Thursday, February 23, 2012

Spending on presidential elections

On Monday, Dave Gilson produced this wonderful graph of spending by presidential candidates since 1860, in constant 2011 dollars. It suggested that we really have seen a departure in the last two cycles, with 2004 being the most expensive race up until that point, and 2008 burying 2004. Jonathan Bernstein points out that inflation-adjusted dollars really don't tell the whole story, since the country has been growing. So I've divided Gilson's cost figures by the number of votes cast* in each election (as recorded by David Leip), producing the graph below:
What surprised me is the stability of the figures over time; in 12 of the 18 elections, the campaigns spent between $2 and $5 per vote cast, in 2011 dollars. That includes most of the elections before the early-70s reforms, suggesting that, contra Jon's point, the reforms did not force campaign spending to artificially low levels. 2004 did see an unusually high level of spending, but not quite as much as was spent in 1968. And 2008 really did see a lot of spending, although 1896 still holds the record.

Like Jon, I don't see any particular problem with increased levels of campaign spending. Most of that spending consists of voluntary donations by wealthy Americans financing ads that inform middle-class Americans about presidential candidates. I fail to see the harm to the Republic.


*Jon suggests that it might be better to measure money spent per eligible voter, rather than per vote cast, since it is the eligible voters that the campaigns are trying to reach. I agree, but figuring out the eligible electorate in many of these years is a tad tricky. Millions of Americans were legally eligible to vote from the 1870s to the 1960s but nonetheless faced taunts, beatings, or death for trying to do so.

Update: See here to see campaign spending as a function of real GDP.

8 comments:

Peter said...

Very cool chart. One thought: between the 1896 election and 1908 election the Tillman Act was passed to prohibit corporations from contributing to campaigns. Mark Hanna was famous for raising large amounts of money from them. The graph suggests the law may actually have been fairly effective in restraining spending.

Seth said...

Good point, Peter! It's hard to know from these data, though, whether 1896 was unusual for the era or whether spending sharply dropped right after it.

Dave said...

Interesting, but I think the measure is largely meaningless, because it doesn't capture non-campaign expenditures (e.g. op-eds, ideologically-driven news stories, or talk radio) which presumably would vastly exceed other expenditures on political speech.

Anonymous said...

Eligible voters is the way to go because turnout varies.

Douglas Knight said...

We're a lot richer than 100 years ago. An inflation-adjusted dollar bought a lot more labor back then. I prefer percent of GDP, as on page 55 of this pdf. This addresses the increasing population, though not the increasing voter turnout.

Jim Glass said...

Inflation adjusting is a terrible way to measure campaign expenditures — it totally ignores growth of national wealth and income, which matter *a lot*, Mr Knight is right about that.

Expenditures per voter is somewhat better, as it partially adjusts for growth -- but it too fails to reflect the increase in income and wealth.

As the nation becomes richer, people can both (1) more easily afford to spend more money on elections, just as they can spend more on everything else, and (2) gain more from doing so, as there is more for the government to direct in their favor, if they can get it to do so.

On its face this would suggest that campaign spending would be expected to increase steadily as a portion of national income.

Yet when adjusting the numbers for GDP, it is remarkable that this *hasn’t* been so.

By % of GDP, the most expensive presidential campaigns:

Yr ......% of GDP
1896 ... 0.0434%
1968 ... 0.0094%
2008 ... 0.0092%
1932 ... 0.0083%
1908 ... 0.0081%
1964 ... 0.0062%
2004 ... 0.0061%
1896 ... 0.0054%
1972 ... 0.0051%

The expected trend of an increasing portion of income being spent on elections is totally lacking.

Which brings to mind a question many political scientists ask -- with such a *vast* amount at stake in a $15 trillion economy, why is so *little* spent on electioneering ... with that amount not even growing in terms of real income?

Seth said...

Cool numbers, Jim. Am I correct that the second appearance of 1896 is actually 1996?

Eric said...

I would like to use your graphic here in a round up where I quote your analysis and link to your post. May I? Let me know at eschulzke [at] desnews [dot] com. Thanks!