The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Working Papers Series Wealth Inequality and Intergenerational Links By: Mariacristina De Nardi Working Papers Series Research Department WP 99-13 Wealth Inequality and Intergenerational Links Mariacristina De Nardi First Draft: January 1998 This draft: February 2000 Abstract Empirical studies have shown that, for many countries, the distribution of wealth is much more concentrated than the one of labor earnings and that households with higher levels of lifetime income have higher lifetime saving rates. Previous models have had diÆculty in generating these features. I construct a computable general equilibrium model with overlapping generations in which parents and children are linked by bequests and earnings persistence within families. I show that voluntary bequests are important to explain the emergence of large estates that characterize the top of the wealth distribution, while accidental bequests are not. In addition, the introduction of a bequest motive generates lifetime saving pro les more consistent with the data. Allowing for earnings persistence within families generates an even more concentrated wealth distribution. A cross-country comparison between the U.S. and Sweden shows that intergenerational linkages are important to explain the upper tail of the wealth distribution also in economies where redistribution programs are more prominent and there is less inequality. Moreover Sweden, with its generous social safety net, has a larger fraction of people with zero or negative wealth. The model is capable of reproducing this feature as well. University of Chicago and Federal Reserve Bank of Chicago. Comments Welcome. I am grateful to Gary S. Becker, Lars P. Hansen, Jose A. Scheinkman, Nancy L. Stokey, and especially Thomas J. Sargent for helpful comments. My work bene ted from many conversations with Marco Bassetto, as well as from his constant support. I also pro ted from discussions with Lisa Barrow, Martin Floden, Alex Monge, Guglielmo Weber, Chao Wei and especially Marco Cagetti. I am thankful to Paul Klein and David Domeij for discussions about the Swedish data. Neither the Federal Reserve Bank of Chicago nor the Federal Reserve System are responsible for the views expressed in this paper. All errors are my own. Email address: nardi@bali.frbchi.org 1 1 Introduction Empirical studies (e.g. Hurst, Luoh and Sta ord [20]; Wol [39]; Lillard and Willis [28]; DazGimenez, Quadrini and Ros{Rull [10]) have shown that labor earnings, income and wealth are signi cantly concentrated, with distributions skewed to the right. However, wealth is the most concentrated of the three variables with a Gini coeÆcient of .72, earnings rank second with a Gini coeÆcient of .46, and income is the most disperse of the three with a Gini coeÆcient of .44. While these empirical regularities are observed in many countries, previous models do not provide a satisfactory explanation of how the observed earnings distribution leads to the observed distribution of wealth (Quadrini and Ros{Rull [35]). In the data, a signi cant fraction of the dispersion in earnings, income and wealth across households is attributable to their di erent positions in the life cycle. Moreover, in the aggregate, the intergenerational transmission of wealth is substantial. Kotliko and Summers [24] calculate that the majority of the current value of the U.S. capital stock (at least 80%) can be attributed to intergenerational transfers rather than to accumulation out of earnings, which is the emphasis of the basic life-cycle model of capital accumulation. Gale and Scholz [12] use direct measures of intergenerational ows and attribute 63% of the current value of the U.S. capital stock to intergenerational transfers. Mulligan [29] shows that intergenerational links are essential to explain the emergence of very large estates. Other empirical work (e.g. Dynan, Skinner and Zeldes [11] and Lillard and Karoly [27]) highlights the fact that households with higher levels of lifetime income have higher saving rates. Carroll [8] shows that it is diÆcult to explain the behavior of these consumers using either a standard life-cycle model or a dynastic model. The goal of this research is to study how wealth is accumulated in a life-cycle economy with intergenerational links and how the characteristics of the accumulation process in uence the distribution of wealth, given the distribution of labor earnings. I consider an incomplete markets life-cycle model with earnings uncertainty, life-span risk, and links between parents and children: the parents care about leaving bequests to their o spring, and the children partially inherit their parents' productivity. In this setup households have several motives to save: to self-insure against income and life-span risk, for retirement and possibly to leave a bequest to their children. The characteristics of the accumulation process impact the life-cycle pattern of wealth accumulation, the dispersion of wealth within cohorts and the overall wealth distribution. I construct a computable general equilibrium model to study how these di erent saving motives help in understanding the dispersion of wealth across households, both in the U.S. and Swedish economies, taking as given the distributions of labor earnings. I also calibrate the model to the Swedish economy to investigate whether intergenerational linkages are important even in economies where redistribution programs are more prominent and there is less inequality. My results show that voluntary bequests are important to explain the emergence of large estates that are usually accumulated in more than one generation and that characterize the upper tail of the wealth distribution in the data, while accidental bequests do not generate more wealth concentration. I also show that the introduction of a bequest motive generates lifetime saving pro les more consistent with the data. Saving over the life cycle is the primary factor in understanding how wealth is accumulated at the lower tail of the distribution, while intergenerational 2 links signi cantly a ect the shape of the upper tail. Moreover, the introduction of a humancapital link in which the children partially inherit the productivity of their parents can generate a yet more concentrated wealth distribution. In this case more productive parents accumulate larger estates and leave larger bequests to their children who, in turn, are more successful than average in the workplace. The cross-country comparison between the U.S. and Sweden shows that intergenerational linkages are also important in economies where redistribution programs are more prominent and there is less inequality. The calibration to the Swedish data also reveals that the model reproduces well the fact that Sweden, with its more generous social insurance net, has a larger fraction of people with zero or negative wealth. Section 2 reviews some related literature, section 3 brie y discusses the main features of the U.S. and Swedish data, and section 4 describes the model. Section 5 is a road map of the experiments that I run in order to understand the quantitative importance of each intergenerational link. Sections 6 and 7 describe the calibration and the results of the various experiments for the U.S. and the Swedish economies respectively. Section 8 discusses other factors that may be important to explain the distribution of wealth and explains how the assumptions I make are likely to a ect my results. Section 9 concludes and discusses some directions for future research. 2 The Literature Previous attempts at studying how the distribution of wealth is determined fall broadly into two categories. The rst group of papers studies overlapping-generations economies where all savings arise over the life cycle.1 The second group of paper studies economies with in nitely lived dynasties. Huggett [19] and Gokhale et al. [13] are the only papers within the rst group to focus primarily on the distribution of wealth. Gokhale et al. [13] aim at evaluating how much wealth inequality arises from inheritance inequality. To do so, they construct an overlapping-generations model and focus on intragenerational inequality of households whose head is age 66. Their model allows for random death, random fertility, assortative mating, heterogeneous human capital, progressive income taxation and social security. All of these elements are exogenous and calibrated to the data. The families are assumed not to care about their o spring, hence all bequests are involuntary. To solve the model, they impose that individuals are in nitely risk averse and that the rate of time preference equals the interest rate. As a consequence, the families in the model have a constant per capita consumption pro le, resulting in a large aggregate ow of bequests from people who die before reaching the maximum lifespan. Moreover, families do not take into account expected bequests when making consumption and saving decisions. Gokhale et al. nd that inheritances in the presence of social security play an important role in generating intra-generational wealth inequality in the cohort they consider. The intuition is that social security annuitizes completely the savings of poor and middle-income people but is a very small fraction of the wealth of richer people, who thus keep assets to insure against life-span risk. In this setup, were a market for 1 Cf. _ _ Imrohoro glu, Imrohoro glu and Joines [21], Hubbard, Skinner and Zeldes [18]. 3 annuities available, rich people would completely annuitize their wealth and no bequest would be left. In Huggett [19], the workers face uninsurable income shocks and uncertain life span. The government taxes bequests at 100% and redistributes them equally to all agents alive. As in most papers that address the distribution of wealth, the skewness is generated by the introduction of a borrowing constraint. The paper succeeds in matching the U.S. Gini coeÆcient for wealth, but the concentration is obtained by having more people with zero or negative wealth and a much thinner upper tail than observed in the actual distribution. The fact that people hit the borrowing constraint too often, leading to a large fraction of people at zero or negative wealth, is a common problem of models with idiosyncratic income shocks. In overlapping-generations models this problem is aggravated by the assumption that young agents are born without wealth and hence need time to accumulate precautionary savings to hedge against income shocks. The proportion of people at zero wealth is less of a problem for the second group of papers, which tries to explain the distribution of wealth in economies populated by in nitely lived dynasties. In this case the precautionary savings have already been accumulated in steady state, hence the borrowing constraints bind less often. These models disregard the fact that the lower tail of the distribution of wealth is mainly comprised of young and old households; they succeed in lowering the proportion of households at zero or negative wealth by treating all agents as if they were middle aged. As for the second group of papers, Krusell and Smith [25] study an economy populated by in nitely lived dynasties that face idiosyncratic income shocks. These dynasties also face a stochastic process for their discount factor and thus have heterogeneous preferences. The discount factor changes on average every generation and is meant to recover the fact that parents and children in the same dynasty may have di erent preferences. Krusell and Smith nd that allowing for di erent discount factors among agents helps in matching the cross-sectional wealth distribution. Casta~neda, Daz-Gimenez and Ros{Rull [9] consider a model of earnings and wealth inequality and use it to study the e ect of tax reforms. Their model economy is populated by dynastic households that have some life-cycle avor; workers have a constant probability of retiring at each period and once they are retired they face a constant probability of dying. They care about their o spring. These newborns enter the model as workers and inherit the family's after-tax capital; in equilibrium their utility is the same as that of old workers. The paper employs a large number of free parameters to match some features of the U.S. data that are considered particularly signi cant, which include measures of the wealth distribution. However, the simple structure of the model does not allow proper accounting for the life-cycle pattern of savings and the role of bequests in generating wealth inequality. Quadrini [34] constructs an in nitely-lived agent model in which agents at each period decide whether to be entrepreneurs or not. Three elements in the model are crucial. First, the existence of capital market imperfections induces workers that have entrepreneurial ideas to accumulate more wealth to reach minimal capital requirements. Second, in the presence of costly nancial intermediation, the interest rate on borrowing is higher than the return from saving, therefore an 4 entrepreneur whose net worth is negative faces a higher marginal return from saving and reducing his debt. Third, there is additional risk associated with being an entrepreneur, hence risk averse individuals will save more. As in Casta~neda, Daz-Gimenez and Ros{Rull [9] the model uses a large number of free parameters to match features of the earnings and wealth distribution. In a recent paper Heer [17] adopts a life-cycle setup in which parents care about leaving bequests to their children. In his framework the bequest motive does not a ect much the distribution of wealth. His results di er from mine because his income process is much less representative of the actual process faced by households and because he assumes that children can perfectly observe their parent's characteristics and wealth. In contrast with the papers that study economies with in nitely lived dynasties, I explicitly model the life-cycle structure, which contributes to a signi cant fraction of the dispersion in earnings, income and wealth across households. Compared to Huggett's [19] paper I add intergenerational transmission of wealth and ability. In contrast with Gokhale et al. [13], and consistent with the data, my model generates higher saving rates for people with higher lifetime income and age-savings pro le consistent with the empirical observations. Compared to Heer [17], my paper does a better job of modeling the earnings process and the bequest motive. It also explores the relevance of intergenerational transmission of ability and applies the model to two countries. Rather than focussing on the wealth distribution, Carroll [8] concentrates on the fact that in the data households with higher levels of lifetime income have higher lifetime saving rates (see Dynan, Skinner and Zeldes [11] and Lillard and Karoly [27]). He shows that neither standard life-cycle, nor dynastic models can recover the saving behavior of rich and poor families at the same time. To solve this puzzle he suggests a \capitalist spirit" model, in which nitely lived consumers have wealth in the utility function. This can be calibrated to make wealth a luxury good, thus rendering nonhomothetic preferences. In my model, nonhomotheticity arises because parents care about leaving bequests to their children (I calibrate this bequest motive taking into account the children's utility of receiving the bequest). This setup allows me to test whether the assumptions I make are consistent not only with the saving behavior of single individuals but also with the wealth distribution as a whole. 3 On the Empirical Facts 3.1 The U.S. Economy Capital Transfer Percentage wealth in the top Percent with output wealth Wealth negative or ratio ratio Gini 1% 5% 20% 40% 80% zero wealth 3.0 .63 .72 28 49 75 89 Table 1: U.S. wealth data. 5 99 5.8-15.0 In table 1 I present various statistics on wealth and wealth distribution in the U.S. The measure of the capital-output ratio depends on the concept of capital one has in mind. The most comprehensive notion of capital typically used includes residential structures, plant and equipment, land and consumer durables. This implies a capital output ratio of about 3 for the period 1959-1992 (Auerbach and Kotliko [3]). A narrower de nition of capital obviously implies a lower capital-output ratio. Stokey and Rebelo [37] exclude land, consumer durables and residential structures owned by the government and obtain a ratio below 2. The notion of transfer wealth as a fraction of total wealth is the one computed by Kotliko and Summers [24]. The authors distinguish two components of total wealth: transfers and life-cycle savings. They compute the transfer wealth component for a person alive at a given point in time as the current value of all non-government transfers received by that person. The current value is computed using the realized after-tax rates of return on wealth holdings. The life-cycle component is the residual one. They estimate that transfer wealth accounts for at least 80% of total wealth. A more recent study by Gale and Scholz [12] uses direct measures of intergenerational ows from the Survey of Consumer Finances and nds that intergenerational transfers account for about 63% of the current value of wealth. The data on the wealth distribution are from Wol [39]. His estimates are based on the 1983 Survey of Consumer Finances. Wealth is de ned as owner-occupied housing, other real estate, cash, nancial securities, unincorporated business equity, insurance, and pension cash surrender value, less mortgage and other debt. Wol makes adjustments to account for consumer durables, household inventories and underreporting of nancial assets and equities. Hurst, Luoh and Sta ord [20] provide data on the wealth distribution using di erent data sources for 1989. They piece together the 1989 PSID household wealth up to the 98.6 percentile and then use the IRS data for the 98.2 to the 99.6th percentile and the Forbes data for the balance of the 32 most wealthy families. They adopt this strategy because the PSID oversamples poor people (Juster, Smith and Sta ord [22] compare the PSID with the SCF data and nd that the PSID is accurate only up to the richest 5% of the population). The wealth measure obtained by Hurst, Luoh and Sta ord is very close to the one adopted by Wol , but they do not perform any adjustment. They nd similar numbers for the wealth distribution. Percentage earnings in the top Percent with Gini negative or coe . 1% 5% 10% 20% 40% 80% zero income .46 6 19 30 48 72 98 7.7 Table 2: U.S. data on gross earnings. Table 2 is computed using data from the Luxembourg Income Study (LIS) data set, which collects income data sets from di erent countries (it is based on the CPS for the U.S.) and makes them comparable. The table is computed using data for households whose head is 25 to 60 years of age, and the de nition of gross earnings includes wages, salaries and self-employment income. We can see from tables 1 and 2 that earnings display much lower concentration than wealth. 6 3.2 The Swedish Economy Capital Transfer Percentage wealth in the top Percent with output wealth Wealth negative or ratio ratio Gini 1% 5% 20% 40% 80% zero wealth 2.0 > :51 .73 17 37 75 99 100 30 Table 3: Swedish wealth data. Table 3 presents various statistics on wealth and wealth distribution as in the U.S. section. The capital/gdp ratio is computed analogously that of U.S. (see Hansson [16]). The wealth distribution, the Gini index and the number of people at zero or negative wealth refers to 1984-85 data (Palsson [32]). The inheritance-wealth ratio number is from Laitner and Ohlsson [26]. They compute the current value of household inheritances in Sweden as a fraction of household wealth; the result they obtain, conditional on the age of the household head varies from a low of .34 (for households 50-59) to a high of .85 (for households 70 and older). Their number for the economy as a whole is .51. However, this number does not include inter-vivos transfers, therefore the actual present value of intergenerational transfers to wealth is higher. Table 4 is also computed from the LIS data set, in the same way described for Table 2. Percentage earnings in the top Percent with Gini negative or coe . 1% 5% 20% 40% 80% zero income Gross earnings .40 4 15 42 68 98 7.6 Table 4: Swedish data on gross earnings. As we can see from tables 1-4, the Swedish distribution of earnings is more equally distributed than the U.S. distribution of earnings, but the Gini coeÆcient for wealth is the same in the two countries (.72 in the U.S. and .73 in Sweden). However, the high Gini coeÆcient for wealth in Sweden results from di erent reasons than in the U.S. In the U.S. the top 1-5% of people hold a large fraction of total wealth, while the fraction of people with zero or negative wealth is relatively small. On the contrary in Sweden the top 1-5% of people do not hold not as much of total wealth, but a much larger fraction of people is at zero or negative wealth. This may be due to the fact that social security and unemployment bene ts are more generous in Sweden than in the U.S., and these social insurance programs are a disincentive to save, especially for poorer people. In fact, individuals for whom social security bene ts are high compared to their life-time income will not save for retirement in presence of a redistributive social security system. Moreover if security nets (such as unemployment insurance) are substantial, precautionary savings will be lower. 7 4 The Model The economy is populated by overlapping generations of people and an in nitely lived government. The agents may di er in their productivity level. The members of successive generations are linked to one another by the altruism of the parents toward their children and the o spring's inheritance of part of their parent's productivity. At age 20 each person enters the model and starts consuming, working and paying labor and capital income taxes. At age 25 the consumer procreates and he cares about leaving bequests to his children when he dies. After retirement the agent no longer works but receives social security bene ts from the government and interest from accumulated assets. The government taxes labor earnings, capital income, and estates and pays pensions to the retirees. 4.1 Demographics During each model period, which is ve years long, a continuum of people2 is born. I assume that each person does not make saving or consumption decisions until he is 20, when he begins working. Thus, I model the agent's behavior starting from age 20 and de ne age t = 1 as 20 years old, age t = 2 as 25 years old, and so on. After one model period, at t = 2, the agent's children are born, and four periods later (when the agent is 45 years old) they are 20 and start working. Since there are no inter-vivos transfers in this model economy, all individuals start o their working life with no wealth. Total population grows at a constant, exogenous, rate (n), and each agent has the same number of children.3 The agents retire at t = tr = 9 (i.e., when they are 65 years old) and die for sure by the end of age T = 14 (i.e., before turning 90 years old). From t = tr 1 (i.e., 60 years of age) to T , each person faces a positive probability of dying given by (1 t ); since death is assumed to be certain after age T , T = 0. The assumption that people do not die before 60 years of age reduces computational time and does not in uence the results because the number of people dying between the ages of 20 and 60 is small. Since I consider only stationary environments, the variables are indexed only by age, t, and the index for time is left implicit. 4.2 Preferences and Technology Preferences are assumed to be time separable, with a constant discount factor. The utility from consumption in each period is given by u(ct ) = c1t =(1 ). Parents care about their sel sh children. The particular form of altruism I consider is called \warm glow": the parents' derive utility from leaving a bequest (net of estate taxes) to their o spring. The utility from leaving a net bequest bt , is (bt ). Considering a more sophisticated form of altruism would increase the number of state variables (already 4 in this setup) and, in some cases, would generate strategic parent/child interaction. 2 In the theoretical sections, I will use the terms \agent", \person", \consumer" and \household" interchangeably. Each household is taken to be composed of one person and dependent children. 3 The number of children is thus n5 if n is the growth rate of the population over 5 years, or n25 if n is expressed in yearly terms. 8 In this economy all agents face the same exogenous age-eÆciency pro le, t , during their working years. This pro le is estimated from the data and recovers the fact that productive ability changes over the life cycle. Workers also face stochastic shocks to their productivity level. These shocks are represented by a Markov process fyt g de ned on (Y; B(Y )) and characterized by a transition function Qy where Y <++ and B(Y ) is the Borel -algebra on Y . This Markov process is the same for all households. The total productivity of a worker of age t is given by the product of his stochastic productivity in that period and his deterministic eÆciency index at the same age: yt t . The parent's productivity shock at age 40 is transmitted to children at age 20 according to a transition function Qyh , de ned on (Y; B(Y )). What the children inherit is only their rst draw; from age 20 on, their productivity yt evolves stochastically according to Qy . An alternative possibility is to assume that agents face heterogeneous income processes (both ex-ante and ex-post heterogeneity as opposed to only ex-post heterogeneity) or \education levels" and that children partially inherit from their parents these di erent income processes. While this is a sensible extension, it would introduce an additional state variable and is left for future research. After retirement, the agents do not work any more but live o pensions and accumulated assets. I assume that children cannot observe directly their parent's assets, but only their parent's productivity when the parent is 40 and the o spring are 15, i.e., the period before they \leave the house" and start working. Based on this information, they infer the size of the bequest they are likely to receive.4 I will discuss the relevance and the qualitative e ects of relaxing all of these assumptions in section 7. The household can only invest in physical capital, at a rate of return r. The depreciation rate is Æ , so the gross-of-depreciation rate of return on capital is r + Æ .5 I assume also that the agents face borrowing constraints that do not allow them to hold negative assets at any time. I assume that the U.S. are a closed economy with an aggregate production function F (Kt ; Lt ) = AKt L1t , where Kt is aggregate capital and Lt is aggregate labor. I instead assume that Sweden is a small open economy, so the interest rate net of taxes is taken as exogenous and equal to that of the U.S. For each country, I normalize the units of labor and of the good so that, in steady state, the average labor income of a worker per period and the wage are 1. 4.3 Government The government is in nitely lived and taxes labor earnings, capital income and estates to nance the exogenous public expenditure and to provide pensions to the retired agents. Labor earnings are taken as exogenous and calibrated to the data, matching the after-tax Gini coeÆcient. Since the U.S. tax system is progressive, this Gini coeÆcient is lower than the one computed from pre-tax labor earnings. In the model, I introduce a constant tax rate l , in order to balance the government budget, while all the progressive features of the tax system are 4 Once again, this is done to keep the number of state variables at a minimum. this linear technology, Æ is irrelevant for the agents in the economy if we hold r xed; it is only important for measuring gross revenues pertaining to capital that are included in the de nition of GNP. 5 Given 9 already re ected in the calibrated earnings distribution. Income from capital is taxed at a at rate a . Estates larger than a given value exb are taxed at rate b on the amount in excess of exb . The structure of the social security system is the following: the retired agents receive a lumpsum transfer from the government each period until they die. The amount of this transfer is linked to the average earnings of a person in the economy. 4.4 The Household's Problem I consider an environment in which, during each period, a t-year-old individual chooses consumption c and risk-free asset holdings for the next period, a0 . The state variables for an agent are denoted as x = (t; a; y; yp), where t is his age, a are the assets he carries on from the previous period, y is the current realization of his productivity process, and yp is the value of his parent's productivity at age 40 until the worker inherits and zero thereafter. This latter variable takes on two purposes. First, when it is positive, it is used to compute the probability distribution on bequests that the household expects from his parent. Second, it distinguishes the agents that have already inherited, for whom we set yp = 0, from those who have not, for whom yp is strictly positive. The agents inherit bequests only once in a lifetime, at a random date which depends on their parent's death. Since there is no market for annuities, part of the bequests the child receives are \accidental bequests," linked to the fact that people's life span is uncertain and therefore they accumulate precautionary savings to o set the life-span risk. The optimal decision rules are functions for consumption, c(x), and next period's asset holding, a0 (x), that solve the dynamic programming problem described below. Let's consider the agent's recursive problem distinguishing four subperiods in his life to clarify the problem he faces in each phase of his life. (i) From age t = 1 to age t = 3, (from 20 to 30 years of age) the agent works and will survive with certainty until next period. Moreover, he does not expect to receive a bequest soon because his parent is younger than 60 and will survive at least one more period for sure. Since the law of motion of yp is dictated by the death probability of the parent, for this subperiod yp0 = yp. n o V (t; a; y; yp) = max u(c) + Et V (t + 1; a0 ; y 0; yp) c;a 0 subject to: h i c 1 + r (1 a ) a + (1 l ) t y h a0 = 1 + r (1 i a ) a c + (1 l ) t y (1) (2) (3) r is the interest rate on assets, and the evolution of y 0 is described by the transition function Qy . 10 (ii) From t = 4 to t = 8, i.e. from 35 to 55 years of age the worker will survive for sure up to next period. However, his parent is at least 60 years old and faces a positive probability of dying any period; hence a bequest might be received at the beginning of next period. Iyp>0 is the indicator function for yp > 0; it is 1 if yp > 0 and zero otherwise. n V (t; a; y; yp) = max u(c) + Et V (t + 1; a0 ; y 0 ; yp0) c;a o (4) 0 subject to (2) and : h i a0 = 1 + r (1 yp0 = a ) a c + (1 l ) t y + b0 Iyp>0 Iyp =0 (5) 0 yp with probability 0 t+5 with probability (1 (6) t+5 ) where Et is the conditional expectation based on the information available at time t. The conditional distribution of b0 is given by b (x; :).6 b represents the bequest distribution a person expects if his parent dies; in equilibrium it will have to be consistent with the behavior of the parent. Since the evolution of the state variable yp is dictated by the death process of the parent, yp0 jumps to zero with probability t+5 (5 periods is the di erence in age between each parent and his children).7 I assume the following processes to be independent: the survival/death of the decision maker; the survival/death of his parent; the size of the bequest received from the parent, conditional on the parent dying; and the future labor income, conditional on the current one. (iii) tr 1, i.e., 60 years old: the period before retirement. The individual faces a positive probability of dying and hence has a bequest motive to save. De ne after tax bequests as b(a0 ) = a0 b max(0; a0 exb ). n V (t; a; y; yp) = max u(c) + c;a 0 t Et V (t + 1; a0 ) + (1 o 0 t )(b(a )) (7) where b 1 (b) = 1 1 + 2 (8) subject to (2), (5) and (6). I choose the functional form for (b) to make it roughly consistent with the child's utility from receiving the bequest. To derive it, assume that the child consumes a constant amount (say the average labor income, 1) in each period, and if he inherits, he consumes the bequest 6 The 7 If probability distribution b depends on x only through t and yp, not through y . yp = 0, eq.(7) implies yp0 = 0 for sure, which we want. 11 in 2 periods, in equal amounts. Under these assumptions, the additional utility he would derive from receiving a bequest b would be: (1 + b2 )1 (1 ) 1 (1 ) (1 + b2 )1 + (1 ) 1 ::: (1 ) + ::: + 2 1 (1 2 1 1 ) (1 + b2 )1 (1 ) : Collecting terms, dropping the constant and de ning 1 appropriately we obtain (b). 1 can be thought of as a measure of how much the parent values the child's utility. I will discuss how I chose 1 and 2 in sections 5 and 6. (iv) From tr to T , i.e. from 65 to 85, after retirement. In the model economy, the agent will not inherit after turning 65 years old because his parent is dead at that time. Moreover, after retirement I assume that people no longer work and just live o pensions and interest. This implies that we can drop two state variables from the retired people's value function, y and yp, and that the only uncertainty the retired agents face is the time of their death. n V (t; a) = max u(c) + c;a 0 t V (t + 1; a0 ) + (1 o 0 t ) (b(a )) (9) subject to (8) and: h i c 1 + r (1 a ) a + p h i a0 = 1 + r (1 a ) a c + p (10) (11) p is the pension payment from the government. The terminal period value function V (T + 1; a) is set to equal (b(a)). 4.5 Transition Function From the policy rules, the bequest distribution and the exogenous Markov process for produc~ (x; ). M~ (x; ) is thus the probability distribution tivity, we can derive a transition function M 0 of x (the state in the next period), conditional on x, for a person that behaves according to the ~ is de ned is (X; ~ X~ ), with: policy rules c(x) and a(x).8 The measurable space over which M X f1; ::::; T g <+ Y (Y [ f0g); X P f1; :::; T g B <+ B(Y ) B(Y [ f0g); 8 For ~ on c; a; b implicit. simplicity of notation I keep the dependence of M 12 n o X~ X [ D n X~ ~ : ~ = X [ d; X o 2 X ; d 2 f;; fDgg where P is the cardinal set of f1; :::; T g, and D indicates that a person is dead. ~ , it is enough to display it for the sets To characterize M L(t; a; y; yp ) (t0 ; a0 ; y 0; yp0) 2 X : t0 t ^ a0 a ^ y 0 y ^ yp0 yp ~ is de ned by M ~ x; L(t; a; y; yp On such sets M ) = 8 > > > > < > > > > : h if x 6= D t It+1t Ia(x)a Iyp=0 + Iypyp t+5 + i b (x; [0; a a(x)])(1 t+5 )Iyp>0 Qy (y; [0; y] \ Y ) if x = D 0 where I is an indicator function, which equals 1 if the subscript property is true and zero otherwise. ~ , notice that t is the probability of surviving into the next period. CondiTo understand M tional on survival, a person currently of age t will be of age t + 1 next period, hence the presence of It+1t. If his parents are already dead, i.e., yp = 0, he cannot receive bequests anymore, and his assets next period are a(x) for sure (as discussed above, this is always the relevant case for people 65 and older). If, instead, his parents are still alive, i.e., yp > 0, they can survive into the next period with probability t+5 ; in that case, tomorrow's assets for the worker will be a(x) and yp0 = yp. Alternatively, the parents may die, with probability 1 t+5 ; under this scenario, the person inherits next period, yp0 = 0, and the probability that next period's assets are no more than a is the probability of receiving a bequest between 0 and a a(x). Qy describes the evolution of income; note that the evolution of income, one's survival and the survival of the parent are independent of each other. Finally, death is an absorbing state. ~ , I can de ne an operator RM~ that maps probability distributions on (X; ~ X~ ): Based on M (RM~ m ~ )(~) Z M~ (x; ~)m ~ (dx); 8~ 2 X~ : This operator describes the probability distribution of nding a person in state x0 tomorrow, given the probability distribution of the state today. Such an operator has a unique xed point, which is the probability distribution that attributes probability 1 to fDg: everybody dies eventually. However, in the economy as a whole, we are not interested in keeping track of dead people, so I will de ne a modi ed operator on measures on (X; X ). Furthermore, it is necessary to take into account that new people enter the economy in each period. The transition function corresponding to the modi ed operator RM is thus: M (x; L(t; a; y; yp )) = M~ (x; L(t; a; y; yp )) + n5 It=5 Qyh (y; [0; y] \ Y )Iyyp n 13 M modi es M~ in two ways. First, it accounts for population growth; when population grows at rate n, a group that is 1% (say) of the population becomes 1=n% in the subsequent period. Second, it accounts for births, which explains the second term in the numerator. If a person is 40 years old (t = 5), his children (there are n5 of them), will enter the economy next period. All of those children have age t = 1 and zero assets.9 Their stochastic productivity is inherited from their parent's at 40, according to the transition function Qyh ; y (which is part of x) is their parent's productivity at 40. The operator RM is thus de ned as (RM m)() Z M (x; )m(dx) 8 2 X RM maps measures on (X; X ) into measures on (X; X ), but it does not necessarily map probability measures into probability measures. Unless the population is at a demographic steady state, the total measure of people alive may grow at rate faster or slower than n, which implies that (RM m)(x) 6= 1 even if m(x) = 1. 4.6 De nition of Stationary Equilibrium A stationary equilibrium is given by: 8 > > > > < > > > > : an interest rate r, allocations c(x); a0 (x), government tax rates and transfers,(a ; l ; b ; exb ; p), a family of probability distributions for bequests b (x; ), and a constant distribution of people over the state variables x: m (x) such that, given the interest rate and the government policy: (i) the functions c and a0 solve the maximization problem described above, taking as given the interest rate, the government tax rates and transfers, and the bequest distribution he expects to receive from his parent, given as a function of his characteristics x; (ii) given a per capita exogenous government expenditure g and the structure of the social security system, the government policy is such that the government budget constraint balances at every period: g= Z h a r a + l t yIt<t r + b (1 9 Since t p Itt r 0 1 ) max(0; a (x) t 1 and a 0, I do not need to include I1t and I0t. 14 i exb ) dm (x); (iii) m is an invariant distribution for the economy, i.e. it is a xed point of the operator RM de ned in subsection 1.4.5: RM m = m : I normalize m so that m (X ) = 1, which implies that m () is the fraction of people alive that are in a state 2 X . (iv) For the U.S., the share of income going to capital is , i.e. (r + Æ ) K = : (r + Æ ) K + w L R Aggregate capital, K , is given by a dm (x). Due to the normalizations, at the steady state, w = 1 and L is the fraction of working age people in the population. Sweden is treated as a small open economy, so r is taken as exogenous.10 (v) the family of expected bequest distributions b (x; ) is consistent with the bequests that are actually left by the parents. Let's now characterize this statement using formulas. De ne rst the marginal distribution of age and income in the population, which is a probability distribution on (f1; :::::; T g Y; P (f1; :::; T g) B(Y )) : mt;y (t;y ) m (fx 2 X : (t; y ) 2 t;y g) 8t;y 2 P (f1; :::; T g) B(Y ) De ne m (jt; y ) as the conditional distribution of x given t and y . For any given (t; y ), m (jt; y ) is a probability distribution11 on (X; X ). For any set 2 X , m (jt; y ) is measurable with respect to P (f1; :::; T g) B(Y ) and is such that Z Xt;y m (jt; y )mt;y (dt; dy ) = m () 8 2 X 8t;y 2 P (f1; :::; T g) B(Y ) The child observes his parent's income at 40. The conditional distribution of the characteristics of the parent at age 40, given an income level yp, is12 m (jt = 5; y = yp). I want the characteristics of the parent at later ages, conditional on his income as of age 40 being yp and conditional on not having died. Denote by l(jt; yp) these conditional distributions. They can be obtained recursively as follows: l(j5; yp) = m (j5; yp) and R ~ x M (x; )l(dx t; yp ) : l( t + 1; yp) t 10 K represents the average capital held by Swedish citizens, which may di er from average capital present in Sweden. 11 fm (jt; y )g is uniquely de ned up to sets of m -measure zero. t;y 12 I use the letter y to distinguish both from y and yp: y plays the role of income for the parent (state variable p p y ) and the parent's income for the child (state variable yp). j 15 j The conditional distributions l(jt; yp) imply conditional distributions of assets la (jt; yp) on (<+ ; B(<+ )) which are given by la (a jt; yP ) l(fx 2 X : a 2 a gjt; yp): Since the probability of death is independent of income and assets, the distribution of assets that are bequeathed by dying parents is the same as the distribution of assets of surviving parents. We thus have b ((t; a; y; yp); a) = t 1 Y l a 2 <+ : n5 ( s=1 s) " 1 8a 2 B(<+ ) 8a 2 <+ a exb aIaex + exb + I (1 b ) a>ex 8y; yp 2 Y; t = 1; :::; T 5 # b b 2 a jt + 5; yp ! (12) In equation (12), I take into account the assumptions made before about the structure of bequest taxation and the assumption that the bequest is distributed evenly among surviving children. I need now to de ne b when t = T 4, which is the last age a person can inherit. Since there are no survivors at age T + 1, I cannot use the survivor's assets to compute the assets that are bequeathed. Instead, let's use the policy function a(x) to de ne: la (a jT + 1; yp) Z X Ia(x)2 l(dxjT; yp) a 8a 2 B(<+ ): With this de nition, equation (12) can be extended to t = T thus the formal requirement of consistency on b . 4 as well. Equation (12) is 4.7 The Algorithm The following steps are used to solve the model: (i) Solve the household's value functions. Assume a functional form for (b) (the utility of leaving a bequest) and start from the last period, T ; next period the agent will be dead for sure, hence he will derive utility only from the bequests he will leave. Solve backward for the value function at T 1. Continue analogously, taking as given the value function for next period until the rst period is reached. The diÆcult part of solving this model is linked to the curse of dimensionality; there are four state variables. To manage this problem, keep track of the value function on a coarser grid (90-150 points) for capital (the grid is not uniform and has more points concentrated at low levels of capital). The maximization problem is solved for a household that starts with an initial level of assets on this grid. However, future investment is allowed to lie on a 16 ner grid; this requires the household to evaluate the value function at points that do not lie on the initial grid, which is accomplished by interpolation. The resulting investment policy is thus de ned on a ner capital grid. Keep track of the transition function and invariant distribution for this economy on the coarser grid. To do so, take the asset level given by the investment policy function, nd the two closest asset levels that include it on the coarser capital grid, and attribute to each of these points a weight according to their relative distance from the original capital level on the investment policy grid. Choose the number of grid points for capital so that the results are neither sensitive to the number of grid points nor to the linear interpolation procedure. (ii) Taking as given the Markov processes for the productivity and productivity inheritance and the agents' policy functions, compute the transition matrix and the associated invariant distribution. Since the agents' policy functions are de ned on a ner grid, we need to map them to the coarser grid used for the value function, transition matrix and invariant distribution. To do so, take the agents' optimal decision, given his state variables, and nd the adjacent values that include his optimal choice in the coarser grid. Then attribute to these two points a weight given by the relative distance between each of the two points and the agent's optimal choice. (iii) Iterate on the tax rate on labor income until the government budget constraint is balanced. (iv) Iterate on bequests until the equilibrium condition described by equation (12) is met. 5 The Experiments To understand the quantitative importance of these intergenerational links, I construct several simulations that I run both for the U.S. and the Swedish economies. I start with an experiment in which the model is stripped of all intergenerational links: an overlapping-generations model with lifespan and earnings uncertainty. The accidental bequests left by the people who die prematurely are seized by the government and equally redistributed to all people alive.13 The idea is to see how much wealth inequality can be generated by the life-cycle structure when only lifespan and earnings uncertainty are activated. The second experiment modi es the rst one: the unplanned bequests left are distributed to the children of the deceased, rather than equally to everybody alive. This experiment is meant to assess whether an unequal distribution of estates is quantitatively important when all bequests are involuntary. 13 This exercise uses Huggett's setup but adapts it to the length of the periods and the productivity process that I use throughout this paper in order to make the results comparable to the other simulations I run. I cannot use the same time period and income process as Huggett, since the simulations with altruism require a higher number of state variables and the model would require huge computing resources to solve. 17 Fixed Parameter t t n g a r p Qy Qyh Calibrated Parameter b exb 1 2 Value * * 1.5 1.2% yearly 19% of GDP 20% 6% 40% average income + + Source(s) Bell, Wade and Goss [6] Hansen [15] Attanasio et al. [2] Econ. Report of the President [31] Econ. Report of the President [31] Kotliko et al. [23] see text Kotliko et al. [23] Huggett [19], Lillard et al. [28] Zimmerman [40] Value 10% 40 years of average earnings .95{.97 -9.5 8 Chosen to Match see text see text capital-output ratio intergenerational transfers share \altruistic feedback", see text Table 5: Parameters for the U.S. economy and their sources. * refers to a vector + see description in the text The third experiment introduces the bequest motive: parents care about their children and leave them bequests. This allows us to see whether the fact that some of the bequests left are voluntary matters. The fourth exercise activates both the bequest motive and parent's productivity inheritance in order to evaluate the importance of the family background. 6 Numerical Simulations for the U.S. Economy Most of the parameters of the model are taken from other sources, while few of them are chosen to match some aspects of the data. I summarize these choices in Table 5. For people older than 60, t is the vector of conditional survival probabilities. The series I use corresponds to the conditional survival probabilities of the cohort born in 1965. People 60 years old and younger survive for sure into the next period. t is the age-eÆciency pro le vector. I take the risk aversion parameter from Attanasio et al. [2] and Gourinchas and Parker [14], who estimate it using consumption data. This value falls in the range (1-3) commonly used in the literature. 18 The rate of population growth, n is set to equal the average population growth from 1950 to 1997, 1.2% g is government expenditure excluding transfers (about 19% of GDP). a is the capital income tax, 20%. r is the interest rate on capital, net of depreciation and gross of taxes. In models without aggregate uncertainty it is commonly chosen to be between the risk free rate and the rate of return on risky assets. I assume an interest rate of 6% so that the capital share of output is about .36. Pensions (p) are such that the social security replacement rate is 40% and the implied government transfers to GDP ratio in the model is consistent with the one reported in the Economic Report of the President [31]. The logarithm of the productivity process is assumed to be an AR(1). I choose its persistence to be consistent with the one used by Huggett [19] adapted to a ve year period model and its variance to match a Gini coeÆcient for earnings of workers of about .43 (Lillard and Willis [28]). The implied autocorrelation parameter is .83 and its variance .41. The logarithm of the productivity inheritance process (for yp) is also assumed to be an AR(1). I take its persistence from Zimmerman's [40] estimates and its variance so that the standard deviation of the logarithm of earnings is in the ballpark provided by Zimmerman [40]. The resulting autocorrelation parameter is .67 and its variance is .42. I convert both the productivity and the productivity inheritance processes to a discrete Markov chain according to Tauchen and Hussey [38]. I use three values for the income process. The resulting income distribution is reported and compared with the data in appendix A. The remaining parameters are chosen to match features of the U.S. economy as follows. b is the tax rate on estates that exceed the exemption level exb . According to U.S. law, each individual can make an unlimited number of tax-free gifts of $10,000 or less per year, per recipient; therefore, a married couple can transfer $20,000 per year to each child, or other bene ciary. For larger gifts and estates, there is a \uni ed credit", i.e., a credit received by the estate of each decedent, against lifetime estate and gift taxes. For the period between 1987 and 1997, each taxpayer received a tax credit that eliminated estate tax liabilities on estates valued less than $600,000. The marginal tax rate applicable to estates and lifetime gifts above that threshold is progressive, starting from 37% (Poterba [33]). However, the revenue from estate taxes is very low (in the order of .2% of GDP in 1985-97) as there are many e ective ways to avoid such taxes (see for example Aaron and Munnell [1]); moreover, only about 1.5% of decedents pay estate taxes. Therefore, in the model I set exb to be 40 times the median income and b to be 10% to match the observed ratio of estate tax revenues to GDP and the proportion of estates that pay estate taxes. I discuss the sensitivity of the model to the choice of these two parameters when describing the results. I use the discount factor, , to match a capital to GDP ratio of 3. In the calibrations in which the bequest motive is activated, I use 1 to get a reasonable share of the bequests to aggregate capital and 2 to make the utility from leaving bequests, (b), roughly consistent with a \truly altruistic model" in the sense that it is reasonably close to the utility of the child from receiving the bequest. Figure 22 compares the function with the true value of receiving the bequest for 19 a 65 year old at the 10% (\poor"), 50% (\median") and 90% (\rich") quantiles of the wealth distribution.14 We see that is a reasonable approximation of the value a truly altruistic parent would derive from the bequest he leaves. However, for the parent of a poor child, the marginal utility of leaving a bequest is higher in the fully altruistic model, and the bene t is more concave. The opposite is true for the parent of a rich child. Table 6 summarizes the results.15 Capital Transfer Percentage wealth in the top Percent with output wealth Wealth negative or ratio ratio Gini 1% 5% 20% 40% 80% zero wealth U.S. data 3.0 .63 .72 28 49 75 89 99 5.8-15.0 No intergenerational links, equal bequests to all 3.0 N/A .64 6 23 64 90 100 17 No intergenerational links, unequal bequests to children 3.0 .37 .65 6 24 65 90 100 17 One link: parent's bequest motive 3.0 .57 .71 12 34 72 93 100 19 Both links: parent's bequest motive and productivity inheritance 3.0 .61 .73 15 38 75 94 100 19 Table 6: Results for the U.S. calibration. 6.1 The Experiment with no Intergenerational Links and Equal Bequests to All The results of this experiment show that an overlapping-generations model with no dynastic links and equal distribution of bequests has serious diÆculties in generating enough skewness to match the distribution of the U.S. wealth. For the parameter values in Table 5, I obtain a Gini coeÆcient that is below what we observe in the data. Moreover, the concentration is mainly achieved by having a lower tail which is too fat and an upper tail which is far too thin. As discussed in the introduction, overlapping-generations models tend to have a large fraction of people against the borrowing constraint. People are born without savings that could be used to absorb negative income or productivity shocks. As a consequence, all the young consumers that get a bad productivity shock hit the borrowing constraint. During their working age, the households gradually accumulate assets both as life-cycle savings for their old age and as a form 14 The functions are normalized by adding a constant to t on the graph. all experiments I exclude 20 year old people from the computations on the wealth distribution. I do so because in this paper I assume that people start o with zero wealth and I therefore do not propose a theory of the distribution of wealth for them. To explain the data for 20 year old people, a theory of inter-vivos transfers would, in my opinion, be required. 15 For 20 of precautionary savings. As a consequence, the fraction of people with zero wealth gradually declines until retirement. At the other end of the distribution, the absence of intergenerational links implies that it is very hard to account for large estates, as a lifetime is too short a period for most households to accumulate such large fortunes. With the current parameterization the top 1% of the population holds just about 5.5% of total wealth. With a richer income process, the top quantiles are somewhat higher, but it is still true that few people have suÆciently high income to accumulate such large estates over a lifetime. With 7 productivity states (instead of 3) the top 1% of people hold 8% of the aggregate wealth (which is still lower than obtained in the model with links and only three income states). My main interest, however, is to show that, with the same income process, experiments with intergenerational links fare much better than an environment in which such links are not present. I the model the richest 2% of people hold about 20 times the average annual labor income in assets or about 7 times the highest annual labor income. In the U.S. data, the richest 2% of people held about 35 years of average labor income in assets in 1994.16 The transfer-wealth ratio is not de ned for this experiment, as bequests are collected by the government and redistributed lump sum to all the population, independently of family links. The overlapping-generations model with no intergenerational links also fails to recover the age-asset pro le observed in the data. All households in the model economy ( gure 1) run down their assets during retirement until they are left with zero wealth at the time we assume they die for sure. This implies a much larger dissaving than we observe in the data ( gure 45), especially for richer households. This also suggests that a substantial fraction of wealth that is accumulated in this economy is linked to the uncertainty over the life span. In such an economy a market for annuities would thrive and would reduce savings signi cantly; by purchasing annuities the households would avoid leaving large accidental bequests when they die in their early retirement years. 6.2 The Experiment with No Intergenerational Links and Unequal Bequests to Heirs In this experiment the accidental bequests left by the deceased are inherited by their own children, rather than being redistributed by the government to all people alive. As we can see from table 6, there is little change in the distribution of wealth. The intuition is simple: some people in the economy inherit some wealth, some other people do not, but nobody cares about leaving bequests. Even high income households do not plan to share their fortune with their children but do so only if they die early in their retirement years. As a consequence, the transfer-wealth ratio is low, no persistence in wealth across families is generated and no large fortunes are accumulated in this economy. Figure 5 depicts the probability of the parent dying for each age of the o spring, conditional on the parent dying before the o spring. In this model economy, the individuals do not die before age 60; therefore the probability of the parent dying before the o spring reaches age 16 Sources: Economic Report of the President [31] and Hurst, Luoh and Sta ord [20]. 21 35 is set to zero. Compare the saving behavior of the average agent who expects to receive some bequest, with his behavior in the hypothetical case in which he does not expect to receive a bequest ( gure 6). The fact that parent's assets are not observable in uences the saving behavior of people that have not yet inherited.17 Since in this economy children do not become orphans before age 40, and everybody attributes positive probability to receiving some bequest until their parent die, the comparison starts for agents of age 40. For them, we can compare the behavior of people whose parent died, and did not leave them any asset (these children do not expect an inheritance anymore), with the behavior of people that still expect to receive something. The top line refers to the age-saving pro le for the average agent in the case in which he does not expect to inherit. We see that the expectation of inheriting, conditional on all other variables being the same, decreases the saving rate of the agent. However these saving rates converge over time because the parent has no bequest motive and by age 90 runs down all of his assets ( gure 4). Therefore the average size of the expected bequest also goes to zero by that time. Since the probability of dying is assumed to be independent of other individual characteristics, the distribution of bequests is simply the distribution of assets among the population of the parent's generation at age 65, rescaled because of population growth (each person has more than one child). Figure 6 shows the strictly positive range of the bequest distribution that an agent faces at 40 years of age should the parent die at that moment, conditional on the observed productivity of his parent at age 40. I do not plot the probability of receiving a bequest of zero because this would make the graph very diÆcult to read. This probability is about 27, 8 and 0% for a 40-year old whose parent had the lowest, middle or highest productivity level at age 40, respectively. Since age 65 is the peak of wealth accumulation, the children of those who die at age 65 (the 40 year olds) are the ones that receive the largest bequest. At this age, the average bequests are, respectively, 2.9, 6.7, and 15.3 years of average labor income, After the parents retire they run down all of their assets by age 90, so the expected bequest declines, and the people whose parents live up to the nal age of the model economy do not receive any bequest. 6.3 The Experiment with only one Intergenerational Link: Bequest Motive The bequest motive leads to a large increase in the concentration of wealth in the economy and explains the emergence of large estates that are accumulated by more than one generation of savers and are transmitted because of altruism. The Gini coeÆcient increases from .64 to .71 and the fraction of the total wealth by the people in the upper tail of the distribution increases signi cantly. The top 1% of the population holds 12% of total wealth, and the top 5% holds 34% of total wealth. The households at the top 2% quantile hold about 25 years of average labor income18 in assets (compared with 20 years in the model with no intergenerational links), or about 8 years of the highest income level in the model. Figure 13 compares the distribution of wealth by age conditional on having or not having 17 All people have a positive probability of getting a bequest, not only those whose parent will actually leave some. 18 The average labor income for an agent in this economy is one over a ve years period, therefore it is .2 yearly. 22 received a bequest.19 The model predicts that the upper tail of the wealth distribution will be mainly made of households that have already received a bequest. The age-assets pro les for various quantiles of the wealth distribution are displayed in gure 10. From this gure we notice a substantial di erence in the main motives that lead the household to save. The median household mainly saves for retirement; the peak in its wealth holdings occurs at age 65 and is about 6.5 times the average annual labor income in the population. If he reaches the age of 85, the median agent consumes all of his assets (before dying at 90) and does not leave any bequest. The median consumer mostly leaves unintended bequests. This becomes clearer when comparing the age-assets pro les for the bottom 10, 30 and 50% with the model with no bequest motive ( gure 1): the pro les are very close.20 Those who are wealthy, either as a consequence of large inheritances or of a successful working life, plan on sharing their luck with their o spring. As we see from gure 10, at the top of the wealth distribution a lot of the accumulation is done especially in order to leave bequests, and large bequests are left even when the parents die in advanced age. Comparing these top quantiles with the ones in the model with no altruistic links, we see how the introduction of a bequest motive produces an age-wealth pro le more consistent with the U.S. data ( gure 45), especially in the second part of the agent's lifetime. In this setup, the absence of an annuity market is a less severe restriction on the behavior of the richest households as most of their savings are primarily accumulated to be left to the next generation. As in the previous simulation, gure 12 shows that the expectation of receiving a bequest reduces the saving rate of the average agent. Here, however, this e ect does not decrease over time. In fact the richest parents do not run down all of their assets by age 90 because of the bequest motive and, over time, the reduction in the size of the expected bequest is balanced by the increased probability of the parent dying. In gure 15 we can see the strictly positive range of the bequest distribution for a 40 year old person, conditional on his observed parent's productivity level, should his parent die during that period. The probabilities of receiving zero bequests are respectively 27, 11 and 0%, for individuals with parents of low, middle and high productivity level. The average bequests expected are respectively 4.7, 7.6 and 15.5 years of average labor income. Even in presence of a bequest motive, the parents run down their assets after retirement, so the expected bequest declines. The fraction of people whose parent lives up to the nal age of the model economy and who do not receive a positive bequest are 93, 87 and 53%, respectively. The average bequest that they expect at that point in life is about 1.4, 1.9 and 5 years of average labor income respectively. 19 The bequest a person has received may be zero if his parent dies with no assets. model without bequest motive generates somewhat steeper pro les until retirement. This happens because to match the capital/output ratio, the model with no intergenerational links requires a higher discount factor ( = :97 compared with = :96). The steeper consumption pro le early in their lives counterbalances the steep decumulation of assets later in life, when the increased probability of dying implies very high e ective discount factors. The bequest motive keeps the e ective discount factor high in old age. As a consequence, we can allow for a lower implying a atter consumption pro le at younger ages. 20 The 23 6.4 The Experiment with both Intergenerational Links Here I explore how the results of the previous section change when parent and children are linked not only by the bequest the parent intends to leave to his children, but also through transmission of productivity. The introduction of an additional link increases the Gini coeÆcient further to .73. This happens because success in the workforce is now correlated across generations: more productive parents accumulate larger estates and leave their bequests to their children who are in turn more successful than average in the workforce. The introduction of a link in the productivity (or \human capital") of di erent generations within a family tends to have two opposing e ects on the accumulation of assets at the tails of the wealth distribution. First, consider the individuals that are at the lowest levels of wealth. These people tend to be the least productive in the workforce. In particular, the people who are least productive at 20 (who tend to be less productive than the others also at later ages) are more likely to have less productive and hence poorer parents. As a result, people with low productivity in their young age will, on average, receive smaller bequests, which will contribute to lowering their pro le of asset accumulation. On the other hand, the anticipation of smaller bequests will lead the same people to save more to compensate for the smaller transfer wealth. The two e ects will be exactly reversed at the upper end of the distribution of wealth. The direct e ect on transfer wealth dominates in our results. At the upper tail, the top 1% of the population hold 15% of total wealth, up from 12% in the previous example, and the top 5% hold 38% of total wealth, up from 34%. The agents at the top 2% of the population hold 29 times the average yearly labor income in wealth and 9 times the maximum level of labor income in this model, compared respectively with 25 and 8 times in the model with bequests only. In gure 21 we can see the strictly positive range of the bequest distribution for a 40 year old person, conditional on his observed parent's productivity level, should his parent die during that period. The probabilities of receiving zero bequests are respectively 29, 11 and 0%, for individuals with parents of low, middle and high productivity level. The average bequests expected are respectively 4.8, 8.1 and 16.2 years of average labor income. Comparing these numbers with the ones for the experiment with bequest motive only, we can see that introducing productivity inheritance has small e ects on the average expected bequests. The introduction of the \human capital" link in the form I consider here leads overall to modest changes in the results of the previous section. The reason for this result might stem from the weakness of the link introduced. In the current model, children inherit from their parent only their initial productivity level (they do so with probability smaller than one, according to the Markov process Qyh ), and then productivity evolves independently and stochastically over time for all agents. This implies that children of poorer households tend to enter in the labor force at the lowest levels of the income process, but expect an improvement later in life and hence tend not to save. To better evaluate the productivity link, I also run an experiment in which the children's initial productivity level is their parent's one at 40 (inheritance with probability one). As a result, the share of wealth held by the top 1% of the population increases by a couple of percentage points, and the Gini coeÆcient for wealth increases to .75. Ideally, one would like the agents to be born with di erent income processes, to recover the fact that more-educated people 24 Fixed Parameter 1 n g a r p Qy Qyh Value * * .95{.97 1.5 -9.5 .8% yearly 25% of GDP 30% 6.86% 50% average income + + Source(s) Statistical Yearbook of Sweden [36] same as U.S. same as U.S. same as U.S. same as U.S. OECD Economic Surveys, Sweden [30] OECD Economic Surveys, Sweden [30] OECD Economic Surveys, Sweden [30] see text OECD Economic Surveys, Sweden [30] see text Zimmerman [40] b exb 2 Value 15% 10 years average earnings 3.6 Chosen to Match see text see text \altruistic feedback", see text t t Calibrated Parameter Table 7: Parameters for the Swedish economy and their sources. * refers to a vector + see description in the text have a permanent advantage over less-educated ones, and that the age-eÆciency pro le tends to be atter for less-educated households, leaving them again to save more in earlier periods. Unfortunately, these considerations would require the introduction of a further state variable in our problem, making computations even more involved. 7 Numerical Simulations for the Swedish Economy As in the calibration of the U.S. economy, most of the parameters for the Swedish economy are taken from other sources, and a few are chosen to match some aspects of the data. I summarize the parameter choices in Table 7. The Statistical Yearbook of Sweden [36] provides the mortality probabilities for people at di erent ages in 1991-1995. In the calibration of the U.S. economy I use the mortality probabilities of people born in 1965 (which are for the most part projected, since these people are still young). Since life expectancy is increasing, the Swedish data underestimate the life expectancy at the various ages with respect to the one faced by people born in 1965 in Sweden. To correct for this problem I use the U.S. data to compute the relative increase in life expectancy for the relevant period. I then correct the Swedish data assuming that the increase in life expectancy is 25 the same in the two countries. However, as a check, I also us the U.S. conditional probabilities in the simulation of the Swedish economy. It turns out that this has a negligible impact on the results even if the life expectancy of Swedish people is about three years longer than that of U.S. people. I take the age-eÆciency pro le t and the preference parameters , and 1 to be the same as those for the U.S. The rate of population growth, n is set to equal the average population growth from 1950 to 1997, .8% The interest rate on capital, net of depreciation and gross of taxes, r, is taken to be 6.86% so that the interest rate net of taxes in the U.S. and Sweden coincide. Pensions (p) are such that the social security replacement rate is 50% and the implied government transfers to GDP ratio in the model is consistent with the one (net of taxes) reported in the OECD Economic Surveys, Sweden [30]. The persistence of the income and productivity inheritance processes are taken to be the same in Sweden and the U.S. Bjorklund and Jantti [7] estimate the degree of intergenerational income mobility in Sweden and do not reject the hypothesis that it is the same as in the U.S. I take the ratio of the variances of the two processes to be the same one adopted for the calibration of the U.S. economy and vary their levels in order to match the Gini coeÆcient for the income process, which in Sweden is somewhat lower than in the U.S. b is the tax rate on estates that exceed the exemption level exb . I take the e ective tax rate to be higher than the one for the U.S., 15%, and the exemption level to be lower, 10 years of average labor earnings. In Sweden taxes are paid on inheritances, rather than on estates, and the revenue from inheritance and gift taxes is approximately .1% of GDP. The statutory tax rate for children's inheritance is higher than in the U.S. (for the rst bracket it is about 50%) and the exemption level is much lower (in the order of $5,000), but there are legal ways, for example bequeathing an apartment or a large rm, of obtaining a much larger exemption level. It is therefore more diÆcult than in the U.S. to de ne the statutory exemption level. The combined choice of b and exb matches the revenues from bequests and gift taxes. As in the calibration of the U.S. economy, I choose 2 to make the utility from leaving bequests (b) roughly consistent with a \truly altruistic" model, in the sense that it is reasonably close to the utility of the child receiving the bequest. The value functions of the Swedish model economy turn out to be more concave than the ones in the U.S. one; therefore the value of 2 that I adopt in this simulation is di erent. As a sensitivity check, I compare the results of the model with both intergenerational linkages when adopting the same 2 used in the U.S. simulation, instead. As for the level of altruism as measured by 1 , it turns out that over the relevant range for assets, the levels of the U.S. and Swedish value function of the bequest receiver are very close. In this sense the assumed intensity of the bequest motive is the same in both economies. The results of the various experiments calibrated to the Swedish economy are summarized in Table 8 and discussed below. 26 Capital Transfer Percentage wealth in the top Percent with output wealth Wealth negative or ratio ratio Gini 1% 5% 20% 40% 80% zero wealth Swedish data 2.0 > :51 .73 17 37 75 99 100 30 No intergenerational links, equal bequests to all 2.1 N/A .64 5 23 64 89 100 24 No intergenerational links, unequal bequests to children 1.9 .38 .67 6 25 67 91 100 26 One link: bequest motive 2.0 .76 .71 8 29 73 95 100 30 Both links: bequest motive and productivity inheritance 2.0 .77 .73 9 31 75 95 100 30 Table 8: Results for the Swedish calibration. 7.1 The Experiments with no Intergenerational Links Compared to the U.S. calibration, people in the Swedish model economy face less earnings uncertainty (to match a lower Gini coeÆcient for earnings, we need to reduce the variance of the income process) and a higher social security replacement rate. The rst element tends to reduce precautionary saving and the second one to reduce life-cycle saving. The model predicts a lower wealth-to-GDP ratio than in the U.S. The ratio predicted by the model is very close to the capital-output ratio in Swedish data. In the rst experiment, the only intergenerational transfers stem from accidental bequests; since there is life-span uncertainty and there are no annuity markets, people accumulate assets to self-insure against the risk of living for a long time. When they die earlier, they leave their assets behind. These bequests are distributed equally to all people in the economy. Looking at the distribution of wealth, we can see that even in an economy in which there is less wealth and earnings inequality and the government redistributes more, the basic version of the model generates an upper tail of the wealth distribution which is too thin; the top 1% of people hold only 5% of total wealth, compared with 14% in the data. Unlike the results for the U.S. model economy, the model for Sweden does not generate too many people at zero wealth. In fact in the data about 30% of the population is in this situation while the model generates 24%. When I assume that involuntary bequests are left to the children of the deceased (third row in table 8), the distribution of wealth is almost the same as in the case in which bequests are evenly distributed to all people alive. This is analogous to what I found for the U.S. 27 7.2 The Experiment with only one Intergenerational Link: Bequest motive As in the U.S simulations, the introduction of the bequest motive helps in generating a more skewed wealth distribution by increasing the share of total wealth held by the rich. The forces discussed for the U.S. model economy that generate this result are also at work for the Swedish model economy. The share of wealth held by the top 1% of people rises from 5% to 8%, and the share held by the top 5% increases from 22% to 29%. Moreover, we can see from gures 35 and 36 that the wealth quantiles of the people who do receive a bequest are signi cantly higher than those who do not get a positive transfer of wealth from their parent. Figure 37 shows the strictly positive range of the bequest distribution for a 40 year old person, should his parent die today, conditional on his parent's productivity at 40 years of age. Conditional on the parent's productivity level from the lowest to the highest, the probability of receiving a zero bequest is 29, 14 and 0% and the average bequests are 2.9, 4.4, and 8.5 years of average labor earnings. Compared to the U.S. simulation, therefore, the number of people expecting to receive no intergenerational transfer is slightly higher, and the average bequest size, for all parent's levels of ability, is lower. 7.3 The Experiment with both Intergenerational Links The introduction of the second linkage, the intergenerational transmission of productivity, helps further in matching the top 20% of the wealth distribution: the share held by the top 1, 5, 10 and 20% increase to 9, 31, 51 and 75%. However, as discussed previously, this linkage is not very strong and hence does not change the results dramatically. Figure 43 shows the strictly positive range of the bequest distribution for 40 year old agents, should their parent die this period. For people whose parent was at the lowest productivity level at age 40, the average bequest is 2.6 years of average labor earnings; for these agents the probability of receiving no intergenerational transfer is 34%. For the individuals whose parent at 40 was at the middle productivity level, the average bequest is 4.4 and their probability of receiving zero bequests is 15%. For those whose parent was at the highest productivity level at age 40, these numbers are respectively 8.7 and 0%. As mentioned in the calibration, I use di erent values for 2 in the U.S. and Swedish model economies. As a sensitivity check I report the results for the calibration of the Swedish economy using the 2 adopted in the U.S. simulations, all other parameters staying the same. In this case, the capital-output ratio is 1.65, the transfer-wealth ratio .46, and the fraction of people at zero wealth 31%. The top 1, 5, 10, 20, 40, 80% hold respectively 8, 28, 46, 71, 94 and 99%. However, with this parameterization, the warm-glow utility of leaving a bequest is much atter than the value function of the children receiving it, even for the richest children. Moreover, the age-saving pro les show fast decumulation after retirement for people at all wealth levels, which is in contrast with the empirical evidence. 28 8 Discussion of the Assumptions In order to make the model manageable and solvable I have made several simplifying assumptions. In this section I discuss the assumptions and their likely qualitative implications on the wealth distribution. It is widely recognized (e.g., Becker and Tomes [4, 5]) that the time and resources that parents devote to children's education are very important in understanding the distribution of earnings and wealth. In this setup the simpli cation that children partially inherit their parents' productivity is meant to recover the fact that education and human capital are closely related to the family background of each person. However, when explicitly modeling human capital investment, the return is commonly assumed to be decreasing, and up to a given level of investment (which may depend on the child's abilities), greater than the rate of return on physical capital. This implies that parents will begin to invest in their children's human capital and then invest in physical capital when the return from human capital reduces to the return on physical capital. In this setup, in the presence of borrowing constraints, the poorest families only invest in their children's human capital, and they may not even be able to invest up to the optimal amount. The richest families not only invest in their children's human capital but also leave them physical capital. Poorer families will tend to have poorer children, thus generating persistence in the lower end of the wealth distribution. At the upper tail of the distribution, rich children might want to save less because they expect large bequests. However, they will also be richer (because of their dominant income process and the bigger transfers they receive from their parents), hence they might want to save more. Depending on which e ect dominates, this will increase or decrease persistence at the upper end of the wealth distribution. Most likely, if the altruism toward one's children is very strong for the richer people, the desire of leaving large estates to children will o set the reduction in saving because of the large bequests received. Which e ect dominates thus depends on how wealth a ects savings at high levels of wealth. As discussed previously, I made restrictive assumptions on the information available to the children on their parent's wealth and income. These assumptions are made for computational reasons but are also likely to a ect the results. In particular, I expect the model in the current version to display fatter tails at both ends of the wealth distribution, compared with a model in which the parent's assets and income are observable by the child. With perfect observability children of poor parents will save more, since they are aware that no bequest will be left to them. On the other hand, children of richer parents will save less. If wealth of poor parents is easier to measure than wealth of rich parents, only the lower tail of the distribution of wealth would become thinner. Another important assumption is that there are no inter-vivos transfers. In the data these transfers often have a compensatory nature: parents tend to give when the children need money the most. This may happen when they go to college, start a new job, get married, buy a house or get a sequence of bad shocks. This assumption is probably most relevant when children are 20 to 35 years of age and are starting o their own life. Allowing for inter-vivos transfers would help reduce the number of people at zero wealth, especially among the young. I take fertility to be exogenous and independent of the agent's wealth. This is likely to be a reasonable assumption for the U.S., but not a good one for other countries, especially 29 developing countries. If poorer families on average are more proli c than richer families then the concentration of the wealth distribution will be increased: poor families have to divide their scarce resources among more children who, in turn, will be most likely poor. Labor earnings are also assumed to be exogenous and people can only invest in a riskless asset. The U.S. data show that there is a noticeable correlation between high wealth and income from running a business. Introducing entrepreneurial choice in the model, for example in the form of investment in a risky asset in presence of minumum investment size and borrowing constraints, would generate more heterogeneity in the people's income processes and more precautionary savings. 9 Directions for Future Research There is considerable debate about abolishing estate taxation. In the U.S., Sweden and many other countries, estate and gift taxes produce little revenue (in the order of 0.1% of GDP) and possibly distort the savings decision of the few people that do most of the capital accumulation in the economy: the rich. I plan to study the e ects of abolishing the estate taxes in the context of this model. I am interested in looking both at the macroeconomic consequences (e.g., what would happen to total capital) as well as the distributional e ects (e.g., would wealth dispersion increase and by how much? Would the poor people become even poorer in absolute terms, or could they also bene t if total capital in the economy increased?). In addition, many rich people are entrepreneurs, and entrepreneurial activity is likely to be an important factor in understanding the distribution of wealth. I plan to study entrepreneurial income and the degree of wealth inequality across di erent countries to better assess the importance of entrepreneurial income in evaluating the wealth distribution. I also plan to use my model to study the demand for annuities. In fact, the market for annuities is remarkably thin. The literature provides several explanations, such as moral hazard, altruism and social security provision. In my setup, social security annuitizes all or most of the savings of poorer people and the bequest motive provides a reason for richer people not to annuitize their wealth completely. It will be interesting to introduce annuities in my setup to study the quantitative importance of these two elements. 30 A Calibration of the income processes As discussed in the calibration section, I convert both the productivity and the productivity inheritance processes to discrete Markov chains according to Tauchen and Hussey [38]. I use three values for the income process. Since I want the possible realizations for the initial inherited productivity level to be the same as the possible realizations for productivity during the lifetime, I choose the quadrature points to be the same for the two Markov processes. Tables 9 and 10 report data on earnings distributions, respectively for the U.S. and Sweden, and compare them with the earnings distributions used in the model. The tables are computed using data for households whose head is 25 to 60 years of age. In row 1 the de nition of gross earnings includes wages, salaries and self-employment income. Row 2 adds social insurance transfers to gross earnings. Row 3 refers to the earnings distribution used in the simulations. Since the setup does not explicitly model other social insurance transfers other than social security, we should compare earnings used in the simulations to data that include social insurance transfers. From this comparison we can see that the model understates both the earnings of the upper tail, and the fraction of people at zero earnings. Percentage earnings in the top Percent with Gini negative or coe . 5% 10% 20% 40% 60% 80% zero income U.S. earnings data .46 19 30 48 72 89 98 7.7 U.S. earnings + social insurance transfers .44 19 30 47 71 88 97 5.3 U.S. simulated earnings .43 13 25 48 72 87 96 0.0 Table 9: U.S. earnings. Percentage earnings in the top Percent with Gini negative or coe . 5% 10% 20% 40% 60% 80% zero income Swedish earnings .40 15 25 42 68 86 98 7.6 Swedish earnings + social insurance transfers .32 13 22 37 62 80 94 1.0 Swedish simulated earnings .32 10 21 40 63 81 92 0.0 Table 10: Swedish earnings. 31 Figures U.S. calibration. Experiment with no links and equal bequests to all 6 5 Wealth 4 3 2 1 0 20 30 40 50 60 70 80 90 Age Figure 1: Wealth .1 .3 .5 .7 .9 quantiles, by age 1 0.9 0.8 0.7 Wealth Gini B 0.6 0.5 0.4 0.3 0.2 0.1 0 30 40 50 60 70 80 Age Figure 2: Wealth Gini for di erent cohorts 32 U.S. calibration. Experiment with no links and unequal bequests to heirs 1 6 0.9 0.8 5 0.7 Wealth Gini Wealth 4 3 0.6 0.5 0.4 2 0.3 0.2 1 0.1 0 20 30 40 50 60 70 80 0 90 Age 30 40 50 60 70 80 Age Figure 3: Wealth .1 .3 .5 .7 .9 quantiles, by age Figure 4: Wealth Gini for di erent cohorts 0.22 0.4 0.2 0.35 0.18 0.3 0.16 0.25 0.14 0.2 0.15 0.12 0.1 0.1 0.05 0.08 0 0 10 20 30 40 Age of the child 50 0.06 20 60 Figure 5: Probability of the parent dying at each age of the child 25 30 35 40 Age 45 50 55 60 Figure 6: Saving for people who expect or not to inherit 33 6 6 5 5 4 4 Wealth Wealth U.S. calibration. Experiment with no links and unequal bequests to heirs 3 3 2 2 1 1 0 25 30 35 40 45 50 55 0 25 60 30 35 40 Age 45 50 55 60 Age Figure 7: Wealth quantiles conditional on not having inherited Figure 8: Wealth quantiles conditional on having inherited 0.06 0.05 Probability 0.04 0.03 0.02 0.01 0 0 1 2 3 4 Bequest Size 5 6 7 Figure 9: Strictly positive range of the expected bequest distribution at age 40, conditional on the productivity of the parent 34 U.S. calibration. Experiment with bequest motive 1 6 0.9 0.8 5 0.7 Wealth Gini Wealth 4 3 2 0.6 0.5 0.4 0.3 0.2 1 0.1 0 20 30 40 50 60 70 80 0 90 30 40 50 Age 60 70 80 Age Figure 10: Wealth .1 .3 .5 .7 .9 quantiles, by age Figure 11: Wealth Gini for di erent cohorts 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 20 25 30 35 40 Age 45 50 55 60 Figure 12: Saving for people who expect or not to inherit 35 6 6 5 5 4 4 Wealth Wealth U.S. calibration. Experiment with bequest motive 3 3 2 2 1 1 0 25 30 35 40 45 50 55 0 25 60 30 35 40 Age 45 50 55 60 Age Figure 13: Wealth quantiles conditional on not having inherited Figure 14: Wealth quantiles conditional on having inherited 0.06 0.05 0.04 0.03 0.02 0.01 0 0 1 2 3 4 5 Figure 15: Strictly positive range of the expected bequest distribution at age 40, conditional on the productivity of the parent 36 U.S. calibration. Experiment with bequest motive and productivity inheritance 1 6 0.9 0.8 5 0.7 Wealth Gini Wealth 4 3 2 0.6 0.5 0.4 0.3 0.2 1 0.1 0 20 30 40 50 60 70 80 0 90 30 40 50 Age 60 70 80 Age Figure 16: Wealth .1 .3 .5 .7 .9 quantiles, by age Figure 17: Wealth Gini for di erent cohorts 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 20 25 30 35 40 Age 45 50 55 60 Figure 18: Saving for people who expect or not to inherit 37 6 6 5 5 4 4 Wealth Wealth U.S. calibration. Experiment with bequest motive and productivity inheritance 3 3 2 2 1 1 0 25 30 35 40 45 50 55 0 25 60 30 35 40 Age 45 50 55 60 Age Figure 19: Wealth quantiles conditional on not having inherited Figure 20: Wealth quantiles conditional on having inherited 0.06 4 0.05 3 2 0.04 1 0.03 0 0.02 −1 −2 0.01 −3 0 0 1 2 3 4 0 5 Figure 21: Strictly positive range of the expected bequest distribution at age 40, conditional on the productivity of the parent 5 10 15 20 25 Bequest Figure 22: Warm glow and true altruism compared: warm glow (solid), poor (dashed), median (dash-dot), rich (dots) 38 Swedish calibration. Experiment with no links and equal bequests to all 6 5 Wealth 4 3 2 1 0 20 30 40 50 60 70 80 90 Age Figure 23: Wealth .1 .3 .5 .7 .9 quantiles, by age 1 0.9 0.8 Wealth Gini 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 30 40 50 60 70 80 Age Figure 24: Wealth Gini for di erent cohorts 39 Swedish calibration. Experiment with no links and unequal bequests to heirs 1 6 0.9 0.8 5 0.7 Wealth Gini Wealth 4 3 0.6 0.5 0.4 2 0.3 0.2 1 0.1 0 20 30 40 50 60 70 80 0 90 Age 30 40 50 60 70 80 Age Figure 25: Wealth .1 .3 .5 .7 .9 quantiles, by age Figure 26: Wealth Gini for di erent cohorts 0.14 0.4 0.35 0.12 0.3 0.1 0.25 0.08 0.2 0.06 0.15 0.04 0.1 0.02 0.05 0 0 10 20 30 Age of the son 40 50 0 20 60 Figure 27: Probability of the parent dying at each age of the child 25 30 35 40 Age 45 50 55 60 Figure 28: Saving for people who expect or not to inherit 40 6 6 5 5 4 4 Wealth Wealth Swedish calibration. Experiment with no links and unequal bequests to heirs 3 3 2 2 1 1 0 25 30 35 40 45 50 55 0 25 60 30 35 40 Age 45 50 55 60 Age Figure 29: Wealth quantiles conditional on not having inherited Figure 30: Wealth quantiles conditional on having inherited 0.06 0.05 Probability 0.04 0.03 0.02 0.01 0 0 1 2 3 4 Bequest Size 5 6 7 Figure 31: Strictly positive range of the expected bequest distribution at age 40, conditional on the productivity of the parent 41 Swedish calibration. Experiment with bequest motive 1 6 0.9 0.8 5 0.7 Wealth Gini Wealth 4 3 2 0.6 0.5 0.4 0.3 0.2 1 0.1 0 20 30 40 50 60 70 80 0 90 30 40 50 Age 60 70 80 Age Figure 32: Wealth .1 .3 .5 .7 .9 quantiles, by age Figure 33: Wealth Gini for di erent cohorts 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 20 25 30 35 40 Age 45 50 55 60 Figure 34: Saving for people who expect or not to inherit 42 Swedish calibration. Experiment with bequest motive 3 3 2.5 2.5 Wealth Wealth 2 1.5 2 1.5 1 1 0.5 0.5 0 25 30 35 40 45 50 55 0 25 60 30 35 40 Age Figure 35: Wealth quantiles conditional on not having inherited 50 55 60 Figure 36: Wealth quantiles conditional on having inherited 0.06 0.05 0.04 Probability 45 Age 0.03 0.02 0.01 0 0 1 2 3 Bequest Size 4 5 6 Figure 37: Strictly positive range of the expected bequest distribution at age 40, conditional on the productivity of the parent 43 Swedish calibration. Experiment with bequest motive and productivity inheritance 1 6 0.9 0.8 5 0.7 Wealth Gini Wealth 4 3 2 0.6 0.5 0.4 0.3 0.2 1 0.1 0 20 30 40 50 60 70 80 0 90 30 40 50 Age 60 70 80 Age Figure 38: Wealth .1 .3 .5 .7 .9 quantiles, by age Figure 39: Wealth Gini for di erent cohorts 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 20 25 30 35 40 Age 45 50 55 60 Figure 40: Saving for people who expect or not to inherit 44 Swedish calibration. Experiment with bequest motive and productivity inheritance 3 3 2.5 2.5 Wealth Wealth 2 1.5 2 1.5 1 1 0.5 0.5 0 25 30 35 40 45 50 55 0 25 60 30 35 Age 40 45 50 55 60 Age Figure 41: Wealth quantiles conditional on not having inherited Figure 42: Wealth quantiles conditional on having inherited 4 0.06 2 0.05 0 Probability 0.04 −2 0.03 0.02 −4 0.01 −6 0 0 1 2 3 Bequest Size 4 5 6 Figure 43: Strictly positive range of the expected bequest distribution at age 40, conditional on the productivity of the parent −8 0 1 2 3 4 Bequest 5 6 7 Figure 44: Warm glow and true altruism compared: warm glow (solid), poor (dashed), median (dash-dot), rich (dots) 45 200 Wealth 150 100 50 0 30 40 50 60 Age 70 80 90 Figure 45: U.S. data, wealth quantiles: .1, .3, .5, .7, .9 References [1] Henry J. Aaron and Alicia H. Munnell. Reassessing the Role for Wealth Transfer Taxes. National Tax Journal, 45:119{143, 1992. [2] Orazio P. Attanasio, James Banks, Costas Meghir, and Guglielmo Weber. Humps and bumps in lifetime consumption. Journal of Business and Economic Statistics, 17:22{35, 1999. [3] Alan J. Auerbach and Laurence J. Kotliko . South-Western College Publishing, 1995. 46 Macroeconomics: . An Integrated Approach [4] Gary S. Becker and Nigel Tomes. An Equilibrium Theory of the Distribution of Income and Intergenerational Mobility. Journal of Political Economy, 87:1153{1189, 1979. [5] Gary S. Becker and Nigel Tomes. Human Capital and the Rise and Fall of Families. Journal of Labor Economics, 4:1{39, 1986. [6] Felicitie C. Bell, Alice H. Wade, and Stephen C. Goss. Life Tables for the United States Social Security Area: 1900-2080. Social Security Administration, OÆce of the Actuary, SSA, 1992. [7] Anders Bjorklund and Markus Jantti. Intergenerational Income Mobility in Sweden compared to the United States. American Economic Review, 87:1009{1018, 1997. [8] Christopher D. Carroll. Why Do the Rich Save So Much? 1998. , 6549, NBER Working Paper [9] Ana Casta~neda, Javier Daz-Gimenez, and Jose-Vctor Ros-Rull. Earnings and Wealth Inequality and Income Taxation: Quantifying the Trade-O s of Switching the U.S. to a Proportional Income Tax System. Mimeo, 1998. [10] Javier Daz-Gimenez, Vincenzo Quadrini, and Jose-Vctor Ros-Rull. Dimensions of Inequality: Facts on the U.S. Distributions of Earnings, Income and Wealth. Federal Reserve Bank of Minneapolis Quarterly Review, 21:3{21, 1997. [11] Karen E. Dynan, Jonathan Skinner, and Stephen P. Zeldes. Do the Rich Save More? Manuscript, Board of Governors of the Federal Reserve System, 1996. [12] William G. Gale and John K. Scholz. Intergenerational Transfers and the Accumulation of Wealth. Journal of Economic Perspectives, 8:145{160, 1994. [13] Jagadeesh Gokhale, Laurence J. Kotliko , James Sefton, and Martin Weale. Simulating the Transmission of Wealth Inequality via Bequests. Mimeo, 1998. [14] Pierre-Olivier Gourinchas and Johnathan A. Parker. Consumption over the Life Cycle. NBER Working Paper, 7271, 1999. [15] Gary D. Hansen. The Cyclical and Secular Behavior of Labor Input: Comparing EÆciency Units and Hours Worked. Journal of Applied Econometrics, 8:71{80, 1993. [16] Bengt Hansson. Construction of Swedish Capital Stocks. Economics, Uppsala University, 2, 1989. Economic Studies, Department of [17] Burkhard Heer. Wealth Distribution and Optimal Inheritance Taxation in Life-Cycle Economies with Intergenerational Transfers. University of Cologne, Germany. Mimeo, 1999. [18] R. Glenn Hubbard, Jonathan Skinner, and Stephen P. Zeldes. Expanding the Life-Cycle Model: Precautionary Saving and Public Policy. American Economic Review, 84:174{179, 1994. 47 [19] Mark Huggett. Wealth distribution in life-cycle economies. Journal of Monetary Economics, 38:469{494, 1996. [20] Eric Hurst, Ming Ching Luoh, and Frank P. Sta ord. Wealth Dynamics of American Families, 1984-1994. Mimeo, 1998. _ _ [21] Ayse Imrohoro glu, Selahattin Imrohoro glu, and Douglas H. Joines. The E ect ot TaxFavored Retirement Accounts on Capital Accumulation. American Economic Review, 88:749{768, 1998. [22] Thomas F. Juster, James P. Smith, and Frank Sta ord. The Measurement and Structure of Household Wealth. Mimeo, 1998. [23] Laurence J. Kotliko , Kent A. Smetters, and Jan Walliser. Privatizing Social Security in the U.S.: Comparing the Options. Review of Economic Dynamics, 1999. Forthcoming. [24] Laurence J. Kotliko and Lawrence H. Summers. The Role of Intergenerational Transfers in Aggregate Capital Accumulation. Journal of Political Economy, 89:706{732, 1981. [25] Per Krusell and Anthony A. Smith, Jr. Income and Wealth Heterogeneity, Portfolio Choice, and Equilibrium Asset Returns. Macroeconomic Dynamics, 1:387{422, 1997. [26] John Laitner and Henry Ohlsson. Equality of Opportunity and Inheritance: A comparison of Sweden and the U.S. Mimeo, 1997. [27] Lee A. Lillard and Lynn A. Karoly. Income and Wealth Accumulation Over the Life-cycle. Manuscript, RAND Corporation, 1997. [28] Lee A. Lillard and Robert J. Willis. Dynamic Aspects of Income Mobility. 46:985{1012, 1978. , Econometrica [29] Casey B. Mulligan. Parental Priorities and Economic Inequality. The University of Chicago Press, 1997. [30] OECD Economic Surveys. Sweden , 1998. [31] Council of Economic Advisors. Economic Report of the President. United States Government Printing OÆce, Washington, 1998. [32] Anne-Marie Palsson. Household Risk Taking and Wealth: Does Risk Taking matter. In Studies in the Distribution of Household Wealth. Slottje and Wol eds, 1993. [33] James Poterba. Estate and Gift Taxes and Incentives for Inter Vivos Giving in the United States. NBER working paper, 6842, 1998. [34] Vincenzo Quadrini. Entrepreneurship, Saving and Social Mobility. neapolis Discussion Paper, 116, 1997. 48 Federal Reserve of Min- [35] Vincenzo Quadrini and Jose-Vctor Ros-Rull. Models of the Distribution of Wealth. Federal Reserve of Minneapolis Quarterly Review, 21:22{36, 1997. [36] Statistics Sweden. , 1997. Statistical Yearbook of Sweden [37] Nancy L. Stokey and Sergio Rebelo. Growth E ects of Flat-tax rates. Economy, 103:519{550, 1995. Journal of Political [38] George Tauchen and Robert Hussey. Quadrature-Based Methods for Obtaining Approximate Solutions to Nonlinear Asset Pricing Models. Econometrica, 59:371{396, 1991. [39] Edward Wol . Estimates of Household Wealth Inequality in the US, 1962-1983. Income and Wealth, 33:251{257, 1987. [40] David J. Zimmerman. Regression Toward Mediocrity in Economic Stature. nomic Review, 82:409{429, 1992. 49 Review of American Eco-