Chartistry

I'm having an argument over at Alas (a blog) about trends in government spending over time, in particular spending on social services. Ampersand, the blog's owner, in a post called "Government spending is not up, up, up", claims that government spending has been nearly flat for almost four decades, and defends a commenter's description of the US social system as having been "starved, looted, privatized, and starved some more---pretty much steadily over the last two or three decades."

This has prompted me to do something I've been meaning to do ever since I took the dark oath that secured my place as a Catallarchist: Make charts, and lots of them. There's a lot of material to cover, so I'm going to do this as a series of posts. Today I'll discuss trends in total government spending. All data are from the Economic Report of the President, unless otherwise specified.

Government spending as percentage of GDP

Ampersand is basing his claim that government spending has been relatively flat since 1968 on the fact that spending as a percent of GDP hasn't changed much over the last 20-30 years. The chart on the right (click to enlarge) shows spending as a percentage of GDP, broken down into Federal spending, state spending, and federal grants to states (not double-counted). While one could argue that spending as a percentage of GDP hasn't really gone anywhere since 1980, and even then only because of the (possibly anomalous) decline in the '90s, it's a bit of a stretch to say that it's been flat since 1968.

All of this is beside the point, though, partly because percentage of GDP is the wrong metric to use. If we were discussing the extent to which government spending burdens the economy, this would be a good metric to use (and it's worth noting that government actually isn't consuming a rapidly-increasing share of the economy). But to determine whether or not social programs are being "starved" and "looted" (not that there's anything wrong with that), real per-capita spending is a better metric. As a country grows wealthier, it should be possible to provide a constant level of government services with a smaller and smaller percentage of GDP, so discussing changes in government spending solely in terms of GDP is misleading because it tends to conceal real increases in the level of funding and services provided.

Real per capita government spending

If we look at spending in real per-capita terms (that is, spending adjusted both for inflation and population growth), we get a radically different picture: Rather than a slight, highly cyclical uptrend, we get a steep, straight, and virtually monotonically increasing graph, with spending more than doubling in the last thirty years. I'm tempted to do a send-up of this paper just by replacing the %GDP figures with real per-capita figures.

Of course, the fact that government spending is up doesn't necessarily mean that social spending is. However, as I'll show in the next installment, not only has total government spending increased significantly, but the increase has been skewed heavily towards social spending.

Share this

Could you do a chart of

Could you do a chart of government spending as a percentage of household income?

Yeah, the use of %GDP is a

Yeah, the use of %GDP is a pretty weird statistical strawman. Is there actually anyone who has this figure in mind when they complain about increasing government spending? It's like arguing that, say, iPod ownership is "flat" because the iPod's percentage market share in MP3 players has remained fairly stable, even though the size of the market itself has expanded dramatically. (Just an example, I have no idea about the figures.)

S.D.: I could give it a

S.D.:
I could give it a shot. Could you elaborate a bit on what you mean by that (Per-capita spending vs. mean household income? Median? Total spending vs. total household income?), and what you want to learn from it?

Matt:
To be fair, it was actually a conservative (libertarian?) commenter, and not Ampersand, who first mentioned the %GDP metric, although I can't imagine why. And like I said, it is a useful metric to use when measuring the economic burden of government spending. It's just not a good metric to use when measuring availability of funding for government services.

Goverment expenditures /

Goverment expenditures / number of households / mean household income x 100

Arguing Ampersand says,

Arguing
Ampersand says, "Government spending is not up, up, up." Brandon Berg argues to the contrary, and does the heavy lifting to back up his position. Of course, I recommend you read and judge for yourself....

Good job.

Good job.

an abject moron can observe

an abject moron can observe the growth just by using anecdotal information, and looking around. obviously this guy and miller read the same orwellian gov't pamphlets, and swallowed it whole.

I'm very suspicious of

I'm very suspicious of conclusions drawn from anecdotal data, and I try to check them against more reliable data when possible. Perception of these things is skewed heavily by the media we consume and the people with whom we interact.

More than once, I've been surprised by things I've found out this way. For example, did you know that real per-capita spending on national defense has been more or less flat over the past 50 years?

Brandon, would you provide

Brandon, would you provide some cites regarding per capital defense spending? Thanks!

Sure. I'll post a chart

Sure. I'll post a chart tonight, with citations. But if you have Excel and know how to use it, you can duplicate it in about five minutes. Just head over to the page with the downloadable spreadsheets from the Economic Report of the President (linked above) to get the tables for real GDP, Federal spending by percent of GDP, and population. Then do a bit of arithmetic and chart out the results. It only has real GDP numbers back to 1959, but since then there's been a slight downward trend, with the high point in 1968 and the low point in 1999.

Also, by "flat," I mean that there's been no strong long-term trend. There have been oscillations.