A CRITICAL READING
Public Campaign Communications Department
May 20, 2003

The General Accounting Office has just released its study of public financing in Maine and Arizona. Entitled “Early Experiences of Two States That Offer Full Public Funding for Political Candidates,” the report documents a number of changes that have taken place since the inception of Clean Elections systems: more candidates running, more contested races, more races decided by competitive margins, a modest increase in citizen confidence in government, mixed changes in the cost of campaigns, a rise in independent expenditures on express political advocacy, and mixed changes in voter turnout. With this data in hand, the authors of the GAO report chose to caution their readers that it was “too early to draw causal linkages to changes, if any, that resulted from the public financing programs.”

That caution is a proper one. With only two elections from which to study Clean Elections’ impact on legislative races (the first of which having been clouded by legal challenges to the implementation of the new system that artificially reduced candidate participation) and only one cycle to examine its impact on statewide races, it is far too soon to draw firm conclusions.

But a close reading of the GAO report suggests that its authors have been perhaps too cautious in their analysis.

The GAO identified five measures for evaluating the public funding programs in Maine and Arizona:

• did they increase voter choice by encouraging more candidates to run?
• did they increase electoral competition?
• did they reduce the influence of special interests and enhance citizen confidence in government?
• did they curb increases in the cost of campaigns?
• did they increase voter participation?

On each of these measures, there are encouraging signs in the GAO’s own research. This is especially the case when one looks at variations between the off-year elections of 1998 and 2002. Significantly more candidates participated in the public funding system in 2002 than in 2000, in both states. The 2000 maiden run was affected by legal and technical uncertainties that deterred many potential participants from using the system. Thus it may be wisest to focus on comparisons between 1998 and 2002 to begin to judge the full effects of the new system.

1. Voter choice
According to the GAO report, there were “no notable changes in the average number of candidates per race” in the two states’ legislative elections. But the report’s statistical comparison chart (p. 19) shows significant increases in 7 out of 8 categories of races between 1998 and 2002 in both states (House and Senate primaries and generals). For example, in Arizona, the average number of candidates in House primaries went from 2.2 to 2.7, and in House general election races from 3.1 to 3.5. At first glance, that may not appear like much of difference, but spread over 60 seats that translates into an additional 30 candidates in the primaries and an additional 24 in the general. In the Arizona Senate, the average number of candidates went from 1.2 in the primaries to 1.4; in general elections it rose from 1.5 to 1.7. With more than half of the participating candidates stating the availability of public funding was a significant factor in their decision to run for office, and the near doubling of the number of such candidates between 2000 and 2002, it seems safe to say that public funding led to a greater number of candidates and hence more voter choices.

The GAO did not examine the impact of public funding on the diversity of candidates running because, it noted in a footnote, neither Maine nor Arizona’s Secretary of State collects data on candidates’ gender or race/ethnicity. That data, however, does exist in the studies done by the non-partisan Maine Citizen Leadership Fund, Arizona Clean Elections Institute, in conjunction with Northeast Action’s Implementation Project. It shows a 12% increase in women candidates in Maine and a 10% increase in Arizona between 1998 and 2002. Women chose to participate in public funding at much higher rates than men in both states, a reflection of their greater difficulty raising funds through traditional methods, and a further sign that the new systems attracted a more diverse candidate pool. As for race, data comparing minority participation in Arizona only covers 2000 and 2002, but it shows a substantial increase in the number of Latino and Native American candidates in 2002. Thirty-seven candidates of color ran for office in 2002, compared to only 13 in 2000. Of those 37 people, 21 opted for public funding.

2. Electoral competition
The first measure of competition is simply the number of races involving two or more candidates running for the same office. Comparing the two off-year cycles of 1998 and 2002, one finds a small improvement in the number of contested races in Maine’s primaries (from 6 to 8%) and a substantial increase in Arizona’s primaries (from 18 to 38%). The data are inconclusive for the two presidential year cycles of 1996 and 2000, but again this may be at least in part due to the newness of the public funding program in both states in 2000. Significant percentages of the contested races in both states included publicly financed candidates. Their impact did not go unnoticed by the GAO’s researchers, who wrote, “If publicly financed candidates had not run in Maine’s 2000 and 2002 primary elections, there would likely have been fewer contested races….[and] since such a high percentage of the contested primary legislative races in Arizona’s 2000 and 2002 elections had publicly financed candidates, as in Maine, there would likely have been fewer contested races if those candidates had not run.”

Incumbent re-election rates are another measure of electoral competition, and there has been only a modest diminution in incumbency since Clean Elections was implemented in both states. But the GAO study does show a decrease in re-election rates in both the Maine and Arizona House of Representatives, though not in the Senate. This may suggest that challengers, who are often newcomers to politics, have a better chance of unseating an incumbent in a smaller district, where the non-financial advantages of incumbency are less salient. Observers in both states report that the entry-level office for most budding politicians is the state representative’s seat, suggesting more competition for those offices. In addition, thanks to term limits, many ambitious politicians are apparently now waiting until a Senate incumbent is termed-out of office, creating an open seat and thus improving their chances of advancement. The effect, then, is to produce higher re-election rates for Senate incumbents, as potential strong challengers are simply biding their time until term limits moves the incumbent out of their way. The GAO also notes that 40% of the House incumbents who were defeated in Maine in 2000 and 71% of those defeated in 2002 lost to publicly financed candidates. Of the two Senate incumbents in Maine who lost their seats in 2000, one was beaten by a publicly financed candidate; in 2002, both defeated Senate incumbents lost to publicly financed candidates.

A final measure of competitiveness is incumbent victory margins. The margin of victory is an indication of the size and scope of a campaign, including the level of debate over pressing issues. Incumbents who win by smaller margins tend to be more accountable to voters once elected. Focusing again on the variation in results between 1998 and 2002, the GAO report shows an increase in the number of races decided by a margin of less than 15% in both of Maine’s houses, and in Arizona’s senate races as well. (Arizona’s House of Representatives has multi-member districts, making comparisons of victory margins unworkable.) For example, in 1998, just 22% of Maine’s senate races were decided by 15% or less; in 2002 that percentage had more than doubled to 50%. The impact on Maine’s house races was smaller, but still significant, increasing from 21% competitive in 1998 to 29% in 2002. The GAO’s researchers tried to further explore the relationship between public funding and competitive races, but offered no firm conclusion, citing “the relatively few variables we used and the limited amount of data available.”

3. Special interest influence and citizen confidence
The GAO reported conflicting responses to its surveys of candidates and interest groups as to the impact of public funding on special interest influence. Not surprisingly, participants in public funding said it made them less beholden to contributors, while non-participants downplayed or denied this impact. Similarly, representatives of powerful interests were divided in their opinions, with most denying that public funding had reduced their influence on the process. One has to wonder why the GAO’s researchers thought they would get objective answers from the very representatives of groups said to have too much influence thanks to their ability to give large campaign contributions! Rather than ask participants in the political process like candidates and lobbyists, the GAO might have sought out the views of more objective observers—such as journalists, academics, and retired elected officials—as to how much public funding was affecting special interest influence. Not having done so, the GAO report offers little insight into the question.

One interesting finding that was downplayed by the GAO researchers surfaced in its polls of voters in both states. While a majority in both said it was too soon to tell if the new systems had reduced special interest influence, in Maine one-quarter of the survey group said it had, compared to 7% who thought it had increased their influence. Even more significant, in both states there was a real improvement in public statements of confidence in government. In Maine, a net of 9% of voters said that the law had greatly or somewhat increased their confidence in government (17% said it did, while 8% said it decreased their confidence); in Arizona, a net of 6% (21% increased confidence minus 15% decreased) were similarly moved. Given how difficult it is to improve public confidence in government in the first place (see Robert Putnam’s book Bowling Alone [Simon & Schuster, 2000] for a comprehensive discussion of the decline in public trust over the last two generations), we ought to hail these modest but real improvements in citizen confidence as a salutary benefit of offering candidates full public funding for their campaigns.

Finally, with only one-third of Maine’s legislature and one-fifth of Arizona’s legislature in 2001-2002 made up of candidates who ran with public funding, it is far too soon to draw any conclusions from surveys of candidates, lobbyists or voters as to the new system’s impact. With approximately double those proportions populating both state legislatures in 2003-2004, researchers would do well to revisit these questions at the end of 2004.

4. Cost of campaigns
On this topic, the GAO results are genuinely mixed. Average spending by legislative candidates dropped in Maine in both 2000 and 2002 compared to the two previous cycles, but it rose substantially in Arizona. Part of this effect is explained in the report by the fact that Arizona’s system makes no distinction in how much public funding a candidate is given for a House or Senate campaign, pegging the amount to the higher historic cost of running for the Senate, which meant that participating candidates for House seats received substantially more in 2000 and 2002 than the average spent by traditionally financed candidates in the past. But this does not explain why the cost of the average Arizona Senate race rose. A possible answer not considered by the GAO is the fact that the number of uncontested Senate seats dropped from 20 (out of 30) in 1998 to 9 in 2002. Contested races invariably lead to higher spending, at least by candidates not bound by spending limits.

As for the impact of public funding on statewide races, at first glance the results appear mixed in Arizona. The average amount spent on the Attorney General, Secretary of State and Corporation Commissioner’s offices dropped between 1998 and 2002, in some cases quite dramatically. Conversely, the cost of the races increased for Governor, State Treasurer, Superintendent of Public Instruction and State Mine Inspector, again in some cases dramatically. The GAO read this as a mixed result, but it can in fact be more easily explained by the fact that a) the governor’s race in 2002 was very competitive and fought down to the wire, and b) in 1998 there was no Democrat running at all for State Treasurer and only last-minute low-spending write-in Democratic candidates for Superintendent of Public Instruction and State Mine Inspector. Had the GAO taken these facts into account, it might have come to a different conclusion about the rise in spending in Arizona.

The fact that the costs of legislative races dropped between 10 and 20 percent in Maine in the two public funding cycles compared to the prior two cycles, and that costs dropped in Arizona by a similar degree between 2000 and 2002, suggests that the new system may indeed be tempering the arms race mentality of fundraising. Certainly for all the candidates who participated in the new system, there is no rising cost of campaigning anymore. Once they qualify for public funding by collecting a large number of $5 contributions, they are no longer on the fundraising treadmill. This salient fact, which is true for 59% of Maine’s current legislators and 36% of Arizona’s, was not mentioned at all in the GAO’s report.

One unequivocal change in a negative direction noted by the GAO is the increase in independent expenditures in both states. These are moneys spent by independent committees to expressly advocate the election or defeat of a candidate. They rose from relatively negligible amounts in 1998 to over a half million dollars in Maine and over $2.6 million in Arizona in 2002. It may be that this occurred because money that would have previously gone directly to candidates, but now cannot be given to participants in the public funding system, is being spent in other ways. Some of this increase may also be explained by the sheer competitiveness of the Arizona gubernatorial race and a tough fight in Maine over control of the evenly divided state Senate. Those two cases attracted the bulk of independent spending noted in the GAO’s report, both of which drew the intense interest of outside parties. Finally, this development is clearly in keeping with national trends showing more money going into independent expenditures and issue advocacy in state and federal races. Thus it is not at all clear how much of this phenomenon is due to the two states’ new public funding systems.

5. Voter participation
The GAO notes that voter turnout in Maine increased 4% in 2000, but held steady in Arizona, compared to the 1996 election. Not included in the report is the fact that turnout in 2002 rose 5% in Maine, compared to 1998, and an even healthier 10% in Arizona. Clearly, changes in voter turnout will be affected by a lot more than the provision of public funding. But without reference to this more recent data on the 2002 elections, the GAO makes an overbroad conclusion, which is that public funding has “no clear link” to increased voter participation in elections. In our opinion, more research into this connection is needed.

In addition, there are many ways to look at voter participation, turnout at the polls being only one easily quantifiable example. Another important element is the extent to which voters are mobilized into supporting candidates. In 2002, voters in Arizona made 90,000 $5 qualifying contributions to support a candidate’s efforts to qualify for public funding; voters in Maine made about 30,000 such contributions. These numbers are triple the number of private campaign contributions made in Arizona and five times the number made in Maine.

Anecdotal evidence from participating candidates also suggesting a greater degree of voter participation in the campaign process. Indeed, the vast majority of participating candidates told the GAO that the program allowed them to spend more time discussing issues with voters, as opposed to raising money. Press reports and statements from participating candidates also show a much greater emphasis on involving voters as grass-roots volunteers in their campaigns.

~

To conclude, the GAO’s report on the early experiences of full public funding in Maine and Arizona demonstrates that these pioneering systems bear further study and engagement. On several measures, including increased voter choice, electoral competition, special interest influence, citizen confidence in government, campaign spending and voter engagement, Maine and Arizona are experiencing real changes, many of them in a positive direction.

With our partners at Northeast Action’s Implementation Project, we plan further analyses on the impact of these laws on disenfranchised constituencies, especially people of color, ethnicity and women, who have been underrepresented in legislatures. In addition, we want to take a closer look at the questions regarding competition and the use of independent expenditures. Lastly, and perhaps our most ambitious effort, will be a study of the impact of Clean Elections/full public funding on the climate for lawmaking. This will include interviews with lobbyists, legislators, former legislators, journalists and others. We do not plan to conduct this until after the close of this session, which is the first session where there are a majority of Clean Elections legislators in each house in Maine and a substantial number in Arizona.