How Well Do Policy Makers Represent the People?

>> Brandice Canes-Wrone: Today we're gonna talk about policy representation in the contemporary United States. Public opinion as Madison said is sort of the sovereign in some level in political systems, it affects who's elected, elected officials respond to public opinion. And in states like California or systems like Switzerland, there are also referenda that affect public opinion, but studying this is really tough.

So around election time I tend to get calls from people like my dad, but other friends and relatives sort of informing me that the polls are wrong and nobody knows what they're doing. And I sort of wonder, I don't know what they think I'm doing most of my time when I'm not in the classroom, we know it's really tough, measuring public opinion is really hard.

Even if you think you have a great estimate of it, you have to be worried that it might be manipulated in some way. Then there's this question of how does it relate when it comes to representation to policy. And that's tough because you're comparing somewhat different things, right, elected officials who are voting on bills and these sort of broad policy polls or questions.

And then there's even the normative question of what do we really want representation to look like? We wouldn't want a system, right, that just followed every issue in some sort of majority opinion way. We care a lot about minority rights, we care about opinion being informed, yet we have a sense that it's a bad thing if policy is too far afield from public opinion.

So today we're actually not gonna talk about that last one that much, except in the Q and A. I'm already gonna try to get through a lot in the first three, but we're gonna try to get through all of how well are mass preferences represented by us policymakers?

We're gonna talk a bit about the measurement of public opinion, obviously there could be a full three hour seminar on how do you measure polls. So we'll save some of that for the Q and A, we're gonna focus on this difference which you saw if you did the readings on responsiveness versus congruence.

Which is when opinion changes, does that affect what policymakers do? And then how congruent is policy with what public opinion, what you might call majority opinion wants at a particular time, has representation declined over time? As a preview of that, the answer as you saw in the readings is yes, but I'm gonna go in a little more detail today with some updated research as well.

Which is it really depends on where you're assessing the date from, so yes, it's declined relative to the peak of American history. But from many periods it actually looks pretty good, so we'll talk a bit about that. There's a lot of talk in the media about whether the affluent have outsized influence, we're gonna spend a bit of time on that particular subject.

And again bringing in even some very recent research, this is a difficult topic I'm kind of close friends with the people, so it was one of those topics. When I heard this was being recorded, I thought, I better be a little careful, close friends with the people who make these claims that the affluent have outsized influence.

When you read their research, it's a little more nuanced, the research itself is very nuanced, the way it gets portrayed in the media is usually not very nuanced. And in addition, there have been some very substantial critiques against it that I think are excellent critiques and that really suggests that the affluent don't.

They might have a little more influence in some areas of policymaking but not nearly as much as is commonly presumed. So we'll talk a bit about that and then finally this, how much can leaders shape or lead mass opinion? And again, the answer is gonna be not as much as you'd think, so this sort of sums up to representation, so say, isn't really as bad as people tend to think it is.

Now again, there's this normative question, what do you think representation should look like? But usually when the critiques come in, it's that policy is not really being responsive to people. So you may not like it when it's responsive to public opinion, that's kind of a separate issue, but it's much more responsive than is commonly presumed.

Okay, so first, what do we mean when we talk about individual public opinion or individual attitudes? And it's worth just pausing and thinking about that, so scholars tend to differentiate between two types of policy positions. That's kind of our policy positions are sort of the standard, do you favor this trade bill, do you not favor it?

There's both current opinion, do you think we should have a trade bill with China? And then there's latent opinion, and latent opinions really, what would you think after the bill is either enacted or not enacted as the policy consequences become clear? And that's not always the same thing, right, so political scientists tend to put more weight than I'd say, and actually not just political scientists.

But elected politicians, we do on latent opinion, cuz that's what you really care about when election time comes. Then there's the prioritization of issues, so you can have a situation where there's largely some agreement between policy and preferences on some issues. But then there's kind of massive disagreement on others, but those are the priorities.

We talk about ideology, which is sort of how interventionist do you think the government should be in the economy, which is obviously related to policy positions. But some political scientists think this is actually sort of more fundamental and sort of causes people's positions on individual issues. And then there's individual opinions of candidates, which are obviously related to these other concepts but can be quite different.

You can have a high favorability of an individual candidate or a low favorability, and like their policy positions or not. You can think they're competent or not competent and have different opinions about their policy positions. So our measurement tends to follow these kind of different conceptions, right? So surveys and polls, which you're probably inundated with requests for, which is why it's so difficult to measure, cuz non response is such a large issue.

Typically, they're a binary choice now between the status quo and a particular policy option. And just to provide a little historical context, we tend to think of the difficulty of measuring public opinion today. Some level it's nothing new, yes, the Internet's new, social media's new, concerns about the pollsters being very many polling agencies being.

Having more partisan bias, you could say is not what the case was in the 1990s in the same way, so in that sense it's new. But when George Gallup founded what's considered the kind of first modern polling agency, all of those concerns were there. Basically, he was the first nonpartisan polling agency, right, and he came in and he found a better way to sample people.

And think of how tough it was to sample people in the 1930s when he was doing this, right? And he got into this business because his mother in law was running for election for secretary of state in Iowa, and he wanted to help her. And he had some background in the entertainment industry trying to predict the success of movies, and he took that and brought it to politics.

So sometimes you hear that Washington's the Hollywood of the east, so there's not complete truth in that but in the polling area there's some truth. So a lot of those problems Gallup was facing are the same ones we're basically facing today. How to get a good sample, how to make sure it's a nonpartisan conception of how the poll is being constructed?

Okay, so we've also got these focus groups and deliberative polls, and this relates to this concept of latent opinion. And actually if you're interested in deliberative polling, right here at Stanford, we have a number and at Hoover, Larry diamond is very active in this. There's a center here that's kind of the leading center on deliverative polling, excuse me.

And what this does is it brings everyone together for a weekend, it's pretty expensive, this is the big downside. And experts are brought in and both sides are presented on a contentious issue. And the people who are brought in, who are designed to be a representative sample, are they debate and discuss the issue.

And then you find out, what would people really think about this if they were presented with both sides over a period of time and really thinking about the issue? So this is valuable, a focus group is kind of a cheaper and quicker way of doing it, where you bring a smaller group of people together for a few hours, often less structured focus group.

But to say what would happen, they might respond this way when a poll's first presented, but what happens once people start hearing what others say about the issue? So these concepts in different ways try to get a latent opinion, then there's a lot of ways we try to measure.

And I hear it's political scientists as well as politicians of course, priorities, and you look at correspondence to officials, obviously. Now social media is captured, you look at memberships and organizations, activism, this is really tough. This is even tougher than that first one or even the second one, which is that obviously this sort of prioritization can be much more easily manipulated.

And politicians are aware of this, if any of you have interned in congressional office. Which I assume many of you have, everybody's very aware that interest groups are trying to present issues as highly salient to your constituents, and it's really salient to the interest group. So it doesn't mean it's impossible to gauge any leverage on this, but it's particularly tough.

Okay, so that's a little bit of a background, and wrapping up the background, this is sort of as in one of the readings, what do we mean when we say that we're measuring representation? And this is really important because often people are talking about two pretty different concepts, right?

So we have responsiveness, which is this as opinion changes, do policymakers adjust what they're doing, and does it seem to matter? And as we're gonna talk about the answer is largely yes, a pretty resounding yes from academic studies of a variety of sorts of a variety of periods.

Then there's this concept of congruence and you might say, well, it doesn't really matter if it's sort of 51% versus 49%. But what if it's 60% of the public supports a policy, shouldn't that be the policy? And there the answer is there's not that much congruence, and that's inherent to the system in some ways when it's a policy that is not the status quo.

So congruence looks pretty good if your goal is to keep the status quo, it doesn't look as strong now. It's not gonna be zero, because as I just said, as opinion moves, you're more likely to see something enacted. But if what you want out of representation is every time 55% of the public supports something you get it.

You're not gonna get that in a system with a separation of powers like the United States, it's not designed to do that. And so you can say you don't like the system, saying the system isn't working well is sort of not very carefully worded. You could say, I would like a parliamentary system where things happen really quickly, that's a legitimate viewpoint, it may not be my viewpoint.

It's not but it's a viewpoint and people have it, and it's not a crazy viewpoint at all. But saying that a system that's designed to bias against the status quo is performing poorly because that's what it's doing kind of isn't a really fair assessment of the system. But I also don't wanna be deceptive and I'll provide some more data that those who sort of talk about how congruence is low are not misinformed.

Okay, so here's the evidence on responsiveness, some people have done this by looking at sort of all issues on these ideological scales that I meant kind of ideology. As I mentioned earlier, what happens if there's a big conservative shift in the policy mood, what happens? And the answer is major laws passed move in a conservative direction and vice versa, of course in a liberal direction, this is about half due to electoral turnover.

So voters come in and they just throw out the people who aren't doing what they want and about half, interestingly, not even due to electoral turnover. So politicians are responsive in both parties, we'll talk briefly about sort of, it's in the readings, this sort of Democrats are responsive but Republicans aren't.

The good news is in the academic literature, that's largely nothing, and accepted viewpoint, not completely unaccepted but largely not accepted, but both parties tend to rationally anticipate. And then the rest of this responsiveness comes from electoral turnover, there's other studies which I've been a part of, that look at individual issues, sort of.

Let's kind of use these polls as best we can and measure across a really large range of issues what's happening when public opinion shifts. And what you see is about a 10 percentage point increase in public opinion on an issue shifts a congressional member's willingness to vote for that.

Even holding their party constant and their pass votes constant by about five to six percentage points, you can say, that's not a lot to me, I get it. But it's there, it's real, it's across a variety of issues, and then we read that cool field experiment paper from New Mexico where the authors go in, write.

And there's this new special session, or there was a special session in New Mexico and a new bill, and so members don't really know, right, what their constituents are thinking. The academics got the public opinion, they informed the members, and the treatment effect is about a ten percentage point shift in the members vote again.

Kind of changes their likelihood of voting against their party by about 10 percentage points, just by informing them of district opinion, right? Just knowing it makes that kind of a difference, so these effects are real, those are just some. Obviously we're not gonna cite everything in the literature, but they give you a flavor of what's out there, this is a very consistent finding.

There's almost no one who says, even the biggest doubters and the biggest critics of the American system don't say this isn't happening, but what they do say is that this congruence level is fairly low. So if you group together everything, changing the status quo, keeping the status quo, all issues, you get about 50% to 60% congruence, now how do you know you have all issues?

You don't but you have issues that there are polls on, so these are actually probably the issues on which you're more likely to see congruence much higher. As I mentioned, if you're keeping the status quo, then changing it much, higher if the respondents party aligns with those in office.

And then again higher as opinion increases, so it looks much better, of course, if 70% of the public supports something, it also is pretty conditional on the office holder. And if you start breaking it out rather than kind of policy, what's policy today? And you look at what individual politicians are supporting, it works exactly as you'd think from your AP government and intro american politics classes.

So this is where I'm sometimes a little sensitive to people thinking we're kind of AP gov and that's what we do. But here what you do find is when we do all this with the numbers and with getting these great samples, it looks exactly as you'd see. So this is research I did a while back with Ken Schatz, who's here at the graduate school of Business.

So we looked at sort of when is it the president seemed to support a popular policy versus not. And even among first term presidents, what you see is it looks really good if they're running for reelection in the next two years. And they're in that range of approval where they have a really good shot at winning, but it's far from slam Dunkin, right?

So below 60% approval, but above 50%, 80% congruence, we're like swinging. But in the first half of their term, it's below 50% regardless of what their approval level is, so that looks pretty bad. It's almost worse than you'd expect by chance, so they're often kind of getting things in early that they know they can't do later in the term when they're running.

So but this is kind of what we'd expect, that people who are running for reelection soon are gonna be catering to public opinion, particularly if they have a good chance to win. And the system was designed this way, again, you might like the system, you might not like it, but that's why we have some places like the house where they're up for reelection every two years.