Wednesday, September 09, 2009

Junk stat of the month, in the junk story of the month. 

Yes, the month is young, but we already have a winner in both categories. A twofer! (Hat tip for this one to Andrew Biggs.)

Jacob Weisberg in Slate says Republicans are trying to kill off old people. Really. I thought he was being sarcastic at first but he doesn't stop going on about it -- no, he's angry.

Weisberg lists several ways in which Republicans are committing geronticide (seemingly everything except Soylent Green) but I'll mention just one of them -- Republicans want to drive seniors to kill themselves by destroying Social Security. He expounds (my emphasis)...

Other GOP policies promote death for senior citizens with more modest incomes. Take the conservative push to privatize Social Security, which George W. Bush proposed and failed to get Congress to pass in 2005.

Social Security has driven life expectancy up and death rates down since it was instituted. It has an especially pronounced impact on suicide rates for the elderly, which have declined 56 percent since 1930.

Had Bush prevailed, we would now be undoing income security for the elderly. Those who gambled on the stock market and lost would be less able to afford medicine, food, and heating for their homes. In aggregate, they'd presumably die younger and commit suicide more often.

Let's put aside some simple facts that Weisberg apparently can't deal with:

That the Bush proposal, far from "undoing" Social Security, rather more modestly proposed to close its $15 trillion funding gap by gradually reducing benefits for the rich. (Does Weisberg prefer closing the gap by increasing taxes on workers to pay benefits to the rich? So Warren Buffett's employees at Dairy Queen will pay tax increases to keep his benefits whole?) ....

That if the Bush proposal had been in force, even after the stock market crash, workers would have received larger benefits than from the status quo ....

And that when today's workers get to retirement age and find their benefits are underfunded by about 25% at a time when taxes can't be raised to pay them (because income taxes have already been raised by 50%), well, hey, what effect is that going to have on their suicide rates?

No, let's put all such issues aside and look at the single claim: enacting Social Security reduced suicide rates among seniors.

Weisberg's only authority for this is a link (indirectly) to a National Bureau of Economic Research working paper by David Cutler and Ellen Meara on the larger subject of mortality rates during the 20th century (of which suicide is such a minor part that it isn't even mentioned in the abstract).

Cutler and Meara say a large decline in suicide rates among the elderly occurred after 1930 and was "coincident" with the increase in income seniors received from Social Security. True enough, but we all know correlation isn't causation (for instance). And as Andrew tells us... "Cutler and Meara don't run any statistical tests themselves; they draw their conclusions entirely from their Figure 14."

Figure 14

OK, now take a look at that chart for a moment. What jumps right out at you from it saying "Ain't no proof of claim here!"

The minor thing is that while a big drop in suicide rates followed 1930, the first monthly benefit check wasn't paid by Social Security until 1940. Though that certainly raises some questions by itself.

The major thing is that the drop in the suicide rate occurred simultaneously across all age groups -- among 65+ers, and those 55 to 64, 45 to 54, 25 to 44, and even 15 to 24(!) -- from the same 1930 peak, in the same pattern for 25 years, only smoothed out somewhat for the lower age groups.

Is the claim really that in 1940 Social Security benefits paid to seniors significantly reduced suicide rates among 30-year olds? (And that in fact Social Security reduced suicides among 30-year olds years before any benefits to seniors were paid at all?)

That seems to stretch credulity (nearly as much as Soylent Green). In which case we should be looking for another cause of the general and simultaneous decline in all suicide rates, across all age groups, in the years following 1930.

And Weisberg should rework his rhetoric and try again.