Psst, Mr. President, want to get re-elected? Here’s an idea: how about big increases in federal spending in key swing state counties, especially those where you are underperforming?

This advice comes courtesy of a new paper by Boston University political scientists Douglas L. Kriner and Andrew Reeves. Their American Political Science Review article,  “The Influence of Federal Spending on Presidential Elections,” finds that, “voters reward incumbent presidents (or their party’s nominee) for increased federal spending in their communities.”

Looking at presidential elections from 1988-2008, the scholars also find that the ballot box bonus for increased county-level federal spending is larger in swing states than in non-competitive states, and pro-government liberals and moderates are more responsive to federal spending boosts than anti-government conservatives, regardless of the incumbent candidate’s party.

More findings soon, but first a quick two-paragraph public service announcement about the data:

Of particular interest to us here at Sunlight is the fact that Kriner and Reeves used data from the Consolidated Federal Funds Reports (CFFR). As our guest blogger Becky Sweger noted last month here on the Sunlight Foundation blog, the U.S. Census Bureau will no longer be publishing the CFFR, which is a major loss.

Though USASpending.gov will continue to report most federal spending (We have a complete analysis at our Clearspending.org website), the loss of a consistent consolidated format means that analyses like these will be harder to do in the future. The loss of CFFR also means an end to several sources of federal spending data, which means less accountability for government spending.

Now back to the findings.

The effect of additional federal spending is not huge, but in a swing state, every little bit counts. Each 1% increase in federal spending yields an increase of 0.0134% of the vote share in competitive states (defined as those where the losing party has averaged at least 45% of the vote share in the three preceding elections) and just 0.00493% of the vote share in non-competitive states.

But if say, you doubled federal spending in a swing-state county, that translates to a 1.13% increase in the vote share, and with the same decline in your opponent’s share, that’s a 2.26% advantage. In the data, though, such an increase is a relatively rare occurrence – a two standard deviation increase is an 80% change. But all it could take is a few of these to flip a swing state.

Another interesting finding is that the electoral bonus for local spending is greatest in counties where the local representative and both senators are of the same party of the incumbent president. As the authors note, “In counties represented by presidential co-partisans, partisan responsibility for a new or expanded federal project is clear and voters reward the president with considerable additional support.”

Also interesting are the partisan effects. Conservatives profess to dislike government, and they are true to their words. They don’t reward presidential spending, regardless of party. Liberals and moderates reward both Republicans and Democrats for government spending.

As Kriner and Reeves note, political scientists have had a hard time showing that members of Congress are rewarded electorally for bringing extra funding to their districts. One review article described a “pattern of non-findings” in measuring the electoral benefit of getting home district earmarks. So this research helps to answer a puzzle: it’s presidents, not members of Congress, who get rewarded for earmarks.

This paper does not investigate whether presidents target federal funding to key counties in swing states for political reasons. There is, however, some interesting research examining political considerations in how funds from the American Recovery and Reinvestment Act were distributed (see: “The Distributive Politics of the Federal Stimulus,” by Jim Gimpel, Frances Lee, and Becca Thorpe). Another interesting paper (“Transparency, Political Polarization, and Political Budget Cycles in OECD Countries,” by James E. Alt and David Dreyer Lassen) finds evidence that incumbent governments in highly polarized countries increase expenditures in election years and decrease expenditures in non-election years, and that incumbent governments in high transparency countries do not run up deficits prior to elections.

Kriner and Reeves are cautious in the implications of their findings: “Republican presidents may continue to reject federal largess in the abstract but, like President George W. Bush, extol specific projects…Democratic presidents with a base ideologically predisposed to approve of increased federal spending may also have incentives to spend generously.”

This is compelling research. And yet another reason why it’s a shame that the CFFR won’t be around in the 2012 election to inform continued research on the effect of federal spending on county voting patterns, especially in the wake of the stimulus bill.