Doug Hibbs sez: It’s a good year for the Democrats

Doug Hibbs writes:

Presidential election outcomes are well explained by just two objectively measured fundamental determinants: (1) weighted-average growth of per capita real personal disposable income over the term, and (2) cumulative US military fatalities owing to unprovoked, hostile deployments of American armed forces in foreign conflicts not sanctioned by a formal Congressional declaration of war. At the end of 2007 weighted-average growth of real incomes during Bush’s second term stood at 1.1 percent per annum. If the same performance were sustained for the rest of the term it might barely suffice to keep the Republicans in the White House, other things being equal. However the economy slid into recession at the beginning of the year and per capita real incomes will most likely decline throughout 2008. Moreover, by Election Day cumulative US military fatalities in Iraq will approach 4,500 and this will depress the incumbent vote by more than three-quarters of a percentage point. Given those fundamental conditions the Bread and Peace model predicts a Republican two-party vote share of 46-47% and therefore a comfortable victory for the Democrats in the 2008 presidential election.

Here are the basic data from Hibbs’s bread-and-peace model:


or this:


See Hibbs’s latest paper for details on his 2008 forecast.

4 thoughts on “Doug Hibbs sez: It’s a good year for the Democrats

  1. There was negative economic growth from 1976-1980? That doesn't sound right to me…. not that I was alive then or anything…

    I must be reading this table from the BEA incorrectly:

    I'm looking at the "chained" data.

    I think this table is telling me that GDP per capita and disposable personal income per capita both increased at a rate of about 2% per year from 1976-1980…

    From the same table (changing the dates), I also get that GDP growth from 1980-1984 was just over 2% annually, and DPI growth was just under 3%. (It doesn't really matter if I use annual 1980-1984 data or Q4 data)…

  2. Alex,

    Hibbs (and other forecasters) use a weighted average of growth, focusing on the most recent years. The idea is that, if the economy improved from 1977-1978, then got worse from 1979-1980, that's perceived as a bad thing. Conversely, a decline from 1981-1982, followed by an improvement from 1983-1984, is perceived positively. The general belief is that voters are responding to the recent trend.

    I agree that my use of the phrase "four years" in the graph label is misleading.

  3. Are the weights chosen for the best fit? If so, then we're fitting the data to multiple growth regressors (however many parameters you have — 4 if you choose one weight per year, 16 if you pick one per quarter, 2 if you pick one "decay" parameter, etc.) That would make the apparent linear fit of the graph much less impressive.

  4. There's been so much discussion lately about how recent growth in incomes is mostly due to growth for the top 1%. Clearly the bread and peace model has incredible fit for just two variables, but I wonder if anyone has tried this model using median income growth. I'd be interested in whether the outcomes as well.

Comments are closed.