This paper examines US trends regarding the intergenerational transmission of economic status, also referred to as intergenerational mobility, meaning the relationship between an individual's socioeconomic status and the status of his/her parents. More specifically this paper seeks to observe any changes in trends regarding this relationship over the past couple of decades. This study leads Lee and Solon (LS) to find no significant evidence of major changes in trends regarding intergenerational mobility in the US.
The data for this study is gathered from the Panel Study of Income Dynamics (PISD), which is administered by the University of Michigan’s Survey Research Center. The PISD is a survey that started in 1968 and consists of information from about 5000 different US families. The sample has been re-interviewed every year up to 1997 and every other year since then. Because the survey tracks an individual from childhood to adulthood, it allows the authors to examine how income is affected by parent income status during childhood.
This analysis considers children born between the years 1952 and 1975. When the children turn 25 their income over that year is the income amount that is compared to the income of their parents during childhood. The final sample of sons contains 11,230 observations for 1,228 individuals, and the sample of daughters contains 12,666 observations for 1,308 individuals.
To gather information on the parents’ economic status to use as comparison, an average yearly income was calculated using the family’s salary for each year while children were between the ages 15-17. For example, the average family income of a child born in 1952 would be the average of the yearly salary from 1967-1969.
The income variation among all the children and their parents is examined, first for sons then for daughters. After plotting the sons’ elasticity measurements (The measurements of variation in income between children and parents) in a time series, the data shows there is no clear trend; meaning intergenerational elasticity has stayed constantly volatile over the years. The same is observed when looking at the data gathered for daughters.
LS then look solely at the data for children who are 30 years old. The rest of the sample is discarded for this part of the analysis, because it is possible that differences across the age groups are masking some trends in certain age groups. However, the authors still observe wildly volatile estimates for intergenerational elasticity. They attribute these findings to sampling error, created by throwing out the rest of the data, rather than extreme volatility amongst 30 year olds.
LS make one more attempt to identify any changing trends in income mobility. They consider the average results in 5-year increments. For example they take the average income of those born from 1952-1956 then those from 1953-1957 and so on. The purpose of this is that it averages out the noise from the sampling error. Looking at the data in this way, again, displays no obvious time trend in intergenerational mobility.
Overall, LS’s analysis yielded measurements of variation in income that were too imprecise to rule out modest trends in either direction. Trends in intergenerational mobility were not found to be increasing or decreasing. Their results suggest that intergenerational income mobility in the US has not dramatically changed over the past two decades.
There are many aspects of this analysis that limit the reliability of the authors’ results. First of all, income is not the best proxy for living, because it is common that very rich families, who are most likely give money to their children when they get older, have relatively low incomes. It is not the very rich that this study is examining but rather the high-income earners. There is a big difference. I’m sure there is a trend between the very rich that have trust funds established for their children, and the socioeconomic status of their children when their children grow older.
Also, measuring children’s income when they are 25 is not an accurate measure of their socioeconomic status. Many “children” this age are still in college, or just starting their careers. A better way to measure their income would be to cut down the sample size and calculate an average income for the children when they are between the ages of 25 and 35. This would give a more accurate representation of their socioeconomic status.
This paper was very short and did not go into enough detail about how the analysis was carried out. Many references are made to the data tables, but not enough information is given as to what the numbers mean.
The authors don’t examine if there are trends specific to certain demographics, such as children of those in the top tax bracket. I believe that more research is necessary in order to identify if there are any changing trends among certain demographics.