The recent fund-raising excesses of President Clinton and Speaker Gingrich have become unlikely stalking horses for public financing of congressional campaigns. Both Clinton and Gingrich were engaged in "party building" activities that had no direct link to any congressional campaigns. Despite this, reformers point to these scandals as evidence that we need to overhaul the entire system of campaign finance. In the words of Ann McBride, president of Common Cause, the goal is to "harness the winds of anger and turn it into a demand for positive action." The president of Public Citizen, Joan Claybrook, cites "the orgy of party spending" as evidence that "immediate efforts to clean up the system [are] absolutely necessary." The McCain-Feingold bill, currently languishing in the Senate, embodies the change campaign finance reformers crave.
McCain-Feingold, also known as the Bipartisan Campaign Finance Reform Act, would outlaw almost all soft-money contributions to political parties–a restriction unlikely to survive judicial review. The Supreme Court has consistently held that contributions for "party building" activities are protected by the First Amendment.
The remainder of the bill–the part that could probably survive judicial scrutiny–places new limits on congressional campaigns. Political action committee contributions would be eliminated. And in case the Supreme Court strikes that provision, the bill lowers maximum allowable PAC contributions from $10,000 per candidate per election cycle to $2,000. McCain-Feingold would also force radio and televison broadcasters to provide free and reduced-rate air time to candidates who abide by spending caps and restrictions on out-of-state contributions.
The leadership of Common Cause, Public Citizen, and the newly formed Campaign for America have made clear that they view McCain-Feingold as an important first step toward their real goal: full public financing of elections. Even though reformers acknowledge that public financing does violence to free speech, they see it as a necessary, drastic measure to end what they feel is the unacceptably dominant role that money plays in politics.
Few conventional wisdoms are so widely and firmly held–and so little supported by evidence–as the chestnut that there exists a vibrant cash-and-carry market for both elected offices and legislative votes. The two fundamental tenets of this "flea market" view of money and politics are 1) special interests buy favors from politicians with campaign contributions, and 2) incumbent politicians then use this dirty money to buy re-election with massive campaign expenditures.
More often than not, the popular press treats these claims as self-evident. In place of rational argument, one is simply offered lists of contributor names and dollar amounts. Visit the Web sites of the Center for Responsive Politics, Common Cause, or Public Citizen, and you will be inundated with such "analysis." When the reformers do make arguments, they point to the ever-increasing amount of money spent on elections, the gaudy spending advantage enjoyed by congressional incumbents, the high re-election rates of incumbents, the dearth of able challengers, and the well-known correlation of contributor interests with the committee assignments and voting records of incumbents.
It is true that over the past 20 years, total campaign spending has increased faster than the rate of inflation. It is also true that during this time the turnover rate in Congress (defeats plus retirements divided by the number of seats) was quite low–less than 10 percent in 1984, 1988, and 1990. In fact, it is more common for incumbents to be unopposed than to be defeated. Further, even when incumbents are challenged, most races are one-sided. Incumbents take advantage of this lack of competition to build up large reserves of campaign cash for use in future elections.
Although those facts make campaign finance reformers hyperventilate, the inferences they draw from them aren't always accurate. For example, the increase in real campaign spending has not caused a decrease in turnover in the House and Senate. Since 1992, the turnover rate in the House has averaged about 20 percent–the highest average turnover rate in more than 40 years. Part of this increase is attributable to redistricting after 1990; retirements and defeats typically increase after redistricting, when some unlucky incumbents are suddenly faced with new and possibly even unfriendly constituents.
Yet current turnover rates are higher now than they were after each of the four previous decennial redistrictings. In fact, over the past 20 years, the turnover rates in both the House and Senate show a slight positive correlation with the average spending advantage of incumbents over challengers. This is the opposite of what the conventional wisdom would predict.
Of course, it is still true that in any given election year, winning candidates spend more than losing candidates. Reformers infer from this that campaign expenditures cause electoral success, but social scientists have been unable to establish a strong causal connection. The fundamental difficulty in measuring the electoral effects of campaign spending is that contributors are more apt to give money to candidates with a greater probability of victory. In 1996, for instance, PACs contributed about $130 million to congressional incumbents and only $26 million to challengers. It is therefore not surprising that winners outspend losers.
Further, such candidate attributes as leadership, integrity, and perseverence are important determinants of both fund raising and electoral success, but are difficult to quantify. Studies that ignore these unobserved factors produce biased estimates that exaggerate the electoral effects of candidate spending. For example, you might observe that candidates who spend more money fare better in elections, but how much of this is actually caused by the money alone? Surely, part of this correlation is due to the fact that the winning candidate was intelligent, articulate, and hard-working. Since those qualities help both fund raising and vote getting, part of the observed correlation is spurious.
The most promising approach to deciphering how large a role money plays in electoral success is to examine changes in candidate vote share associated with changes in candidate spending. To the extent that the other factors that affect elections are constant over time, they will not affect changes in fund raising or vote getting. Only factors that change matter. For example, my research has indicated that House incumbents who transfer onto powerful committees or are promoted to leadership positions receive an extra $50,000 in campaign contributions, all else constant. This windfall translates into higher campaign spending but does not lead to any significant increase in vote share. I have examined several other "spending shocks" for House incumbents, but they also produce no statistically significant increases in vote share. This suggests that at least for incumbents, marginal spending is unimportant. Challengers, of course, might be a different case.
The 2nd Congressional District in Utah provides a unique case study on the effects of challenger spending. In 1992, Democrat Karen Shepard beat Republican Enid Greene in a race for an open seat.
In a 1994 rematch, however, Enid Waldholtz (Greene had gotten married in the interim) defeated the incumbent Shepard. Soon after the election, it was discovered that the Waldholtz campaign had spent almost $2 million in stolen funds. (Because of the resulting scandal, Waldholtz declined to stand for re-election in 1996.) This provides an ideal experiment, since the infusion of illegal cash was not caused by Waldholtz's political acumen, but rather by the activities of her then-campaign manager and husband (and now convict), Joe Waldholtz.
In 1992, Waldholtz received only 36 percent of the vote. In 1994 she won with 56 percent. If we attribute all of the 20-percentage-point increase in Waldholtz's vote to this ill-gotten windfall, then it seems that an extra $100,000 in spending yields 1 percent of the vote. However, other factors also contributed to the Waldholtz victory. First, Shepard's surprisingly liberal voting record alienated many voters (she even needed a police escort at one district meeting). Second, 1994 was a good year for Republicans in general: The average electoral swing to Republicans challenging Democratic freshman was six percentage points.
This leaves an unexplained increase in Waldholtz's vote percentage of at most 14 points. In the 1992 race, a conservative third party candidate, Merrill Cook, polled 18 percent of the vote, but no major independent candidates ran in 1994 (Cook won the seat in 1996, running as a Republican). If only half of Cook's voters in 1992 went for Waldholtz in 1994 (a very conservative estimate), then the unexplained swing to Waldholtz shrinks to just five percentage points. If all of this swing in vote share is attributed to the illegal influx of campaign spending, then it seems that an extra $100,000 of campaign spending buys at most 0.25 percent of the vote.
That case study is provocative, and also in line with other research. Steven Levitt, an economist at Harvard University, has examined all repeat meetings of House candidates (such as Waldholtz-Shepard) from 1978 to 1990. Levitt has found that an extra $100,000 in candidate spending (whether spent by an incumbent or challenger) leads to an increase in the candidate's vote share of about 0.2 percentage points. I have replicated Levitt's analysis using more recent data and have found similar results.
These recent studies cast serious doubt on the notion that elections can be bought. If we take Levitt's estimate of 0.2 percentage points per $100,000, then to buy just 10 percent of the vote in a typical House district, you would need to spend $5 million. Newt Gingrich spent a little more than that in 1996 and won by 15 percentage points (but his opponent spent over $3 million). Only four other House candidates spent as much as $2 million in 1996, while the average House incumbent spent only about $650,000. This suggests that only very close elections can be bought. But few elections are that close. For example, if every challenger had been given an extra $100,000 to spend in 1996, then Levitt's model predicts that two additional House incumbents would have been defeated. Extending this to the previous five House elections, only nine more races would have turned out differently.
In a 1996 study, Ohio State University political scientist Tim Groseclose and I found still more evidence that campaign spending is not all that the conventional wisdom supposes. Like many other researchers before us, we discovered that stockpiles of unspent contributions, or incumbent war chests, had no deterrent effect on challengers in House races. What was novel about our study was that we also considered the possible deterrent effect of an incumbent's personal wealth. Many reformers have called for a ban on both war chests and self-financing, since neither of these funding sources is limited by current law.
The conventional wisdom holds that potential challengers shy away from taking on well-heeled incumbents, which confers an unfair advantage to the wealthy. But we found no evidence that an incumbent's personal wealth affects either the presence or quality of challengers. By itself, this lack of deterrence is puzzling. Yet these findings are quite consistent with the studies of campaign spending already discussed. Challengers are not deterred by incumbent war chests or personal fortunes because money simply isn't that important in congressional races.
But if challengers aren't deterred by war chests and personal wealth, why are there so few good challengers in congressional races? The obvious answer is that even good challengers would lose. Even when congressional approval ratings were at an all-time low, most survey respondents indicated that they would vote to re-elect their own representatives and senators. It isn't hard to understand why most incumbents are so safe from competition.
Indeed, it's a little like the reason men do not ask Maria Shriver out on dates: It isn't because Arnold Schwarzenegger can outspend them; it's just that he was chosen through a competitive process and for now he's doing a satisfactory job. And though Arnold may have some incumbent advantage, the main reason for his success is that Maria really likes him. Similarly, congressional turnover rates are low because politicians are such able and attentive suitors to their constituents.
Even though campaign spending isn't the key to re-election, that doesn't mean candidates don't work very hard to raise funds. Just because campaign ads don't have much of an effect doesn't mean that anything else candidates do is more effective. But it is ridiculous for reformers to claim that dollars, not votes, elect legislators. I don't vote for a candidate just because he spends the most money, nor do you. It's insulting to think so little of our fellow citizens that we believe elective offices go to the highest bidder.
If campaign spending really doesn't alter the electoral prospects of a candidate, why do so many citizens, clubs, unions, and corporations make contributions? The quick answer: They don't. Only 5 percent of voters make political contributions of any kind. Further, the total amount spent on political campaigns per election cycle is less than consumers spend over a similar period on chewing gum, greeting cards, or yogurt. If legislation is for sale, it seems unlikely that the power to coerce millions would be priced so low. Even the PACs are pikers; PAC contributions are typically in the range of $500 to $2,500, well below the legal maximum of $10,000 per candidate for each election cycle.
Further, PAC contributions accounted for only 34 percent of total congressional campaign receipts in 1988 and 1990; the share fell to less than 30 percent for 1992 through 1996. Perhaps the decline is linked to the realization that, despite several studies, there is no systematic evidence that legislators cast their votes in response to campaign contributions. PAC contributions mostly function not as bribes but as donations to like-minded legislators. If PAC contributions were meant as bribes, they wouldn't be very effective ones. Given the limited effectiveness of campaign spending, a $500 PAC contribution probably won't elicit much gratitude from a member of Congress.
Contrary to the conventional wisdom, congressional campaign finance is not out of control. Money is nowhere near the dominant force that reformers claim. Once we realize this, proposals to limit freedom of speech in the name of cleaning up elections, such as the McCain-Feingold bill, become less appealing. Private campaign finance does not have nearly as severe a corrupting effect on elections as many claim. Though the reformer lobby should greet this revelation as good news, somehow I doubt it will.
Jeff Milyo (firstname.lastname@example.org) is an assistant professor of economics at Tufts University and a postdoctoral fellow in political economics at Harvard and the Massachusetts Institute of Technology.