In order to examine the impact of inheritances on wealth over time, it is necessary to look at household wealth over time. The challenge with examining households is that many of them shift and change: singles marry, married couples divorce, divorced people remarry. While families are experiencing these changes, wealth is being shifted around. This makes it difficult to tease apart the independent contribution of inheritances to wealth. Using the Panel Study of Income Dynamics (PSID), this analysis looks at the impact of inheritances on wealth over time. In order to overcome the challenges of families forming and dissolving and reforming, we analyze a restricted set families where the same married couple heads the household over the entire study time.
Since 1984, the PSID has collected data on household wealth and whether or not a household has received a lump sum or inheritance. We define household wealth or net worth as a household’s total assets less their total indebtedness. The PSID collected wealth and inheritance information in 1984, 1989, 1994, 1999, 2001, 2003, 2005, 2007, 2009, 2011, and 2013. The PSID asks respondents to report up to three inheritances of at least $10,000 received after the last wave of the survey. From this, we can construct a complete picture of large inheritances that people receive from the period 1984 onward. In 1984, people are asked to report on all the inheritances they’ve received up to 1984, but because it might be difficult for people to remember the timing and amount of large gifts a long time in the past. For this reason, we do not consider any gifts received prior to 1984.
I restricted the sample to only the married heads of households in 1984 who, in 2013, are married to the same person. The PSID imputes wealth when the various components of wealth are missing. There was 1 case that had missing wealth data in 1994, this household was dropped from the analysis in that year. This leaves a sample of 867 households, with a total of 9,464 observations over all of the time points. These families are not observed at every time point between 1984 and 2013. In total, 73 households are not observed at one of the time points. Out of these 867 households, 500 (57.6%) do not inherit over this time period, and 367 (42.3%) do inherit. At baseline, there are significant differences between these two groups, as displayed in Table 1 below.
Table 1: Descriptive baseline (1984) statistics for inheriters and non-inheriters (1984-2013)
Stratified by Lifetime Inheritance
No Inheritance Inheritance p test
n 8678 7671
wealth2adj (median [IQR]) 113275 [44644, 284314] 143235 [67085, 337001] 0.018 nonnorm
wealth1adj (median [IQR]) 39189 [12919, 145684] 44952 [15675, 155239] 0.211 nonnorm
inchhadj (median [IQR]) 79751 [54422, 108428] 87828 [61555, 122340] 0.005 nonnorm
agehd (median [IQR]) 38 [30, 47] 36 [32, 42] 0.071 nonnorm
agesp (median [IQR]) 36 [29, 44] 35 [30, 39] 0.164 nonnorm
racehd_rec (%) <0.001
White 7806 (90) 7341 (96)
Black 579 ( 7) 39 ( 1)
Other 293 ( 3) 291 ( 4)
racesp_rec (%) <0.001
White 7788 (90) 7401 (96)
Black 555 ( 6) 39 ( 1)
Other 335 ( 4) 231 ( 3)
educhd (%) <0.001
LessHS 1533 (18) 297 ( 4)
HS 3022 (35) 2185 (28)
SomeCollege 1990 (23) 1761 (23)
Bachelor 1441 (17) 2487 (32)
Graduate 627 ( 7) 938 (12)
educsp (%) <0.001
LessHS 1316 (15) 217 ( 3)
HS 4355 (50) 3197 (42)
SomeCollege 1530 (18) 1796 (23)
Bachelor 1269 (15) 1881 (25)
Graduate 190 ( 2) 562 ( 7)
empstathd_rec (%) 0.021
Employed 8042 (93) 7370 (96)
Unemployed 342 ( 4) 69 ( 1)
Retired/Disabled 222 ( 3) 101 ( 1)
Other 72 ( 1) 131 ( 2)
empstatsp_rec (%) 0.859
Employed 5273 (61) 4855 (63)
Unemployed 4 ( 0) 0 ( 0)
Retired/Disabled 71 ( 1) 35 ( 0)
Homemaker 3242 (37) 2700 (35)
Other 88 ( 1) 81 ( 1)
Note: n's are weighted
At baseline, inheriters had higher levels of total wealth, wealth derived from financial assets, and household incomes. Inheriting households were more often white, and had heads and spouses who were more likely to have gone to college.
Looking at the inheriters versus the non-inheriters in terms of overall wealth progression, a clear pattern emerges. For all levels of the net worth distribution, inheriters start out with a small net worth advantage, but end up with much higher net worth than non-inheriters. The difference in both initial and final wealth of inheriters versus non-inheriters increases as you go up the net worth scale.
Because the characteristics of inheriters are different than non-inheriters and because the starting points of inheriters is larger than non-inheriters, it is important to determine how much of the difference in net worth is attributable to inheritance and how much is attributatble to these socio-economic factors. Mixed effects quantile regression is employed to determine the difference in the change curves of inheriters vs. non-inheriters when age of head, race of head, employment status of head, employment status of spouse, educational status of head, household income, and wealth in 1984 is added to the model. Because wealth in 1984 is added as a control, the regression is run on all observations from 1989-2013. Controlling for wealth in 1984 allows for both groups to have a similar starting point.