"Growth Has Been Good for Decades. So Why Hasn’t Poverty Declined?" @Neil_Irwin

Dependency Rate May Explain Growth/Poverty Disconnect

Income Inequality

Lynn Chapman
June 5, 2015

Neil Irwin's question about economic growth and poverty rates may have a fairly simple yet potentially alarming answer. His chart (below) shows the actual poverty rate declining sharply from the late 1950's until the mid-1970's. The poverty rate then roughly flattened for the next 30+ years despite economic growth. He quotes several knowledgeable sources who link economic growth to reductions in poverty, and asks "Growth Has Been Good for Decades. So Why Hasn’t Poverty Declined?"

Obviously the data suggests that economic growth does not automatically translate into lower poverty despite the adage that "a rising tide lifts all boats". So what is missing in the linkage between economic growth and poverty rates?

The answer may well be a simple factor called the dependency ratio. The Census Bureau calculates the dependency ratio based on age groups in the population. This dependency ratio is based on the assumption that the young (under 18 years old) and the old (over 65 years old) tend to consume more than they produce and are therefore "dependent" on the output of those between the ages of 18 and 65 for the bulk of their consumption. The Census Bureau then calculates this ratio based on the age cohorts across the entire population, over time.

In comparing the Census Bureau dependency ratios (below) with the actual poverty ratios (above), we find an interesting correlation over time. During the general period when the poverty rate was falling sharply, the dependency ratio was also falling sharply. Then during the period when the poverty rate was relatively stable, the dependency ratio was also relatively stable. 

This is not a particularly surprising outcome. Dependency ratios are calculated on a national level, but they exist in the real world on a household-by-household basis. Just like poverty. Poverty is more common among households with a high dependency ratio. For example a household made up of a single mother with three small children - a higher dependency ratio household - is much more likely to be living in poverty than a household of two fully employed adults with two children - a lower dependency ratio household. As the dependency ratio declines nationally, one would expect that the poverty rate would decline as well. This would explain at least some of the decline in the poverty rate during the 1960's and early 1970's.

The stability of the poverty rate between the mid 1970's and about 2010 - a period of time when the dependency ratio was also relatively stable - may suggest that economic growth doesn't necessarily impact the poverty rate. It may be that a rising tide lifts all boats PROPORTIONATELY. Which is to say the poor may be better off than they were before, but the percentage of the population living in "poverty" may remain the same despite economic growth at the aggregate level. 

The alarming part of this line of thinking is that if the poverty rate is more closely tied to the dependency ratio than it is to economic growth, then we have already entered an unavoidable period of rapidly increasing poverty rates. As shown in the chart above, the dependency ratio has just begun a period of rapid increase that will continue for the next 15-20 years. These numbers are not conjecture, they are simple math based on the current age profile of the population. There is a chance that the birth rate could drop faster than forecast in these numbers thus putting downward pressure on the dependency ratio, but the increase in the dependency ratio is due primarily to Old Age rather than to Youth. The fact is, this forecast of a sharply rising dependency ratio has a low margin of error over the next 15-20 years. 

One saving grace in this otherwise alarming picture is that the dependency ratio is based on age and not income. An obvious antidote to sharply rising poverty rates is for the young to start working sooner, and the old to continue working longer. Both of these trends will likely happen to some extent, but only if the jobs are available. One risk is that if job growth does not increase faster than the dependency ratio, the old will merely displace the 18-65 year old workers and the effective dependency ratio will not improve even as the old retire at a later age. 

Neil Irwin's question is one of the more pivotal questions facing economic and social policy makers today. Demographic changes have the power to trump almost all economic policy initiatives - just ask Japan and many of the other developed countries. And elevated poverty rates carry with them a tremendous social as well as economic cost to society - just ask the many developing countries who are trying to create a strong middle class and a strong economy. 

Does anyone want to do a back-of-the-envelope calculation of what will happen to our poverty rate if the loose historical correlation between poverty rates and the dependency ratio holds true over the next 15-20 years of the Census Bureau forecast?