(This essay was published in Hong Kong Economic Journal on 30 April 2014)
There is a general perception that income inequality has increased and intergenerational income mobility has declined in many parts of the world. Intergenerational income mobility refers to a child’s chance of moving up in the income distribution relative to his or her parents. Most people intuitively believe that income inequality and intergenerational income mobility must be related – that if intergenerational income mobility is declining then income inequality must be increasing over time. What is the evidence from around the world? And what is the evidence in Hong Kong?
Many countries have collected empirical data on the dispersion of income across individuals and households through surveys and population censuses, so we know quite a bit about what has happened to income inequality over time. The evidence shows that measured income inequality has increased over the past 30 years in many societies that have experienced fast economic growth. In general, household income inequality has risen faster than individual income inequality over this period. This is true of the rich societies in North America and Europe, but also among the emerging economies of the less developed world. The same thing has happened in Hong Kong (see Figure 1).
Research and Empirical Evidence
Much less is known about intergenerational income mobility over the same period because data linking the income of children to their parents in these societies is not so readily available. A recent study published in 2014 by a team of economists from Harvard University, the University of California at Berkeley, and the US Department of Treasury has provided the best empirical study of intergenerational income mobility in the US, combining administrative records, population censuses and surveys to link data across generations (see Chetty et al., 2014 – I shall elaborate on their main results below).
In Hong Kong, the evidence is mixed. A research study of intergenerational education mobility in Hong Kong by Professor Chou Kee-lee ( 周基利教授 ) appeared in news reports 15 months ago, showing that access to higher education opportunities had deteriorated between 1991 and 2011. This finding was considered the first piece of empirical evidence on the intergenerational transmission of poverty in Hong Kong, focusing on education mobility for individuals aged 19 and 20 years old.
However, as I report below, I have found data from the population censuses from 1976 to 2011 to tell a different story: there is no compelling evidence to show that intergenerational education mobility has declined over time in Hong Kong.
US Income Mobility Unchanged
First, let me turn to the study by Raj Chetty, et al. to look at the evidence in the US and the methodologies used. They analyzed trends in mobility for the 1971-1993 birth cohorts using administrative data on more than 50 million children and their parents. Their main results measured mobility by estimating (1) the correlation between parent and child “income percentile ranks” and (2) the probability that a child would reach the top fifth of the income distribution conditional on his/her parents’ income quintile.
It is important to note that Chetty, et al. use “income percentile ranks” to measure intergenerational income mobility instead of income directly. The parents’ income is compared with the income of others in their own generation and is ranked on a percentile basis from 0 to 100. Likewise, the income of the child is ranked within his or her own generation. Mobility is based on the relative positions in the income distribution, for example, a child’s prospects of rising from the bottom to the top quintile.
This approach avoids the measurement problems that arise when average incomes increase over time or income inequality in the population increases. For example, if one defines mobility based on the probability that a child from a low-income family (e.g., the bottom 20%) reaches a fixed upper income threshold (e.g., $100,000), then mobility by this measure would increase over time if average income keeps rising. Conversely, if income inequality increased over time, the difference in expected incomes between children born to low (e.g., bottom-quintile) versus high (top-quintile) income families would be magnified and mobility would fall because a child’s income depends more heavily on her parents’ position in the income distribution today than in the past.
Intergenerational mobility measured in “income percentile ranks” removes all these problems and is therefore a much better measure of relative intergenerational income mobility over time.
Using these rank-rank measures, Chetty et al. found intergenerational mobility has remained unchanged in the US in recent decades. The relationship between parent and child percentile ranks for income has been highly stable.
Their empirical estimates showed that the quantitative effect of parents’ “income percentile rank” on a child’s “income percentile rank” (the estimated rank-rank correlations in intergenerational education mobility) was unchanged at around 0.34 over time. This effect can be interpreted as the difference in the mean percentile rank of children from the richest families versus children from the poorest families – in other words, the incomes of children from the richest families are about 34 percentile ranks higher than those of children of the poorest families.
They also provide evidence of quite a large degree of mobility – which, importantly, runs counter to the public perception that intergenerational mobility has deteriorated in recent decades. The chances of moving from the bottom fifth to the top fifth of the income distribution were no lower for children entering the labor market today than they were in the 1970s.
For example, the probability that a child of parents in the bottom fifth of income redistribution would reach the top fifth was 8.4% for children born in 1971, compared with 9.0% for those born in 1986. Children born to the highest-income families in 1984 were 74.5 percentage points more likely to attend college than those from the lowest-income families. The corresponding gap for children born in 1993 was 69.2 percentage points, suggesting that if anything, intergenerational mobility may have increased slightly in recent cohorts.
However, they have also found that while rank-rank measures of mobility remained stable, income inequality has increased over time. This is consistent with previous research on income inequality trends in the US. Income inequality as measured by both the dispersion of individual incomes among young adults aged 25 or 30 years old, and the relative income gap between graduates with a university degree and those with only high school diplomas, has increased over time. In other words, the economic consequences of the “birth lottery” – the parents to whom a child is born – are larger today than in the past.
A useful visual analogy for these contrasting findings is provided by Chetty et al. They ask us to think of the income distribution as a ladder, with each percentile representing a different rung (see Figure 2). The rungs of the ladder have grown further apart (inequality has increased), but children’s chances of climbing from the lower to higher rungs have not changed (rank-rank mobility has remained stable).
This result may appear to be surprising to those who intuitively believe that inequality and mobility must be correlated. However, much of the increase in income inequality in recent decades has come from the extreme upper tail, such as the top 1%, a group not typically associated with mobility. In fact, growing evidence shows that mobility among the top 1% is very high even within the same generation. Individuals tend to drop in and drop out of the top 1% within a few years.
Chetty et al. further showed that upward mobility varied considerably across different geographical areas in the US. They discovered that five factors accounted for the differences in mobility: segregation, inequality, quality of education, social capital, and family structure.
Hong Kong Education Mobility Unchanged
Reliable income data for estimating upward mobility in Hong Kong is not available, as I do not have access to administrative records like those used by Chetty et al. As the next best approximation I have used data on years of schooling to link parent and child and estimate upward mobility. I construct, in the same spirit as Chetty et al., “schooling percentile ranks” of parents and children in their own generation to measure the relative economic position of an individual. Economists have long learned that schooling is a very reliable measure of a person’s future income.
To estimate the rank-rank correlations in intergenerational education mobility, I performed an analysis of all young men and women between the ages of 25 and 29 who lived with their parents over the period 1976-2011, using the population census datasets. The empirical estimates are quite similar to those obtained by Chetty et al. (see Figure 3).
Figure 3 shows that the estimated rank-rank correlations in intergenerational education mobility have declined rapidly over time for the cohorts of children born in each of those periods during 1946-61 from 0.441 to 0.327. This implies that although upward mobility was low at the beginning of this period, but it was improving very rapidly over the one and a half decade. The estimates for the cohort born in 1971-86 is quite stable at around 0.31-0.32, which is about the same level as that for the cohort born in 1956-61 at 0.33. This suggests that the cohorts born during 1971-86 have experienced the same amount of upward mobility as those born in 1956-61. Furthermore the magnitude of these estimates is quite similar to those obtained by Chetty et al. for the US.
The estimated rank-rank correlations in intergenerational education mobility for the cohort born in 1961-71, on the other hand, improved dramatically as the estimates fell to 0.255 and 0.259, respectively. Why did this happen? There were two factors. First, there was considerable emigration of the middle class in the 1990s leading up to 1997. The total number of persons who emigrated is estimated to be 416,500. This opened up education opportunities for those who stayed in Hong Kong. Their access to university places increased, especially for children whose parents had lesser schooling. Second, opportunities for university places expanded rapidly in the latter half of the 1980s as the government began to expand university places, in part as a means to restore confidence in Hong Kong in the face of emigration of the middle class.
It is reasonable to surmise that the dramatic improvement in upward mobility for lower income families between 1986 and 1996 was a temporary phenomenon – a consequence of less competition for school places and, especially, university places. This period is now behind us. The significance of having such an unusual period of opportunity is that we have to reinterpret the previous finding by Professor Chow Kee-lee. Rather than conclude that upward mobility deteriorated in Hong Kong between 1991 and 2011, it is more appropriate to conclude that the cohort born during 1971-81 had an exceptionally good run of luck due to high emigration rates and the expansion of tertiary education. Subsequent cohorts did not have such good fortune.
A more plausible interpretation of upward mobility in Hong Kong is that it improved dramatically for the cohort born in 1946-56 and remained stable afterwards. The cohort that was born in 1961-71 lucked out as upward mobility received a huge boost due to middle class emigration in the lead up to 1997. Since then upward mobility in Hong Kong has returned to the same stable level.
Raj Chetty, Nathaniel Hendren, Patrick Kline, Emmanuel Saez, and Nicholas Turner, “Is the United States still a land of opportunity? Recent trends in intergenerational mobility,” Working Paper 19844, National Bureau of Economic Research, January 2014
Raj Chetty, Nathaniel Hendren, Patrick Kline, and Emmanuel Saez, “Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States,” Working Paper 19844, National Bureau of Economic Research, January 2014
Nineteenth essay in the series on Rekindling Hong Kong’s Magic