by increasing or decreasing the 2
The results of the empirical model in Chapter 3 show that households with higher income uncertainty are less likely to have a child at a point in time. household saving show that saving rates are higher for married couples with no children and lower for those with children. The 1983-89 panel of the SCF is used to examine the interaction of income uncertainty and changes in the number of children on the saving behavior of households at diﬀerent stages of the life cycle. Income uncertainty actually reduces savings of the households with low or very high wealth holdings and does not aﬀect the saving behavior of other households. are not consistent with the predictions of the precautionary saving model that suggests agents faced with uncertainty about future income increase their savings. Using a life cycle model that incorporates precautionary motives for saving. this chapter extends the empirical work on precautionary saving. having an additional child decreases savings of households with young heads and increases savings of those with older heads. By examining the implications of uncertainty on the fertility decisions of households and incorporating fertility decisions as a motive for household saving behavior. however.S. The ﬁndings. This ﬁnding is consistent with the life-cycle theory of saving and consumption and shows that household composition is an important factor 3 . The data on U. Also.cost of homeownership. Chapter 3 investigates the relation between household saving and fertility decisions. Precautionary saving models predict that uncertainty about future income may cause households to reduce their current consumption in order to raise their stock of precautionary saving.
Chapter 4 examines the eﬀect of ﬁnancing children’s college education on household savings. 90 percent of dependent undergraduate’s parents contributed ﬁnancially to the costs of their children’s education. Of those contributing to their children’s college costs in 1987. According to the 1996 National Postsecondary Student Aid Survey. Also. savings for college education increases with the age of 4 . First. Understanding the eﬀect of ﬁnancing children’s college education on household saving behavior is important for at least three reasons. parents contribute a signiﬁcant amount to their children’s college expenses. the quality-quantity model of fertility behavior assumes that parents have preferences both for the expenditure per child and the number of children. The college ﬁnancial aid system imposes an implicit tax on the savings of households that are potentially eligible for ﬁnancial assistance. about 65 percent reported using some previous savings. This chapter uses the amount of parental expenditure on children’s college education as a measure for child quality. Chapter 4 estimates the household’s expected expenditures on children’s college education and investigates the eﬀect of expected college expenses on household savings. Using the actual college expenditures reported in the 1983-86 SCF. Third.of life-cycle savings. Second. an analysis of ﬁnancing college education and family size highlights an important aspect of the quality-quantity model. The results show that parents save for college expenses of their children. Given the rapidly rising cost of college tuition. families who save for college reduce their eligibility for ﬁnancial aid.
5 .the household head. These results are consistent with the predictions of the life-cycle theory of saving and consumption that households save in advance for expected expenses to smooth their consumption.
the literature has focused on the impact of demographic variables such as the eﬀect of age.Chapter 2 Do Children Aﬀect Household Portfolio Allocation? 2. risky assets. For example. Conversely. It is likely that children living in the household aﬀect the way a household allocates its wealth across diﬀerent assets such as owner-occupied housing.1 Introduction Empirical studies of household portfolio composition have identiﬁed large diﬀerences in portfolio allocation choices of diﬀerent demographic groups. race and gender of the household head on the portfolio composition. Jianakoplas and Bernasek . 1 6 . they may hold most of their ﬁnancial assets in riskless See Poterba and Samwick . King and Leape . and Sund`n and Surette  e for gender eﬀects. and Ioannides  for age eﬀect. and interest-bearing accounts.1 The inﬂuence of children living in the household on the portfolio composition has not been yet discussed. Chiteji and Staﬀord  for race. households with children may purchase more housing than households with no children or they may have a higher probability of owning a home. So far. Parents may choose to invest part of their household portfolio in stocks to meet the rising costs of a college education.
I analyze a model in which households decide on portfolio shares for diﬀerent assets jointly with the tenure choice (the decision of owning or renting) and the consumption demand for housing services. Using data from the 1989.form to decrease their families’ exposure to risk. 1992.S. this chapter investigates the eﬀect of children on household portfolio composition. then changes in housing programs or tax deduction rules for mortgage interest payments inﬂuence their portfolio allocation by increasing or decreasing the cost of homeownership. 1995 and 1998 SCF. Speciﬁcally. I focus on how the number and age of children living in the household aﬀect (i) the homeownership decision. and (iii) the housing expenditure of homeowners and renters. households with children may decrease the portfolio share for other assets considerably while they increase the portfolio share for housing. 7 . paying particular attention to the impact of children on the demand for housing services and homeownership decision. It has also important policy implications. (ii) the portfolio shares for housing and the other assets that homeowners and renters hold. Understanding the size of the impact of children on household portfolio allocation is intrinsically interesting. Also. households have generated signiﬁcant concern in the last twenty years. as the result of higher consumption demand for housing. If households with children allocate a larger share of their portfolio to owneroccupied housing. Low levels of retirement savings of U. The failure of households with children to invest suﬃcient assets in retirement accounts may lead to a lower retirement wealth.
Exceptions are the theoretical model of Brueckner . Brueckner analyzes the behavior of homeowners. households cannot separate the level of consumption of housing services from investment in housing as an asset. Wolﬀ  uses the 1983. and both report that owner-occupied housing accounts for about 30 percent of household assets. 65 percent of households are homeowners. The optimal level of owner-occupied housing for households may be higher than the optimal level for households that are only interested in long run returns. and King and Leape  examine the 1960-62 Michigan Surveys of Consumer Finances.S. Explaining the size of the portfolio eﬀect allows a better understanding of the cost of children. Owner-occupied housing diﬀers from other types of wealth in its dual role as both a consumption good and an investment good. and the value of an average homeowner’s property is 60 percent of its total assets.Most U. In his model. and the ownership of their principal residence determines the level of consumption of housing services. 2 See Henderson and Ioannides  and Berkovec and Fullerton  8 . According to the 1995 SCF.2 In the presence of tax distortions and transaction costs. 1992 1995 SCF. households hold a large portion of their wealth in the form of owner-occupied housing. Households with children are likely to have a higher demand for housing services and the consumption constraint can be even more binding. 1989. its impact on the portfolio choice between housing and other assets has not been discussed much. While the dual role of housing has been recognized. the general equilibrium model of Berkovec and Fullerton  and the numerical analysis of Flavin and Yamashita .
The results of his model show that when the constraint imposed by housing is binding. His model analyzes the resulting distortion of the eﬀect of this investment constraint on the portfolio choice of homeowners.  treat the presence of children in the household as endogenous and ﬁnd that a 10 percent increase in the probability of having a child raises the likelihood of homeownership by 2. Flavin and Yamashita use numerical methods to calculate the mean-variance eﬃcient frontier. This chapter extends the previous studies of portfolio choice by examining the eﬀect of both consumption and investment motives on the portfolio share for housing and other assets.5 percent. Their results show that the portfolio constraint imposed by the consumption demand for housing causes a life-cycle pattern in the portfolio shares for stocks and bonds such that the ratio of stocks to net worth increases as the household head gets older. Their simulation concentrates on the eﬀect of taxes on the tenure choice and owner-occupied housing. For example.  show that 9 . households decide on tenure and quantity of housing taking both consumption and investment motives into account. Robst et al. Neither of these studies explicitly analyzes the determinants of the consumption demand for housing and the portfolio share for housing. the homeowner’s optimal portfolio is ineﬃcient in a mean-variance framework. The literature on housing demand has recognized the role of children on the tenure choice and the demand for housing services. Harun et al.an investment constraint requires that the quantity of housing owned is at least as large as the quantity of housing consumed. In Berkovec and Fullerton.
an additional child increases the probability of owning a home by around 8 percent. deﬁned beneﬁt pensions and mutual funds. Ihlanfeldt  reports housing demand estimates obtained separately from two samples-recent movers and nonmovers. Demographic characteristics such as age.S. and race of the household head are shown to be signiﬁcant factors that reduce the level of information cost that would be suﬃcient to 10 . 401(k)s. households typically invest in only a few of the assets available in the economy. according to the 1995 SCF. Goodman and Kawai  ﬁnd that larger households prefer more housing. Among recent movers. only 41 percent of households held stocks directly or indirectly in IRAs. U. however. Many studies have investigated the reasons that most households choose to hold incomplete portfolios. marital status. The information cost of monitoring and managing a portfolio is suggested as an important reason for holding riskless assets. little systematic treatment of children has appeared in the estimation of tenure choice and housing demand. For example. the importance of the current and expected family size diﬀers between owners and renters: while renters demand more housing with an increase in family size and expectation of an additional child within the next nine months. However. these variables do not aﬀect the housing demand of homeowners. The results of the previous studies show that dependent children have some impact on the demand for housing. as noted in Goodman . Besides housing. After controlling for the household size. their results show that the presence of children in school has either an insigniﬁcant or a negative eﬀect on the demand for housing.
In the conditional demand equations. and they estimate equations for both the probability of owning an asset and its demand conditional upon ownership. The theoretical model developed in the chapter shows how the portfolio constraint imposed by the consumption demand for housing aﬀects the portfolio shares for housing and other assets. His results show that household characteristics such as age and education of the household head are signiﬁcant in explaining the probability of owning stocks.discourage households from investing in risky assets. The empirical model compares the portfolio allocation of homeowners to that of renters. the eﬀect of age and marital status appears to be signiﬁcant only for some of the assets. Children living in the household have not been the focus of any study examining the portfolio choice of households. Bertaut  uses the 1983-89 SCF to analyze the eﬀect of household characteristics on portfolio allocation. however. Their ﬁndings show that age and marital status of the household head signiﬁcantly aﬀect the probability of asset ownership. taking into account the eﬀect of children on the consumption demand for housing. Chiteji and Staﬀord  link independent young African-American adults back to their parents. Using the Panel Study of Income Dynamics. The results show that the number of children has a positive and signiﬁcant eﬀect 11 . King and Leape  analyze a model in which investors choose to hold incomplete portfolios. Their ﬁnding is that parents who held stocks are more likely to have children who hold stocks as young adults. For example. This chapter aims to do so by examining the eﬀect of the number and the age of children on household portfolio choice.
Children living in the household also aﬀects the portfolio choice of renters.on the probability of owning a home.2. Section 2.5.2 introduces the theoretical model and discusses the empirical speciﬁcation of the model. Section 2. The main conclusion of the chapter is that homeowners shift their resources from retirement accounts to housing with an increase in the number of children. As a result of the portfolio constraint imposed by the housing demand of children. The remainder of this chapter is organized as follows. A summary of the ﬁndings and concluding remarks are presented in Section 2. homeowners decrease the portfolio share in retirement accounts while they increase the portfolio share in housing. The consumer maximizes a multiperiod utility function. Controlling for the number of children and other variables. homeowners with all children older than age 13 invest a greater share of their portfolio in vehicles and other real estate and a smaller share of their portfolio in housing.4.2 2. The number of children also increases the housing demand of homeowners. Following Brueckner  and 12 . 2. and how much to allocate to other risky assets.3 describes the data set and the variables used in the empirical work. Renters invest a smaller share of their portfolio in interest-bearing accounts with an increase in the number of children.1 The Model Theory This section examines the behavior of a consumer deciding whether to rent or own a home. The estimation results are reported in Section 2.
and h ≥ 0. hc ) + δE[V (R + y)]. then she holds owner-occupied housing (h > 0) and is constrained to consume the same amount of owner-occupied housing in her portfolio (hc = h). (2... If the consumer purchases a house. U gives the utility from the current consumption. I assume that third and subsequent periods are buried in the indirect utility function given remaining wealth at the beginning of the second period.2) 13 . 1. The dollar amount of asset j purchased is denoted aj . with a0 being the riskless asset. housing services (hc ). A consumer in this economy is assumed to obtain utility from the current consumption of a single nondurable good (c). and owner-occupied housing earns rh .1) where y is future labor income. E gives the expected utility. The only source of uncertainty is assumed to be from returns on J + 1 assets and owner-occupied housing (h). j = 0. V is an indirect utility function. The j th asset earns a gross return of rj . Short selling is ruled out for all assets including housing. . j = 0. J.Henderson and Ioannides . 1. The consumer’s objective function can be written as follows: U (c. J... and δ is the discount factor. The ﬁrst period budget constraint is given by J c=w− p o hc h − j=0 aj . so that aj ≥ 0. (2. and consumption in future periods that depends on the random total return R from the investment portfolio. .
h The total return of the portfolio is given by J R = rh h + j=0 rj aj . j = 1.. . (2. h = 0 in equations (2.7). and θjh is the covariance of returns between asset j and housing. For homeowners.6) and (2. are the variances of rh and rk . For renters. (2. j = 1.J.4) where po r is the price of a unit of housing for renters.6) and the standard deviation J J K σ = (θhh h + 2 j=1 2 haj θhj + j=1 k=1 aj ak θjk )1/2 . .3) If the consumer rents a house. (2. 14 . 2.where w is her initial wealth and po is the current price of a unit of housing. the return on housing and the return on other assets are assumed to be normal variables with the expected values rh and rj . the total portfolio return R is a normal random variable with the expected value J R = rh h + r0 a0 + j=1 r j aj (2. The total return of the J portfolio is given by R= j=0 rj aj .5) since h is equal to zero for renters.. θjk is the covariance of returns between asset j and k. (2.7) where θhh and θjj . respectively. J. then the ﬁrst period budget constraint is given by J c = w − po hc − r j=0 aj .. In the model..
J. For both homeowners and renters. hc ) + δ V (R + σz + y)φ(z)dz.8) where φ(.6) and (2. .1) in terms of R. the asset levels aj . In the second stage. (2.7).. that maximize (2.5). (2.2).1) subject to (2.. In the ﬁrst stage.) is the standard normal density function.. a household determines whether to own or rent a house: H = 1 if Xh β1 + ε1 > 0 = 0 otherwise. J. and shares of wealth to allocate to each asset j (sj ). (2. that maximize (2. .8) subject to (2. 1. and the standard normal variable z as follows: U (c... 2.6) and (2. σ. and to owner-occupied housing (sh ) is modeled as follows.2 Empirical Model The joint determination of whether own a house (H=1) or not (H=0). I rewrite the objective function (2. how much to spend on housing (Eh ). The consumer also decides on c∗ . . j = 0..3).7) and decides to own or rent a house comparing the utilities in two outcomes. First. 1. are chosen optimally with hc and σ held constant.. J. c j (2. The consumer’s problem is to choose c∗ .2. j = 0.4). this problem can be solved in two stages. and a∗ . (2.9) where Xh is a vector of year dummies and characteristics that are associated 15 . hc (and thus σ) is chosen optimally.Following Fama and Miller  and Brueckner . 1. (2. j = c j 0. h∗ and a∗ . The empirical model described in the next section focuses on the interaction between these two stages of decision making. h∗ .
as a regressor in estimating (2. j = 0. 1. and also the housing expenditure: j = 0.10) (2. In the second stage. log Eh = Xc βoc + εoc .. J sj = Xβoj + εoj sh = Xβh + εh If owner. εoc . βh . The two stage method described in Lee and Trost  is used to estimate the model...11) where X and Xc are vectors of household characteristics and year dummies. 16 .. Similarly. (2. J. . and the error terms in equations (2..11) are assumed to have a joint normal distribution. 1.. the household decides on the share of portfolio allocated to each asset and housing. j = 0. j = 0. respectively.9) . 1.. . In the ﬁrst stage. εrj . I use ˆ ˆ φ(X β1 )/Φ(X β1 ). and εrc are the error terms. J sj = Xβrj + εrj sh = 0 If renter. Separate equations are speciﬁed for homeowners and renters. 1. and ε1 is an error term. βoj . βrj ..10) for homeowners. J. εh . .11). a probit model of the tenure choice in equation (2.. . log Eh = Xc βrc + εrc . where φ and Φ are probability density and cumulative distribution of the standard ˆ ˆ normal distribution. φ(X β1 )/(1 − Φ(X β1 )) is used as a regressor for renters in estimating (2.9) provides an estimate of β1 . Second.(2... β1 is a parameter vector.. βoc and βrc are the parameter vectors to be estimated. and εoj .with the probability of owning a house.
in 1992. Each survey consists of a representative sample of the U. certiﬁcates of deposit. 3) RETIRE includes IRAs. 3 17 . 1. saving accounts. 401(k)s. in 1995. all types of bonds.299. The survey contains detailed information on household portfolios. 5) VEHICLE is the value of all the vehicles the household owns. and other assets like arts and precious metals. cash value of life insurance.2. call accounts. Keogh. For owners.S. a triennial survey conducted by the Federal Reserve Board.S. and in 1998. Investments in businesses are not included in total assets because they generate an income that is diﬃcult to separate from earnings.3 Data The data for this study are taken from the 1989. and demographic characteristics. The consumption demand for housing is computed for renters and homeowners as follows. the cost of housing services depends on In the 1989 SCF.143 households. 2) STOCK includes all assets held in stocks. and other deﬁned contribution plans.309 households. and 7) OTHER includes trusts. and mutual funds. money market deposit accounts. population and a supplement of high-wealth households drawn from Internal Revenue Service ﬁle of high-income returns. 1.3 Total assets are grouped into six categories: 1) ACCOUNT includes all holdings of checking accounts. 6) RESTATE includes the market value of seasonal residences and other property.519 out of 4. 1995 and 1998 SCF. population.409 out of 4. 4) HOUSE is the market value of owner-occupied housing. income. the supplement consists of 866 out of 3.480 out of 3.906. 1992. 1. The SCF constructs sample weights to blend the supplements with the area-probability sample to get a more representative sample of the U.
r. the interest rate (r). the mortgage interest payment (m). The calculation of marginal tax rates is described in Appendix A.015 for each of the sample years. is the annual inﬂation rate calculated using the CPI-U deﬂator. I make several assumptions. households that neither rent nor own their homes are excluded for lack of information to cal18 . maintenance and depreciation costs (d). The inﬂation rate. ρ. π. Since marginal tax rates are not reported in the SCF. the income tax rate (τ ). For renters. and the rate of increase in house prices. the rate of increase in the nominal price of housing (ρ) and the overall inﬂation rate (π).the gross value of the residence (G). To calculate the housing expenditure by using equation (2. I assume an annual rate of depreciation of d=0.12) This formulation assumes that homeowners claim tax deductions for property taxes and mortgage interest payments. The interest rate. the annual rental expenditure reported in the SCF is used as the consumption demand for housing. is the rate of increase in the median sale price of houses in that year. Following Henderson and Ioannides .12).1. is assumed to be the interest rate on treasury bills. First. I impute them using detailed account information on the sources of income and demographics for each household. Property tax rates and mortgage interest payments are reported in the SCF. A few restrictions are imposed on the sample. The housing expenditures (Eh ) of homeowners are then deﬁned as Eh = [(1 − τ )r + d + (1 − τ )τp − (ρ − π)]G − mτ. the property tax rate (τp ). (2.
6 However.807 households in 1989.1 percentile of the weighted wealth distribution in the 1989. marital status (MARRIED) and gender (FEMALE) of the household head and the fraction of homeowners (HOMEOWN).1 percent weighted wealth holdings in each wave of the SCF are dropped. households with the highest 0. A household is assumed to be a renter if it rents all or part of the farm/ranch/apartment/house/mobile home in which it lives. I take the estimated earnings of the household head and the spouse at the age of 45 and an individual-speciﬁc eﬀect. 5 Of the remaining households. (ii) it owns both the mobile house and the site. 209 and 193 were in the 0. 1992.4 Second. The variables are described in detail in Appendix A.989 observations. 116. 1995 and 1998. or (iii) it owns part or all of the farm/ranch on which it lives on. 4 19 . As a proxy for permanent income.509. 1992. Therefore. 317 and 309 households were neither renters nor owners and were dropped from the sample.5 The ﬁnal sample consists of 13.773 and 3. 2.1 shows the summary statistics for all the variables used in the estimation. Table 2.3. households with female heads are headed by single females.culate housing expenditure. and 1998 SCF. to avoid the inﬂuence of extreme outliers on the regression. In 1989. 127. Sample demographics show the age of the household head (AGE). 1992. 183. 1995. 214. 1995 and 1998. The calculation of permanent income follows King and Dicks-Mireaux  and is described in Appendix A. 3. respectively. The same pattern is true for permanent income (INCOME). The calculated expenditure of housing consumption (Eh ) A household is assumed to be a homeowner if (i) it owns the house/apartment that it lives in or owns it as a part of a condo. 6 The SCF deﬁnes the head of the household to be the husband for all married households. 3.900.2. respectively. both mean and median wealth (ASSET) have risen since 1992. most of which have not changed much over time. respectively. a co-op or a townhouse association.
83 in 1989 to 0. Table 2.5 percent in 1998. The increases in ACCOUNT. STOCK and RETIRE in 1998 oﬀset the decline in HOUSE. age. representing 39.was higher for homeowners in 1992 than in other years due to the decline in house prices in that year. The percentage of households with all children older than age 13 (CHAGE13) has stayed the same since 1992. HOUSE is the most important asset. VEHICLE and RESTATE.75 in 1995 and stayed the same in 1998.3 presents housing expenditures of homeowners and renters in 1998. there is a steady growth in the portfolio share for STOCK and a steady decline in the portfolio share for RESTATE since 1989. the composition of households’ portfolios reveals the importance of housing as an asset. Assets in these accounts increased from 5. The average number of children (NCHILD) living in the household declined from 0.6 percent in 1998). Table 2.3 percent in 1989 to 11.2.4 percent of total assets in 1998. First.7 percent of total assets in 1989 to 10. the share for RETIRE increases sharply. As shown in Table 2. The second largest asset in the households’ portfolios is VEHICLE (18.2 presents interesting changes in household portfolio structures over time.2 percent in 1995.2 percent in 1998 due to an increase in the portfolio share for saving accounts. wealth and children (the number of children living in the household) groups. 20 . The portfolio share for ACCOUNT declined from 14. but it rose to 13. The ﬁrst column shows the share of households in diﬀerent income. followed by ACCOUNT. This suggests that households have substituted ﬁnancial assets for nonﬁnancial assets. Second.
5 show the household portfolio composition in 1998 by household permanent income. First.4 shows the portfolio shares of assets that homeowners and renters hold. VEHICLE is the third largest asset (7. The housing expenditures of renters and homeowners also increase with income. Among households with wealth below $250. The percentage of households who are homeowners increases with income.5 percent of total assets) followed by ACCOUNT (26.4 and 2. Tables 2. respectively. The average housing expenditure is $7. wealth and the age of the household head.8 percent of total assets) following 21 . however. This is due to an increase in the value of residences and also to the tax deduction for property taxes and mortgage interest payments that decrease the opportunity cost of homeownership.030 for renters. wealth and the number of children in the household. there are marked diﬀerences in household portfolios of renters and owners. For homeowners.000. For renters. Average housing expenditures for homeowners and renters are presented.0 percent). in the remaining two columns of the table.9 percent. and the number of children.000 and income below $50. The ﬁrst row of Table 2. VEHICLE is the most important asset held (41. renters spend more on housing than owners. for renters. the expenditure on housing declines after the age of 65. accounting for 57.The second column indicates the percentage of each of these groups that are homeowners. For homeowners. reaching a peak among households with two children. It also increases with the number of children. Since the primary residence is the largest part of homeowners’ wealth. age of the household head. it declines after age 50. wealth.042 for homeowners and $6.
2 percent). First. For higher levels of income. For example. among homeowners that have wealth exceeding $1 million. The portfolio shares for other assets such as STOCK. 86. accumulation in STOCK relative to other assets increases over age 65.2 percent of total assets while housing accounted for only 22. For example. of the households with income below $15. as shown in Tables 2. Table 2. Another noteworthy ﬁnding is that the portfolio shares for STOCK and RETIRE for both homeowners and renters rise with income.6 percent.7 percent are homeowners.000.6 percent of total assets in housing. the fraction of households who are homeowners increases. the share of the portfolio allocated to STOCK rise at a rapid rate with wealth. Also. This suggests that households with heads over age 65 substitute 22 . For homeowners.4.3 and 2. Also. Not surprisingly.000.9 percent of their total assets in housing. the share of the portfolio allocated to RESTATE and for all households.7 percent are homeowners holding 75. portfolio composition of households with heads over the age of 65 diﬀers considerably from other age groups’ portfolios. the portfolio share for ACCOUNT almost doubles both for homeowners and renters over the age of 65 compared to 50-64 year old group. Several ﬁndings are worth noting. but they hold only 42.4 also presents the life cycle patterns in household portfolios.RETIRE (10. RETIRE and RESTATE are almost equal for renters and owners. in contrast. we observe striking diﬀerences in the composition of portfolios by the level of wealth. while the housing share of portfolio declines. STOCK is the most important asset category with a share equal to 25. 42. Of the households with income above $100.
Homeowners invest a smaller share of their portfolio in interest-bearing accounts and stocks with an increase in the number of children. Finally. Table 2. Also. Table 2.3 percent for those with three or more children. and age groups. The table indicates a strong relation between children and the share of portfolio allocated to housing. the presence of children increases the share of the portfolio allocated to vehicles.5 reveal striking diﬀerences in portfolio structures across income.liquid assets for nonﬁnancial assets.5 looks at the link between children and shares of assets in both renters’ and homeowners’ portfolios. For example. The ﬁrst is their eﬀect on the choice of tenure. but it stays steady after age 65. age and wealth are similar. the relative changes in portfolio shares of assets by income. and table 2.5 shows the portfolio shares by the number of children living in the household. Second.0 percent of the wealth for households with no children. Children are likely to aﬀect the portfolio structures in two ways. wealth. The results indicate that the number of children living in the household aﬀects the portfolio shares for assets and 23 . The portfolio share for owner-occupied housing increases with the number of children. Tables 2.3 investigates the eﬀect of children on the tenure choice. and 65. 60. the portfolio share for HOUSE declines with age among the households headed by persons below age 65. While portfolio composition diﬀers considerably between renters and homeowners.9 percent for households with 2 children.4 and 2. and the second is their eﬀect on asset shares of portfolios conditional upon ownership. housing accounts for 56.
Age and age-squared of the household head are included to capture a possible change in portfolio behavior related to the life cycle. STOCK.the probability that a household owns a home. Previous research also indicates that a household’s marginal tax rate (MRT) has an eﬀect on its asset allocation decisions. Dummy variables indicating the number and the age of children living in the household are included in X. and the disturbance covariance matrix is singular. 410 report zero wealth holding. Thus.7 I exclude those households from the sample and correct for sample selection. HOUSE. Portfolio choice theory has shown the importance of age.4 Estimation and Results The resulting set of equations constitutes an endogenous switching model in the form of a multivariate regression model. 2. permanent income and wealth in determining the asset shares in household portfolios. RETIRE. VEHICLE. OTHER. Portfolio shares of the J + 1 assets and housing sum to one. I drop one group of assets. 111 in 1992. and RESTATE in the estimation of the model. The other variables in X are chosen to be consistent with previous empirical studies. The empirical model below investigates the eﬀect of children on both asset shares and homeownership decision.898 households. and 106 in 1998 had zero wealth holding. 100 in 1995. Moreover. Of 13. and include ACCOUNT. 24 . Then I solve for the parameters of OTHER from the other equations. the marital 7 93 households in 1989.
status and the gender of the household head and willingness to undertake risky investments (RISKY) may also aﬀect the household’s asset allocation. All variables that enter X are also included in Xc and Xh , with two exceptions. First, the marginal tax rate aﬀects the tenure choice and homeowners’ expenditure on housing since homeowners can claim tax deductions for mortgage interest payments and property taxes. However, the marginal tax rate is not expected to aﬀect the housing expenditure of renters. Thus, marginal tax rate is not included in Xc . Second, willingness to undertake risky investment does not enter Xc because it has an eﬀect on the tenure choice regarding the investment motive but not on the expenditures on rental housing. In addition, the vector Xh includes the race of the household head. Table 2.6 presents the estimates of the probit model of equation (2.9). The estimates of the homeownership equation are consistent with previous studies. As a household’s permanent income rises, the probability of homeownership increases. Age of the household head increases the probability of ownership until age 74. The coeﬃcients for WHITE and MARRIED are significant and positive, indicating that at the sample mean, households with white heads are 10.2 percent more likely to own than households with non-white heads, and those that are married are 26.1 percent more likely to own than those that are not. The coeﬃcients on the variables showing the number of children are positive and signiﬁcant. Households with one child are 6.3 percent, and those with two children are 10.8 percent, more likely to own relative to households with no children. The probability of owning starts to decrease 25
after the second child, household with three or more children are only 9.6 percent more likely to own relative to households with no child. The probability of being a homeowner also increases with the household’s marginal tax rate, suggesting that the tax-deductibility of property taxes and mortgage interest is more valuable at a higher marginal tax rate. Tables 2.7- 2.10 show the coeﬃcients and the standard errors for each of the seven asset equations and the housing expenditure equation for homeowners. Permanent income has signiﬁcant but small marginal eﬀects on the structure of homeowners’ portfolio. The share of the portfolio allocated to RETIRE, HOUSE and VEHICLE increase with income, while the share allocated to ACCOUNT, STOCK and RESTATE decreases with income. Higher levels of wealth are associated with higher shares in ACCOUNT, STOCK, RESTATE, OTHER, and lower shares in HOUSE and VEHICLE. The marginal eﬀect of wealth on the share allocated to STOCKS, HOUSE and RESTATE is large. A 10 percent increase in assets would increase the share of the average portfolio allocated to STOCK by 0.62 percentage point. A similar increase in assets would induce 1.25 percentage point decrease in HOUSE and 0.66 percentage point increase in RESTATE. Age is an important determinant of portfolio shares in a homeowner’s portfolio, and the results in Table 2.7 and 2.8 reveal a quadratic relationship in terms of age. Portfolio shares for RETIRE, HOUSE and RESTATE increase with age, reaching a peak at the age of 50, 63 and 50, respectively. Portfolio shares for ACCOUNT and STOCK, however, decrease with age until the age of 26
50 and 43, respectively. This relation between age and portfolio shares suggests that the structure of a household’s portfolio changes when the household head reaches middle age. For example, households headed by persons above the age of 45 start substituting liquid assets for nonﬁnancial assets such as HOUSE and RESTATE. The coeﬃcients on the number and age of children suggest that the presence of children plays a signiﬁcant role on the portfolio structure of homeowners’. Several results are of particular interest. First, relative to households with no children, households with one child have a 5.6 percent higher portfolio share of HOUSE, controlling for age and permanent income. Similarly, households with two and three or more children have 8.9 and 9.2 percent greater portfolio shares in HOUSE. Second, the portfolio shares for ACCOUNT, RETIRE, and VEHICLE decrease with an increase in the number of children. Controlling for the number of children, households with all the children older than age 13 hold a smaller portfolio share in HOUSE and a greater share in VEHICLE and RESTATE. Finally, homeowners that are willing to undertake risky investments hold a greater share of risky ﬁnancial assets, such as STOCKS and RETIRE, and a smaller share of less risky assets, such as ACCOUNT and HOUSE. All other things held constant, the portfolio shares allocated to ACCOUNT and RESTATE have declined in 1998. Households have substituted STOCK, RETIRE and VEHICLE for the other asset categories since 1995. An increase in the marginal tax rates leads to an increase in the portfolio share allocated 27
and a lower share for VEHICLE. RETIRE and RESTATE.7-2. Selfselection occurred in households’ tenure choice. RESTATE and OTHER and a decrease in the share for ACCOUNT and VEHICLE. homeownership would not have the same eﬀect on renters.11) for renters. and the share for VEHICLE is signiﬁcantly higher for households with three or more children. Since 1995. For these assets. The quadratic relationship observed between the shares of assets in homeowners’ portfolio and the age of the head holds true for the ﬁnancial assets in a renter’s portfolio. More permanent income is associated with a higher share for ACCOUNT.0 percent higher in renters’ portfolio.to HOUSE and VEHICLE. The estimates of the Mills ratios for renters are signiﬁcantly diﬀerent from zero 28 . the 1998 portfolio share for RETIRE is 5.9 and 2. Tables 2. while the portfolio share for ACCOUNT and STOCK decreases until the age 40 and 43. The coeﬃcients on the selection terms in equations for ACCOUNT. STOCK. It leads to a decrease in the share allocated to ACCOUNT. respectively. The portfolio share for RETIRE increases with age until the age of 58. Compared to 1989. The eﬀect of children is less pronounced for renters than for homeowners. for example. RETIRE and HOUSE for homeowners are all statistically signiﬁcant. renters have shifted toward RETIRE in their portfolio.10 report coeﬃcients of the selectivity variables.10 present the estimates of the equations (2. the share for ACCOUNT decreases. RESTATE and OTHER. As renters have two or more children. Tables 2. An increase in total assets leads to an increase in the share for STOCK. should they choose to buy homes.
having more children increases the housing expenditures of homeowners by only 3. For homeowners. On the other hand. and RESTATE. there were no signiﬁcant diﬀerences in the average behavior of the two groups prior to home purchase. For both renters and owners. RETIRE. The housing expenditure of homeowners increases 8. I mean a household headed by a white married.11 presents the estimates of shares for assets that a typical homeowner holds. By a typical household. After the second child. the negative selectivity bias for renters’ implies the reverse: renters spend less on housing compared to average household of the sample had it chosen to rent.9 percent higher housing expenditure than homeowners with no child. Homeowners with one child have 11.10 present the estimates of the housing expenditure equation. I use the estimated coeﬃcients and the variables of the model to calculate the portfolio share for each asset by the number of children and the age of the household head.8 and 2. Table 2. the signiﬁcance and the same sign of the selection terms indicate that self-selection occurred in a hierarchical sorting: the positive selectivity bias indicates that those who own a house spend less compared to average household had it chosen to own. the expenditure on housing increases with the number of children.for ACCOUNT.3 percent with the second child. all of the children in the household are younger 29 . but the number of children has no eﬀect on renters’ expenditure. This implies that other than in regards to these three assets.2 percent. The last two columns in Tables 2. The age of the children in the household has no eﬀect on the housing expenditure of renters nor homeowners.
The chapter examines the impact of children on the homeownership decision and the constraint of consumption demand for owner-occupied housing. One contribution of this chapter is to study the eﬀect of the portfolio constraint imposed by the consumption demand for housing on the portfolio shares in housing and other assets. and the share allocated to RETIRE becomes the second largest in the portfolio. an increase in the number of children increases the probability that a household owns a home. The household head is willing to take risky investments and holds mean wealth ($188. 2.5 Conclusion Using the 1989.160) and permanent income ($46. Using a 30 . VEHICLE is the second most important asset in the portfolio when the household head is 30 years old. children have two eﬀects on the portfolio structure of households. more is invested in RETIRE. Second. At all ages. children change the demand for each asset.than age 13. 1995 and 1998 SCF. this chapter investigates how the number and the age of children living in the household inﬂuence the portfolio composition of households. As mentioned above. The portfolio shares of assets calculated in Table 2. 1992.690) and has a 15 percent marginal tax rate. As the household head reaches middle age. The number of children has a negative eﬀect on the portfolio share for RETIRE. First. and its importance in the portfolio increase with the number of children living in the household. HOUSE is the most important asset. conditional on the tenure choice.11 include both of these eﬀects.
This result suggests that. the consumption demand for housing is higher than the investment demand. the ratio of housing to total assets increases as the number of children increases. households are saving enough for retirement. An important implication of the ﬁndings of this chapter is that the constraint imposed by the consumption demand for housing decreases the share of portfolio allocated to retirement wealth as the number of children in a household increases. Since households cannot separate the level of consumption of housing services from their investment in housing as an asset. However. for households with children. As homeowners have more children. the chapter compares the determinants of portfolio allocation of homeowners to that of renters. The results show that the number of children living in the household has a signiﬁcant eﬀect on the tenure choice and on the housing demand of homeowners. Therefore. Considerable research has focused on whether U. the portfolio share for ﬁnancial assets such as interest-bearing accounts and retirement accounts decreases. the policies that change the cost of housing and aﬀect ownership decision inﬂuence not only the portfolio share for owner-occupied housing but also the portfolio share for retirement assets. One direction for further research is to include the liabilities and bor31 . the ratio of retirement accounts to total assets in renters’ portfolios does not signiﬁcantly decrease with the number of children.switching regression model that takes into account the consumption demand for housing.S. and the portfolio share for housing increases.
The impact of children on the portfolio share for housing may be an important determinant of household mortgage debt. Most households ﬁnance their home purchases with mortgage debt. 32 .rowing constraints of households into the model of portfolio choice.
968 46.58 0.59 0.65 0.695 47.75 0.807 0.985 6.12 0.66 0.8 0.131 5.658 222. The text deﬁnes total assets.9 0.59 0.684 92.27 0.65 0. permanent income and net worth.61 3.83 0.28 0.750 0. 2) All dollar values are reported in 1998 dollars.3 0.97 48.59 0.525 101.664 6.319 50. All variables are deﬁned in Appendix A.97 48.815 258.28 0.2.28 0.75 0.154 0.054 49.64 0.1: Descriptive Statistics by Year 1989 Income and Assets INCOME ASSETS (Mean) ASSETS (Median) MRT Eh Demographics AGE MARRIED FEMALE NCHILD CHAGE13 HOMEOWN RISKY Number of observations % with positive wealth 1992 1995 1998 47. 1989-1998.660 12.80 0.97 48.12 0.509 0. Notes: 1) Tabulations are weighted using sample weights.158 0.Table 2.5 0.900 0.191 92.55 3.328 206.50 3.829 116.773 0.151 203.11 0.14 0.97 Source: Survey of Consumer Finances. 33 .51 2.164 0.
432 0. 34 .068 0. and OTHER.059 0.130 0.Table 2. 2) The text deﬁnes the assets called ACCOUNT. 1989-1998.415 0. HOUSE. STOCK.410 0.094 0. RESTATE.050 0.043 0.394 0. VEHICLE.197 0.067 0.072 Source: Survey of Consumer Finances.112 0.208 0.196 0.2: Mean Asset Shares by Year 1989 Portfolio Shares ACCOUNT STOCK RETIRE HOUSE VEHICLE RESTATE OTHER 1992 1995 1998 0.043 0.076 0. Notes: 1) Tabulations are weighted using sample weights.132 0.057 0.186 0.076 0.072 0.105 0.143 0.047 0. RETIRE.053 0.059 0.
46 14. 1998 Eh 1998 dollars %HH %HO HO RR All households 100 66.456 11.29 4.90 12.38 29.748 Age Under 35 35-49 50-64 Above 65 Wealth Below $50K $50K-100K $100K-250K $250-1000K Above 1000K Children CHILD0 CHILD1 CHILD2 CHILD3 22.764 6.Table 2.489 32.90 19.03 21.26 21. 1998.78 7.46 72.17 15.3: Expenditure on Housing.931 6.24 86. 2) HH represents all households.002 6.16 12.35 42.72 3.883 Above $100K 8.645 61.09 64.587 95.078 7.677 6.391 Source: Survey of Consumer Finances.08 22.263 5.28 9.183 7.976 7.89 6.496 6.866 4.475 5.12 80.564 5.486 95.741 7.49 36.024 6.973 6.065 7.42 78.22 7.042 6.72 78.29 68.79 64.40 34.50 51. 35 . HO represents homeowners and RR represents renters.847 15.195 5.438 8.69 3.72 14.43 9.55 6.293 $15-30K 22.546 80.030 Income Below $15K 10.803 8.69 4.378 $50-100K 29.04 5.93 1.400 5.77 67.555 93.843 9.54 5.22 64. Notes: 1) Tabulations are weighted using sample weights.081 $30-50K 29.081 6.
092 0.025 0.135 0.047 0.014 0 0.759 0.047 0.007 0.039 0 0.101 0.056 0.046 0 0 0 0 0 0.485 0.088 0.091 0. 1998 ACCOUNT STOCK RETIRE HOUSE VEHICLE HO RR HO RR HO RR HO RR HO RR 0.630 0.137 0.028 0.071 0.004 0.028 0.075 Age Under 35 0.068 0.162 0.295 0.010 0.020 0.221 0.645 0.587 0.083 0.127 0.093 $15-30K 0.608 0.281 0.162 0.078 0.157 0.694 0.102 0.097 0.234 0.436 RESTATE HO RR 0.489 0.022 0.078 0.086 0.019 0.205 0.109 36 All households Income Below $15K 0.374 0.147 0.027 0.091 0.238 0.212 0.087 0.559 0.042 $50K-100K 0.201 0.040 0.050 0 0.073 $250-1000K 0.089 0.199 0.415 0.019 0.541 0.190 0.011 0.047 0.252 0.730 0.015 0.055 0.359 0.017 0.054 $100K-250K 0.059 0.750 0.049 0.107 0.454 0.074 0.143 0.071 0 0.183 0.089 0.072 0.112 0.077 0.064 0.192 0.032 0.453 0.149 0.048 50-64 0.049 0.052 0 0.287 0.002 0.260 0.256 0.091 0.270 0.054 0.083 0.046 0.086 0.069 0.051 0.062 Above $100K 0.401 0.049 0.033 0.068 0 0.061 0.021 0 0.047 0.128 0.111 0.062 0.064 0.065 0.055 0.166 0.490 0.135 0.075 continued on the next page.018 0.041 0.080 0.4: Mean Asset Shares. .043 0 0.149 0.535 0.579 0 0.251 0.080 0.056 0.090 0.061 0.129 Wealth Below $50K 0.132 0.172 0.082 Above 1000K 0.040 35-49 0.022 0.062 Above 65 0.111 0 0.151 0.181 0.445 0.213 0.122 0.226 0.413 0.Table 2.031 0.066 $50-100K 0.090 $30-50K 0.293 0.389 0.158 0.165 0.122 0.112 0.
1998.044 0.495 0.195 0.074 0.115 0.122 0.Table 2.577 0.019 0.062 0.044 0.110 0.653 0 0 0 0 0. Notes: 1) Tabulations are weighted using sample weights.048 0.117 0.085 0. The text deﬁnes the assets called ACCOUNT.052 0.040 0.098 0. 3) HO represents homeowners and RR represents renters.088 0.070 0.5: Mean Asset Shares.609 0.038 0.049 0.061 0.040 0.032 0.087 0. HOUSE. .087 0. VEHICLE. 2) All dollar values are reported in 1998 dollars. STOCK.045 0.102 0.471 0.038 0.200 0.098 0.026 0. and RESTATE.044 0.222 0.373 0.089 0. RETIRE.069 0.055 0. 1998: Continued ACCOUNT STOCK RETIRE HOUSE VEHICLE HO RR HO RR HO RR HO RR HO RR RESTATE HO RR 0.512 Children CHILD0 CHILD1 CHILD2 CHILD3 37 Source: Survey of Consumer Finances.288 0.560 0.
004 ** 0.040 -0.2.034 ** 0.005 ** 0.014 -0.118 0.041 * -0.047 ** 0.030 ** 0.007 0.579.Table 2.261 0.005 ** 0.049 ** 0. 2) Variables are deﬁned in Appendix A.040 ** -0.747 0.044 ** 0.030 0.053 0.137 ** 0.000 MTR RISKY WHITE YEAR92 YEAR95 YEAR98 Notes: 1) ** indicates signiﬁcance at 1 percent level.102 -0.080 0.042 ** 0.137 0. and * indicates signiﬁcance at 5 percent level.341 0.052 0.304 0.052 ** 0.108 0.072 0.153 ** 0.010 2.085 0.096 0.029 -0.6: Results from Probit Estimation HOMEOWN Coeﬃcient Standard Errors Marginal Eﬀects -4.063 0.174 0.000 0.050 CONSTANT AGE AGE2 /100 MARRIED FEMALE CHILD1 CHILD2 CHILD3 CHAGE13 INCOME/10. The number of observations N=13.287 0.193 0.046 0.143 0. 38 .404 0.210 0.
013 0.190 0.027 0.020 0.006 -0.029 0.307 0.056 0.004 0.003 0.002 -0.001 0.010 0.049 HOUSE Coef SE 1.006 0.003 0.002 -0.003 -0.003 ** YEAR92 -0.021 0.009 -0.089 0.006 * CHILD2 -0.009 0.008 0.003 ** YEAR98 -0.006 0.7: Results: Asset Shares and Housing Expenditure of Homeowners ** ** ** ** ** ** ** ** 39 ** ** ** ** * * ** ** ** ** ** ** ** ** ** ** ** * ** ** ** * ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ACCOUNT Coef SE CONSTANT 0.004 0.006 -0.002 -0.005 0.049 0.013 0.004 -0.455 0.198 0.003 -0.515 0.007 0.016 0.006 0.002 0.024 0.007 -0.001 0.Table 2.003 -0.063 0.087 0.023 0.005 0.004 0.022 0.112 0.009 0.001 * MTR -0.092 0.026 0.021 0.001 0.001 -0.003 0.005 CHILD1 -0.001 0.013 0.007 0.004 0.005 0.020 0.006 L INCOME -0.023 0.007 0.007 ** CHAGE13 -0.003 -0.056 0.012 0.008 0.004 0.002 0.001 -0.001 0.018 ** RISKY -0.002 0.001 ** MARRIED -0.008 0.031 0.006 0.067 0.003 -0.024 0.003 0.016 0.004 0.020 -0.002 ** L ASSET 0.001 -0.029 0.009 0.002 0.054 RETIRE Coef SE -0.004 ** MR:home -0.006 0.021 0.019 0.063 0.007 -0.006 -0.007 -0.020 -0.004 0.011 0.063 0.001 ** AGE2 /100 0.432 0.006 -0.037 0.008 0.062 0.010 0.077 0.027 0.001 0.006 0.069 -0.008 0.032 0.007 -0.398 0.003 0.002 0.007 0.006 ** CHILD3 -0.004 0.003 ** YEAR95 -0.006 ** FEMALE 0.009 0.003 0.008 -0.008 0.026 0.003 -0.008 0.021 0.012 0.115 0.031 * STOCK Coef SE -0.023 0.005 -0.005 0.011 -0.021 0.005 0.010 0.005 0.007 -0.006 0.002 0.033 0.000 0.010 0.027 0.001 0.061 0.018 ** MR:+ wealth -0.007 0. .010 ** * ** continued on the next page.039 VEHICLE Coef SE 0.009 0.000 0.010 0.006 -0.002 -0.059 0.007 0.013 0.005 0.001 0.024 0.054 ** AGE -0.125 0.005 -0.
008 ** CHILD3 -0.025 0.001 0.220 0.001 0. .204 0.2.120 ** ** ** ** ** RESTATE Coef SE CONSTANT -0.001 ** MARRIED -0.503 0.234 0.003 -0.119 0.014 0.024 0.005 YEAR92 0.007 * L INCOME -0.002 -0.025 MR: + wealth 0. and * indicates signiﬁcance at 5 percent level.023 0. All variables are deﬁned in Appendix A.006 -0.018 0.009 FEMALE -0.001 0.052 0.177 0.158 0.004 0.116 0.010 0.189 0.012 0.025 0.Table 2.568 0.008 0.001 -0.025 -0.181 0.003 ** L ASSET 0.584 0.002 ** 2 AGE /100 -0.024 0.080 ** AGE 0.004 -0.010 0.004 0. RESTATE.015 -0.018 0.068 0.004 0.023 0.002 0. STOCK.006 0.036 Notes: 1) ** indicates signiﬁcance at 1 percent level. RETIRE.017 0.005 ** MR:home -0.007 0.011 ** 0.036 0.005 -0.019 0.004 -0.009 ** CHILD1 -0.008 CHILD2 -0.307 0.002 0.008 0.004 -0.002.043 0.015 0.002 0.017 0.002 0.025 0.005 -0. VEHICLE.002 ** MTR -0.003 0.032 0. 2) The text deﬁnes the assets called ACCOUNT.004 0.003 0.017 0.000 0.246 0.005 -0.021 0.004 0.202 0.004 -0.001 0.005 * YEAR95 0.025 ** RISKY -0. and OTHER.094 0.008 CHAGE13 0.008 0.007 ** * * 1. The number of observations N=10.066 0. MR represents Mills Ratio.005 YEAR98 -0.001 0.010 0.018 -0.004 0.393 -0. HOUSE.053 OTHER Coef SE -0.049 -0.8: Homeowners: Continued ** ** ** ** ** ** ** ** ** 40 log Eh Coef SE * -0.015 0.
004 0.9: Results: Asset Shares and Housing Expenditure of Renters ** ** ** ** ** ** ** * ** ** ** ** ** ** * ** ** ** 41 ACCOUNT Coef SE CONSTANT 0.173 AGE -0.014 0.007 0.015 * YEAR98 0.005 0.051 0.017 0.009 0.017 CHILD2 -0.793 0.014 0.006 0.008 -0.037 RETIRE Coef SE ** -0.021 0.038 0.003 0.002 ** 0.001 0.012 YEAR92 -0.010 0.120 0.017 0.070 0.014 ** L ASSET -0.004 -0.019 0.015 0.059 0.033 0.002 0.019 -0.001 0.002 ** MARRIED -0.059 0.012 0.033 0.011 0.014 0.035 VEHICLE Coef SE 2.020 0.017 0.013 0.011 0.079 * RISKY 0.080 0.154 ** 0.010 0.004 0.019 ** CHILD3 -0.008 0.012 0.039 STOCK Coef SE -0.005 0.030 0.025 0.000 0.103 -0.021 ** FEMALE 0.035 0.008 0.052 0.000 0.006 -0.019 0.003 0.002 0.285 0.072 0.016 0.064 ** 0.019 0.014 0.015 -0.032 0.026 -0.003 0.027 0.243 0.319 0.Table 2.029 0.028 0.018 0.001 0.011 -0.014 ** 0.034 0.054 0.007 0.001 0.053 ** continued on the next page.015 0.109 0.002 ** -0.003 ** AGE2 /100 0.015 -0.021 ** CHAGE13 -0.026 -0.005 0.023 0.007 0.014 YEAR95 -0.014 0.022 0.020 -0.008 -0.381 0.015 0.014 MR:home -0.027 0. .050 0.001 -0.019 0.068 0.089 0.074 0.026 0.010 0.022 L INCOME 0.243 -0.010 0.001 0.029 0.013 0.023 0.204 0.079 0.061 0.013 * CHILD1 -0.048 -0.130 0.019 0.024 0.015 0.086 0.011 ** 0.004 ** MTR 0.077 0.031 0.038 * MR:+ wealth 0.097 -0.039 0.018 0.
189 0.014 -0.033 ** MR: + wealth -0.020 0.004 0.003 ** L ASSET -0.001 0.027 0.069 0. 2) The text deﬁnes the assets called ACCOUNT.030 0.010 0.005 0.030 0.014 0.285 -0.029 0.000 0.014 0.10: Renters: Continued ** ** 42 log Eh Coef SE 4.008 RISKY 0.020 0.002 0.004 -0.010 L INCOME 0.004 0.064 0.010 CHILD1 0.017 0.012 0.365 -0.002 -0.102 0.033 -0.011 0.020 0.001 0.015 FEMALE -0.022 0.024 MR: home 0. MR represents Mills Ratio.035 -0.035 0.011 -0.010 * YEAR98 -0. RESTATE.003 -0.020 -0.012 CHILD3 -0.011 CHILD2 -0.037 0.001 0.002 2 AGE /100 -0.038 0.038 ** Notes: 1) ** indicates signiﬁcance at 1 percent level and * indicates signiﬁcance at 5 percent level.030 0.001 0.006 ** ** RESTATE Coef SE CONSTANT -0. All variables are deﬁned in Appendix A.017 0.118 0.038 0. VEHICLE.014 -0. .018 0. The number of observations N=3.2.007 0.044 0.173 0.054 MTR -0.061 0.009 * YEAR95 -0.032 0.005 0.036 0.067 0.014 CHAGE13 -0.009 YEAR92 -0. HOUSE.130 * AGE 0.045 0.083 0.038 0.001 0.073 ** 0.021 0.073 * ** ** ** OTHER Coef SE 0.109 0.022 0.012 0. and OTHER.062 0.003 -0.004 0.030 0.577.004 0.125 -0.Table 2.042 0. RETIRE.001 0.580 0.011 0.017 0.014 -0.005 0.334 0.315 0.365 0.005 ** -0.002 * MARRIED 0.110 0.014 0.016 0. STOCK.014 -0.
HOUSE.047 0.650 0.111 0.088 0.047 0.626 0.056 0.642 0.044 0.602 0.079 0.100 0.105 0.064 0.038 0.534 0.043 0.061 0.132 0.Table 2.049 0.122 0. and RESTATE.114 0.128 0.103 0.033 0.053 0.552 0.095 0.617 0.057 0.049 0.094 0.054 0.055 0.090 0. RETIRE.089 0. 43 .036 0.590 0.040 0.594 0.11: Portfolio Shares for Assets by the Number of Children and Age CHILD0 AGE=30 ACCOUNT STOCK RETIRE HOUSE VEHICLE RESTATE AGE=40 ACCOUNT STOCK RETIRE HOUSE VEHICLE RESTATE AGE=50 ACCOUNT STOCK RETIRE HOUSE VEHICLE RESTATE 0.037 0.056 0.048 0. STOCK.064 0.099 0.058 0.030 CHILD2 CHILD3 0.063 0.086 0.112 0.047 0.043 CHILD1 0.058 0.141 0.093 0.611 0.096 0.038 0.120 0. VEHICLE.055 0.044 0.043 0.049 0.101 0.043 0.036 0.577 0.053 Notes: The text deﬁnes the assets called ACCOUNT.070 0.102 0.023 0.607 0.
1 Introduction Many recent studies have recognized the role of precautionary motives on household saving behavior. Hubbard et al. In many household-level data sets. Dynan  and Starr-McCluer  ﬁnd lit1 2 See Zeldes . tracking the ageearnings proﬁle. however. Skinner . Yet.  and Carroll .2 For example. Deaton  and Browning and Lusardi  give a list of empirical puzzles. Carroll  shows that this kind of consumption proﬁle is consistent with a precautionary saving model in which individuals face uncertainty about their future earnings. consumption proﬁles over age are hump-shaped. 44 .Chapter 3 The Eﬀect of Precautionary Motives on Household Saving and Fertility 3. As an extension to the traditional life-cycle model.1 Precautionary saving models predict that uncertainty about future income may cause households to reduce their current consumption in order to raise their stock of precautionary saving. the standard life-cycle model suggests that households smooth consumption and spread resources across periods of high and low income. empirical work on the strength of precautionary saving has provided mixed evidence. Kimball . these models are able to explain some of the empirical consumption puzzles.
and.  and Lusardi  ﬁnd more support for the precautionary motive.tle or no evidence for precautionary motive. Browning and Lusardi  and Carroll et al. this chapter extends the empirical work on precautionary saving. the precautionary motive includes saving to protect the well-being of children against income ﬂuctuations. By examining the implications of uncertainty on the fertility decisions of households and incorporating fertility decisions into household saving decisions.  for the details.3 One problem that has not been mentioned in the literature is that all of these empirical models try to explain the eﬀect of income uncertainty on household savings. Most of a household’s saving motives can be grouped into one of three categories: life-cycle motives. and bequest motives. household income or the age of the head might aﬀect household saving and fertility simultaneously. ﬁnding an appropriate instrument. fertility might be aﬀected by uncertainty or income ﬂuctuations. and incorporating the restrictions of the theoretical model. 3 45 . whereas Carroll and Samwick . For example. precautionary motives.  suggest that the mixed results might be due to the diﬃculties in empirically testing for precautionary saving. given precautionary and other motives. ﬁnally. that is. the bequest motive includes saving to leave assets to children. ignoring the eﬀect of uncertainty on household composition. Yet the causal eﬀect might go in the opposite direction. This chapter takes account of the fact that children are endogenous along with the The problems include proxying certainty. It seems reasonable that these motives are aﬀected by the presence of children. Furthermore. See Browning and Lusardi  and Carroll et al. the life-cycle motive includes saving for children’s education.
the eﬀect of income uncertainty on fertility over the life-cycle. all of the four reasons reveal a hump shape: saving for ‘rainy days’ peaking in the 41-50 age group.1 percent. retirement.  for a survey of life-cycle fertility models. More than 32 percent reported that ‘rainy days’ were an important motivation for saving. The second most frequent reason was saving for retirement.4 For example. with 18 percent. This chapter also addresses a neglected topic in the childbearing literature. The most frequently reported reason for saving was to increase resources for ‘rainy days’ such as unemployment and unexpected needs. and saving for the education of children peaking between age 31-40. respectively. This suggests that the relative importance of saving for each motive depends highly on the composition and the life-cycle stage of the household. When disaggregated into age groups. The proportion of households citing saving for children’s educational expenses and home purchase were 5. saving for retirement peaking between age 51 and 60. saving for a home purchase peaking below age 31. 46 . Most life-cycle fertility models incorporate some types of uncertainty.3). Wolpin  estimates a dynamic stochastic model of fertility within 4 See Hotz et al. Table 3.7 and 4.saving behavior when estimating the eﬀect of children on savings. namely.1 presents the proportion of households citing the following motives -‘rainy’ days. buying a home and education of childrenas the most important reasons for saving in the 1983 SCF (data come from the panel of 1983-89 SCF and is discussed at length in section 3.
an environment where infant survival is uncertain.6 This chapter examines whether income uncertainty is associated with lower fertility and higher savings. however. However. treating children as a durable good the demand for which is found to respond to increases in unemployment risk (like other durable goods in Dunn. Dunn  ﬁnds consumers respond to increases in the unemployment risk by postponing purchases of a home or a vehicle. and does not aﬀect savings of the rest of the population. the time path of the husband’s income. Yet the prediction of the precautionary view of savings is not validated: income uncertainty actually reduces savings of households with either high or low wealth holdings. and consider a number of uncertainties such as the outcome of the contraceptive eﬀort.5 Hotz and Miller  integrate the life cycle fertility and labor supply. Using the data from the panel of 1983-89 SCF. and transitory shocks to the wife’s wage. and thus the variance of income does not appear in the decision function. This chapter also examines whether having a child has an eﬀect on Wolpin  presents a model in which income is stochastic but his model also assumes that households have quadratic utility. this chapter can be viewed as a combination of those two prior works. I ﬁnd that households with higher income uncertainty are less likely to have a child. None of these studies. The ﬁnding is consistent with previous studies that found little or no eﬀect of precautionary motive on savings. 6 Becker  suggests that children can be viewed as durable goods yielding psychic income to the parents. there is evidence that income uncertainty has a direct eﬀect on fertility and family size. In a study that addresses whether unemployment risk is an important factor in the timing of the purchase decision of durable goods. have speciﬁcally analyzed whether uncertainty about earnings is a signiﬁcant factor on the choice of whether or not to have a child. even after controlling for the fact that saving is endogenous to the fertility behavior.) 5 47 . Thus.
household savings. The results show that having a child appears to reduce savings of households with young heads and to increase savings of those with middle-aged heads. The remainder of this chapter is organized as follows. Section 3.2 examines both the theoretical and the empirical model. Section 3.3 describes the data set and the variables used in the empirical work. The empirical results are reported in Section 3.4, and a summary of the ﬁndings with conclusions are provided in Section 3.5.
The Relationship between Fertility and Saving
Households are assumed to maximize a lifetime utility function that is
additively separable over time. The utility of household i at age t depends on the number of children, Mit , and a composite consumption good, Cit :
β t U (Mit , Cit )
where β is the discount factor and T is the time of death. The household faces two decisions at each period: whether to have a child, and how much to consume. If parents give birth to a child at age t, then ∆CHILDit = 1, and = 0 otherwise. The number of children at age t, Mit , is the sum of all births until age t. The household is able to borrow and lend across time periods at a real interest rate. Savings at age t, Sit , depend on the household income, the cost of consumption good, and the cost of children. The household income is assumed 48
to be stochastic. Thus the household faces uncertainty about future income. Depending on the utility function, income uncertainty can aﬀect the fertility and consumption decisions of the household. This utility maximization problem, in general, is intractable and does not deliver closed-form solutions without imposing structural assumptions concerning the utility function. This makes deriving testable implications impossible, even for a two-period model. The construction of the model, however, shows how fertility and saving decisions can be determined simultaneously. The lack of testable implications from the theoretical model allows me to examine a general form of saving and fertility behavior. For the empirical speciﬁcation, I assume that the level of savings of a household i at time t, (Sit ), is a linear function of the variability of the household’s income (Φit ), birth of
s a child (∆CHILDit ), and a set of observable variables (Xit ) that measure s the life-cycle stage of the household. The matrix Xit includes the number of
children living in the household, permanent and transitory income and other household demographics. Permanent income is deﬁned as the expected income for year t conditional on the demographics of the household, and transitory income is deﬁned as the diﬀerence between realized and expected income for year t. Savings of a household i at time t can be thus represented as:
s Sit = γ0 + Φit γ1 + ∆CHILDit γ2 + Xit γ3 + u1it
where γ0 , γ1 , γ2 and γ3 are the parameters to be estimated, and u1it is an error term representing unobservable variables. 49
The precautionary saving model predicts that saving is increased by a combination of a positive third derivative of the utility function and uncertainty about the future income. Therefore, a positive value for γ1 is implied by a utility function with a positive third derivative (as with constant absolute risk aversion (CARA) or constant relative risk aversion (CRRA) utility functions). For a quadratic utility function (for which the third derivative is zero), saving behavior does not respond to income variability, and in this case, γ1 should be zero. The life-cycle model suggests that a household that gives birth to a child at time t saves less (due to an increase in necessary consumption). Households with younger heads may save even less with an additional child because their current (expected) income is less than the annuity value of their lifetime income, and the diﬀerence between their income and expenditure is even greater. Such a model suggests γ2 should be negative, and the coeﬃcient of the interaction of ∆CHILDit with the age of the household head should be positive. The childbearing decision of a fecund household is speciﬁed as a function of Φit and a set of household speciﬁc variables that aﬀect the preferences
c for a child, Xit . A household is considered to be fecund if the wife is younger
than age 49 or if the head of the household is a female younger than age 49. The decision to have an additional child is represented as
c ∗ ∆CHILDit = η0 + Φit η1 + Xit η2 + u2it
as suggested by the precautionary saving model. This data set contains detailed information on household assets. η1 and η2 are parameters to be estimated. Maddala  shows that the resulting estimates of the coeﬃcients are consistent. Then I estimate the equation (3. then they should also reduce their ‘consumption’ of children. and u2it is an error term representing unobservable variables. I restrict the sample to the fecund population and get an estimate γ2 of γ2 by using the ˆ probit ML method for the equation (3. liabilities. This implies that households with higher income variability are less likely to have a child. 3.where ∗ ∆CHILDit = 1 if ∆CHILDit > 0 = 0 otherwise. where η0 . The model is estimated using a two- stage estimation procedure described in Maddala . If consumers react to increases in uncertainty by cutting down their consumption.3 Data The data set used for estimation is the 1983 and 1989 panel of the SCF. Note that the model is identiﬁed even if u1it and u2it are not independent s c and Xit includes all the variables in Xit . The 1983 SCF interviewed a 51 .2) by OLS for all of the sample after substituting γ2 for γ2 for the fecund population ˆ and 0 for the other households. income and characteristics in 1983 and 1989.3). First. and η1 should be negative.
bonds. Total liabilities include mortgage debt. 103 households and 1. where the ﬁrst includes liquid assets (checking assets.sample of 4. This information could be used to exclude both realized and unrealized capital gains. Keogh accounts. cash value of life insurance and the later includes residential property. Total assets is the sum of ﬁnancial assets and nonﬁnancial assets.7 Household saving is derived as a ﬁrst diﬀerence in net worth between 1983-89. 497 of them were reinterviewed in 1989. substantial inconsistencies are observed between reported net investments in assets and measured changes in holdings. saving accounts. The 1989 SCF also asked households to report major changes in asset holdings since 1983. The 1983 SCF consists of a dual sample. Net worth (NWORTH) is the total value of household’s assets minus its total liabilities. mutual funds. trusts. However. The ﬁrst saving measure. a list sample was drawn from tax information provided by International Revenue Service. other loans for property. 7 52 . individual retirement accounts. which makes it diﬃcult to distinguish between active and passive saving. other real estate. call accounts. money market deposit accounts. home equity. stocks. certiﬁcates of deposit and saving accounts). vehicles and other real assets like art and precious metals. which will be called SAVE1. An oversample of 438 high-income households came from this list in 1983. automobile loans. and 361 of them were reinterviewed in 1989. and this amount is divided by six to get the annual household saving. credit card debt. business equity. In addition to a standard multi-stage area probability sample. includes capital gains. loans. balances outstanding on lines of credit and loans on consumer durables. See Kennickell and Starr-McCluer  for a general description of the 1983-89 panel.
I regress log income on age. I kept the value of the primary residence constant. To remove the predictable component of income growth. ‘In [the preceding calendar year] how much was the total income you (and your family living here) received from all sources. The income measure comes from the question. 9 The income measure includes both capital and non-capital income. dummies representing asset holdings in 1983. I deﬁne two measures of income uncertainty. before taxes and other deductions were made?’ Income of the households for 1983. education.The inconsistency seems to be lower for home purchases (Kennickell and StarrMcCluer ). 1987 and 1988 are drawn from the 1983-1989 panel. All values are converted to 1989 dollars using the Consumer Price Index Research Series Using Current Methods(CPI-U-RS). and 1985 are drawn from the 1986 wave of SCF which was conducted with a large subset of 1983 respondents using a shorter questionnaire. year dummies. The ﬁrst measure assumes that households have knowledge about their future income and expect their income to change over time as household characteristics change. 103 households in the 1983 SCF. To exclude the capital gains. 822 were reinterviewed in 1986 using a shorter questionnaire. 1984. The precautionary saving model predicts that income risk regarding capital income might have a diﬀerent eﬀect 53 . I adjusted SAVE1 as follows to obtain a measure called SAVE2: whenever a household did not buy or sell a house that was the family’s primary residence. 2. 1986. household demographic variables and age-interaction terms.8 Income values for 1982. Using the panel dimension of income observations in the data. 8 9 Income uncertainty at- Of the 4.
11 Another income variability measure. 1986 and 1989. Most studies use instrumental variables for the uncertainty proxy using information on occupation. The mean of the reported income over the 1982-88 period (MEANINC) is also used as another measure of income. Therefore. using instrumental variable estimators is not useful when the ﬁrst stage instruments are poor. while transitory income (TRANSINC) is the mean of residuals from the earnings equation.11 This measure assumes households have no information to forecast future income aside from their current income. 10 Female labor supply decisions are correlated with household fertility decisions. 54 . variability measures like VLI and VRLI might be poor proxies for uncertainty. I control for the employment status of the spouses and female heads in the earnings regression. However. The empirical results hold true for this measure too. In addition. education and industry. is the coefﬁcient of variation of log income. However.tributed to each household is equal to the variance of residual log income (VRLI) for the 1982-88 period. households probably expect their income to change over time and know when some of these changes will happen. Dummies representing the amount of assets that households hold by 1983 are included in the regression to control for this eﬀect. this information is only available for 1983. ﬁnding an appropriate instrument on household saving behavior than that of earnings.10 Household permanent income (PERINC) is deﬁned as the mean of predicted income over the seven year period. which is not reported in this chapter. The second measure of uncertainty is the variance of log income for the 1982-89 period (VLI). Unfortunately. VRLI may suﬀer from the same deﬁcit if income change is due to a factor that the household has information about but is not controlled for in the income regression. Not excluding such expected changes biases this VLI measure of uncertainty upward. As pointed out in Lusardi  and Browning and Lusardi .
divorce. All variables are described in detail in Appendix B.2 illustrates the composition of the sample in detail. 299 households out of 1. Of the remaining 1.479 experienced a major change in family composition and were dropped from the sample. households that experienced a change in composition such as marriage. 13 The sample design in 1983 speciﬁcally excluded households with the heads under the age of 22. However.) This eliminates the income variability or net worth change caused by family separation or family creation. 84 were dropped because of outlying net worth or saving values and 66 were dropped because of missing income values. this exclusion or a similar one is necessary when working with means which are aﬀected by outliers. those households with more than three missing or non-positive income values are dropped. 12 55 . 000 is somewhat arbitrary. separation or the death of either head or spouse are excluded.. The sample selection criteria for the sample are as follows. 035 households with the heads between the age of 22 and 88 in 1983. The ﬁnal sample consists of 1. In the panel 1983-89 SCF.to exclude for identiﬁcation is problematic. Therefore.13 Table 3. 000. The variable ∆CHILD indiThe cut-oﬀ net worth of $10 million and saving of $600. A household is only included in the analysis if it remained intact between 1983 and 1989 (i. See Kennickell and Starr-McCluer  for details.12 To calculate an accurate measure of income uncertainty.e.1. I use VRL and VRLI without an instrumental variable estimator.180 households. I also exclude those households with net worth greater than $10 million in 1983 or 1989 or for whom the absolute value of the change in net worth per year is more than $600.
According to the SCF data.14 I refer to the households that had a child as households with an additional child.611 in 1983. respectively.e.6 percent of the fecund households had a child between 1983 and 1989.179 less than the mean net worth of the rest of the fecund houseOnly 3. The average net worth for households with an additional child is $82. I use a probit model and a dummy variable to indicate the fertility choice instead of using a count data model.2 provides the variable means by household fertility of the fecund households.0 percent had three.cates the fertility of the household between 1983 and 1989. mostly married (89. Columns (1) and (2) provide the variable means and standard deviations of all of the households in the sample and the fecund households. and saved more compared to the other households in the sample. Fecund households are headed by younger persons. Columns (3) and (4) of Table 3. Therefore.1 percent).3 percent of the families experienced more than one birth during that time period-2. Also. fecund households are faced with higher income uncertainty than the rest of the sample. Most of the diﬀerences in net worth and saving between the two groups can be attributed to the fact that these two groups are at diﬀerent stages of their life cycles. 18.3 percent had two children and 1. 14 56 .are plausible: households with an additional child are younger. comparisons across the two groups of households . i. which is $37. have higher expected income and have a higher number of young (0-6 years old) children in 1983..with and without an additional child . but had higher income between 1983 and 1989. ∆CHILD = 1 if the household experienced at least one birth of a child. Among fecund households. had less net worth in 1983.
from 67. This suggests that households at the tails of the income distribution face higher uncertainty. 000. we observe that households who had a child are faced with lower income uncertainty (0. Uncertainty estimates are greater for households with mean income below $10. while fecund households without an additional child saved $7. 000 and above $60. According to SAVE1 and SAVE2. 462. 648 and $6. they saved $11. when we compare the income uncertainty of the two groups.0 percent in 1983 to 72. The remarkable diﬀerence in the housing tenure choice of the two groups shows the link between the decisions of having a child and purchasing a house. the bottom 25 percent of the distribution faces a lower income variability than the households in the 25-50 percent of 57 .7 percent in 1983 and rose to 77.8 percent in 1989. the increase is insigniﬁcant. 690.132 versus 0. respectively. The same argument is true for net worth and SAVE2: households in the bottom 25 and top 10 percent of the net worth distribution in 1983 have the highest income variability. On the other hand.holds.3. Also. Households in the bottom 25 and top 10 percent of the SAVE2 distribution face higher income variability than the rest of sample. When households are grouped by SAVE1. The measures of income uncertainty by household characteristics are given in Table 3. 170 and $8. The homeownership proportion for the rest of the fecund sample is higher in 1983 but compared to the households with an additional child.191). households with an additional child saved more than the rest of the fecund sample.5 in 1989. The homeownership proportion among households with an additional child was 53.
households with young and middle age heads that had a child face lower income variability.4 Estimation and Results Table 3. Almost 36 percent of the households with heads below age 31 in 1983 had a child during the following six year period. The right-hand variables include factors that are expected to aﬀect the demand for a child.5 shows the results of the probit analysis of the fertility de- cision of the fecund sample. and 1 percent among the age 40 and above group.4 represents household saving. Households who had a child between 1983-89 are diﬀerent from other households in terms of their saving. Considering the income uncertainty. 3. income and income variability. Table 3. regardless of the uncertainty measure. When grouped according to the number of children. The dependent variable is ∆CHILD = 1 if the household had a child between 1983-89. income and income uncertainty by childbearing decisions and the age of the household head. versus 16 percent of the households with heads between age 31-40. Diﬀerence between the savings of households with and without an additional child increase as the age of the household head increases. and their permanent income is higher. 0 otherwise. for whom it reaches its highest value.the SAVE1 distribution. the estimates of income uncertainty decrease as the number of children living in household in 1983 increases. 58 . income uncertainty is lower except the top 10 percent of the distribution. Households that had a child save more. regardless of how savings were measured. For other SAVE1 groups.
column (3) uses VRLI and mean income. whereas older households with small children are more likely to experience another birth. middle (MIDDCH) and older (HIGHSCH) children in 1983 and the interaction terms for age (AGE83×HOWN83 and AGE×YOUNGCH). The probability of having another child is lower for a household that has a full-time working spouse or that is headed by a white person. PERMINC and TRANSINC). column (2) uses VLI and mean income and ﬁnally. A married household is 8 to 10 percent more 59 . race of the household head (WHITE). The analysis in column (1) uses VRLI as the income uncertainty measure and permanent and transitory income as the income measures. The coeﬃcient of age is highly signiﬁcant and negative. an income risk measure (VRLI and VLI).5 show that other things being equal. respectively. number of young (YOUNGCH). age (AGE). the probability of having another child declines with income variability (regardless of the measure) and the number of children in each age group living in the household. Evaluated at the sample mean values. The results in table 3. marital status (MARRIED). The signs of the age interaction terms imply that older homeowners are less likely to have a child.6 percent. number of adults living in the household (NADULT). a 0. homeownership in 1983 (HOWN83).namely. indicating that being one year older reduces the probability of having another child by 1 percent.1 increase in VRLI and VRL decreases the probability of having a child by 0. a dummy indicating whether spouse works full time at paid employment in 1983 (SPFULLT). a household income measure (MEANINC.5 and 0.
both permanent income and mean income in columns (1) and (2) are signiﬁcant. the number of adults and children living in the household (NADULT and NCHILD). the income uncertainty interaction terms show whether or not . Similarly. income (PERMINC. ∆CHILD. Finally. 60 . The top 10 percent and bottom 25 percent net worth holdings in 1983 are included to address the saving behavior of the wealthy and the not wealthy. a measure of income uncertainty (VRLI and VRL).5.7 use the same income and uncertainty measures as columns (1)-(3) in table 3.6 and 3. The predicted probability of having a child. the probability of having a child seems to increase with income. a dummy indicating having 1983 net worth in the top 10 percent and bottom 25 percent (NWORTH90 and NWORTH25). Estimates of the saving equations are presented in tables 3. age of the head (AGE). TRANSINC and MEANINC).7 for SAVE1 and SAVE2. namely. transitory income in column (1) has a negative eﬀect and mean income in column (3) is insigniﬁcant. the behavior of the wealthy and the not wealthy are diﬀerent than the rest of the population. However. a self-described expectation to leave a bequest (BEQUEST83). compared to households headed by an unmarried person.likely to have another child. Columns (1)-(3) in tables 3.6 and 3. the change in the number of adults between 1983-1989 (∆NADULT). age interaction terms (AGE×NCHILD and age×∆CHILD) and income uncertainty interaction terms (VRLI (VRL)×NWORTH25 and VRLI (VRL)×NWORTH90). in terms of the eﬀect of uncertainty. is included as a right-hand variable with other factors that might aﬀect the saving behavior.
evaluated at the sample average of VRLI. However.162.The results for two measures of savings are quite similar (SAVE1 in Table 3. around $7. For example. Income uncertainty reduces savings of the households in the top 10 percent and bottom 25 percent of the wealth distribution and does not aﬀect the rest of the population. regardless of the measure.162 save about $15.6 and SAVE 2 in Table 3. and it is signiﬁcantly lower than the estimated propensity to save out of permanent income. which is 0.6 show that households in the top 10 percent of the wealth distribution and with VRLI of 0. households in the top 10 percent of the wealth distribution save almost $11.6. the results in Table 3. Saving also increases with income. let us look at the eﬀect of the number of adults living in the household. Before we examine the eﬀect of children on savings. Both SAVE1 and SAVE2 reduce with the number of adults living in the household.7).132-$8.447.500 less than the rest of the population as a result of an increase in income uncertainty. Having 1983 net worth in the top 10 percent is associated with higher levels of SAVE1 and SAVE2 in all speciﬁcations.6 and Table 3.7 do not the support the idea that households save a higher fraction of transitory income. respectively.34.000 more than the rest of the sample. Changes in the number of adults between 1983 61 .24 in column (1) of Table 3. which is 0. The results in Table 3.500 less than the rest of the population whereas households in the bottom 25 percent of the wealth distribution save $3. The estimated coeﬃcient of the propensity to save out of transitory income is 0.
however being one year older and having an additional child increases savings.555-$12. Also.and 1989. Having an additional child reduces savings. we observe that the eﬀects of the children and age interaction terms decrease but do not disappear. Households with children save less when the household head is below age 35 and save more above that age. however. The same is true for the number of children living in the household.7. That is another impact of children on household savings.6 and 3. When we control for permanent income as in column (1) of Tables 3. do not aﬀect SAVE1 but appear to reduce SAVE2. Controlling for the number of children already living in the household. The overall eﬀect of having an additional child on household savings depends on the age of the head: households with heads younger than age 29 save less compared to households with heads age 29 and older. age does not aﬀect the savings behavior of those without children. This chapter also estimates average savings of households who did not have a child between 1983-89 and compares it with what they would have saved if they had chosen to have a child. households expecting to leave a bequest save signiﬁcantly more (around $12. For this. and the fertility decision is modeled as an endoge- 62 . This result highlights the importance of the interaction between household composition and the age of the household head. the sample is restricted to only fecund households.796 more) than household that do not expect to leave a bequest.
nous switching model. depends on the age of the household head. even after controlling for several demographic characteristics.6. average SAVE1 of the households that did not have a child is around $12. I take into account the fact that fertility decisions are endogenous to household saving decisions.8. The direction of the response. 3. changes in the number of children and children already living in the household reveal a signiﬁcant eﬀect on household savings. Overall.5 Conclusion This chapter estimates the eﬀect of the precautionary motive on house- hold fertility and savings by relating income uncertainty to the changes in the number of children and household net worth. In estimating this eﬀect.133 according to the results of the three regressions in Tables 3. The results show that households would have saved around $2.15 The results are given in Table 3. Finally. implying that younger households save less whereas older 15 See Maddala  for the models with self-selectivity.066 less if they had chosen to have a child.695-13. however. Income uncertainty actually decreases savings of the households with high or low wealth holdings and does not aﬀect the saving behavior of the rest of the population. This ﬁnding suggests that the overall eﬀect of children on household saving is negative.297-4. 63 . The empirical results suggest that income uncertainty directly aﬀects the probability of having a child.
The main ﬁnding of this chapter is consistent with the life-cycle theory of saving and consumption. the ﬁndings are not consistent with the predictions of the precautionary saving model that agents faced with uncertainty about future income increase their savings. 64 .households save more with an increase in the number of children. Household composition is an important factor of life-cycle savings. At the same time. the age eﬀect on savings disappears. After controlling for the number of children living in the household and the expectation of leaving a bequest.
052 0.323 0.362 0.111 0. Number of observations: 1035.192 0. retirement. 1983 All By Age Below 31 31-40 41-50 51-60 61-70 70 and over Rainy Days Retirement Home 0.017 0. population as a whole.250 0.057 0.016 0.345 0.011 0.Table 3.047 0.065 0.1: Saving Motives by Age Groups.230 0.S.326 0.383 0. Observations are weighted to reﬂect the U.004 0.047 0.117 0.176 0.206 0. 65 .000 Source: Survey of Consumer Finances.273 0.289 0.000 0.041 Children 0. 1983-1989. buying home and education of children respectively. Notes: The table reports the proportion of households citing the selected motives as the most important reason for saving as ‘rainy days’.010 0.122 0.
297 0.437 0.162 0.672 0.302 0.487 0.556 0.189 38.334 0.725 0.135 0. All dollar values are in 1989 dollars.734 0. All variables are described in Appendix B.747 0.7 0.490 0.767 1.6 13.624 0.001 0.4 13.055 0.780 0.186 2.670 0.537 0.645 0.191 0.460 0.652 0.2: Descriptive Statistics by Household Fertility Decision SAVE1 SAVE2 MEANINC PERMINC NWORTH AGE EDUC WHITE MARRIED NCHILD YOUNGCH MIDDCH HIGHSCH ∆CHILD NADULT BEQUEST HOWN83 HOWN89 VRLI VLI All HH 7699 6080 37668 36339 140628 45.795 0.180 0.791 1.246 0.8 12.289 0.407 0.891 0.778 0.456 0.092 2.725 0.9 0.132 0.151 N 1035 509 422 87 Note: ∆CHILD=1 if the household had a child between 1983 and 1989 (0 otherwise).Table 3.537 0.845 0.1 0.112 2.1.120 0.4 0.938 0.516 2.424 0. 66 .169 Fecund HH ∆CHILD = 0 8282 7648 6868 6462 43109 44122 40171 39921 112861 119790 36.788 0.197 ∆CHILD = 1 11170 8690 38852 41210 82611 30. The table reports means of the variables.372 0. Observations are weighted using the sample weights.1 12.
265-52.2449) (0.2657) Above $60.446 15.1468 0.1803 0.490-7.2117 (0.3230) $127.518-314.426) MEANINC Below $10.1194 (0.2579) $7.1284 (0.0 0.169 (0.1296 (0.9087) NWORTH Below $10.000-59.1269 (0.6504) $10.3279) SAVE1 Below (-$1.162 (0.000-29.818 10.2269) (0.9431) (0.1889 (0.2371) $1.999 36.1996) (0.5530) NCHILD No Children 55.2299 0.2281 (0. and VLI is the variance of log income.3980) Above $24.1.4780) $ 10.3234 (0.1832 (0.5537) $ 1.0 0.2458) 0.056 25.5836) (0.0988 (0.0 0.0 0.078)-1.0 0. VRLI is the variance of residual log income.2330 (0.125 25.818 15.0 0.4940) (0.2290 0.1716 (0.1794) Above $30.5066) 1 Child 17.1479 (0.2013 0.078) 25.0 0.1559 0.2134 0.3: Mean Income Uncertainty by Household Demographics % HH VRLI All 100 0.0998 0.4092) (0.1027 0.7 0.0 0.7 0.1715 (0.3334) $52.492 15.232 25.000 12.1330 0.265 25. All variables are deﬁned in Appendix B.1436 0.1855 (0.1088 (0.1 0.4232) $10.1527 (0.999 37.2920 0. 67 .1095 0.1808) (0.3333) (0.492 10.1363 (0.0 0.2740) (0.0902 (0.1056 0.726.0977 (0.2171) 2 and more 27.000 14.4963) 0.446 10.4472) (0.1 0.Table 3. Observations are weighted using the sample weights (N=1305).489 25.3376) VLI 0.057-30.0 0.2210) (0.5245) 0.1184 0.2 0.1912 0.6357) (0.2274) (-$1.2642) $30.126-24.0 0.10.233-12.2374 (0.2738) (0.725 25.0 0.1451 0.0 0.2 0.2825) (0.1312 (0.4881) (0.7517 25.1707 (0.6129) (-$739)-1.2391 (0.6786) (0.5512) SAVE2 Below (-$739) 25.2086) Above $314.0 0.3524) (0.2973) Note: Standard deviations are given in parentheses.433) 0.0 0.2713 (0.1462 (0.3362 0.0 0.
0812 0.1369 0.5 4. 68 .2376 VLI 0.4: Savings.257 43.1507 20.272 9.1408 0.636 27.0841 56.331 23.3064 0.619 3.8 10.220 34.812 6.1727 3.2920 6. and VLI is the variance of log income.528 4. Income and Income Uncertainty by Age and Fertility AGE Below 31 ∆CHILD 0 1 31-40 0 1 41 and above 0 1 %HH SAVE1 SAVE2 PERMINC VLRI 12.390 52.422 0. VRLI is the variance of residual log income.1.477 40.1 4.487 0.899 0.1839 0.5 24.2261 Note: ∆CHILD=1 if household had a child between 1983-1989 (0 otherwise). All variables are deﬁned in Appendix B.9 11.1774 0.Table 3.049 5.3 8.769 0.884 36.628 0.912 0.
20 Pseudo R2 0.0001] [ 0. .385** 0.0920] 1.131 0.392** 0.0008] [-.0612] [0.062 -160.543 0.0553] [-0.886** 0.0095] [-.0787] [-0.154** 0.627** 0.564 HOWN83 2.047 -0.175** 0.178** 0.566 [-0.060 -160.000 0.0829] [-0.225** 0. Number of observations N=509.132** 0.077 0.3338] [-0.980 -0.0001] [ 0.749** -0.1037] [-0.0625] -0.418 0.0844] [-0.130** 0.1041] [-0.0577] (3) Coef StdE 1.998 -0.565 1. YAGE is YOUNGCH×AGE.420 YAGE 0.000* 0.953** 0.409 0.236 [-0.886** 0.186** [-.068 YOUNGCH -1.073 SPFULLT -0.31 0.225** [-.049 Likelihood -152.00 0.006 TRANSINC/1000 -0.357 0.232 HAGE -0.31 0.369 -0.019** 0.0094] [-.0675] -0.0613] 0. ** indicates signiﬁcance at 5 percent level.228 WHITE -0.14 0.203** (1) Coef StdE 1.019** 0.545 0.130** 0.166** 0.768 0.175** 0.027** 0. Marginal eﬀects are given in the brackets.410 0.369 [-0.1038] [-0.001** 0.623** 0.3340] [-0.001 MEANINC/1000 MARRIED 1.208** 0.5: Probit: Fertility Decision of Fecund Households (2) Coef StdE 1.024** Note: Coef reports coeﬃcients and StdE reports standard errors.0426] -0.0541] -0.35 0.024** 0.0594] 69 CONSTANT AGE VLRI VLI PERMINC/1000 0.001 1.060 -2.384** 0.060 [-0.0845] [-0.397 0.047 [-0.024** 0.0550] [-0.368 -0.448 0.021** -0.063 0.366 -0.020** [-.130 0.0786] [-0.001 1.2375] -2.020** -0. and HAGE is HOWN83×AGE.134 0.229 -0. and * indicates signiﬁcance at 10 percent level.063 0.714 NADULTS 0.720* -0.208** 0.154** 0.717* -0.1040] [-0.627** 0.1297] [ 0.389 0.331 HIGHSCH -0.844 MIDDCH -0.001* [ 0.0099] [-.Table 3.
035.Table 3.6: Regressions of SAVE1 on Income Uncertainty with Endogenous Fertility Decision (1) Coef StdE 19212 19254 2250 2252 343 3520 26829 12463 ** -21824 6015 ** -71146 29321 ** 2180 2441 -500 3222 36903 11508 ** -464 3230 37290 11447 ** -22195 5914 ** -71517 28439 ** (2) Coef StdE 15228 19373 (3) Coef StdE 15240 19296 2258 2365 70 -208 294 -134139 93086 4256 3304 343 102 ** 240 59 ** -7554 3838 ** -22542 11934 * 646 320 ** 12654 5751 ** -7664 5177 .22 -21218 5702 ** -63955 26900 ** -131 306 -131 305 -185235 83042 ** -186203 83131 ** 6412 2952 ** 6474 2950 ** CONSTANT VRLI VRL NWORTH25 NWORTH90 VRLI× NWORTH25 VRLI× NWORTH90 VLI× NWORHT25 VLI× NWORHT90 AGE ∆CHILD AGE ×∆CHILD PERMINC TRANSINC/1000 MEANINC/1000 NADULTS NCHILD AGE× NCHILD BEQUEST83 ∆ NADULT R2 255 53 ** -7132 3916 ** -24479 11818 ** 698 319 ** 12916 5780 ** -7860 5120 * .22 ** ** ** ** ** Note: Coef reports coeﬃcients and StdE reports standard errors. ** indicates signiﬁcance at 5 percent level. and * indicates signiﬁcance at 10 percent level. Number of observations=1.22 257 53 -7244 3908 -24818 11791 707 319 12952 5786 -7845 5112 . .
18 223 57 -8264 3929 -24089 11699 686 319 11736 5772 -10074 5321 . . and * indicates signiﬁcance at 10 percent level. ** indicates signiﬁcance at 5 percent level.035.Table 3.7: Regressions of SAVE2 on Income Uncertainty with Endogenous Fertility Decision (1) Coef StdE 23200 19365 1501 2042 712 3511 24093 12070 ** -20171 6222 ** -61036 29489 ** 1583 142 32041 2137 3237 11432 ** 182 3244 32484 11382 ** -20204 6023 ** -60981 28519 ** (2) Coef StdE 17753 19528 (3) Coef StdE 17718 19448 1550 2091 71 -257 293 -126939 91894 3945 3267 292 100 ** 210 63 ** -8479 3843 ** -21818 11811 * 624 320 * 11545 5760 ** -9888 5374 .18 -19347 5780 ** -53719 26977 ** -161 306 -160 305 -175653 82756 ** -177076 82940 ** 6075 2966 ** 6149 2969 ** ** ** ** ** * CONSTANT VRLI VRL NWORTH25 NWORTH90 VRLI× NWORTH25 VRLI× NWORTH90 VLI× NWORTH25 VLI× NWORTH90 AGE ∆CHILD AGE×∆CHILD PERMINC/1000 TRANSINC/1000 MEANINC/1000 NADULT NCHILD AGE× NCHILD BEQUEST83 ∆NADULT R2 221 56 ** -8168 3935 ** -23842 11737 ** 679 320 ** 11712 5766 ** -10090 5329 * .18 Note: Coef reports coeﬃcients and StdE reports standard errors. Number of observations=1.
154 10.8: The Eﬀect of a Change in the Fertility Decision on SAVE1 Fecund HH E(SAVE1|∆CHILD=0) E(SAVE1|∆CHILD=1) N (1) (2) (3) 13. 72 .672 8.527 10.Table 3.133 12.695 12.375 422 422 422 Notes: E(SAVE1|∆CHILD=0) denotes average SAVE1 of the households that did not have a child between 1983-1989 and E(SAVE1|∆CHILD=1) denotes average SAVE1 of the households had they chosen to have a child.
According to the 1996 National Postsecondary Student Aid Survey (NPSAS).900 (Choy and Henke ). While the percentage was lower for those in the lower income group (income below $35.000) it was 98 percent (Presley and Clery ).000). Of those contributing to their children’s college costs. According to the 1987 NPSAS. First. and 80 percent reported using some current income. parents contribute a signiﬁcant amount to their children’s college costs.1 Introduction The purpose of this chapter is to analyze an important life-cycle saving motive: saving for children’s college education. and the average amount of their support was about $3. about 65 percent reported using some previous savings.Chapter 4 Saving for Children’s College Education 4. 90 percent of dependent undergraduate’s parents contributed to their children’s college costs. 65 percent of the parents contributed a positive amount to their children’s college costs as a gift. Gale and Scholz  estimate that the annual ﬂow of parental 73 . Understanding the eﬀect of ﬁnancing children’s college education on household saving behavior is important at least for three reasons. for those in the higher income group (income above $70. Using the 1983-86 SCF.
However.441. families who save for college reduce their eligibility for ﬁnancial aid. contributions to children’s education yield a wealth of $1. Second.5 billion. which is 12 percent of the aggregate net worth in 1983. Edlin .contributions totaled about $35 billion. the focus has been on calculating the ﬁnancial aid tax and measuring its negative impact on household asset accumulation. the ﬁnancial aid tax rate on capital income can be as high as 50 percent. Gale and Scholz  convert the ﬂow of college support to a stock of wealth using steady-state assumptions. as shown in Long . Using the data on actual expenditures on children’s college education. this chapter examines the eﬀect of anticipated educational expenses on household savings. According to their estimation. To date. Long  ﬁnds that the eﬀect of the ﬁnancial aid tax on asset holdings is smaller than the eﬀect in the prior literature. The college ﬁnancial aid system imposes an implicit tax on the savings of households that are potentially eligible for ﬁnancial assistance. According to Edlin  and Feldstein . Dick and Edlin  and Long  have recently examined the adverse eﬀect of the means-tested student aid process on household asset accumulation. Dick and Edlin  use data on ﬁnancial aid awards to calculate a marginal tax rate and ﬁnd that families with children attending average-priced colleges face a ﬁnancial aid tax ranging from 2 percent to 16 percent. Using alternate but also plausible assumptions. the results in Edlin  and Feldstein  depend on a variety of assumptions such as the number of children enrolled in college. anticipated college costs and the amount of aid that is received and so on. Feldstein . 74 .
diﬀerent forms of parental expenditure such as children’s schooling. they ﬁnd an inverse relationship between family size and children’s schooling. The analyses in Willis  and Becker and Lewis  show that parents with few children have substituted quality for quantity. are negatively related to subsequent levels of inheritance. and test predictions of their model using the veterans sample of white male twins and the sample of their adult oﬀspring. Speciﬁcally. Behrman et al. The estimates in Tomes  conﬁrm the prediction of the quantity-quality model that bequests and children’s income are negatively related to family size. Tomes  empirically tests whether parental bequests of wealth and human capital investments represent substitute forms of parental transfer. The results of his model conﬁrm that investments in children’s human capital. The estimates in Tomes  show that family size and children’s schooling are jointly determined.Third. In the empirical investigation of this model. Using the National Longitudinal Survey of the High School Class of 1972.  develop a model relating children’s schooling to family size. Steelman and Powell  investigate the relationship between the structure of the sibling group and parental ﬁnancial support for children’s college education. Without unequal access to schooling. which are measured by children’s income and years of schooling. they analyze the inﬂuence of size and ordinal position of siblings on the like- 75 . child care and bequests have been used as the qualitative measure. the quality-quantity model of fertility behavior assumes that parents have preferences both for the expenditure per child and the number of children. with and without equal access to ﬁnancing for education.
Saving for children’s education is the third most important saving motive after saving for retirement and ‘rainy days’ and accounts for 9. an analysis of ﬁnancing college education and family size highlights an important aspect of the quality-quantity model. A number of studies have analyzed motives for saving such as saving for retirement. using Japanese household data.1 percent of gross saving. Steelman and Powell  argue that later-born children are more favored relative to earlier-born ones due to the family life cycle. Their results show that the number of siblings signiﬁcantly decreases both the likelihood and amount of parental contribution to children’s college education.2 In addition.2 percent of gross saving. The results of their analysis show that retirement and precautionary motives account for 25. Their ﬁndings also show that the importance of each saving motive depends on the age and the life-cycle stage of the household. ordinal position alters parental support in favor of later-born children. Horioka and Watanabe  analyze the amount of gross saving and dissaving for each of twelve motives including saving for retirement. saving for ‘rainy days’ and saving for bequests and inter vivos transfers. The data set used in the chapter does not provide information on the ordinal position of the child attending college. 1 76 . respectively.1 Given the rapidly rising cost of college tuition.lihood and amount of parental support. bequests. Parents have more resources when later-born children reach college age. It would be of interest to investigate this eﬀect on the level of parental support using the information on household savings. This chapter also uses the amount of parental expenditure on children’s college education as a measure of child quality. education and so on. 2 See Browning and Lusardi  for a survey of the literature. emergencies. Moreover.7 and 28.
’ is the most cited reason. The sample includes households with nonretired heads and spouses (The SCF and restrictions on the sample are discussed in Section 4. income ﬂuctuations and bequests have motivated substantial research. His results show that households smooth their consumption into the academic year and do not cut their consumption in the 6-9 months before the academic year starts. His ﬁndings are consistent with the life-cycle theory of consumption and saving. While 35.Although saving for retirement. Using the Consumer Expenditure Survey. One exception is Souleles . ‘rainy days. The SCF contains a question that asks the household’s most important reason for saving.’ home purchase and children’s education. Table 4. medical and dental expenses.5 percent of households list ‘rainy days’ as the most important reason for saving.1 shows the percentage of households reporting that they cannot or do not save. The table provides the responses of the sample used in this chapter. 15. 77 .3 percent list retirement and 5.3).3 percent list education as the most important reason for saving. education of children. buying a home and other reasons as the most important reason for saving. saving for ‘rainy days. Among the households saving for retirement. ‘rainy days’ (emergencies and unemployment). Souleles  examines consumption of households as they pay for the college expenses of their children. and buying durable household goods. The last column of Table 4. the motive of saving for children’s education has not been much investigated.1 shows the percentage of households in the 1983 survey citing retirement. taking vacations and so on. Other reasons for saving include saving for ordinary living expenses.
Controlling for the number of children. the eﬀect of the number of children on the percentage of households reporting saving for children’s education disappears. As the number of children increases. and the percentage of households citing retirement as the most important reason increases.8 percent). The percentage of households in the 25-75 percentile of the wealth distribution citing ‘rainy days’ as the most important saving motive is higher than the percentage of households in the lower and higher wealth groups. For example. 24. Among the households in the bottom 25 percentile of the wealth distribution.3 percent vs. among households with 1 or 2 children. while only 2.The percentage of households saving for ‘rainy days. An almost equal number of households with one or two children and with three or more children report saving for children’s education as the most important reason. 5.1 also shows the percentage of households citing each saving motive by the number of children and net worth in 1983.’ retirement and other reasons show a systematic trend relative to the total number of children. the percentage of households saving for retirement increases with wealth. This table shows that the number of 78 . a higher percentage of the households with one or two children report saving for children’s education than those with three or more children (8. Among the households in the higher wealth groups. Table 4.7 percent of those in bottom 25 percentile of wealth distribution report saving for retirement. the percentage of households citing ‘rainy days’ and other reasons as the most important reason decreases.4 percent in the top 25 percentile report saving for retirement.
The data from the 1983-86 SCF is used to estimate two equations in which the dependent variables are household savings and educational expenses. Households with higher income and wealth expect to have higher educational expenses. The results show that an increase in the number of children decreases the per child college expenditures paid by households by approximately by $317 in 1986 dollars.000 in children’s college expenses saves $8. a household with a 43 year old head expecting to have $2. the eﬀect of anticipated educational expenses on household savings are estimated. Also. In this chapter. households save for their children’s college expenditures. and the amount of savings increases with the age of the household head.000 more than it would had it not expected to have any college expenses. I introduce life-cycle savings into the quality and quantity model of fertility and derive predictions concerning the eﬀect of expected educational expenditures on household savings. The remainder of this chapter is organized as follows. Further. I also obtain predictions concerning the simultaneous determination of family size and college expenditures per child. the empirical ﬁndings provide an answer to why saving is concentrated among wealthier households. Other things constant. and they save in advance for these expenses.2 an79 . Section 4. We continue to observe this eﬀect even after controlling for the household wealth.children has a signiﬁcant eﬀect on saving motives. Using the actual college expenses reported in the SCF. The results are consistent with the predictions of the Life-Cycle Theory of saving and consumption that households save in advance for children’s college expenditures.
5 estimates the determinants of college expenditures and uses these estimates to investigate the eﬀect of expected college expenditures on household savings. investment to each child’s education. n) subject to c1 = y1 − A c2 = y2 + (1 + r)A − πen 80 (4. the return on accumulated assets (1 + r)A and second period wage income y2 are divided between consumption c2 and paying for children’s college education e. Section 4. Section 4. c2 .2 A Model of Saving for College This section considers a world in which individuals (parents) live for two periods. and the number of children to maximize U = U (c1 . chooses to have n children.3 describes the 1983-86 SCF. Parents choose ﬁrst-period consumption. In the ﬁrst period. per capita college investment is assumed to be equal for all n children. Section 4.2) (4. Finally.1) .3) (4. For simplicity.6.4 provides a framework for the empirical analysis of the interaction between savings and college expenditures.alyzes a model of the quality-quantity interaction of fertility with household savings. second-period consumption. a couple earns y1 . 4. In the second period. and the family consumes together c1 and saves A to earn interest at the rate of r. a summary and conclusions are presented in Section 4. e.
6) (4.4) where the three choice variables are accumulated assets (A). which results in an increase in accumulated assets. educational expenses (e) and the number of children (n).2) and (4.7) where U1 and U2 are the marginal utility of consumption in the ﬁrst and second periods. Substituting (4.3) into (4. respectively. y2 + (1 + r)A − πen. 81 .8) where γ is the elasticity of substitution. Ue is the marginal utility of children’s education. Since the right-hand side is a constant. and Un is the marginal utility of family size. This expression implies that an increase in educational expenses decreases the second-period consumption relative to ﬁrst-period consumption. equation (4. a decrease in second-period consumption is likely to decrease the ﬁrst-period consumption. y2 + (1 + r)A − πen (4. If the utility function is CES with equal elasticity of substitution between all arguments.5) can be written as follows: 1 y1 − A = (1 + r) γ−1 .5) (4.1) yields the following unconstrained maximization problem: U = U (y1 − A. e. n) (4. The ﬁrst-order conditions are −U1 + (1 + r)U2 = 0 −πnU2 + Ue = 0 −πeU2 + Un = 0 (4.where π is the price of education.
6) and (4.To extend the analysis to account for uncertainty.3 The combination of a positive third derivative of the utility function and uncertainty about future income reduces consumption in the ﬁrst period. household saving can be associated with two diﬀerent saving motives: saving for uncertainty about future income (precautionary saving) and saving for children’s education. then U2 is a convex function. Solving the consumer’s problem yields the following equation −U1 + (1 + r)E1 [U2 ] = 0 where E1 represents the expected marginal utility of consumption in the second period conditional on all information available in ﬁrst period. When uncertainty about future income is assumed. This interdependence implies an inverse relationship between the number of children and educational expenses. In this case.7). If the third derivative is positive. 3 82 . E1 [U2 ] exceeds U2 [E1 ]. The interaction of the quality and quantity dimensions of choice is reﬂected in the fact that the marginal costs of education and family size depend on the level of each other in equations (4. let us assume that the second period wage income y2 is stochastic. The empirical speciﬁcation of the model described below controls for precautionary saving while it estimates the eﬀect of educational expenses on household savings. This condition shows that greater uncertainty is linked to greater saving in the ﬁrst period when the third derivative of the utility function is positive.
SAVE1. The college expenditure variable is the outlay of college education per child.824 households and a high-income supplement of 438 households. This variable includes children of previous marriages living with former spouses. income and demographic characteristics. In 1986. the data does not diﬀerentiate between children away in college or living on their own and with former spouses. including those not living in the household. The ﬁrst measure.822 of these households were reinterviewed. The variables used in the empirical analysis are classiﬁed into four groups: fertility. college expenditure. 4 83 . Unfortunately. The fertility variable (CHILD) is the number of children of either the respondent or spouse. The respondents were also asked how many years of college their children completed from 1983-85. 2. liabilities. Household savings are measured in two ways. I use the number of children attending college and the number of years they attended to normalize college expenditures. The SCF contains detailed information on household assets. respondents were asked if they had any children attending college from 1983-85 and if they had any college expenses on the behalf of their children.4 The college expenditure variable (COLLEXP) captures the quality dimension associated with the expenditure per child. savings and other control variables.4.3 Data The empirical analysis uses data from the 1983-1986 SCF. The 1983 survey contains interviews from a random sample of 3. In 1986.
I estimate household permanent and transitory income. certiﬁcates of deposit and saving accounts). and the demographic characteristics associated with tastes (urban residence. Total liabilities include mortgage debt. and whether or not the household head is willing to undertake risky investments). 5 84 .is the change in net worth between 1983 and 1986 divided by the number of years. trusts.5 SAVE1 includes the realized and unrealized capital gains. stocks. business equity. In order to exclude unrealized capital gains on the primary residence. other real estate. Using the reported household income for 1982. 1984 and 1985. gender of the household head and other household characteristics. marital status. education. Other controls include variables that aﬀect savings. Transitory income (TRINC) Total assets is the sum of ﬁnancial assets and nonﬁnancial assets. balances outstanding on lines of credit and loans on consumer durables. money market deposit accounts. cash value of life insurance and the later includes residential property. home equity. automobile loans. bonds. These are age. the value of the primary residence in 1983 is kept constant. Household permanent income (PERINC) is deﬁned as the predicted income in 1985 obtained from regressing the log of total income on age. and the educational level of the spouse. mutual funds. call accounts. race. vehicles and other real assets like art and precious metals. educational expenditures and fertility decisions. where the ﬁrst includes liquid assets (checking assets. loans. gender and the educational level of the household head. Keogh accounts. 1983. reasons for borrowing and saving. Whenever a household did not buy or sell a house that was the family’s primary residence. race. Net worth (NWORTH) is the total value of household’s assets minus its total liabilities. other loans for property. individual retirement accounts. a second measure of savings (SAVE2) is used. credit card debt. saving accounts.
806 and according to second measure. Households with retired household heads are assumed to be in the life-cycle stage of dissaving.931 and $3. Table 4. The average household net worth in 1983 is $81. These restrictions leave us with a sample containing 1.is the diﬀerence between reported income in 1985 and estimated permanent income. respectively. The sample is restricted to families that did not change composition from 1983-86. Estimating the relationship between savings and educational expenditures is complicated for families who experienced a major change in composition such as marriage and divorce. the average expenditure is $2.005. Table 4. Appendix C.000 are excluded to avoid the diﬃculty of modeling the relationship between educational expenditures and savings. For households with nonzero college expenditures. it saves $4. The average household in the sample is headed by a forty two year old married high school graduate and includes two children. Also. households with family income above $100. The head of the median household reports that it is all right to borrow money for educational expenses.575 in 1986 dollars. The sample is also constrained to include only the households with nonretired household heads and their spouses if the head is married.296. According to the ﬁrst measure of savings. Permanent and transitory incomes in 1985 are $25.811.690 households.3 presents average household savings and college expenses by 85 . the typical household saves $5.2 presents summary statistics of the variables used.1 gives a detailed deﬁnition of the variables used in the estimation of the model.
3 also breaks down savings and college expenses by the number of children in college and net worth in 1983. 25-75 percentile and top 25 percentile) spend less per child as the number of children attending college increases. However. Households in the top 25 percentile of the wealth distribution with two or more children in college save on average $18. Table 4. Households in all three wealth groups (bottom 25 percentile. However. As the number of children in college increases. the number of children in college is inversely related to the expenditure per child as predicted by the quantityquality model. The data show that households continue to save while children are in college. household savings increase with the number of children in college. Savings of households with children in college increase with net worth.10) . Except the households in the bottom 25 percentile of the net worth distribution with two or more children attending college.077.the number of children attending college. household savings increase with the number of children in college. The data show that college expenses increase with net worth.4 Empirical Speciﬁcation The simultaneous relation between educational expenses (e∗ ) and house- hold savings (a) is speciﬁed as follows: e∗ = δ1 ni + x1i κ1 + u1i i ai = η2 e∗ + x2i κ2 + u2i i 86 (4. college expenditures per child decreases. more than twice as much the households in the same wealth group with one child in college.9) (4. 4.
The theoretical model derives predictions concerning the eﬀect of the completed lifetime fertility on the educational expenses.e. u2i ) are assumed to be randomly drawn from a 2-variable distribution with E(ui ) = 0 and E(ui ui ) = . η2 > 0.6 Information on educational expenses is available only if the household has a child attending college and if the household spends a positive amount The model is also estimated using the predicted number of children obtained from the Poisson model. I use estimates of the parameters of a Poisson regression model to construct the completed fertility proﬁle when the household head is 55 years old. Another prediction of the model is that educational expenses increases household savings. κ3 is a vector of parameters. is obtained to estimate the expected educational expenses. i. and κ1 and κ2 are the vectors of parameters to be estimated. n. The structural disturbances ui = (u1i . The model predicts that an increase in the number of children decreases the anticipated and actual educational expenditure. the data on household fertility gives the number of children ever born to a household headed by a person of a certain age. However. and φi is an age speciﬁc factor. 6 87 .e.11) where x3i is a vector of demographic characteristics. The expected completed family size is given by E[n] = exp(x3i κ3 + φi ) (4. the completed fertility. i.where x1i and x2i are the vectors of exogenous variables. The empirical results hold true for this measure of fertility too. δ1 < 0. Therefore.
4. a dummy indicating whether the spouse works for a full time job in 1983 (FSPOUSE). I estimate a Tobit model for the educational expenditures of the households with children attending college. I obtain expected i i educational expenses as follows. and a dummy indicating whether the household does not live in a SMSA area (NSMSA). Estimates of the coeﬃcients in Table 4.on ﬁnancing her education. and education of the spouse (HIGHSCHSP and COLLEGSP). Married households have more children. Then I use the estimates of those parameters to construct the proﬁle of anticipated educational expenses. First. Let gi = 1 indicate that the household has a child attending college. martial status (MARRIED). The right-hand variables include household demographics expected to aﬀect the number of children: namely. age (AGE). gender (FEMALE).5 Estimation and Results Table 4. Controlling for permanent income. Then e∗ is observed to be ei if e∗ > 0 and gi = 1.4 contains estimates of the Poisson regression model of the fertility equation. gi = 0 indicate that none of the children are attending college. However. households headed by high school and college graduates have fewer children than those headed by persons without a high school degree.4 are consistent with previous studies. permanent income (PERINC). An increase in the permanent income increases the number of children. 88 . race (BLACK) and the education of the household head (HIGHSCH and COLLEG).
As predicted by the quantity-quality model. SCHEDUC and SHOME). SEMERG. and permanent and transitory income (PERINC and TRINC). Households with children attending college between 1983-86 are included in the estimation of the Tobit regression. The instrumental estimate of the coeﬃcient on the number of children is almost three times as large as the OLS estimate (-$459 vs. 338 had a child attending college between 1983-86. namely.controlling for marital status. a dummy indicating whether or not the household head believes it is all right to borrow money for college expenses of children (BEDUCAT) and dummies indicating the most important reason for saving (SRETIRE. the amount of college expenditure decreases with the number of children.690 households in the sample. I use the estimates of the regression to predict the completed household fertility when the household head is 55 years old (CHILD).9). The partial derivative of the expected college expenditure with respect to the number of children is calculated at the mean values of the estimated number of children (CHILD) and other explanatory 89 . Columns 1 and 2 contain the results with CHILD and columns 3 and 4 contain the results with CHILD.03. -$187). Of the 1.5 reports estimates of the equation (4. The average CHILD is 3. age (AGE) and education (COLLEG) of the household head. Table 4. and 252 reported contributing a positive amount to their children’s college expenses. The right-hand variables also include other factors that might aﬀect the college expenditures. households with spouses working full-time and with high school and college degrees have fewer children.
Households citing saving for children’s education as the most important reason for saving spend more than other households. respectively. While the average contribution of households in the bottom 25 percentile of wealth distribution is $1. an additional child results in a drop of $317 in expected college expenditures at the mean of values. the average contribution of the top 25 percentile is $3.12) where σ is the estimate of the standard error. I use estimates of the Tobit model and the expected completed fertility. ˆ Using the approximation.093 per child. Households with heads who believe that it is all right to borrow for educational expenses have higher expenditures. CHILD. Estimated contributions of the households with children in college are very close to the actual expenses.6 show SAVE1 and SAVE2 for households with 90 . and δ1 and κ1 are the estimates of δ1 and κ1 . Table 4. ˆ σ (4. The last two columns of Table 4. The average COLLEXP is $1.6 presents actual and estimated college expenses by household net worth in 1983. Estimates in Table 4.variables as follows: ˆ ˆ δ1 ∗ Φ((δ1 ni + x1i κ1 )/ˆ ). The amount of contribution to children’s college education increases with wealth.5 show that increases in permanent and transitory income increase the level of expenditures for educational expenses.436.789 per child. to calculate the expected college expenditures (COLLEXP ). Φ is the standard normal cumuˆ ˆ lative distribution. Households citing saving for retirement and buying a home as the most important reason for saving spend less on children’s education.
Explanatory variables include age (AGE).000 between 1893-86. the coeﬃcient of expected college expenditure (COLLEXP ) is negative and the coeﬃcient of age interaction term (AGE×COLLEXP ) is positive. Interestingly. The estimates of SAVE1 and SAVE2 are very similar. gender (FEMALE) of the household head. Table 4. In estimates of both equations. households in top 25 percentile of the wealth distribution save almost ﬁve times more if they have a child in college. However. the data in Table 4. a dummy indicating whether or not the household had a windfall greater than $3. Households in the bottom 25 percentile of wealth distribution save signiﬁcantly less than those without children in college. a dummy indicating whether the household head is willing to take risky investments (RISKY). indicating that an increase in expected college expenditure raises 91 .and without children in college.3. permanent and transitory incomes (PERINC and TRINC). which are retirement and emergencies (SRETIRE and SEMERG).6 show that wealthier families contribute more to their children’s education and continue to save while their children are in college. households with greater wealth save more if they have a child attending college. nonurban residence (NSMSA) and dummies indicating household net worth in 1983 (NWORTH25 and NWORTH75). and the number of children attending college between 1983-86 (NCHCOLL). two other reasons for saving. Columns 1 and 2 contain the estimates for SAVE1.7 presents the eﬀect of expected college expenditures on household savings. and columns 3 and 4 contain the estimates for SAVE2. Similar to Table 4.
I mean a household in the 25-75 percentile of the wealth distribution. had it not expected to contribute a positive amount.493 more and households in the top 25 percentile save $11. Figure 4. Households citing saving for retirement as the most important reason save more.398 less than those in the middle of the wealth distribution . the number of children attending college does not signiﬁcantly decrease household savings. The saving behavior of a household with a child in the ﬁrst year of college in 1983 can be quite diﬀerent from a household with a child ﬁnishing up college in 1983. Finally.1 shows the eﬀect of the age of the household head on SAVE1. Also. showing that households save approximately 39 percent of their transitory income.833 more than other households.savings after age 28. household savings are calculated in ﬁve year intervals. the data does not have detailed information on the years that children were attending college between 1983 and 1986. households with heads who are willing to undertake risky investments save $7. By typical.7. The eﬀect of transitory income on both measures of savings is positive and signiﬁcant. the ﬁgure ﬁrst calculates savings of a typical household expecting to contribute $2. Permanent income increases both SAVE1 and SAVE2. However. households in the bottom 25 percentile of the wealth distribution save $2. saving for emergencies does not signiﬁcantly aﬀect savings. citing a motive other than retirement or emergencies as the most 92 . Unfortunately. Finally.000 to college expenses and compares it to what it would have saved. This result does not necessarily mean that households are not saving for children’s college education. For each age group. Using the estimates in Table 4.
I introduce life-cycle savings into the quality and quantity model of fertility and derive predictions concerning the eﬀect of educational expenditures on household savings. this will increase its saving by $4. who is not willing to undertake risky investments and did not receive a windfall greater than $3. This striking result is due to the assumption that this household is assumed not to cite saving for retirement as the most important reason. that this household starts saving for retirement when the household head is 43 years old. This ﬁgure only shows that controlling for other factors. and it increases with the age of the household head. If we assume. 4. The results in Table 4. If the household does not expect to contribute to children’s college expenses.7 show that the eﬀect of saving for retirement on household saving is positive and raises household savings by $4. The results show that savings of the household with an anticipated $2. the eﬀect of anticipated college expenses on savings is positive and signiﬁcant. The eﬀect of expecting to contribute $2000 on household savings is $8.000 college expenses increase with age.000 between 1983-86.000 at the age of 43. The household is assumed to have average permanent and transitory incomes for their age group. savings decline to zero at the age of 43. for example.894. I also obtain predictions 93 . Saving motives change with age and household composition.894. headed by a male.6 Conclusion This chapter examines the eﬀect of saving for children’s college edu- cation on household savings.important saving motive.
this present chapter examines the eﬀect of college expenditures over the life-cycle and ﬁnds that most of the saving done by wealthier households can be attributed to saving to ﬁnance their children’s college expenses. The results are consistent with the predictions of the life-cycle theory of saving and consumption that households save in advance for expected expenses.000 at the age of 43. I analyze the eﬀect of educational expenditures on two diﬀerent measures of savings. The results are also consistent with the ﬁndings in Souleles . households smooth consumption into the academic year and do not cut consumption in the 6-9 months before the academic year starts. By focusing on household savings. Using the actual college expenditures reported in the 1983-86 Survey of Consumer Finances. The amount of savings for college expenses increases with the age of the household head. which are the change in net worth between 1983 and 1986.concerning the simultaneous determination of family size and college expenditure per child. The main ﬁnding of this chapter is that households save in advance for children’s college expenditures. and the change in net worth excluding the capital gains on primary residence. Other things constant. The model uses the expected expenditures and other control variables that aﬀect savings to estimate an equation of savings. which show that despite large college expenses. I estimate expected expenditures on children’s college education. 94 . the diﬀerence between savings of households with and without college expenses can be as high as $8.
341 0. Notes: This table reports the proportion of households citing the selected motives as the most important reason for saving.391 0.002 0.369 0.102 0.068 0.023 0.064 0.049 0.042 0.029 0.404 0. SEMERGE: saving for ‘rainy days.000 0.1: Saving Motives By the Number of Children SRETIRE 0.355 0.063 0.087 0.027 0.058 0.057 0.057 0.000 0.023 HH 0.016 0.074 0.380 NOSAVE 0.339 0.310 0.066 0. 1983.455 0.351 0.345 SEMERG SCHEDU 0.043 0.330 0.053 0.007 0. .032 25-75p 75 to 100p Source: Survey of Consumer Finances.337 0.341 0.401 0.035 SOTHER 0.043 0.381 0.060 0.223 0.120 0.083 0.377 0.397 0.279 0.082 0. SRETIRE: saving for retirement.543 0.278 0.153 CHILD 0 1-2 3 or more 0.053 SHOME 0.009 0.382 0.244 0.004 0. SHOME: saving to buy a home.049 0. SOTHER: saving for other reasons and NOSAVE: cannot/does not save.115 0.079 0. Tabulations are weighted using the sample weights.383 0. The number of observations N=1690.’ SCHEDU: saving for the education of children.212 0.010 0.042 0.Table 4.471 0.027 0.009 0.440 0.000 0.033 0.070 0.013 0.016 95 0.469 0.264 0.037 NWORTH 0-25p 0 1-2 3 or more 0 1-2 3 or more 0 1-2 3 or more 0.021 0.
48 BEDUCAT 0.35 WINDF 0.24 0.47 2.65 0.08 0.41 33402. Notes: Tabulations are weighted using sample weights.1. 96 .10 FEMALE 0. Deviation CHILD 2.17 0.86 PERINC 25931.75 TRINC 3296.14 COLLEXP> 0 2005.11 12491. All variables are described in Appendix C.13 0.37 NSMSA 0.36 SAVE1 5806.43 HIGHSCH 0.45 Source: Survey of Consumer Finances.43 14.33 MARRIED 0.53 0.Table 4.28 0. 1983-86.13 35397.68 AGE 42.50 BLACK 0.86 0. All dollar values are reported in 1986 dollars.2: Descriptive Summary of Variables Variables Mean Std.28 RISKY 0.32 15127. The number of observations N=1690.08 161860.47 SAVE2 4811.43 COLLEG 0.64 NWORTH 81575.24 0.32 2817.
All dollar values are reported in 1986 dollars. 97 . 1983-86.Table 4.3: Savings and College Expenses by the Number of Children in College SAVE1 SAVE2 COLLEXP NCHCOLL 0 1 2 or more NWORTH 0-25p 5041 6661 12357 4206 5277 10317 0 2236 1657 0 1 2 or more 0 1 2 or more 0 1 2 or more 3762 4961 1829 5804 6032 7707 4870 7745 18077 3705 5005 1695 4348 5305 5695 4551 5312 15577 0 905 797 0 1882 951 0 2937 2355 25-75p 75-100p Source: Survey of Consumer Finances. The number of observations N=1690. Tabulations are weighted using sample weights. Notes: NCHCOLL shows the number of children attending college between 1983-86.
98 .062 ** 1690 2.022 0. Notes: ** indicates signiﬁcance at 5 percent level.128 ** 0.389 0.134 ** 0.43 -2971.293 Source: Survey of Consumer Finances.145 ** 0.032 0. 1983-86.106 0.038 0. Variables are described in Appendix C.408 0.078 0.062 ** 1.717 0.48 0.126 ** -0.Table 4.097 ** 0.057 ** -0. Error -1.461 0. and * indicates signiﬁcance at 10 percent level.004 ** -0.435 0.038 0.039 ** -0.930 0.4: Poisson Regression: Number of Children CONSTANT AGE FEMALE FSPOUSE HIGHSCH COLLEG BLACK MARRIED NONSMSA PERINC/1000 HIGHSCHSP COLLEGSP N OBS Mean of dependent variable Log L R2 Coeﬃcient Std.964 0.090 ** -0.1.
7 ** 1391.8 ** -1103.Table 4.5: Tobit Estimates of College Expenditure Equation Coeﬃcient Std.9 ** -4903.0 9.0 ** 72.7 1467.5 18.4 1589.5 22.1 611.8 -372.6 -407. Error -601.4 23.2 2524.8 232.6 133.9 2912.9 ** -187.1 72. 99 .3 134.7 * 997.1 -4936.3 2444.9 -458.1 1063.6 -1124.0 ** 2905.5 472.0 409.2 19. Error Coeﬃcient Std.7 338 .9 -45.2 88.9 557.5 551.81 CONSTANT AGE CHILD CHILD PERINC/1000 BEDUCAT SRETIRE SEMERG SCHEDUC SHOME COLLEG TRINC/1000 SIGMA N OBS Proportion of + observations Log L ** ** * ** ** * ** ** -2429.6 408.0 0.5 -1335.4 92.8 468. All variables are described in Appendix C.4 555.8 -7.70 Source: Survey of Consumer Finances.746 -2429.8 473. and * indicates signiﬁcance at 10 percent level. 1983-86. Notes: ** indicates signiﬁcance at 5 percent level.9 ** 67.6 1452.1.3 609.1 550.4 10.
6: College Expenditures and Savings by the Number of Children in College NWORTH CHCOLL 0-25p 0 1 25-75p 0 1 75-100p 0 1 %HH COLLEXP 23. Notes: Tabulations are weighted using sample weights. The number of observations N=1690.52 0 5.29 1436 44.02 0 7. 100 . 1983-86.65 0 1.59 2334 17. CHCOLL=1 if the household has a child attending college between 1983-86 (0 otherwise).93 3093 COLLEXP 1278 1445 1614 1960 2219 3064 SAVE1 SAVE2 3800 3732 2802 2989 5570 4256 8864 6712 3731 3599 15904 12709 Source: Survey of Consumer Finances.Table 4. All dollar values are reported in 1986 dollars.
8 ** 6820.2 5041.2 1820.8 4.9 139.3 0.0 1271.5 0.6 407.7: Eﬀect of Anticipated College Expenses on Savings ** ** ** ** ** ** ** * ** ** ** 101 CONSTANT AGE AGE2 PERINC/1000 TRINC/1000 a COLLEGEXP a AGE×COLLEGEXP SRETIRE SEMERG NWORTH25 NWORTH75 WINDF RISKY FEMALE NSMSA NCHCOLL R2 SAVE1 Coeﬃcient Std. All variables are described in Appendix C.5 -1310.0 7833.0 382.9 1728.1 4894. Error 22618.2 3.107 SAVE2 Coeﬃcient Std.8 2783.1. Notes: ** indicates signiﬁcance at 5 percent level.5 2493.8 140.3 2954.6 10944.3 0.1 2796.1 3139.6 ** -9329.5 2409.0 1642.0 5607. a Predicted value of the variable from Tobit regression of educational expenditures.2 3422.3 416. .6 2026.084 Source: Survey of Consumer Finances.1 4.0 2460.6 388. and * indicates signiﬁcance at 10 percent level.8 3349.3 2016.2 ** -1284.6 -315.7 ** 0.Table 4. 1983-86.8 .3 * 3028.6 1516.5 1325.3 ** 309.0 -790.2 ** -9.5 .2 3.4 2538.4 2434.3 10.8 160.3 ** 10. Error ** 22327.2 -10.9 11080.0 -11398.9 2474.8 761.8 163.4 ** 336.1 * 4679.8 382.8 3199.3 -774.8 -615.
068 − | 28 | 33 | 38 | 43 • | 48 | 53 Age • • −5.1: The Importance of Educational Expenses on Savings SAVE16 10. • savings of a household with no college expenses. 736 − • • • | 23 −1. 045 − • 3. 791 − ? savings of a household with $2000 college expenses. 102 .Figure 4.
Appendices 103 .
households are assumed to claim standard deductions instead of itemizing deductions. tax-exempt interest. The SCF does not contain information on some possible deductions such as medical expenses.1 Estimating Marginal Tax Rates The marginal tax rate of each household is computed using the tax Form 1040 and the information on sources of income. business income and farm income. state and local income taxes. and age of the household head and the spouse. dividends. Subtracting the standard deduction and exemptions from the AGI 104 . In determining ﬁling status and personal exemptions. I use the information on marital status. royalties. Thus. All married couples are assumed to ﬁle a joint return.Appendix A Appendix for Chapter 2 A. number of dependents. rents. The SCF collects information on many components of total income. including wage and salaries. taxable interest. alimony received. The sum of household income from all sources gives the adjusted gross income (AGI). Components of income such as other gains and IRA distributions that are not reported in the SCF are set to zero. job expenses and moving expenses.
CHILD0 CHILD1 CHILD2 CHILD3 =1 if no children are living in the household. Consumption demand for housing. Marginal tax rate of the household. I then apply the appropriate tax rate schedule to calculate the household’s tax liability.yields the taxable income.3. A. The marginal tax rate is computed by running this method twice . =1 if three or more children are living in the household. See Appendix A. INCOME ASSET MTR Eh Total assets of the household.1. 105 . =1 if only one child is living in the household. it is the opportunity cost of owning a house. =1 if the household head is a single female. =1 if two children are living in the household. Number of children younger than age 22 who live in the household. =1 if the household head is married.2 Name Deﬁnition of Variables Description Estimated earnings of the household head and spouse at the age of 45. See Appendix A. AGE MARRIED FEMALE NCHILD Age of the household head in years. The diﬀerence in total tax liabilities divided by 100 gives the marginal tax rate.once with AGI and then with AGI minus 100. For homeowners.
εpi is an unobservable variable measuring characteristics such as ability (εpi 2 has zero mean and variance of σs ). =1 if the household head is white. =1 the household is a homeowner. and c(AGEi ) is a cohort eﬀect. =1 if the household head reports that he is willing to take risky investments.3 Estimating Permanent Income The measure of permanent income is constructed using the method outlined in King and Dicks-Mireaux . A. and the second is transitory changes in earnings. This measure is deﬁned as predicted earnings at the age of 45 plus an individual-speciﬁc eﬀect. The ﬁrst is due to the movements along the age-earnings proﬁle over the life cycle. (A. =1 if the household is included in the 1998 survey. =1 if the household is included in the 1995 survey. βp is the parameter vector.1) where Zi is a vector of observable characteristics. Thus.CHAGE13 HOMEOWN WHITE RISKY =1 if the youngest child is older than age 13. earnings 106 . The permanent income Y for individual i is deﬁned as Ln Yi = Zi βp + εpi − c(AGEi ). YEAR92 YEAR95 YEAR98 =1 if the household is included in the 1992 survey. Observed earnings are assumed to diﬀer from permanent income in two ways.
(A. c(AGEi ). and is assumed to be uncorrelated with εpi ).1) and (A. I need the estimates of βp .2) and estimate the resulting earnings equation using each wave of SCF separately.) measures the log of the age-earnings proﬁle. AGEit is the age of the respondent and uit is the log of the transitory component (uit has zero mean 2 and variance of σu . The same procedure is used for spouses. Since age-earnings proﬁle e(AGEit − 45) and c(AGEi ) cannot be identiﬁed for this estimation. I assume that α = 0. to get an estimate of εpi .3) 2 2 2 where α = σs /(σs + σu ). permanent income is calculated from Zi βp . and this provides the ˆ estimate βp . I calculate the minimum variance estimator of εpi using εpi = α(εpi + uit ). For heads with zero ˆ earnings.2) where e(. King and Dicks-Mireaux  use outside data to impose a cohort eﬀect. ˆ is zero.in year t are Ln Eit = Ln Yi + e(AGEit − 45) + uit . To construct an estimate of permanent income. I assume that the cohort eﬀect. Instead. ˆ (A. with one exception. εpi and c(AGEi ). Following King and Dicks-Mireaux . Finally. The selectivity-adjusted earnings functions are estimated for the sample consisting of individuals with nonzero earnings. Earnings equations are estimated separately for household heads and spouses.5. I combine (A. Their permanent income is adjusted for 107 .
108 . ˆ where Yi is the permanent income estimate.non-participation at diﬀerent stages of the life cycle as follows: ˆ Yiw = Yi P rob(Ei > 0). and the probability of nonzero earnings is computed for each spouse from the probit estimates. Household permanent income is the sum of the estimates of permanent income for the head and spouse.
VRI MEANINC PERMINC TRANS NWORTH NWORTH25 Variance of log income. VRLI Variance of residual of log income from the earnings equation. AGE Age of the household head in 1983. Mean of the residual income from the earnings regression. NWORTH90 =1 if the household in the 10 percent of the net worth distribution in 1983. Mean of reported income between 1982-88. Mean of the predicted income from the earnings regression. 109 . SAVE2 First diﬀerence in net worth between 1983-89 controlling for capital gains in home prices divided by 6. Household net worth in 1983.Appendix B Appendix for Chapter 3 B. =1 if the household in the bottom 25 percent of the net worth distribution in 1983.1 Name SAVE1 Deﬁnition of Variables Description First diﬀerence in net worth between 1983-89 divided by 6.
Number of children between age 0-6 in 1983. =1 if the household owns a home in 1983. =1 if the spouse is working fulltime in 1983. =1 if the household head is white. 110 . =1 if the household had an additional child between 1983-89.EDUC WHITE MARRIED NCHILD YOUNGCH MIDDCH HIGHSCH ∆CHILD Years of education of the household head in 1983. Number of children between age 13-18 in 1983. Change in the number of adults between 1983-89. Number of children between age 7-12 in 1983. Number of children living in the household in 1983. =1 if the household owns a home in 1989. =1 if the household head is planning to leave a bequest. =1 if the household head is married. NADULT ∆NADULT BEQUEST HOWN83 HOWN89 SPFULLT Number of adults living in the household in 1983.
PERINC TRINC SAVE1 SAVE2 Predicted 1985 household income. =1 if the household head has a high school degree. Amount of expenditure on the college education of a child in 1986 dollars. 111 . =1 if the household head is female. NWORTH NWORTH25 Net worth in 1986.1 Name CHILD COLLEXP Deﬁnition of Variables Description Number of children ever born to the household head. Diﬀerence between net worth in 1986 excluding the capital gains on primary residence and net worth in 1983 divided by 3.Appendix C Appendix for Chapter 4 C. Diﬀerence between net worth in 1986 and 1983 divided by 3. AGE FEMALE HIGHSCH Age of the household head in 1983. =1 if the household is in the bottom 25 percentile of the wealth distribution. NWORTH75 =1 if the household is in the top 25 percentile of the wealth distribution. Diﬀerence between total income in 1985 and permanent income.
Number of children attending college between 1983-86. BEDUCAT =1 if the household head thinks it is all right to borrow for education. =1 if the spouse is working at a full-time job. 112 . =1 if the household head is married. =1 if the household head is willing to undertake risky investments.COLLEG =1 if the household head has a college degree. SRETIRE SEMERGE SCHEDU =1 if retirement is the most important reason for saving. =1 if emergencies are the most important reason for saving. SHOME =1 if saving to buy a home is the most important reason for saving. COLLEGSP FSPOUSE BLACK MARRIED RISKY =1 if the spouse has a college degree. =1 if children’s education is the most important reason for saving. HIGHSCHSP =1 if the spouse has a high school degree.000 between 1983-86. NSMSA NCHCOLL =1 if the place of residence is not in a SMSA. =1 if the household head is African-American. SOTHER =1 if the household cited another reason as the most important reason to save. WINDF =1 if the household received a windfall greater than $3.
Family re- sources. theories and micro facts. Bertaut. 113 . Robert A.  Martin Browning and Annamaria Lusardi. Pollak.  Carol C. 1998. taxes.  Gary S. Review of Economics and Statistics. Becker. households: Evidence from the 1983-1989 Survey of Consumer Finances. Lewis. Cambridge: Harvard University Press. A Treatise On the Family. Journal of Political Economy.  James Berkovec and Don Fullerton. 1992. housing. 100(2):390–429. 97(2):398–419. 34(4):1796– 1855. Journal of Political Economy. On the interaction between quantity and quality of children. Behrman. Stockholding behavior of U. 1989. 1996. 1991. 1973.  Jere R. Household saving: Micro A general equilibrium model of Journal of Political Economy. family size.S. 80(2):263–275. 81(2):S279–S288. and Paul Taubman.Bibliography  Gary S. and portfolio choice. and access to ﬁnancing for college education. Journal of Economic Literature. Becker and Gregg H.
 Christopher D. Samwick. 1994.  Ngina S. 1997. Jan K. Staﬀord. 89(2):377–380. Consumption and investment motives and portfolio choice of homeowners. 80(3):410–19. 1997. Technical Report 15. Krane. 1998. 1999. 41(1):41–71.  Christopher D. 15(2):159–180. How important is precautionary saving? Review of Economics and Statistics.  Christopher D. Carroll. Carroll. Carroll and Andrew A. 1997. Board of Governors of the Federal Reserve System. 109(1):111–148. The nature of precautionary wealth. Carroll and Andrew A. Samwick. How does future income aﬀect current consumption? Quarterly Journal of Economics. 114 . Finance and Economics Discussion. Chiteji and Frank P.  Christopher D. and Spencer D. American Economic Review. Carroll.  Christopher D. Journal of Monetary Economics. Karen D. Portfolio choices of parents and their children as young adults: Asset accumulation by African-American families. Unemployment risk and precautionary wealth: Evidence from households’ balance sheet. Quarterly Journal of Economics. 1999. 112(1):1–56. Journal of Real Estate Finance and Economics. Buﬀer stock saving and the life-cycle/permanent income hypothesis. Brueckner. Dynan.
Edlin. Understanding Consumption. Choy and Henke R. Miller. The Theory of Finance. Deaton.  Karen Dynan.S. 1992.  Wendy E. National Center for Education Statistics. The implicit taxes from college ﬁnancial aid. Technical Report 48. Susan P. IL. Owner-occupied housing and the composition of the household portfolio over the life cycle. Washington.  Marjorie Falvin and Takshi Yamashita. Dunn. University of California. F. Board of Governors of the Federal Reserve System. 115 . U. Journal of Public Economics. 1993. Dryden Press. 65(3):295–322. 1998. Robin. 1993.  Eugene.  Angus S. Hinsdale.  Aaron Edlin. 1997. Fama and Merton H. Technical Report 1998. and Durable Goods Purchase Decisions. San Diego. Unemployment Risk. Dick and Aaron S. Parental ﬁnancial support for undergraduate education. 7(2):143–158. New York: Oxford University Press. Finance and Economics Discussion.  Andrew W. Technical report. How prudent are consumers? Journal of Political Economy. 101(6):1104–13. 1972. Department of Education. 1992. Precautionary Saving. Is College Fiancial Aid Equitable and Eﬃcient? Journal of Economic Perspectives.
116 .  Donald R. 20(1):83–102. 12(2):214–237. 107(442):537–552.  William G. Ioannides.  J. College scholarship rules and private saving. Why do people save? A micro-analysis of motives for household saving in Japan. Regional Science and Urban Economics. American Economic Review. Journal of Urban Economics. 1994.  Allen C.  Allen C. Hendershott. 1983. Gale and John Karl Scholz. Housing decisions of American youth. Economic Journal. Permanent income. Vernon Henderson and Yannis M. American Economic Review. 1982. 21(2):228– 241. 1995. and demand for housing: New evidence. Haurin. Journal of Economic Perspectives. 35:28–45. A model of housing Journal of Urban Economics. Demographics of individual housing demand. 1990. Vernon Henderson and Yannis M. Ioannides. 1997.  J. tenure choice. hedonic prices. Goodman. and Dongwook Kim. Goodman and Masahiro Kawai. 85(3):552–66.  Charles Yuji Horioka and Wako Watanabe. Patric H. 1994. Intergenerational transfers and the accumulation of wealth. Journal of Urban Economics. 73(1):98–113. Martin Feldstein. Owner occupancy: Investment vs consumption demand. 1987. 8(4):145– 160.
Miller. Joseph Hotz and Robert A.  Richard Deitz John Robst and KimMarie McGoldrick. Glenn Hubbard. 29(2):219–229. 1999. Jacob Alex Klerman.Willis. 36(4):620–630. An empirical investigation of alternative approches to estimating the equilibrium demand for housing. 1998. 103(21):360–399.  Arthur B. 1988. Applied Finance. and Robert J.  Keith Ray Ihlanfeldt.  Yannis M. 1981. Joseph Hotz. Income variability. 1997. chapter The Economics of Fertility in Developed Countries. Are women more risk averse? Economic Inquiry. An empirical analysis of life cycle fertility and female labor supply. uncertainty and housing tenure choice. Journal of Political Economy. Economics. Zeldes. Pre- cautionary saving and social insurance.  V. 117 Journal of Urban . pages 275–347. Jonathan Skinner. Household saving and portfolio change: Evidence from the 1983-89 SCF panel. 56(1):91–118.  R. Dynamics of the composition of household asset portfolios and the life cycle. V. Jianakoplas and Alexandra Bernasek. Regional Science and Urban Economics. Handbook of Population and Family Economics. 1997.  Nancy A. Review of Income and Wealth. Econometrica. 1992. 1995. 43(4):381–399. Kennickell and Martha Starr-McCluer. 9(1):97–105. Ioannides. and Stephen P. 2(3):145–159. Amsterdam: Elsevier.
Leape. 1978. 92(366):247–267. Econometrics.  Miles S. American Economic Review. life-cycle.  Mark Long. 88(2):449–453. 1997. 1982. On the importance of the precautionary saving Journal of motive. Economic Journal. 118 .S. 69(2):155–193.  G. Dicks-Mireaux. Arthur B.  Mervyn A.  Mervyn A. Maddala. 1998. Asset holdings and the Journal of Business and Economic Statistics. Retrospective reporting of household wealth: Evidence from the 1983-89 Survey of Consumer Finances. 58(1):53–73. Limited Dependent and Qualitative Variables in Econometrics. L. Trost. 1990.  Lung-Fei Lee and Robert P. 15(3):452–63. 8:357–382. Kimball.  Annamaria Lusardi. Journal of Public Economics. The impact of asset-tested college ﬁnancial aid on household savings. King and L-D. Kennickell and Martha Starr-McCluer. Econometrica. King and Jonathan I. 1998. Estimation of some limited dependent variable models with the application to housing demand. 2001. 1983. Cambridge University Press. Precautionary saving in the small and in the large. Wealth and portfolio composition: Theory and evidence.
The family. Acquiring capital for college: The constraints of family conﬁguration. Journal of Monetary Economics. Clery. 1981. Department of Education. 1997. Gender diﬀerences in the alloe cation of assets in retirement savings plan. 1996. 22(2):237–55. American Economic Review. U. Souleles. Poterba and Andrew A. Surette. Journal of Public Economics. Middle income undergraduates: Where they enroll and how they pay for their education. life cycle consumption and precautionary savings.  Martha Starr-McCluer.S. 89(5):928–958. Household portfolio allocation and over the life cycle.  Nicholas S. 1989. National Bureau of Economic Research. 119 .  Lala Carr Steelman and Brian Powell. College tuition and household savings and consumption. Journal of Political Economy. National Center for Education Statistics. 2000. American Sociological Review. 54:844–855.  Annika E.  Nigel Tomes. Technical Report 6185. Sund´n and Brian J.  Jonathan Skinner. Risky income. Health insurance and precautionary savings. James M. Washington. Technical report.  Jennifer B. 1988. American Economic Review. Presley and Suzanne B. 1998. Samwick. 86(1):285–295. 2001. 77:185–207. 88(2):207–211. inheritance and the intergenerational transmission of inequality.
 Robert J. An estimable dynamic stochastic model of fertility and child mortality. Journal of Economic Perspectives. Quarterly Journal of Economics. Willis. 1989. Optimal consumption with stochastic income: Deviations from certanity equivalence. Wolﬀ. A new approach to the economic theory of fertility behavior. Wolpin. 19(2):209–234. 1984. 120 . Journal of Political Economy. 1973. A model of fertility and children’s schooling. 1998. Journal of Political Economy. Nigel Tomes. 104(2):275–98. Recent trends in the size distribution of household wealth. Zeldes. 92(5):852–874. Economic Inquiry.  Edward N.  Stephen P. 1981.  Kenneth I. 12(3):131–150. 81(2):S14–S64.
121 . where she received a g c Master of Arts degree in Economics in June 1997. Permanent address: 834 Main Street Apt. Eﬀective August 2002. IN 47901 This dissertation was typed by the author. she accepted an assistant professor position at Purdue University. the daugh¨ ter of Onder Yilmazer and Necla Yilmazer. g c She began her graduate studies at Boˆazi¸i University. 1970. B Lafayette. She later continued her education at the University of Texas at Austin. She received her Bachelor of Arts degree in Business Administration from Boˆazi¸i University in January 1994. Turkey on June 2.Vita Tansel Yilmazer was born in Izmir.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.