Sabato's Crystal Ball

Why Outside Spending Is Overrated

Lessons from the 2014 Senate Elections

Alan I. Abramowitz, Senior Columnist, Sabato's Crystal Ball February 19th, 2015


The Koch brothers and their network of wealthy conservative donors recently announced that they intend to spend almost $900 million on the 2016 elections. This level of spending by a group operating independently of any candidate or political party would be unprecedented in American politics. In fact, it would exceed the combined spending by the Democratic National Committee and the Republican National Committee during the 2012 election cycle. Understandably, this announcement reinforced concerns among Democrats and liberals that spending by the Koch brothers and other conservative groups could give Republican candidates a crucial advantage in key House and Senate contests and in the race for the White House.

Since the Supreme Court’s 2010 Citizens United decision opened the door to spending by Super PACs funded by unlimited contributions from corporations, unions, and wealthy individuals, there has been a dramatic surge in spending by outside groups on federal elections. In 2012, according to data compiled by the Center for Responsive Politics, groups not affiliated with any candidate or party spent over $1 billion on the presidential and congressional elections, more than three times the amount that such groups spent in the previous presidential election year. In 2014, outside groups spent over $565 million, almost twice the amount that such groups spent in the previous midterm election year.

Spending by outside groups strongly favored Republican candidates in 2012. According to the data from the Center for Responsive Politics, conservative groups spent over $700 million on the 2012 presidential and congressional elections while liberal groups spent less than $300 million. Despite this huge advantage in outside spending, however, candidates favored by conservative groups generally fared poorly in 2012. In the presidential race, conservative groups supporting Mitt Romney outspent liberal groups supporting Barack Obama by $418 million to $131 million but Obama still defeated Romney. In the battle for control of the Senate, the balance of outside spending favored Republicans over Democrats by a narrower margin of about $150 million to $115 million. Nevertheless, Democrats won almost every competitive race and actually added two seats to their Senate majority.

The impact of outside spending in 2014

While there is little evidence that conservative outside groups were successful in influencing the results of either the presidential election or key Senate contests in 2012, some observers believe that their efforts may have had a greater impact on the 2014 midterm elections, especially in the crucial battle for control of the Senate. Conservative groups spent more than $250 million in support of Republican Senate candidates in 2014, including almost $35 million in North Carolina, more than $33 million in Colorado, and more than $31 million in Iowa. Republican candidates won all three of those races and almost every other key Senate contest, scoring a net gain of nine seats in the upper chamber.

There is no question that 2014 was a very good year for Republicans, especially in the U.S. Senate elections. The question remains, however, to what extent did outside spending by conservative groups contribute to GOP gains in the 2014 Senate elections? On that score, there are good reasons to be skeptical. For one thing, liberal groups did a decent job of matching conservative groups in 2014 when it came to outside spending. Table 1 displays total spending by outside groups on the 10 Senate races with the largest amount of total outside spending.* These 10 races accounted for over 90% of spending by outside groups on the 2014 Senate elections. Altogether, liberal groups spent about $215 million to support Democratic candidates in these 10 states while conservative groups spent about $224 million to support Republican candidates. In four of these races, including North Carolina and Colorado, outside spending by liberal groups was actually greater than outside spending by conservative groups. And even where one side had an advantage in spending, as in Georgia or Michigan, it appears that both sides had more than enough money to get their messages across to the voters, especially when party spending and candidate spending are considered along with outside spending.

A glance at the data in Table 1 reveals another important feature of spending by outside groups: liberal and conservative groups tend to spend money on the same races. For all 34 two-party contested 2014 Senate elections (Alabama and Kansas’ races didn’t have candidates from both major parties), the correlation between spending by liberal groups and spending by conservative groups was an extraordinarily strong .94. Moreover, liberal and conservative groups also tend to spend money on the same races that the national parties are spending money on. For the same 34 Senate contests, the correlations between spending by liberal groups and spending by the Democratic and Republican Senate campaign committees were .96 and .92 respectively. The correlations between spending by conservative groups and spending by the Democratic and Republican Senate campaign committees were .93 and .91 respectively.

These spending data indicate that liberal and conservative outside groups along with the national parties were all pouring money into the same relatively small set of races that were considered competitive, and there was almost perfect agreement about which races those were. Moreover, even though they are not legally allowed to “coordinate” with each other, the outside groups and parties track each other’s spending closely: If they see other groups or parties spending money on a race, they spend money; if they don’t, they don’t. This can be seen from the fact that neither outside groups nor the parties spent a significant amount of money on the Virginia Senate race — a race that shocked everyone by turning out to be extremely close.

The fact that vast sums of money were being spent by liberal and conservative groups along with the national parties on the same small set of Senate races probably limited the impact of such spending. Not only was one side’s spending generally matched by the other side’s spending, but the sheer volume of spending probably exceeded the point of diminishing returns in many of these states. For example, after each side had spent $30 million on attack ads in a small state like Iowa, it’s hard to believe that an additional $1 million in spending on attack ads by either side was going to have much impact on the Hawkeye State electorate — except perhaps causing more Iowans to turn off their televisions.

Along with relatively balanced spending by both sides and diminishing returns on spending, there is another important reason to be skeptical about the influence of outside spending on the results of these Senate contests — the powerful influence of partisanship. In the current era of electoral competition, partisanship exerts a very strong influence on the outcomes of elections from the presidency down to the state and local level and Senate elections are no exception. In 2014, the correlation between the Democratic Senate candidate’s vote margin and Barack Obama’s vote margin in 2012 was an astonishing .89. With Maine excluded — Sen. Susan Collins (R) faced minimal opposition in her race and ran far ahead of Mitt Romney — the correlation goes up to .95.

In order to measure the effects of outside spending and party spending on the outcomes of 2014 Senate races, I conducted a regression analysis with the Democratic Senate margin as the dependent variable and four independent variables: the difference between pro-Democratic and pro-Republican outside spending in millions of dollars, the difference between Democratic and Republican Party spending in millions of dollars, incumbency (coded as +1 for Democratic incumbents, 0 for open seats, and -1 for Republican incumbents), and the 2012 Democratic presidential margin. The results of this analysis are displayed in Table 2 for the 34 Senate contests with Democratic and Republican candidates.

Table 2: Results of regression analysis of 2014 Senate election results

Sources: Center for Responsive Politics and data compiled by author

This regression equation does an excellent job of explaining the results of the 34 contested Senate races, explaining 86% of the variance in the Democratic Senate margin. Among our four independent variables, presidential partisanship was by far the strongest predictor of the election results. The estimated unstandardized regression coefficient for the presidential partisanship variable means that for every one percentage point increase in Barack Obama’s 2012 margin in a state, the Democratic Senate candidate in 2014 could expect to receive an increase of about .8 percentage points in his or her margin.

Incumbency also had a significant impact on the outcomes of Senate races. According to these results, on average, a Democratic or Republican incumbent could expect to receive a nine-point boost in margin compared with a Democratic or Republican candidate in an open-seat race. The estimated constant, or intercept, in the regression equation indicates that there was a significant Republican tide in 2014 — a Democratic candidate for an open seat in a state in which Barack Obama and Mitt Romney received equal shares of the vote in 2012 would be predicted to lose by a margin of more than six percentage points in 2014. However, these results indicate that neither outside spending nor party spending had a significant impact on the results of these Senate races. The estimated coefficients for the spending difference variables are both negligible in size and actually in the wrong direction. Our results indicate that after controlling for state presidential partisanship and incumbency, relative spending by outside groups and political parties had no discernible impact on the Democratic candidate’s margin in these contests.

Conclusions and implications for 2016

Republicans made major gains in the 2014 Senate elections but the findings reported here indicate that outside spending by conservative groups had little or nothing to do with those gains. The main reason why Republicans did very well in 2014 was that Democrats were defending far more seats than Republicans and many of those seats were in states that normally favor Republicans based on recent presidential voting patterns. Democrats lost all seven of their seats in states carried by Mitt Romney in 2012 even though Democratic candidates enjoyed an advantage in outside spending in several of those races.

The factors that limited the impact of outside spending in 2014 are very likely to be present in the 2016 elections as well. In the large majority of states, the winners of the presidential and Senate elections will be determined by the relative strength of the parties in the state. In the last four presidential elections, 40 of the 50 states have supported the same party in each contest, and there is little reason to expect anything different in 2016. In the 2016 Senate elections, Democrats are likely to gain at least a few seats simply because Republicans will be defending a large number of seats in blue states that they picked up in the 2010 midterm election. Notwithstanding the plans of the Koch network to spend almost $900 million on the 2016 elections, neither party is likely to enjoy a substantial advantage in spending in the relatively small number of competitive states that will decide the presidential election or control of the Senate.

The fact that outside spending is unlikely to have a major impact on the results of the 2016 elections does not, of course, mean that Americans should not be concerned about the vast sums of money pouring into American elections since the Citizens United decision. What we are seeing today is essentially an arms race between the two major parties with Democrats struggling to keep up with spending by the Koch brothers and other conservative billionaires. Even if Democrats and their liberal allies are able to remain financially competitive, it’s reasonable to question whether having both of our major parties increasingly dependent on financial largesse from the super-rich is a healthy development for American democracy.

*The Kansas Senate election is not included in this analysis because there was no Democratic candidate. Conservative groups spent more than $10 million to support Republican incumbent Pat Roberts while liberal groups spent almost $7 million to support independent Greg Orman.

Alan I. Abramowitz is the Alben W. Barkley Professor of Political Science at Emory University and a senior columnist for the Crystal Ball. His most recent book is The Polarized Public: Why American Government Is So Dysfunctional. Follow Alan on Twitter @AlanIAbramowitz.