I believe that politicians have some discretion to set policy, and that they should use that discretion to enact the substantively best policies they can, taking account of political and policy constraints. Political constraints include the need to satisfy voters and win elections, the status-quo bias in public opinion, low levels of political trust, and the limited policymaking capacity of our institutions. (In the words of political scientist John Kingdon, “Congress is easily fatigued”. The same goes for the executive branch.) Policy constraints include the difficulty of identifying beneficial policies, the risk of unintended consequences, and problems of implementation.
Let’s call this view pragmatism. Pragmatism in this sense is committed both to 1) achieving substantively good values and 2) realism about politics and policy.
I want to contrast this theory with accounts by two well-known political writers, Matthew Yglesias and Ibram X. Kendi. First up today, Yglesias.
Yglesias has an interesting but ambiguous post up on his substack about Afghanistan. I’m interested in the normative theory of politics he appeals to, especially the role of public opinion. His argument goes like this (this is my paraphrase, except for the quotes):
Withdrawing U.S. forces from Afghanistan was correct on the merits
Generally, politicians should “say and do things that poll well”, a position he calls “popularism”
Withdrawing from Afghanistan polled well and seemed relatively safe politically
In fact, however, media coverage “has been hysterical and lacking in perspective.”
This hysterical, unbalanced coverage was surprising – to journalists, to Yglesias, and (Yglesias suspects) to the Biden White House
As a result of unbalanced and misleading media coverage, Democrats may do poorly in the midterms, and the entire episode may end up reinforcing hawkish and damaging foreign policy myths: “And that’s the problem with doing the right thing regardless of the consequences — if you lose the political fight, you tend to end up losing the substantive fight over time anyway.” Also this: “But my fear is that you only get to take so many political risks in life, and having rolled the dice and lost politically on Afghanistan, it’s now less likely that Biden will challenge the hawkish establishment in other areas like Iran policy that are probably more important.”
The only real alternative to the Biden withdrawal would have been to keep American forces in Afghanistan indefinitely: “This is the can-kicking logic that led to Obama’s incomplete withdrawal and that governed Trump’s decision-making until he was a lame duck . . . the objective political incentives favored it.”
And his conclusion: “. . . I think it makes sense for politicians to be hesitant to run big risks for the sake of doing the right thing.”
It’s not entirely clear that he thinks withdrawal was a mistake, all things considered, but at the very least he nods in that direction (while also acknowledging that if Biden had not withdrawn, he (Yglesias) would have criticized this decision).
So, what should we make of this argument, especially his qualified endorsement of “popularism”?
As I said, I believe that politicians should try to enact the substantively best policies they can, taking account of political and policy constraints, including public opinion and the risk of unintended consequences. We live in a democracy, so responsible leaders should and generally will design and prioritize policies to maintain public support. This is especially important when the political opposition is committed to some really bad policies and is a threat to democracy to boot. But within these constraints politicians should also focus on the merits of different policy choices. If wage subsidies are the best way to help low income workers, but minimum wages are more popular, it might make sense to support minimum wages rather than wage subsidies (assuming you think minimum wages are beneficial). And there are very real limits on the capacity of the President and the Congress to get things done, so it is critical to prioritize effective and substantively important policies, not just policies that happen to poll well with voters.
This is not popularism, understood simply as saying and doing things that poll well, although I think it is actually close to what Yglesias believes.
Yglesias seems to qualify his support for popularism in two ways that bring it closer to my view. First, he endorses a version of popularism in which politicians should support popular policies after taking account of what political scientists sometimes call “latent preferences”, the preferences that voters will hold after a policy has been implemented.
Second, Yglesias assumes there are limits on political capital (you only get to take so many risks) and a real risk of unanticipated consequences. If unanticipated consequences tend to be negative and deplete your ability to move your agenda forward, then it is best to do nothing unless a policy is really important. (Status quo bias in public opinion may also limit how much political leaders can do and necessitate prioritization even if the risk of negative unanticipated consequences is low.)
In this qualified version of “popularism” public opinion plays an important role – you only push for policies that can garner and maintain public support – but within this constraint you still prioritize policies by considering their merits, and their importance relative to other policies.
Thus, I think Yglesias largely endorses pragmatism as I defined it above. When he argues for taking public opinion seriously, I think he is most concerned with the politics of progressive idealists who do not always acknowledge that some progressive ideas are unpopular. I agree that this is a problem, but it is still important to focus on the merits of policies. When Yglesias endorses doing and saying things that poll well, it sounds like he is endorsing unrestrained pandering on behalf of politicians. I don’t think this is an attractive model of representation in general, and in the end I do not believe it is what he really supports.