Conventional wisdom holds that politicians do not keep their promises. According to the 2006 Role of Government IV International Social Survey Program (ISSP) survey, only 26 per cent of Canadians think that “their elected representatives make the effort to keep their election promises.” Canadians are not the only ones to hold this opinion. Out of 21 countries that participated in the survey, the average percentage of respondents stating that politicians keep their promises was 20 per cent. Israel had the lowest percentage (8 per cent) and Switzerland the highest (37 per cent).

Political leaders accuse one another of lying on a regular basis and media stories frequently portray political leaders as promise-breakers. Moreover, political scientists are reluctant to embrace the notion that candidates actually carry out their campaign promises. In their classic work Absent Mandate, Harold Clarke and collaborators affirm that “absent mandates are likely to be the rule, not the exception. Elections decide who shall govern, but rarely the substance of public policy.” For his part, Anthony King, Canadian-born political scientist at the University of Essex, affirmed that “party manifestos are empty and meaningless documents having a virtual random relationship to what a party will do in office.”1

Parties keep their promises

Recently a team of researchers decided to systematically test whether Canadian governments indeed ignore their own campaign promises. Answering this question is important. The fulfillment of election promises is at the heart of democratic accountability. Citizens should expect that the winning party will mobilize the resources necessary to fulfill the mandate on which it has been elected, and judge the government accordingly. The underlying logic and expected outcomes are clear at least in a majoritarian system of government such as Canada’s: the commitments that the ruling party promises to fulfill are the winning issues and policy options, those that led the party to victory in the first place.

To do an empirical “reality check,” our research team collected the pledge fulfillment records of seven successive Canadian governments from 1993 to 2015. We first manually recorded specific individual pledges contained in the officially sanctioned election program of the wining party at each election. Pledges are defined here as statements containing unequivocal support for proposed government policy actions or outcomes. The “unequivocal support” part of the definition stipulates that a statement must contain an explicit promise to do something. Statements in which parties promise to “consider” or “look into” specific policy actions do not qualify. The second part of the definition stipulates that pledges contain proposed government policy actions or outcomes that are testable. This requires that the election program author, not the researcher, provide the criteria that serve as the basis for assessing whether the pledge is fulfilled.2

Once recorded, individual pledges are classified in three categories: “fulfilled”, “fulfilled in part” and “unfulfilled.” To be classified as fulfilled, a pledge has to be followed by a subsequent government action (a law, a regulation, a policy statement, a budget allocation, a treaty or an agreement that has been passed). A pledge is rated “fulfilled in part” when the corresponding action is a compromise or is in the process of being completed, but below the standard or beyond the timeline set by the promise wording. Pledges are classified as “not fulfilled” when they are not followed by a government action.

They are then classified as “at least fulfilled in part” or “not fulfilled” for analysis purposes as well as comparability with studies by researchers in other countries that use the same methodology. Pledges are coded and matched with government actions by teams of trained research assistants. Intercoder reliability tests are conducted, and intercoder agreement is usually well above 90 per cent.

Figure 1 displays the proportion of pledges fulfilled at least in part by each successive governing party between 1993 and 2015. The rate of pledge fulfillment by the Liberals has varied between a low of 53 per cent for Jean Chrétien’s first government (elected in 1993), and a high of 77 per cent for his third government (elected in 2000). The rate for the Conservatives has varied between a low of 62 per cent for Stephen Harper’s second government (elected in 2008), and 84 per cent for his third government (elected in 2011).

Out of a total of 828 pledges recorded over the period, 565 (68.2 per cent) were at least partly fulfilled and 253 (31.8 per cent) unfulfilled. The Liberals made 392 pledges in the four elections between 1993 and 2004, and fulfilled 252 of them at least in part between 1993 and 2006 (average fulfillment rate of 64.3 per cent). The Conservatives made 436 pledges in the three elections between 2006 and 2011, and fulfilled 313 of them at least in part between 2006 and 2015 (average fulfillment rate of 72.5 per cent). The difference between the Conservatives’ and Liberals’ score disappears almost entirely if we remove the very low Liberal score in 1993, an outlier. Let’s consider the factors that shape the likelihood of pledge fulfillment.

One possible factor is majority government. According to conventional wisdom, majority governments can more easily carry out their campaign pledges than minority governments. This is not necessarily the case. Minority governments often take advantage of informal voting coalitions, allowing them to conduct their business as effectively as minority governments. Paul Martin’s minority government (2004) was better able to fulfill its pledges than two of Jean Chrétien’s majority governments that preceded it (1993 and 1997). In fact, minority governments appear to fulfill fewer pledges than majority governments simply because majority governments last longer.3 The average pledge fulfillment score for the four majority governments in our data set is 70 per cent, a mere three percentage points above the average score for the three minority governments.

Another factor with some impact is reelected governments. Paul Martin’s minority government was able to fulfill a relatively high proportion of its campaign pledges in part because it was a reelected government. Reelected governments are better positioned than newly elected governments to fulfill their campaign pledges because they can obtain public-service advice and expertise in the process of drafting their election platforms and, subsequently, in the day-to-day operations of government.4

Reelected governments may also take advantage of projects already in the legislative “pipeline” when the election was called, an option unavailable to newly elected governments. This is dramatically illustrated by the high score of Stephen Harper’s Conservative government elected in 2011. During the 2011 campaign, the Conservatives promised that if reelected they would immediately pass several bills which had died on the order paper when they were defeated by a vote of no confidence. The newly elected Conservative majority government fulfilled no fewer than 23 promises in the first three weeks of its mandate.5 By contrast, the 1993 Chrétien government was a newly elected government. As such, it did not have the benefit of experience in governing and the continuation of established policy agendas.

A third factor is budget balance. The likelihood of pledge fulfillment increases when the budgetary situation is favourable. This is exemplified by the high pledge fulfillment scores of the 2000 Chrétien and 2011 Harper governments, under both of which there was higher overall GDP growth and lower deficits than under any other government in this sample.

In contrast, faced with the 1992 recession, the first Chrétien government reneged on the fulfillment of several key promises to expand social programs (childcare, social security) in its “Red Book” election program.6 The second Harper government, which came to power on the eve of the Great Recession of 2008, provides further evidence of this. Anxious to preserve its fiscal position, the Harper government cut the budgets of several departments, which prevented the fulfillment of a number of election promises. It is no coincidence, thus, that these are the two governments with the lowest pledge fulfillment scores in our sample.

It is also possible that differences in the characteristics of individual pledges influence the likelihood of pledge fulfillment. The likelihood of fulfillment could depend on the degree of policy change that a pledge will produce if enacted and who is affected by the changes. “Transactional” pledges such as the promise to introduce a tax credit for families that send their children to art school would have a greater chance of being fulfilled precisely because they have limited effects beyond the clientele they target and consequently fewer “veto points” from which powerful organized interests can block their enactment. By contrast, “transformational” pledges, such as the promise to change the electoral system or to legalize cannabis, would be less likely to be successfully enacted because of their potentially profound impact on society.

Last but certainly not least, single-party governments such as the ones in Canada fulfill larger percentages of campaign pledges than coalition and divided governments like the ones in Italy or the United States. Figure 2 displays the percentage of pledges fulfilled in each of the 12 countries in a recent study of pledge fulfillment.7 Governments, we can see, keep 63 per cent of their campaign pledges on average. Countries above this average (U.K., Sweden, Portugal, Spain and Canada) have mostly single-party cabinets; countries below it (Germany, United States, Netherlands, Bulgaria, Austria, Italy, Ireland) have had mostly divided or coalition governments. Single-party governments are better equipped than coalitions to implement specific agendas based on a set of coherent policy ideas and can better manipulate the political process to fulfill their campaign promises.

The emphasis on promises during election campaigns is exemplified by Jean Chrétien’s Red Book in the 1993 election. The Red Book was unusual in how specific it was. Party programs before 1993 tended to contain few substantive promises and many vague statements of principle. It was also rare in Canada to have an entire program released at once. The Red Book was also unusual in that it set out costs for each promise. Moreover, the Liberals continued to release their programs in a similar format in subsequent elections and other parties followed suit.

The emphasis on promises once elected is illustrated by Stephen Harper personally reminding his cabinet ministers of the party’s campaign commitments in his mandate letters to them. Justin Trudeau went one step further by publishing the mandate letters in the media after being elected in 2015. Trudeau also created a “results and delivery” unit within the Privy Council Office to ensure that government ministers are working effectively to fulfill the party’s election pledges.8

Not only is the frequency at which parties fulfill their campaign pledges much higher than what we would expect on the basis of conventional wisdom, but differences over time and across countries can be explained in predictable ways.

Do voters know when parties keep their promises and when they don’t?

If Canadian governments manage to keep most of their promises, why do so many citizens believe the opposite? Democracy requires that citizens’ preferences be informed by what policymakers actually do.

Is the problem, then, one of political ignorance? There is an extensive body of literature, going back to Walter Lippmann, to support this conclusion. In their recent bestseller, Christopher Achen and Larry Bartels argue that the primary sources of citizens’ political attitudes are social identities and group attachments, not policy preferences or thoughtful assessments of political facts. Add to the recipe negativity bias: when citizens make inaccurate evaluations of politicians’ performance in office, they are more likely to err on the side of negative than positive verdicts.9

Our research team recently checked whether Canadian citizens’ evaluations of the fulfillment of election pledges are as inaccurate and negative as pictured in the literature. Our study asks respondents to the 2015 Canadian Election Study (CES) survey to evaluate the fulfillment of seven campaign promises made by the Harper Conservatives during the 2011 election, including a pledge to give tax credits for children’s cultural activities (fulfilled), a pledge to end the long-gun registry (fulfilled) a pledge to purchase CF-35 fighter jets (unfulfilled) and a pledge to limit the terms of senators (unfulfilled).10

Note that overall, the 9,157 fulfilled pledge verdicts (64 per cent) far exceed the 5,246 unfulfilled pledge verdicts (36 per cent), indicating that Canadians have a much more favourable opinion of their government’s pledge fulfillment performance than conventional wisdom suggests. Canadians’ evaluations are also much more accurate than what conventional wisdom leads us to expect. Although verdicts of unfulfilled turn out to be accurate in only 42 per cent of cases, verdicts of fulfilled are accurate in 84 per cent of cases. Given that there are more fulfilled than unfulfilled pledges in our sample, there are overall many more accurate (68 per cent) than inaccurate (32 per cent) individual verdicts.

This is not so different from results in other countries. Table 2 displays the relationship in Ireland, Portugal, Sweden, the U.K. and the United States. The results are very similar, with the proportion of accurate verdicts smaller by a few percentage points than in Canada.

How then to explain the surprisingly high proportion of positive and accurate pledge fulfillment verdicts? Are we really so well informed? A plausible explanation starts from the observation that the 2006 ISSP and the 2015 CES survey questions are not comparable. The ISSP survey question asked respondents whether they thought politicians keep their promises in general whereas the 2015 CES survey questions asked about specific promises. Specific pledge fulfillment questions tend to elicit more positive and accurate answers than questions about promise keeping in general. Specific pledge questions provide respondents with a factual point of reference that they can use to arrive at a reasonable verdict.

People informed about political facts will answer questions about the fulfillment of specific pledges more accurately than people not so informed. This was confirmed in our results: political information has a significantly positive effect on the accuracy of specific evaluations. In contrast, questions about politicians’ promise-keeping performance in general do not provide such reference points and thus, as our results show, these questions are bound to elicit answers that rely mainly on the respondents’ negative biases and stereotypes.

Moreover, when considering whether their government has fulfilled a specific campaign promise, citizens who are unsure of the correct answer have several shortcuts that they can use to help them respond,11 such as party identity or even personal characteristics (sex, race, regional ties) that they share with political leaders. Of course, heuristic shortcuts are not based on factual observation so there is always a risk that they may generate misinformation. Nevertheless, on balance we find that they help people reach more accurate than inaccurate verdicts.

Citizens can hold politicians accountable

Our research demonstrates that the conventional wisdom that parties don’t fulfill election promises does not pass the test of empirical scrutiny. To put it simply, despite expectations to the contrary, parties more often than not keep their election promises. And this applies to both majority and minority governments in Canada as well as the other countries studied. Observable and predictable patterns suggest that the likelihood of high scores of pledge fulfillment increases when the government is reelected and governs as a single-party executive with favourable budget balances. It applies especially to pledges that benefit specific clienteles and, we believe, those that are given greater media attention.

We also find that Canadians are able to assess correctly whether or not a particular pledge was fulfilled, but are far more negative than justified by the facts when asked in general about whether governments keep their promises. Clearly, the survey questions chosen matter. To better assess citizens’ capacity to safeguard democratic accountability, pollsters should ask specific questions about pledges, not general ones about promise keeping by politicians.

Politicians should take these findings to heart. Citizens are capable of evaluating whether politicians have kept their promises and holding them accountable in this regard. Democracy would benefit if both politicians and citizens paid closer attention to the substantive content of party platforms before, during and after elections.

Continue reading “Do Politicians Keep their Election Promises, and Does it Matter?”