Friday, November 21, 2008

Van de Water: Immigration and Social Security

Paul Van de Water of the Center on Budget and Policy Priorities writes on the affect of increased immigration on Social Security solvency. Here are the key findings:

  • Increases in immigration tend to improve Social Security's financial status, and decreases in immigration tend to worsen it.
  • More immigration would likely eliminate only a small portion of Social Security's long-term shortfall.
  • The impact of immigration on Social Security's finances is modest and should not be a major factor in setting either immigration or Social Security policy.

The whole piece is well worth reading. Paul was the Assistant Deputy Commissioner for Policy at the Social Security Administration at the time I was the Associate Commission for Retirement Policy, which pretty much made Paul my boss (and a good boss, smart analyst and overall great guy he is).

I don't disagree with Paul's conclusion that immigration has a modest positive impact on Social Security financing, although I also wouldn't be surprised if in the future we concluded the positive effects were modest or even slightly negative. The reason is that modeling of immigration's effects on Social Security is still a work in progress, so I don't feel 100 percent confident in the results currently out there.

Paul notes that immigrants differ from native born in a number of ways – earnings, fertility, mortality, and so on. Unfortunately, the current modeling of immigration by the Social Security Trustees doesn't really capture these differences: immigrants are assumed to have the same earnings and life expectancies as native born, so in a sense they're treated as equivalent to an increase in the birth rate even when they might not be.

CBO's modeling of immigration is better, in that their model better captures the earnings of immigrants – a large number of low earners coupled with a small number of high earners.

A big issue that both SSA and CBO are now better capturing is the effect of immigrants who come to the U.S. but later leave. Many of these folks will leave "stranded taxes," meaning they paid taxes while here but won't be around to collect benefits. I suspect SSA somewhat overestimates the positive effects of these immigrants, since assuming they have the same earnings as native born will both reduce the rates of return of those who stay through retirement, as well as increasing the amount of stranded taxes from those who return home and don't collect benefits.

Here's a general analytical crutch I use when thinking about immigration and Social Security financing: imagine that immigrants and their offspring (kids, grandkids, etc.) formed a separate Social Security system in which they paid taxes and later collected benefits. For immigration as a whole to benefit Social Security financing, this separate Social Security program for immigrants need to be more than solvent, collect more in taxes than in pays in benefits, even over the long term. Initial immigrants are likely money losers, since they tend to have truncated careers, low earners, and longer life spans. If these folks receive internal rates of return in excess of the trust fund bond rate – around 3 percent over inflation – this implies that they pay less into the system than they collect from it (in present value) and, as individuals, they probably hurt solvency. But immigrants tend to have more kids than native born, meaning that they generate additional tax payers in the future. The question here is what the next generation looks like, in terms of their earnings, fertility, disability incidence, mortality, and so on. Assuming the first generation of immigrants may be money losers, this next generation needs to be money producers for the program. I don't know what the answers are here.

In any case, I think Paul basically has it right, but there's more to be done on the technical side and I wouldn't be too surprised to see some slightly different results in the future.


No comments: