First, the obvious: I can't make a 0.1% increase per decade from the late-mid '80s to around 2008 go away. It's not 0. It's not huge, but it's not 0.


Second, the less obvious. Take that spike in 2008. Even if the $ amount of entitlement spending had held absolutely flat, Krugman's chart would have shown a spike. With an increase in spending, the spike is larger.

Now look at the mid-late '90s. Two things were happening. Welfare reform (reforms tend to flatten growth, but then in hindsight they're ignored as it's confidently claimed that there would have been no growth). And, for that chart more importantly, the GDP increased by something like 4% per year on average (a bit more, actually). So if entitlements had increased by 3% per year, adding up to a 13% increase by the end of the 4th year, that chart would still show a decrease. In other words, unless entitlement growth exceeds GDP growth, the trend line has to be down. It's not. Entitlement growth has exceeded, on average over the period from '89 to '08, GPD growth by a small amount.


Critical thinking requires that I consider the likely counterarguments. (It's easy to win an argument with yourself. You always go unchallenged because you're both the choir and cheering on victories over the defeated serried ranks of strawmen.)

The cleanest argument would be population growth. From '89 to '08, population growth was around 23%. US GDP growth was upwards of 35%. GDP > population growth means that the trend line was flattened on this chart, and considerably so. The sharp downturn in the late '90s comes not only from welfare reform, but was also helped by the GDP's increasing at twice the rate of population growth.

A messier argument involves demographics--but this wouldn't help Krugman's argument. It would merely say, at best, that any increase in entitlements is appropriate since the cohorts to be helped by the entitlements increased--children, for instance. But childhood poverty decreased from 1989 through 2008 (more so through 2007--2008 saw a sharp uptick). In other words, instead of "there was no increase" the argument would be "of course there was an increase, and properly so." That he makes this argument for times of recession and doesn't consider possible structural changes in society and the economy goes to his arguing a pettily partisan point: A quick ego-boosting win at the expense of the truth is more important than truth at the expense of a quick ego-boosting win. (This, btw, is pure Krugman. He writes for his audience. Yes, this is an insult and should be taken as such.)

After this messier argument we get into the underbrush where nasty-to-wage arguments breed in their warrens: % of population that's under 18, appropriateness of measures of poverty , family structure changes, "hidden" entitlements that aren't perceived by the recipient as cash but count against a funding source as cash, appropriateness of GDP measures, appropriateness of the CPI, etc.