Are citizens good judges of government performance? Evaluations of promise keeping by governing parties

Only a minority of UK citizens think that politicians generally keep their election promises. However, research by Robert Thomson and Heinz Brandenburg indicates that the public are  good at evaluating whether politicians have kept specific policy pledges, but this is affected by their pre-existing levels of distrust in politicians and in particular parties.

Letter to The Guardian, 2012. Picture: Brett Jordan via a (CC BY 2.0) licence

According to democratic theory, parties make promises to voters during election campaigns – promises to take certain actions and achieve certain outcomes – and attempt to fulfil those promises if they enter government office after elections. The making and keeping of campaign promises is one of the main ways in which elections are supposed to have an observable impact upon what governments do. Recent research has shown that governing parties are actually quite reliable in keeping their election promises, and nowhere more so than in the UK with its first-past-the-post electoral system, which frequently produces single-party governments that fulfil upwards of 70% of their election pledges. Yet, only a minority of UK citizens think their elected representatives keep campaign promises.

In our recent article, we address this puzzle with a survey experiment. The British Election Study included a set of questions we proposed about election pledges in the first wave of their 2014–17 Internet Panel. We asked questions about six specific pledges made by the 2010 coalition partners, Conservatives and Liberal Democrats. These included pledges that were unfulfilled (the famous tuition fee pledge of the Liberal Democrats or the Conservatives’ pledge to reduce migration from hundreds of thousands to tens of thousands), two that were at least partially fulfilled (the Conservatives’ pledge to increase health spending year-on-year) and two that were fully fulfilled (both parties’ pledge to scrap ID cards and biometric passports). These questions allowed us to test whether respondents were aware of the (non-)fulfilment of electoral promises, and how strongly awareness depends on political knowledge.

But the experimental element we added allowed us to go further than that. We split our sample in two, and posed different types of questions to each group. The treatment group was asked about ‘promises made by one or both of the parties that afterwards formed the government’ while the control group was asked about ‘policy proposals made in 2010’. The very subtle difference was that the treatment group was primed to think about parties and politicians when answering the question while the control group was not. The idea behind this experiment was to raise the salience of ‘heuristic thinking’ in the treatment group. We suspected that if we made respondents think about parties and politicians when answering questions about election promises, the answers would not be simple reflections of political knowledge, but affected by whether or not respondents trusted parties. We expected those with low levels of political trust to bring the common stereotype of promise-breaking politicians to bear on their answers, and therefore to say that promises were broken, even if the promises were kept. By contrast, we expected those with high levels of trust to have positive expectations of governing parties’ behaviour, and therefore to say that promises were kept, even if they were broken.

What we found was that respondents are actually highly capable of distinguishing between promises that were kept and those that were broken. We simply asked respondents whether they thought each of the promises had been kept or broken, without providing any additional information on what the government had actually done or not done in relation to each promise. We found that respondents were significantly more likely to say that a promise was kept if it was in fact kept, compared to those broken promises. By contrast, in other research, when people are asked general questions about whether they think politicians keep their promises, their answers are generally very negative. We also found that people who are more knowledgeable about politics are better at evaluating accurately, but even those with very little political knowledge are more often than not able to distinguish a fulfilled from an unfulfilled pledge.

Our experiment with the wording of the question yielded the most interesting finding. It showed that trust in politics really matters. People who are more trusting of political parties are more likely to evaluate pledges as fulfilled than people with less trust. But trust matters considerably more when, through question wording, people are primed to think of parties and politicians when considering the fate of electoral promises. Figure 1 shows how the trust effect in the treatment group (the party promise treatment) differed from that in the control group (the proposal treatment). Distrust, or the stereotype of promise-breaking politicians, had a large effect in the treatment group, which was primed to think about parties and promises. Those with low trust were significantly less likely to say that a pledge was fulfilled, even if it was actually fulfilled, if they were primed to think about parties’ promises than if they were not primed in this way. For those with high levels of trust, the experimental treatment made no difference.

Figure 1: The effect of trust as a heuristic on citizens’ evaluations

Probability of a positive evaluation (i.e. evaluating a pledge as partially or fully fulfilled) of a pledge that was in fact at least partially fulfilled. Unbroken lines are point estimates; broken lines are 95% confidence intervals.

Citizens are therefore good judges of government performance, capable of accurately assessing both promise keeping and breaking. This ability to judge accurately is important because it means that citizens as voters have the capacity to reward and punish parties for keeping or breaking their campaign promises. However, their assessments are also significantly affected by pre-existing levels of trust or distrust. The effect of trust is important because it acts as one of the heuristics that inform citizens’ assessments of what governments do and achieve when they evaluate performance.

This post represents the views of the authors, and not those of Democratic Audit. It first appeared on the PSA’s Political Insight blog  and is based on the article ‘Trust and Citizens’ Evaluations of Promise Keeping by Governing Parties’ in Political Studies.

About the authors

Robert Thomson is Professor of Politics at Monash University.

Heinz Brandenburg is Senior Lecturer in Politics at the University of Strathclyde. 


Similar Posts

Comments are closed.