From Coronavirus to the economy, almost everything you know about the performance of Boris Johnson’s government is wrong – commentators are using phantom metrics
Commentators overestimate government influence over policy outcomes and misconceive true indicators of government performance
Nearly all analyses of the Johnson government, and governments more generally, assume a link between performance and voter feedback. In the case of the Johnson government, commentators might write that declining support reflects failures of the levelling-up agenda or management of the pandemic. But intriguingly, evidence suggests that voter and commentator evaluations of government performance have little basis in reality. Writing in the prestigious American Political Science Review, Dynes and Holbein recently demonstrated that voters struggle to hold governments accountable. Performance metrics are outside the control of politicians, Democratic and Republican state administrations performing equally well in areas such as the economy, education and crime. Such results echo longstanding findings in political science. Conducting a meta-analysis of earlier studies, Imbeau and colleagues concluded that the ‘average correlation between the party composition of government and policy outputs is not significantly different from zero’.
This raises intriguing questions. If voters and commentators struggle to assess performance metrics, which metrics do they evaluate? Eclectic factors sway voters, researchers demonstrating the influence of sporting events, natural disasters and shark attacks (!), but the metrics used by commentators are more interesting; journalists and academics are professional analysts. In this essay, I argue that many commentators use phantom metrics. Such metrics ostensibly measure government performance, but overestimate government influence over policy outcomes. Moreover, phantom metrics misconceive true indicators of government performance, often evaluating noise associated with government presentation, media coverage and government incumbency. Relatedly, such analyses overlook subtle metrics such as agenda coordination and management of the civil service. I develop this argument with reference to the UK Johnson government. Such trends affect all governments, yet the UK is an interesting case. Many studies of the country stress the limits of partisan influence, often underlining the maturity of the political system and civil service. But paradoxically, a vocal commentariat uses phantom metrics, feeding the biases of voters.
Though commentators are not slow to pass judgement on the Johnson government, few appreciate the limits of government influence upon policy. Economic policy is a prime illustration, research showing that partisan differences are often illusions. For example, the levelling-up agenda involves multiple levels of government, the civil service being influential and many councils being Labour controlled. Coronavirus is a classic competence issue, nearly all agreeing on the need to reduce fatalities and the NHS and civil service playing key roles. For these reasons, partisan effects have been small; within the United Kingdom, countries have adopted similar approaches. On first examination, this analysis seems counterintuitive. But on reflection, it becomes more reasonable. For example, why do commentators seldom discuss the role of the civil service and local government?
When government performance cannot be properly evaluated, phantom metrics emerge. Quality of presentation is one. Crucially, the Johnson administration has projected a poor image. On Coronavirus, this has been the case. Beginning with the Barnard Castle fiasco, senior figures have compromised the integrity of government policy. Contrasts with politicians in other countries are conspicuous. Notwithstanding certain scandals elsewhere, other political classes have behaved better than the Johnson government. Elite disregard for lockdown is a policy failure, making wider non-compliance more likely. Yet such incidents are comparatively minor; aside from limited government influence over policy outcomes, the true metrics of government competence are much more complex, many being concealed from public view; for example, agenda coordination and management of the civil service are crucial. But significantly, official hypocrisy gives the impression of incompetence, harming the image of the government. Electoral conditions heighten perils, red wall voters placing high value on the observation of rules. At the next election, perceptions of pandemic mismanagement may damage the government.
Critically, Coronavirus is a preeminent policy area, voters and commentators parsing policy for signs of efficacy. There is a similar logic in economic policy, observers paying attention to high-profile agendas. Indeed, research suggests that party influence is more apparent in visible policy areas. But on the crucial levelling-up agenda, the Johnson government has made errors. Retreats on high-speed rail lines were mistakes, such infrastructure being visible and high-profile. Had the government undertaken these projects, it might have demonstrated commitment to northern voters. But as with Coronavirus, this was partly a presentational failure. It is difficult to make general conclusions about levelling-up, the agenda being multi-faceted and subject to diverse influences.
Given the importance of presentation, the media plays a crucial role. As left-wingers emphasize, the print media is friendly to the Conservative Party. This is less significant than it was once, newspaper circulation declining, yet remains a significant advantage. Had Jeremy Corbyn presided over the Coronavirus crisis, the media would have adopted a more critical line. On the other hand, incumbency disadvantages the government. When governments are in power, particularly during crises, they are forced to implement unpopular measures. The Johnson government cut universal credit by £20, pushing hundreds of thousands into poverty. Had Labour been in power, they would likely have undertaken a similar measure. Certain studies find that left-wing parties more reliably cut welfare, reflecting greater trust among low-income voters. Of course, this does not mean that one may not oppose the universal credit cut; I was against this measure. But to some extent, the notion that this was a government failure is a phantom metric, all governments making unpopular decisions.
Institutional configurations underpin such processes. Conservative parties tend to dominate countries with majoritarian electoral systems and liberal economies, such as the UK. Reflecting this, UK institutions tend to promote Conservative rule; the media is a prime example. Of course, some metrics are genuine. As I have emphasized, the Johnson government deserves censure in certain areas. Moreover, academic work on this topic has problems. Quantitative studies overlook nuance, associated with analysis of multiple cases; voters may be attuned to subtleties, judging governments accordingly. There are other reasons to oppose certain politicians. When figures such as Donald Trump hold power, they set illiberal tones, endangering minorities and threatening constitutions. Relatedly, heads of government sometimes make monumental decisions, often in the field of foreign policy. If politicians are unfit to make these decisions, as was the case with Trump, this is exceptionally serious. Politics matters desperately, particularly when extremists are close to power.
Despite these caveats, phantom metrics are a significant problem. On one level, we should regret that much analysis is untrue. Though journalists are the worst offenders, academics are not immune. Given discrepancies between scholarly findings and popular perceptions, academics should correct misperceptions. To some extent, phantom metrics are inevitable, reflecting the complexity of government and agency bias of humans, but their scale can be reduced.
On another level, phantom metrics heighten tribalism; those who are convinced of government responsibility tend to be quicker to anger. But as we know, tribalism can be damaging, weakening faith in institutions and sowing discord. Unfortunately, this behaviour partly lies beyond rational argument. As I have written before, tribalists resemble sports fans. The material stakes of politics are lower than tribalists claim, yet tribalists are emotionally invested in outcomes, à la sports fans; this encourages emphasis on partisan differences, much of which is untrue and damaging. Tribalists will doubtless continue this behaviour, modern culture and attitudinal structures underpinning this. But others must object to it, even when disagreement provokes ire.
If you enjoyed reading this, do think about subscribing! Subscription is free – all it means is that you’ll receive a weekly email. But every new subscriber makes me very happy 😊 😊 😊