Prior to the 1992 presidential elections, Democrats in both the House and the Senate passed legislation limiting spending in United States House elections to $600,000, while providing $200,000 in communication vouchers. The purpose of the bill was not so much to clean up congressional campaigns as it was to embarrass then-President George Bush. In vowing to veto the legislation, President Bush allowed Democrats the opportunity to appear to be on the side of "good" government, without having to face the potential repercussions of actually implementing campaign reform.(1)
After Bill Clinton's victory in the 1992 presidential election, Democrats were confronted with a president who promised to sign campaign finance reform. Yet, the possibility that campaign reform actually would be enacted made congressional Democrats considerably more cautious. As described by Congressional Quarterly Almanac (Campaign finance overhaul dies, 1995, p. 32):
The long history of the legislation was rich with evidence that many Democrats in both chambers shared GOP objections to establishing a system that would provide congressional candidates with federal subsidies. Other Democrats, particularly in the House, were deeply, if privately, opposed to an overhaul of the financing system that had protected their seats and majority status for years.
While the reform effort in the 103rd Congress eventually was killed by a Republican-led filibuster, Democrats "set the stage for defeat by waiting until the eleventh hour to come up with a compromise version of a bill they had previously maintained would be a top priority in Clinton's first two years" (Campaign finance overhaul dies, 1995, p. 32).
Congressional Republicans, for their part, have been consistent in their opposition to any version of campaign finance reform that includes spending limits. Prior to 1994, Republicans argued that imposing spending limits would damage their chances of gaining control of the legislative branch of government. In the aftermath of the 1994 elections, congressional Republicans have not eased in their criticisms of spending limits and, by extension, public financing. Speaker of the House Newt Gingrich, for example, has argued that more rather than less should be spent in congressional campaigns. Others have joined in the chorus, noting that considerably more is spent each year on product advertisements than is spent on political advertisements (see, e.g., Smith, 1995).
A superficial glance at the 1994 congressional elections would seem to support the Republican view of campaign finance reform. Gimpel (1996, p. 10), for example, observed that funding for Republican challengers was "up some 42% over 1992." Similarly, Babson and St. John (1994) reported that corporate political action committees (PACs) were more willing to support Republican challengers in the 1993-1994 election cycle than in past election cycles, while other reports indicated that the 1994 congressional elections set a record for the total amount of money spent in a midterm election.(2) While such evidence suggests a strong correlation between the money spent by Republican candidates in 1994 and the Republican takeover of Congress, it is not clear whether the Republican "earthquake" of 1994 would have been curtailed significantly had spending limits (with or without public financing) been in place. Eismeier and Pollock (1996, p. 81), for example, observed that "money continued to flow into congressional campaigns in much the same way in 1994 that it had for the last several election cycles - despite significant signs of upheaval in the months before the election." As a result, "If more Republican money had flowed into these marginal races instead of into the coffers of safe incumbents, the damage to Democrats could have been worse" (Eismeier & Pollock, 1996, p. 86).
Still, the failure of the Democrats to pass meaningful campaign finance reform and the increased fundraising activities of Republican challengers prior to the 1994 elections does lead to an interesting question: Could the Democrats have avoided the 1994 landslide had campaign finance reform been in place prior to the 1994 elections? The purpose of this article is to explore the likely effect of campaign finance reform on the 1994 House elections. Because the 1994 elections witnessed the dramatic Republican takeover of the House of Representatives, they provide the ideal setting to examine the likely effects of campaign finance reform. Based on Republican arguments against campaign finance reform, we would expect that had campaign finance reform been in place prior to the 1994 elections, it would have served to protect incumbent Democrats. According to this logic, the failure to pass campaign finance reform very well may have cost the Democrats majority control of Congress.
We approach this topic by specifying and estimating a regression model aimed at predicting the Democratic percentage of the vote at the district level. The results from this model then are used to examine the likely outcome of the 1994 congressional elections under a number of hypothetical spending scenarios. To develop the clearest possible sense of what might have occurred under different variations of campaign finance reform, we consider a broad range of potential reforms (i.e., spending limits, matching funds, full public financing, and partial public financing) across a wide range of potential limits ($100,000-$1,000,000). We should note, however, that while we are considering a broad range of potential reform scenarios, including spending limits and public financing of congressional campaigns, we have not accounted fully for all types of reform. For example, we have not estimated how reforms aimed at controlling the effects of soft money or independent expenditures would affect congressional elections, nor have we estimated how a ban on political action committee contributions would alter the electoral process.
Overall, the results of these simulations indicate that while spending limits would have reduced the number of Republican victories in 1994, Republicans most likely would have gained control of the House of Representatives even with relatively modest spending limits in place. In addition, the results indicate that under various scenarios involving a combination of public financing and spending limits, campaign finance reform might have aided rather than undermined the Republican takeover of Congress. To ensure that the pattern of results observed in 1994 was not simply a reflection of a highly dramatic and atypical election year, we also have conducted simulations based on data from the 1992 House elections. While the results across election years were far from identical, the 1992 results confirm our general conclusion that Republicans may have benefited significantly from the enactment of campaign finance reform.(3)
While there are some notable exceptions (Goidel & Gross, 1996; Green & Krasno, 1988; Krasno & Green, 1993), most of the work in political science has tended to support the Republican view of spending limits (see, e.g., Jacobson, 1980, or Sorauf, 1992): That is, spending limits serve to inhibit electoral competition. There is less agreement regarding the likely effect of public financing, although one commonly expressed view is that public financing in combination with spending limits also would inhibit electoral competition (Jacobson, 1980). A more mixed assessment of public financing was found by Abramowitz (1991), who contended that while public financing might increase competition substantially, doing so would require an unrealistically high level of public funding. On a similar note, Sorauf (1992, p. 215) contended that "reasonably generous spending limits accompanied by a substantial public subsidy might not greatly hurt challengers." Overall, the campaign finance literature paints a mixed assessment of public financing, where, in order to improve electoral competition, the negative consequences of spending limits would have to be balanced by substantial public funding.(4) (For contradictory evidence, see: Goidel & Gross, 1996; Krasno & Green, 1993).
More systematic analyses of individual states that have adopted some form of public financing also have yielded inconsistent results. In his review of state public financing schemes, Alexander (1992, p. 146) observed that "while excessively low spending limits can cause public financing to work to the advantage of incumbents, evidence exists that public funding programs have provided significant benefits to challengers." This mixed review of state public financing schemes can be found in other works as well. Mayer and Wood (1995) found little evidence that the system of public financing adopted by Wisconsin in 1977 has resulted in increased electoral competition. Donnay and Ramsden (1995), on the other hand, contended that while a cursory glance at Minnesota's system of public financing might indicate that competition actually has decreased over time, introducing a fuller range of controls indicates that public financing has stimulated electoral competition. On a slightly different note, research by Ruth Jones (1981, p. 358), investigating the partisan implications of public financing at the state level, found that while systems of public financing tend to benefit the majority party, overall "any public funding may well be more beneficial to the minority than to the majority party."
One of the problems in these state-level analyses involves the question of what standard should be used to gauge the success (or failure) of campaign finance reform.(5) If, for example, we find that competition is greater in states that have adopted some form of public financing than in states that have not, can we then conclude that public financing increases competition? Not if we have reason to believe that competition would have been greater in these states even without public financing. Similarly, a decline in electoral competition following the adoption of a system of public financing does not indicate necessarily that public funding has been a failure, if it can be argued that the decline in competition would have resulted even if public financing had not been adopted.
One viable alternative involves the use of simulations to estimate the likely effects of different reform scenarios. This approach is appealing for a number of reasons. First, it is flexible enough to allow the testing of multiple reform scenarios. State-level analyses are limited to testing the efficacy of the reform proposal currently in place. Consequently, such analyses are limited in their ability to determine the likely effect of slightly different variations of the same general reform proposal. For example, while we might be able to determine that Wisconsin's system of public financing has resulted in decreased competition since its inception in 1977 (Mayer & Wood, 1995), we are unable to determine how slight variations in their system of public financing would alter the level of competition. Second, while the results of the simulations are dependent on the assumption of the model employed (Goidel & Gross, 1996), simulations provide a more stable foundation for making inferences than inferences based on the regression model alone.
Finally, recent work by Gelman and King (1994) presented a forecasting model that incorporates both the individual district vote prediction and the uncertainty of the prediction into estimates of aggregate seat loss. As Gelman and King (1994, p. 517) observed, "This avoids the unreasonable assumption that election outcomes are exactly determined and can be forecast without error." The analysis presented here incorporates the work of Gelman and King (1994) by first computing the probability that the Democratic candidate won the election and then summing these probabilities across all congressional districts. These probabilities then are used to estimate the number of Democrats that would be expected to be defeated under a given hypothetical scenario.(6)
Data and Methods
The bulk of this analysis is based on simulations of a number of hypothetical campaign finance reform scenarios. To conduct the simulations, it was necessary first to run a basic regression model. In the analysis, the Democratic percentage of the vote is seen as a function of candidate spending, incumbency, nonincumbent candidate quality, and the underlying partisan division of the district. Our unit of analysis is the aggregate, district-level, Democratic percentage of the vote per House election. Because a great deal of the previous literature has been devoted to the important distinction between incumbent and challenger spending, interaction terms between incumbency and candidate spending also were included in the model (see, e.g., Green & Krasno, 1988; Jacobson, 1978, 1980, 1985, 1990). More formally, the model can be specified as follows:
Dem% = a + [b.sub.1] (Dem$) + [b.sub.2] (Rep$) + [b.sub.3] (Dinc) + [b.sub.4] (Rinc) + [b.sub.5] (DemQual) + [b.sub.6] (RepQual) + [b.sub.7] (Partisan) + [b.sub.8] (Dem$ X Dinc) + [b.sub.9] (Rep$ X Rinc) + e
where Dem% is Democratic percentage of the vote; Dem$ is the natural logarithm of Democratic spending; Rep$ is the natural logarithm of Republican spending; Dinc is Democratic incumbency, coded 1 if there is an incumbent Democrat and 0 otherwise; Rinc is Republican incumbency, coded 1 if there is an incumbent Republican and 0 otherwise; DemQual is candidate quality, measured on a 5-point scale where 0 indicates no previous political experience, 1 indicates nonelective political experience, 2 indicates local political experience, 3 indicates state legislative experience, and 4 indicates statewide experience or former service as a representative; RepQual is measured identically to DemQual for Republican nonincumbent candidates; and Partisan is the underlying partisan division of the district as indicated by the percentage of the vote received by Clinton in 1992.
Two points are in order regarding the model. First, the model explicitly incorporates the distinction between incumbent and nonincumbent expenditures made in the previous literature (see, e.g., Green & Krasno, 1988; Jacobson, 1978, 1980, 1985, 1990). Second, the model taps the underlying partisan division of the district using previous presidential vote rather than previous congressional vote.(7) Unlike the previous congressional vote, the previous presidential vote should not be as contaminated by the effects of congressional incumbency and consequently should give a better indication of the underlying partisan leanings of the district. Third, ordinary least squares analysis is used to estimate the model.(8)
Data for this analysis were collected on all contested elections in 1994 and in 1992, though we limited the current analysis to only those elections in which both candidates spent at least $500.(9) The data on candidate spending, Democratic percentage of the vote, and incumbency were taken from Congressional Quarterly's Politics in America 1996 (Duncan & Lawrence, 1995) and The Almanac of American Politics 1994 (Barone & Ujifusa, 1995). Data on candidate quality were taken from the special editions of Congressional Quarterly Weekly Report devoted to the 1994 and 1992 congressional elections.(10) The results of the initial regression analysis are displayed in Table 1.
As can be seen in Table 1, the model provides an impressive fit to the data across election years. Looking first at the results for the 1994 elections, values of both [R.sup.2] and adjusted [R.sup.2] indicate that the model explains nearly 88% of the variance in the Democratic party vote. Moreover, with the exception of the variable tapping Democratic nonincumbent quality, every variable in the equation is statistically significant. While this nonfinding was unexpected given past research on the electoral importance of nonincumbent political experience (see, e.g., Bond, Covington, & Fleisher, 1985; Jacobson, 1989; Krasno & Green, 1988), given the anti-Democratic mood of the electorate in 1994, it is not altogether surprising.
Also worth noting is the relatively high value for the coefficients associated with the variables tapping Democratic and Republican incumbency. While it may seem that these values are highly inflated, keep in mind that the main effect in this case indicates the value of incumbency for those incumbents who spent no money in their reelection attempt. Since we would assume that these hypothetical incumbents were not challenged, the expected vote should be near 100%.(11)
Table 1 OLS Regression of Democratic Vote on Candidate Spending Variable Name 1994 Estimate (SE) 1992 Estimate (SE) Democratic spending 3.15 (.40)(**) 2.58 (.52)(**) Republican spending -3.19 (.29)(**) -3.49 (.37)(**) Democratic incumbent 51.32 (10.7)(**) 16.81 (11.15) Republican incumbent -33.90 (14.7)(*) -58.32 (9.12)(**) Democratic quality -.09 (.34) .58 (.49) Republican quality -.97 (.31)(**) -.39 (.43) District partisanship .38 (.03)(**) .35 (.04)(**) Democratic incumbent X Democratic spending -3.44 (.80)(**) -1.02 (.86) Republican incumbent X Republican spending 2.06 (1.10)(*) 3.97 (.71)(**) Constant 30.31 (6.44)(**) 47.45 (7.82) [R.sup.2] .8796 .7297 Adjusted [R.sup.2] .8763 .7230 Notes: * p [less than] .05; ** p [less than] .01
Finally, the effects of spending for Democratic and Republican nonincumbent candidates are nearly indistinguishable. Each dollar spent by a nonincumbent Democrat appears to be worth about the same as each dollar spent by a nonincumbent Republican. For Democratic incumbents, however, the effect of spending money appears to have been considerably less than for Republican candidates. According to our estimates, spending by Democratic incumbents appears to decrease their share of the vote. Spending by Republican incumbents, on the other hand, appears to have benefited their candidacy, though the effect of spending by Republican incumbents is considerably less than the effect of spending by nonincumbent candidates. The findings regarding the effect of spending for incumbent and nonincumbent candidates are consistent with the previous literature in that nonincumbent spending appears to be considerably more important than incumbent spending. However, the results also indicate that the effect of incumbent spending may differ according to whether the incumbent is a member of the party advantaged at the midterm or a member of the party disadvantaged at the midterm.
Overall, the results based on the 1992 elections are somewhat less impressive. First, we failed to find any significant effects of nonincumbent candidate quality on the Democratic percentage of the vote. In part, this may reflect the electoral environment of 1992. Although Republicans were hopeful that congressional redistricting would allow the party to make considerable gains in 1992, President George Bush's sagging public approval ratings limited Republican gains to a mere 10 seats. As a result, while redistricting favored the Republican party, short-term political forces placed them at a disadvantage. Congressional redistricting also may help to explain why the 1992 regression explains less of the variance than the 1994 model. While the 1992 model explains over 72% of the variance in Democratic percentage of the vote, this is considerably less than nearly 88% explained by the 1994 model. Presumably, this reflects the greater instability exhibited frequently in elections immediately following congressional redistricting. Finally, although the difference for Democratic incumbents was not statistically significant, spending by incumbents was less effective than spending by nonincumbents.
On the basis of these initial regression results, we ran simulations to estimate the likely effect of spending limits, matching funds, partial public financing, and full public financing on the electoral fortunes of Democratic congressional candidates. To estimate the effects of the respective spending scenarios on the outcome of the 1994 elections, we computed two related measures: (a) the aggregate number of Democrats predicted to receive less than 50% of the vote; and (b) the aggregate number of Democrats predicted to lose based on a summation of each individual probability that the Democratic candidate won the election (Gelman & King, 1994; Goidel & Gross, 1996). Because the second measure of Democratic vulnerability incorporates uncertainty into the aggregate predictions, the presentation of the results focuses on this estimate of Democratic seat loss. The actual estimates of Democratic seat loss for 1994 and 1992 can be found in the tables presented in the Appendix.
In creating the hypothetical spending scenarios, we attempted to use as wide a range of reform proposals and limits as possible. For the simulations designed to estimate the effects of spending limits, we set any candidate spending greater than the designated limit equal to the limit. The designated limits ran from a low of $100,000 to a high of $1,000,000. For the simulations of matching funds, we first multiplied candidate spending by two; if this modified spending variable was greater than the designated limit, we truncated candidate spending to the limit. As in the analysis of spending limits, the limits ranged from $100,000 to $1,000,000. For the simulations of partial public financing, we first created a modified candidate spending variable by adding the value of the communication voucher; if this modified spending variable exceeded the designated limit, we then set it equal to the limit. The limits in this particular portion of the analysis ranged from $400,000 to $1,000,000, while the vouchers ranged from $100,000 to $300,000. Finally, for the simulations of full public financing, we set candidate spending equal to the designated limit, which ranged from $100,000 to $1,000,000.
To provide a baseline for comparing the results of the simulations, we first computed estimates of Democratic seat loss for a model assuming no change in existing campaign finance laws. According to this baseline prediction, in 1994, the Democrats would have lost 57 seats had the elections conformed exactly to the expectations of our regression model.(12) We should note that this baseline prediction slightly overpredicts the actual number (53) of seats lost by the Democratic party. In 1992, had the elections conformed exactly to our regression model, Democrats would have gained 4 seats. In reality, Democrats lost 10 seats during the 1992 elections.
As can be seen in Figures 1A and 1B, spending limits in 1994 would have had the effect feared by most opponents of campaign finance reform. That is, had spending limits been imposed prior to the 1994 elections, Republicans would have won fewer Democratic seats. In fact, only with limits set as high as a million dollars would the Republican victories in 1994 approach the number of victories predicted in the baseline model. The results also indicate, however, that if one takes into account the regression model's tendency to overpredict Democratic seat loss, Republicans still would have been likely to take control of the House with limits set as low as $300,000. As such, while the general pattern of findings tends to support the argument that spending limits would have limited the number of Republican victories in 1994, it appears unlikely that spending limits, except at very low levels, would have kept the House under Democratic control.
As can be seen in Figure 1B, spending limits in 1992 would have reduced the number of seats predicted to be gained by the Democratic party, though it also is worth noting that the overall effect of spending limits appears to be marginal. Presumably, this reflects the fact that, under spending limits, Democratic challengers and open-seat candidates would have been less successful in their efforts to win office during the 1992 elections. Taken in context of the findings from 1994, these results seem to indicate that the imposition of spending limits, without some accompanying form of public financing, most likely would limit the success of the party favored by short-term electoral forces.
If spending limits would have limited Republican victories in 1994, the effect of matching funds is less clear [ILLUSTRATION FOR FIGURE 2A OMITTED]. With limits set at relatively low levels, Republican victories would have been reduced significantly. As was the case with spending limits, even with matching funds at relatively low levels, Republicans likely still would have taken control of the House. At higher levels of matching funds, Republicans actually would have benefited from some form of public financing. According to our estimates, Republicans would have won two additional seats had matching funds been set at $400,000. At higher levels of public funding, Republicans would have won even more House seats. For example, with matching funds and limits set at $800,000, we estimate that Republicans would have won an additional 10 seats during the 1994 elections.
If the effect of matching funds on electoral competition is somewhat unclear in 1994, the results from 1992 are unequivocal. Clearly, according to the results presented in Figure 2B, matching funds would increase the number of seats expected to be gained by the Republican party. At very low limits, Republican prospects would increase from a loss of 4 seats to a gain of 11. Moreover, with matching funds up to a limit of $500,000 to $600,000, Republicans would have been expected to gain 25 House seats. While Republican prospects diminish somewhat with matching funds set at higher levels, overall it appears that Republicans would have benefited significantly had a system of matching funds been in place prior to the 1992 elections. More broadly, it appears that matching funds tend to benefit the minority, although the effect is contingent upon the level of matching funds and the electoral context (i.e., the election year).
Full Public Financing
The effect of full public financing on the 1994 elections is similar to the effect of matching funds, although the overall effect is magnified somewhat. Even with full public financing, competition would be diminished if spending limits were set at a relatively low level - less than $200,000 according to Figure 3A. Once public financing extends beyond these relatively low levels, however, Republican candidates would have benefited tremendously. According to our estimates, if full public financing were set at $400,000, Republicans would have gained an additional 13 House seats in 1994. If full public financing were set at $600,000, we estimate that Republicans would have gained an additional 21 House seats.
Similar to the effect of matching funds in 1992, had full public financing been in place prior to the 1992 elections, Republican candidates would have been the beneficiaries [ILLUSTRATION FOR FIGURE 3B OMITTED]. For example, had full public financing supported House candidates up to a limit of $500,000, Republicans would have been expected to gain 20 House seats. Consistent with the results from 1994, it appears that public financing would have facilitated rather than limited Republican efforts at gaining control of the House of Representatives.
Partial Public Funding
The most impressive results of this analysis are based on assumptions of partial public financing that most resemble the version of campaign finance reform that failed to become law in 1993. The results of our analysis indicate that campaign finance reform that involves public grants (or vouchers) in coordination with spending limits would have increased significantly the number of Republicans elected in 1994 and 1992.
Looking first at the results from 1994, in each of the scenarios considered, we estimate that Republicans would have gained additional seats had campaign finance reform been in place prior to the 1994 elections [ILLUSTRATION FOR FIGURE 4A OMITTED]. For example, in the version of campaign finance reform that most resembled the version considered by the House in 1993 - vouchers of $200,000 and spending limits of $600,00 - the simulations indicate that Republicans would have won an additional 14 House seats. Even with lower limits ($400,000) and more modest vouchers ($100,000), the number of Republican victories would have increased slightly.
In general, the results from 1992 are consistent with this general pattern [ILLUSTRATION FOR FIGURE 4B OMITTED]. Republicans would have benefited under each of the scenarios. However, the results from 1992 differ in that Republicans would have maximized their gains had public subsidies been combined with relatively low limits. In 1994, on the other hand, Republicans would have maximized their gains by combining public subsidies with relatively high limits.
Overall, the results indicate that spending limits would have limited the number of seats won by the Republicans in 1994. Except at very low levels, however, the effect would not have been drastic enough to have saved Democratic control of the House of Representatives. Moreover, simulations based on data from the 1992 House elections indicate that, at least in some election years, Republicans might benefit from the enactment of spending limits. While our evidence is relatively limited on this point, we would suggest that spending limits, without some form of public financing, inhibit the electoral prospects of the party advantaged by short-term electoral forces.
Spending limits in combination with some form of public financing, we would contend, enhance the electoral prospects of the minority party. Except at very low levels, Republican candidates would have benefited had a system of matching funds or full public financing been in place prior to the 1994 elections. In 1992, Republicans would have benefited from a system of matching funds or full public financing even with relatively low spending limits in place. Finally, even with relatively modest vouchers and relatively low spending limits, partial public funding would not have diminished Republican gains. In fact, more substantial vouchers and higher limits would have amplified rather than diminished the Republican earthquake of 1994.
Discussion and Implications
Prior to 1994, Republicans feared that the implementation of campaign finance reform, assuming that it involved a combination of public grants and spending limits, would inhibit their ability to gain control of the U.S. House of Representatives. While the increased fundraising success of Republican challengers and the overall increase in campaign spending during the 1993-1994 election cycle would seem to provide supportive evidence, our analysis indicates that campaign finance reform would not have limited significantly the number of seats won by the Republican party in 1994, and actually may have resulted in an increase in the number of seats won. At a more general level, our analysis tends to support the state-level research of Ruth Jones (1981), who notes that on balance, public financing benefits the minority party more than the majority party.
In an electoral system in which challengers find it difficult to raise the funding necessary to mount a serious campaign against an incumbent, the minority party necessarily is disadvantaged. Even in the wake of the 1994 electoral "earthquake," defeating an incumbent remains a formidable task. Jacobson (1996, p. 10), for example, reported that, fortunately for Democrats, "the electoral value of incumbency remained at a high level in 1994." One consequence of the strength of incumbency is that hopes for significant party gains rest primarily on those seats left open by opposition party retirements. Absent a large number of majority party retirements (as was the case in 1994), the potential of the minority party to make substantial gains in House elections necessarily is limited.
Having finally overcome the obstacles created by their minority party status, Republicans are now able to reap the benefits of majority party status. As Eismeier and Pollock (1996, p. 95) observed:
In the early days of the 104th Congress Republicans held a record number of fundraisers.... For tactical and strategical reasons, therefore, money is likely to pour into Republican coffers for 1996. There will be more than enough money to defend Republican seats in the House and Senate and go after open seats. There will also be a substantial pool of venture capital available to credible Republican challengers taking on the large number of Democrats in the House who lost the aura of invincibility in 1994.
For Democrats, the failure to act on campaign finance reform may have been incredibly shortsighted. As we have demonstrated in this paper, in all likelihood campaign finance reform would not have halted the Republican takeover of the House. However, those corporate political action committees that preferred Democratic incumbents because of their majority status now should prefer Republican incumbents even more.(13) As Eismeier and Pollock (1996, p. 95) noted:
Democratic survivors will probably be able to fund adequately their 1996 campaigns. But it will require more hustling, an unappealing prospect that may urge some into retirement. Between defending incumbents and attempting to win back seats lost in 1994, the resources of labor PACs will be stretched very thin. Even with the support of labor PACs, Democrats have lagged behind Republicans in money for challengers.... Republican incumbents may become entrenched, and money to defeat them, or for that matter to defend Democratic incumbents, more and more scarce.
Overall, a system of campaign financing that places challengers at a disadvantage also places the minority party at a disadvantage. In many ways, the obstacles confronting Democrats as the minority party are the same as those that confronted Republicans. Even with Democrats in control of Congress, however, the proliferation of business-oriented political action committees, both in number and in total contributions, has created a system that theoretically should have benefited the Republican party. Democratic incumbents were able to adapt and prosper under this system of campaign financing, but they benefited only so long as they could remain in the majority. Now that Democrats are in the minority, their fundraising ability should be undermined significantly, as should their ability to win back control of the House of Representatives.
Data and documentation for replicating this article can be obtained from the authors after May 1998. All statistical analyses presented in this paper were conducted using SHAZAM.
We would like to thank Adam Smith, Ron Lockaby, Don May. and Reginald Simmons for assistance in collecting the data. An early version of this paper was presented at the 1996 Annual Meeting of the Midwest Political Science Association, Palmer House Hilton, Chicago, April 18-20, 1996.
1 Krasno and Green (1993) also made the point that President Bush's impending veto encouraged Democrats to pursue campaign finance reform aggressively. Writing prior to the failure of campaign finance reform in 1994, Krasno and Green (1993) noted that the 1993-1994 period would present the Democrats with the opportunity to pass meaningful campaign finance reform, if they could overcome Republican and academic objections that spending limits diminish competition.
2 For example, Congressional Quarterly Weekly Report (June 3, 1995, 53(22), 1595-1597) reported that congressional candidates raised a record $740.6 million during the 1994 election cycle.
3 While some readers might prefer that a longer time series be included, we have focused our attention on just two election years, 1994 and 1992. Theoretically, this reflects our interest in the 1994 elections and how reform might have affected the dramatic Republican takeover of the House of Representatives. Including the 1992 data allows us to examine how closely the results based on the 1994 elections are mirrored in a less dramatic election year. While it would be possible to extend the analysis to other congressional election years (i.e., 1990, 1988, etc.), we would contend that one would observe similar patterns in other election years.
4 Other works also have reflected this general view regarding the efficacy of public financing. Magleby and Nelson (1990, p. 176), for example, concluded that "public financing and limitations would foster increased competition only if the limitations were set high enough to permit challengers to become visible." Other works have taken this logic a bit further by suggesting that public financing be offered to candidates without accompanying spending limits (Malbin, 1984).
5 We should note that there is a growing body of research investigating the effects of campaign spending at the state level (see, e.g., Gierzynski & Breaux, 1991, 1993; Stonecash, 1990; Thompson, Cassie, & Jewell, 1994).
6 According to Gelman and King (1994), the probability that the Democratic candidate won the election can he computed as follows:
Prob (Democrat Wins) = ([Y.sup.Pred] -.50)/[SE.sup.Pred Y]
Summing these probabilities across congressional districts then should give a reasonable approximation of the number of Democrats expected to win, or in this case the number of Democrats expected to lose.
7 We did test for additional interactions between nonincumbent candidate quality and candidate spending. Because only one of the interactions was significant and because inclusion of the interactions did not improve the overall fit of the model, we dropped these interactions from the final model.
8 While there still is debate regarding the appropriate estimator of a model that includes campaign spending, we have decided to use OLS to estimate our model. In the previous literature, models that have employed OLS generally have been less supportive of campaign finance reform (e.g., Abramowitz, 1991; Jacobson, 1980), while models that have employed more sophisticated estimation procedures have been more supportive of reform (e.g., Goidel & Gross, 1996; Green & Krasno, 1988; Krasno & Green, 1993). For this reason, the results presented in this analysis should be biased in favor of the status quo and, by extension, against campaign finance reform. In addition, the purpose of the present article is to forecast the number of Democrats expected to lose in a number of hypothetical reform scenarios. Our purpose is not to identify the number of incumbents expected to lose, nor to attribute causal responsibility to incumbent or challenger spending. Because we are interested primarily in forecasting rather than in making causal inferences, and because the previous literature indicates that OLS provides estimates that tend to he most supportive of the status quo, we believe that OLS is a suitable estimator for the purposes of this analysis.
9 Overall, this left 334 valid cases available for analysis in 1994 and 375 valid cases in 1992.
10 Congressional Quarterly Weekly Report, 52 (41), 1994.
11 These incumbents are "hypothetical" because we limited the sample to contested races where both candidates spent at least $500.
12 This figure is computed by first estimating the number of seats the Democrats were expected to lose based on the results of the regression model and then subtracting the number of seats held by the Republicans prior to the 1994 elections. Of the 334 total seats included in the analysis, 123 were held by Republicans prior to the 1994 elections. The baseline model then predicts that Democrats will lose 180 of 334 contested elections in 1994, resulting in a net Republican gain of 57 seats.
13 Salant and Cloud (1995), for example, reported that in the period immediately following the 1994 elections business-related PACs gave $581,949 to congressional candidates. Eighty-nine percent of this funding went to Republican candidates.
Abramowitz, A.I. (1991). Incumbency, campaign spending, and the decline of competition in U.S. elections. Journal of Politics, 53, 34-56.
Alexander, H. E. (1992). Financing politics: Money, elections, and political reform. Washington, DC: CQ Press.
Babson, J., & St. John, K. (1994). Momentum helps GOP collect record amounts from PACs. Congressional Quarterly Weekly Report, 47, 3456-3459.
Barone, M., & Ujifusa, G. (1995). The Almanac of American Politics 1994. Washington, DC: National Journal.
Bond, J. R., Covington, C., & Fleisher, R. (1985). Explaining challenger quality in congressional elections. Journal of Politics, 47, 510-529.
Campaign finance overhaul dies. (1995). Congressional quarterly 1994 almanac. Washington, DC: Congressional Quarterly Inc.
Congressional Quarterly Weekly Report. (various issues). Washington, DC: Congressional Quarterly.
Donnay, P. D., & Ramsden, G. P. (1995). Public financing of legislative elections: Lessons from Minnesota. Legislative Studies Quarterly, 20, 351-364.
Duncan, P. D., & Lawrence, C. C. (1995). Congressional quarterly's politics in America 1996: The 104th congress. Washington, DC: CQ Press.
Eismeier, T. J., & Pollock, P. H., III. (1996). Money in the 1994 elections and beyond. In P. A. Klinkner (Ed.), Midterm: The elections of 1994 in context (pp. 81-98). Boulder, CO: Westview Press.
Gelman, A., & King, G. (1994). A unified method of evaluating electoral systems and redistricting plans. American Journal of Political Science, 38, 514-554.
Gierzynski, A., & Breaux, D. A. (1991). Money and votes in state legislative elections. Legislative Studies Quarterly, 16, 203-217.
Gierzynski, A., & Breaux, D. A. (1993). Money and the party vote in state house elections. Legislative Studies Quarterly, 18, 515-534.
Gimpel, J. G. (1996). Legislating the revolution: The contract with America in its first 100 days. Boston, MA: Allyn & Bacon.
Goidel, R. K., & Gross, D. A. (1996). Reconsidering the "myths and realities" of campaign finance reform. Legislative Studies Quarterly, 21, 129-150.
Green, D. P., & Krasno, J. S. (1988). Salvation for the spendthrift incumbent. American Journal of Political Science, 32, 844-907.
Jacobson, G. C. (1978). The effects of campaign spending in congressional elections. American Political Science Review, 72, 469-491.
Jacobson, G. C. (1980). Money and congressional elections. New Haven, CT: Yale University Press.
Jacobson, G. C. (1985). Money and votes reconsidered: Congressional elections, 1972-1982. Public Choice, 48, 7-62.
Jacobson, G. C. (1989). Strategic politicians and the dynamics of U.S. House elections, 1946-1986. American Political Science Review, 83, 773-793.
Jacobson, G. C. (1990). The effects of campaign spending in House elections: New evidence for old arguments. American Journal of Political Science, 34, 334-362.
Jacobson, G. C. (1996). The 1994 House elections in perspective. In P. A. Klinkner (Ed.), Midterm: The elections of 1994 in context (pp. 1-20). Boulder, CO: Westview Press.
Jones, R. S. (1981). State public campaign finance: Implications for partisan politics. American Journal of Political Science, 25, 342-361.
Krasno, J. S., & Green, D. P. (1988). Preempting quality challengers in House elections. Journal of Politics 50, 920-936.
Krasno, J. S., & Green, D. P. (1993). Stopping the buck here: The case for campaign spending limits. Brookings Review, 17-21.
Magleby, D. B., & Nelson, C. J. (1990). The money chase: Congressional campaign finance reform. Washington, DC: The Brookings Institution.
Malbin, M. (1984). Looking back at the future of campaign finance reform: Interest groups and American elections. In M. Malbin (Ed.), Money and politics in the United States: Financing elections ia the 1980s (pp. 232-276). Washington, DC: Chatham House Publishers.
Mayer, K. R., & Wood, J. M. (1995). The impact of public financing on electoral competitiveness: Evidence from Wisconsin, 1964-1990. Legislative Studies Quarterly, 20, 69-88.
Salant, J. D., & Cloud, D. S. (1995). To the '94 election victors go the fundraising spoils. Congressional Quarterly Weekly Report, 53, 1055-1059.
Smith, B. A. (1995). Faulty assumptions and undemocratic consequences of campaign finance reform. Yale Law Journal, 105, 1049-1091.
Sorauf, F.J. (1992). Inside campaign finance: Myths and realities. New Haven, CT: Yale University Press.
Stonecash, J. M. (1990). Campaign finance in New York Senate elections. Legislative Studies Quarterly, 15, 247-262.
Thompson, J. A., Cassie, W. E., & Jewell, M. E. (1994). A sacred cow or just a lot of bull? Party and PAC money in state legislative elections. Political Research Quarterly, 47, 223-237.
[TABULAR DATA FOR TABLE 1 OMITTED]
[TABULAR DATA FOR TABLE 2 OMITTED]
[TABULAR DATA FOR TABLE 3 OMITTED]
Table 4 Simulated Effects of Partial Public Financing 1994 Predicted Outcome 1992 Predicted Outcome Sum of Probabilities Sum of Probabilities Voucher = $100,000 Spending Limit 400,000 61 8 500,000 63 8 600,000 65 7 700,000 66 6 800,000 66 5 900,000 66 5 1,000,000 67 5 Voucher = $200,000 Spending Limit 400,000 66 14 500,000 69 13 600,000 71 12 700,000 72 11 800,000 73 10 900,000 73 10 1,000,000 73 10 Voucher = $300,000 Spending Limit 400,000 69 17 500,000 72 17 600,000 74 15 700,000 76 14 800,000 77 14 900,000 77 13 1,000,000 78 13 Note: Cell entries are the number of Democrats predicted to lose when each individual probability that the Democrat lost is summed across all districts. Parentheses indicated the number of seats the Democrats would lose in the hypothetical scenario.
Donald A. Gross is a professor in the Department of Political Science at the University of Kentucky.
Todd G. Shields is an assistant professor in the Department of Political Science at the University of Arkansas.
Robert K. Goidel is an assistant professor in the Department of Political Science at Indiana State University.…