Intro. [Recording date: February 19, 2019.]
Russ Roberts: My guest is economist Jacob Vigdor.... Our topic for today is the minimum wage, and in particular the minimum wage in Seattle and your efforts to measure its impact and the challenges of that. Let's start with some history. What's happened in Seattle with the minimum wage? It's kind of at the vanguard of living wage, minimum wage legislation.
Jacob Vigdor: Yes. Back in 2013, there was a lot of local activity surrounding the minimum wage. And it started in a little town south of here called Sea-Tac. It's where the airport is. There was an organizing campaign back in 2013 to get a ballot initiative together to raise the minimum wage to $15 an hour immediately. And that ballot initiative was successful. So, as of early 2014, the small city of Sea-Tac, Washington, adopted the first $15 minimum wage in the nation. Then, on the heels of that success, the other thing that happened in the election of 2013 you had a Mayor and a City Council in the City of Seattle that campaigned on the issue of raising the wage to $15 an hour. So, those candidates were successful. And so what happened after that was a few months of negotiation. So, the Mayor of Seattle at the time, Ed Murray, decided to get some labor groups and some business groups in a room together and kind of hash out a deal to figure out what kind of policy to put into place to raise the minimum wage to $15 an hour. And so, they did it. They passed a law, in June of 2014. It involved a slower phase-in schedule. And different phase-in schedules for businesses of different sizes. So, large businesses hit the $15-an-hour level in 2017. The smaller businesses just hit $15-an-hour this year.
Russ Roberts: And, of course, there are a lot of people who made predictions about what would happen. In particular, there's been a focus on the fast food industry and, because I follow people--probably a mistake, but I do--who care about this and who are constantly ranting about it--there have been both triumphant claims that, 'Uhp, see! It didn't hurt anybody. It only helped people.' Versus the opposite, which said, 'Uhp, it's decimated this particular group, or this particular industry.' Or if one of those claims is made, people on the other side would explain why it still was an open question. And obviously people care a lot about this. You have been involved in trying to assess the impact. Talk about who has sponsored that research and how you got involved.
Jacob Vigdor: Sure. Yeah. I actually moved to Seattle to take this job in July of 2014. So, just the month after the City Council passed the ordinance. I had been reading about the Seattle minimum wage in the news, and one of my first days on campus I had lunch with one of my new colleagues, Mark Long here in the Evans School, and I asked him if anybody was doing anything about studying this minimum wage initiative. And he told me that there was a little group getting together. And he invited me to join them for a meeting. So, I did. That would have been in late July, early August of 2014. And it turns out that the City of Seattle had put out a request for proposals. Because they were commissioning a study of the impact of the minimum wage. And so, I joined a group here at the U. of Washington that was putting a bid together for that contract.
Russ Roberts: And, you won that bid, presumably.
Jacob Vigdor: Well, we were the only bidder. It was kind of a funny story. Because, you know, usually a request for proposals will say things like, 'Well, here's the budget that we're looking for.' Or, 'Here's the [?] we have to award.' There was no such information in Seattle's RFP [Request for Proposal]. So, we looked at the scope of work that the City had put into this Request for Proposals. And we just sort of penciled out how much that would cost to get all that work done. And we put together a bid for $1.7 million dollars. And then, right after we submitted the bid, the Mayor put out his proposed budget. And in the Mayor's budget we saw that there was a line item for this study for $100,000.
Russ Roberts: 'Nyeh[?]. Perfect.
Jacob Vigdor: Yeah. But, being the sole bidder, what ended up happening is that the folks in City Hall just sort of called us up and said, 'Okay. Let's get together and figure something out.' So, we went in and negotiated. They actually managed to scrounge up a little bit more money than the $100,000 they had initially thought of spending. And then we told them that, 'Well, we can supplement whatever money you give us with foundation grants. So, why don't we just sort of get this project started with a little bit of seed money from the City; and then we'll go out and see if we can raise some more money?'
Russ Roberts: And you did. And you've been studying the results in a couple of different places, if I understand it correctly. Is that right? A couple of different studies.
Jacob Vigdor: So, what we've done, over the past few years--it's a multi-faceted, multi-disciplinary effort to study the minimum wage. The part that gets most of the media attention is the data analysis component, where we're looking at unemployment insurance records to try to figure out what's going on in the low-wage labor market here. But we also got a lot of survey and interview work. We've been talking to business owners and managers. And we've also been talking to 50 low-income families that have been trying to make ends meet on the basis of work here in Seattle. And, what's kind of nice is that these different facets of the study kind of corroborate each other and help us paint a more complete picture of what's going on.
Russ Roberts: Now, before you got involved with this study--let's go back to when you arrived in Seattle, even--you presumably had some idea about the minimum wage and its impact. Talk about that and how that has affected your analysis of it all.
Jacob Vigdor: Like many economists, I had come into this thinking that there were certain issues with the minimum wage. I thought that it was a not-very-well targeted way of trying to deliver higher incomes to low-income families. I thought that for every dollar that you delivered to a family in poverty, you were delivering additional dollars to teenagers and other workers that we weren't really that concerned about. I was also concerned that the minimum wage would largely be paid for largely by a tax on food. With the idea being that if you look through the supply chain of food, from farms to grocery stores to restaurants, all of those sectors of the economy happen to be places where there's a large amount of low-wage labor. And also those are sectors of the economy where the profit margins are pretty low. So, I was concerned that this would end up being passed along to consumers. So that's sort of where my priors were coming into it.
Russ Roberts: And some of those consumers, of course, are low-wage, poor folks who are going to pay more for something that's really important.
Jacob Vigdor: That's right. So, if you think about the incidence--you know, what share of a family budget goes to food--that's going to be a very regressive tax, so to speak.
Russ Roberts: And, measuring this--the effects on either the labor market or the price of food--is, like every economics measurement problem, made more complicated by a whole bunch of factors, one of which is that there are a whole bunch of factors that affect labor markets and prices, and employment, and wages; but also by the fact that this was an issue that had been discussed for some time. So, the other challenge, of course, is that some of the effects may take place before the actual legislation goes into effect, because people are anticipating it.
Jacob Vigdor: That's right.
Russ Roberts: So, what's the--how did you cope with that? Tell us about the methodology you've used and how you've used more than one kind in trying to assess--let's start with the most salient issue that people worry about, which is employment.
Jacob Vigdor: Yeah. Well, the thing that really helped us out, and what really changed my mind about whether this project was really worth spending a lot of time on: Washington State has administrative data that captures how many hours people work in a calendar quarter. And, that was kind of a game-changing piece of data for us, because it meant that we could actually compute what people's hourly wages were. Whereas most studies that have gone before this have not really had access to information on how much people earn per hour, on sort of a systematic basis, throughout the economy. So, our data came from Washington State. And so basically our strategy came down to comparing the employment patterns in Seattle to comparing employment patterns in other parts of Washington State. And so that was a central challenge for us--is trying to figure out what parts of Washington State give us the best idea of what might happen in Seattle if the minimum wage had never gone up.
Russ Roberts: And how did you go about doing that?
Jacob Vigdor: Well, we tried a number of different methods to start out. So, the classic way to do this in the minimum wage literature is to compare the place that raises the minimum wage to someplace next door. And that goes back to the Card and Krueger paper in 1994 which compared Pennsylvania and New Jersey. So we tried that: We tried comparing the city of Seattle to the area just outside the city and to counties that were a little bit further away. We found some issues with those methods--with the standard kind of border-discontinuity methods--because they didn't pass falsification tests--
Russ Roberts: Explain.
Jacob Vigdor: That steered us to using synthetic control methods.
Russ Roberts: But explain what you mean by falsification tests. What was wrong with the, say, suburban areas outside Seattle where the legislation wasn't in place?
Jacob Vigdor: Yeah. Yeah. So, the basic idea of the falsification test--you know, if the areas right outside of Seattle are a good control for the city of Seattle, then what you do and what we did, was try to run the same methodology but using data from before the minimum wage increased. And then, there, we would expect to see effects of zero. Because, there really is no difference in minimum wage between the quote-unquote "treatment" and quote-unquote "control" regions. But when we did that, using Seattle versus the areas immediately outside Seattle, we saw that there were these large-effect estimates even when there was no policy on the horizon. So, that led us to think that, underlying trend difference between the City of Seattle and the immediate suburbs of Seattle that make the suburbs kind of a poor match for the city.
Russ Roberts: So, this could be an example like--just to make it concrete--employment could have been growing in Seattle while it was shrinking in the suburbs. Which would make it--it could be growing for a particular type of labor while it was shrinking in the suburbs for that type of labor, that level of skill, that segment of the labor force. And then you wouldn't be able--you essentially wouldn't be controlling for differences between the two areas. Too much else was going on.
Jacob Vigdor: Yeah. That's right. And, you know, if you want to tell a bigger story about what's going on, since the 1980s, a lot--if you go back 50 years and track the time period from, say, 1960s to the 1980s, that was a period of time when the suburbs were growing and the cities were shrinking. And Seattle was definitely part of that trend. But then, since the 1980s, people are actually coming back to the city. And the city is enjoying a lot of population growth. The labor market is doing very well here. And so, the city is actually doing better than the suburbs by a lot of measures.
Russ Roberts: 13:18 And since they don't match each other in various types of differences that might matter for the labor market, you weren't able to use those as your effective control group.
Jacob Vigdor: That's right. We didn't really feel confident about that methodology. And so we tried something different: the Synthetic Control Method.
Russ Roberts: And explain that.
Jacob Vigdor: So, when you do Synthetic Controls, basically you are just looking around--in our case the State of Washington--and trying to identify parts of the State of Washington that have a strong track record of matching the employment trends in the City of Seattle. And, what we discovered when we did that is that the best matches for the city were not really the inner-ring suburbs that were closest to Seattle. They were actually on kind of the outskirts of the greater Puget Sound, metropolitan region. So, what you might call the ex-urbs. So, these were places that were actually growing kind of fast, because they were reflecting the population growth of the region. And that was the best match for a city which is also exhibiting population growth. So, in the State of Washington, the synthetic control pointed us towards areas that were down near Olympia, on the south end of the sort-of urbanized region. And areas up near Everett and Marysville on the north end of the region. And some parts of the state near Spokane, as well. And so these--using about 10 years' worth of data from before the minimum wage went into effect, we were matching the employment trends--the peaks and the valleys--over a period of time from about 2005 to 2014.
Russ Roberts: And, of course, you didn't know anything about why they matched, or why they matched reasons that were coincidental or important. This was just the data talking. Correct?
Jacob Vigdor: Yeah. This is kind of a black-box kind of methodology. It's not designed to give you the story. It's designed to sort of illuminate for you: 'Okay, these are the regions that really minimize the trend differences between Seattle and the control region.' And, it's only ex-post, after you see the parts of Washington State that sort of lit up in this process, that you can then kind of infer a story about, 'Okay, what makes these regions the better match compared to, say, what's right across the city limit?'
Russ Roberts: One example that comes to mind is that if you think that a lot of the impact of the minimum wage is going to show up in the food industry, you'd ideally want to have the matching areas have similar patterns of restaurant growth, expansion of employment in the food industry. Right? You wouldn't want to have--even if the labor markets are similar, you'd be somewhat uneasy if you had the data if Seattle was adding 10% in its number of restaurants a year but the suburbs--the ex-urbs--you were matching were shrinking. Or vice versa.
Jacob Vigdor: Yeah. And I think that the pattern in Seattle that we really wanted to match was a pattern of pretty strong growth. With the exception--just like every place else, Seattle's labor market shrank during the Recession--between 2007 and 2009 or so. But, in the period since about 2010, this city was growing really rapidly. And that's measured both in terms of population and in terms of employment. And, so, we were just looking around Washington State, trying to find other geographic regions that could match that pattern of growth.
Russ Roberts: And the reason for doing that is that if the minimum wage could have an impact, but it wouldn't show up in the form of fewer jobs, but not as many jobs added in that type of growth. Right? You are trying to create a what-if that you can't literally observe.
Jacob Vigdor: Sure. Yeah. So, if employment had been growing at the rate of, say, 10% a year, and then the growth rate declines to 5% a year, well, you know, you could look at the data and say, 'Well, employment is still growing.' But then, if it's growing slower than it used to, then maybe we want to infer that that's a negative impact of minimum wage. And so, what we did was compare Seattle to these other regions that have the same track record of strong growth.
Russ Roberts: And, so tell us what you found.
Jacob Vigdor: So, when we did this methodology, what we found was, first of all, the minimum wage did appear to raise wages. Which is--sort of a--that's what we expected to see. But when we looked at employment, we actually saw a reduction. And the reduction was actually a little bit larger in magnitude when we looked at hours worked rather than just simple head-count measures of employment. So, basically, what we're coming up with and the revised estimates that we've put together suggest that wages went up about 3%, as a result of the minimum wage increase; but then hours were down about 6 or 7%.
Russ Roberts: So, when you say 3% you mean for the people you expect to be affected or overall the whole population? How do you define--
Jacob Vigdor: Yeah--this was another challenge--
Russ Roberts: How do define who is affected? Is it just people who are earning within the band of their current wage and what the wage is going to be or the people just above it? Who you would also expect potentially to have an impact.
Jacob Vigdor: Yeah. So there's this kind of this existential question of what is a low wage job. And, if a low-wage job all of a sudden has its wage increased, does that make it no longer a low-wage job? What we decided to do was to look at the data and try to figure out if there was a point in the hourly wage distribution above which we really didn't see any action as a consequence of the minimum wage increase. And we ended up choosing the threshold of $19 an hour. So, when we talk about the low-wage labor market, we are really focusing on jobs paying less than $19 an hour. And that's sort of inflation-adjusted to 2015 dollars. We've experimented with different thresholds ranging anywhere down from, you know, $15 an hour all the way up to $25. I think we even tried up to $40. And the results are pretty robust to our selection of threshold--
Russ Roberts: So, when you say--
Jacob Vigdor: But that's what--yeah, go ahead.
Russ Roberts: So, when you say, incomes, or wages, or earnings I guess would be the right phrase--earnings are up 3%, or now wages are up 3%, is that for workers earning $19 and less?
Jacob Vigdor: Yeah. In the segment of the labor market with hourly wages under $19, we saw a 3% increase in hourly wages.
Russ Roberts: And for a full-time worker who didn't--well, some of those workers, of course, who got a wage increase, were able to keep the same hours, I assume. And some didn't. Can you talk about that?
Jacob Vigdor: Yeah. So, I think one of the important things about the low-wage labor market is, overwhelmingly part-time work. Some of the work that we've done with the data, we've seen that amongst people earning the lowest wages, only about 10% of them are looking like they work 40 hours a week, consistently, for, you know, periods of more than 6 months. So, it's overwhelmingly part-time work. And we see that in general there are hours decreases. The hours decreases appear to be concentrated amongst the least experienced workers. And, in fact, the work that we've done suggests that a lot of the impact of the wage policy has been not necessarily affecting people who already had jobs, but it's reducing the rate at which new workers enter the labor market.
Russ Roberts: And that's consistent with other work I think that's been done out of--is it Texas A&M? and Jonathan Meer's? Is that what they found, also?
Jacob Vigdor: Yeah. So, they're--it's kind of interesting that the minimum wage literature is very bifurcated. And there are studies that will show you that there's not very much impact whatsoever. But then there are other studies that show that there are these negative impacts, and yeah, that what we are pulling up here is consistent with a lot of prior studies, actually, sort of, that have shown, in these prior studies examined the impact of minimum wage increases on teenage work. Teenage earnings, teenage employment. And there have been more consistent negative estimates there. So, our work is consistent with that.
Russ Roberts: And that's your point about lower skilled workers. And I assume--well, let me ask you this: Do you also have data on multiple jobs? When you say people are working part time, are some of them working multiple part-time jobs? Are you able to tell that?
Jacob Vigdor: Yeah. Our data have names and Social Security numbers for everyone. We can track how much they are working across all the jobs they hold. So long as those jobs are kind of above the table, and they also need to be jobs that are traditional wage-earning: you get a W-2 from your employer, type of jobs. The gig economy, if you are a contractor and not an employee--if you get a 1099 and not a W-2, that employment is not in our data.
Russ Roberts: And you said above the table. So, somebody who is getting cash to work on a landscaping project on a Tuesday morning is not going to show up in your data.
Jacob Vigdor: Yeah. That's correct. Of course, another reason that someone might not show up: If you own a business and you are paying people, you know, the work that you put in, in terms of owning a business, is not reported, either.
Russ Roberts: But you are probably not a low-wage worker in that situation. Or is that not the case?
Jacob Vigdor: You might have some people who are in the low-wage labor market who transition. So, for example, if you start out doing landscaping work and then you decide that you want to start your own landscaping business, then you might transition from doing wage work to being a business owner.
Russ Roberts: And not make so much money for a while, until you get your business started, and--yeah.
Jacob Vigdor: Yeah. Well, you could make a little money or a lot of money. But once you become a business owner then you stop reporting your earnings to the UI[?] system, and therefore you are invisible to us.
Russ Roberts: So, when you say that hours were down 6 or 7%, can you give us an idea what that means exactly? Because as I said, as I hinted at--I'm sure it's true; that's the average. It's not every worker. So, what does that number mean? How should I interpret that?
Jacob Vigdor: Yeah. So, if you want to think about a low-wage worker, think about someone who is working maybe 20 hours a week. So, our numbers are suggesting that for the typical person who is working 20 hours a week to start, then, when the minimum wage goes up, in particular to $13 an hour as of the beginning of 2016, they're maybe getting on average between 18 and 19 hours a week. So, it's a relatively small impact. And we do show evidence that there's a lot of heterogeneity in that effect: That more experienced workers, they may be seeing some reduction in hours, but it's going to be a smaller amount. It's basically less than 5%. So, if you are starting out at 20 hours a week, you are still getting at least 19, and, you know, maybe averaging a little bit more than 19 hours a week. It is the less experienced workers--the people who had a less of a wage history at the time minimum wage went up--that they are seeing hours reductions that exceed the average. They might be on the order of 10%.
Russ Roberts: And what's the net impact on earnings? Which is hours times wage.
Jacob Vigdor: So, when we look at the low-wage labor market overall, what we're picking up is the amount of money paid out in the low-wage labor market declined. But that's, of course, lumping together. When we looked more specifically at the trajectories of individual workers with differing levels of experience, we found that the more experienced workers were coming out ahead. On average they were taking home--not necessarily taking home, but their paychecks were reflecting an extra about $20 a week. The less experienced workers who at least had a job to start with, were more or less breaking even: their increase in hourly wages was being pretty much offset by a reduction in hours. And then, the big losses in terms of much lower pay would be amongst the workers who hadn't even entered the labor market yet when the minimum wage started to increase, because they were finding it harder to find any work at all.
Russ Roberts: Help me think about those people. So, I'm a--maybe I've been out of the labor force, or maybe I've just turned, I've just left school. And I'm looking for work; I don't have a lot of skills, and I'm trying to get a job as a dishwasher or a gardener--helping somebody with gardening or contracting. And, you're saying I'm not going to find the work to start with? Is that the issue?
Jacob Vigdor: Yeah. So, there are lots of job openings in Seattle. But the general pattern is that employers are emphasizing experience. And, you know, as economists we tend to talk about skills. But the Number 1 skill that you can have as a dishwasher is experience being a dishwasher. So, typically, if you are a dishwasher in a restaurant kitchen, you are going to be using equipment; and if you have some familiarity with that equipment already, you don't need to be trained. You can be productive starting from Day 1 on the job. And that's really what employers are emphasizing. They are really interested in having people who have familiarity with the job already, who do not need to be trained on the job. Because on-the-job training is an investment, and at $15 an hour that investment doesn't make sense from the business owner's perspective.
Russ Roberts: Well, that's your supposition. And you said you augmented some of the statistical work with some interviews and, we might call them [?] sociological case study approach. Is that what people say? Or, is that your guess? And even if they say it they may not really--it just could be a story that they're telling, right?
Jacob Vigdor: It could be a story that they're telling. But, the story is consistent with the data that we're picking out. But basically, one of the things that we see in the low-wage labor market, the turnover rates are very high. And particularly amongst teenage workers, many of whom are just looking for a little bit of summer employment or just a few hours. From the employer's perspective, that's a riskier proposition. To hire someone who is young, doesn't necessarily have a work history, they are not necessarily going to be reliable. You know--that proposition is highly sensitive to the wage. That's at least the story that we hear from business owners; and it's consistent with the data that we're picking up. Which is: If you only have to pay a teenager $9 an hour, or $10 an hour, it's a very different proposition than paying them $15. Because of the risk involved. And, the labor force attachment is kind of weak. I had one business owner tell that with teenage workers, sometimes they'll come in and they'll say, 'Oh, can I have Friday night off?' And the manager says, 'Well, no. I need you to work Friday night.' And then the response is, 'Well, then I quit.' So, if you are dealing with a segment of the workforce for whom the labor force attachment is that weak, then any kind of investment you make in trying to train that worker may not yield much payoff if they are going to be quitting after a couple of weeks.
Russ Roberts: And the argument is that at $9 an hour, it's cheaper if I lose a couple of hours of training versus $15, especially if I don't get anything after that.
Russ Roberts: So, you are confirming all my priors. Which I should just get out on the table for listeners, although some of you, I'm sure, have an idea what they are. I have often argued that the minimum wage is a tax on low-skilled workers to help other low-skilled workers. Exactly in the way you've laid it out: That, the workers with the least experience and the least skills, who most need a chance to get on the labor market ladder, are--those rungs were cut off for them. There is a benefit to people who keep their jobs or who keep their hours, or who have small reductions in hours. And that--it's kind of a brutal way to improve the skill--excuse me--the income of low-skilled people. Now, on the other side, you could argue--I wouldn't, but you might--that, 'Well, this is exactly what we want. Because we don't want to help teenagers. The whole idea of this was to help poor families and people who have low skills, who are trying to support their children. And, as a result, it's not so bad that teenagers can't get work. They're going to get back into school: 'So, they don't make any summer money. What's the big deal?'
Jacob Vigdor: Yes. You could argue that what we have in the low-wage labor market is a combination of some workers who have a family to feed and are really looking for money that they need to survive, competing against workers who--you know, really they're trying to gain experience. That's really important for them. If they make some money, 'Hey, that's great.' But they don't need the money to survive. And so the consequence--you know, if you think about it: If you are trying to get a job in part so you have some experience to put on your resume so that it's easier for you to get more work down the road--um, you know, your reservation wage--if you have all of your sort of life covered by your parents or by your family, your reservation wage is going to be really low. And so--
Russ Roberts: Reservation wage being the amount you are willing to accept to just take that first hour.
Jacob Vigdor: Yeah. That's right. That's right. I mean, so you have a lot of the young workers, college-age workers, who do internships. And unpaid internships. And part of the message there is they really value getting the experience; and the money, you know, they can do without. So, if are one of these workers that is trying to make ends meet on the basis of this type of work--you actually are an adult, you are on your own, no one else can support you--um, and you are going into this competition against people who are really willing to work for just about any wage--it's going to make things more difficult for you. So, you could look at the minimum wage as a way of--so, if we think of the minimum wage as a policy that basically prices the least experienced workers out of the market, if we think that those least-experienced workers are going to be disproportionately people who are going to be fine in the long run--because, as you said, they are going to stay in school; they are going to invest in their own human capital, and they are going to end up getting jobs that are much higher-paying in the long run. Maybe that's okay. I think that if you look around the city of Seattle--Seattle is the city with the second-lowest percentage of teenagers in the population in the United States. Only San Francisco has a lower percentage of teenagers in the population. So, what this means, is Seattle is a little bit different from the rest of the country. If we think that the minimum wage is a policy that prices teenagers out of the market, the interesting little factoid about Seattle is that we don't have a lot of them to start with.
Russ Roberts: That's fascinating. And I guess the other concern I would have is, this is pretty short run. It hasn't been in place very long. There hasn't been that much time for employers to respond. And of course, the minimum wage is actually going up to $16 dollars--is that January 1st this year?
Jacob Vigdor: Yeah. For the large employers in Seattle, the current minimum wage is $16 an hour.
Russ Roberts: And so, it will be interesting to revisit some of these numbers in 2-3 years, to the extent that we can. We have to be able to match and be confident we are matching these workers well with others that we can compare them to. But the--other question, which would give you information, would be the effects on restaurant formation. So, a lot of--there's a lot anecdotal evidence, which, I, even though it confirms my biases I try to ignore that some restaurant closes and says, 'Yeap. We just couldn't pay the minimum wage.' Those stories run now and then, and again, on Twitter. People who think like me wave those around. And the people who don't think like me ignore them. Or explain that it's just one restaurant. So, what do we know systematically, if anything, about those kind of issues in Seattle? Which would be important?
Jacob Vigdor: Yeah. One of the most interesting conversations that I had regarding the restaurant industry: I got a call, shortly after I signed on to this study, from a person who was the CEO [Chief Executive Officer] of the Washington Restaurant Association. He wanted to meet, and he wanted to just kind of bend my ear and get my perspective about--what perspective we were taking. Basically, he just wanted to see, 'Okay, are you guys just sort of hook, line, and sinker going to write a study where the conclusions are foregone conclusions? Or are you really doing this on the level?' And so the interesting part of that conversation was, I sat down with this CEO of the Washington Restaurant Association, and the first thing he said to me is, 'We're going to be fine. Our members of this Association--the minimum wage, it's not going to break them.' And the reason why, he said, is because, 'there are so many strategies that we have to basically reduce our labor.'
Russ Roberts: Right.
Jacob Vigdor: And he proceeded to tick off about 8 different specific business strategies that a restaurant owner can use to cut back on their use of low-wage labor. And these include things like, instead of hiring a prep cook to chop vegetables for you, you just order chopped vegetables. And, so, that eliminates that job from your kitchen.
Russ Roberts: And, you order them from an area that doesn't have a $13 or $15 minimum wage, so they are cheaper than they would be if you did it yourself.
Jacob Vigdor: Yes. Those vegetables might be chopped in Mexico, for example. Another thing that he mentioned, and this is something that we see quite a bit in Seattle--so, there's a move away from table-service restaurants to order-at-the-counter type restaurants. So, if you are imagining a restaurant where you go, you sit down, someone comes to take your order--that person is on the clock; you have someone delivering your food--that person is on the clock; you have someone bussing your table--their hours are on the clock. The new style of restaurant in Seattle, and actually one of the closest restaurants to my house did this transition over the past couple of years. You go; you order at the counter. They call out your number or your name when the food is ready, so you are the one transporting the food to your table. You are the one bussing your table when you are done. And so basically what had been tasks accomplished by low-wage workers on the clock are now being accomplished by an unpaid person--which is the customer. So, little tweaks like this--that's how restaurants have coped. And it's absolutely true that the restaurant industry in Seattle is, by and large, doing fine. There are some closings; but there are quite a number of openings as well. Our efforts to try to understand whether the minimum wage has impacted the business opening and closing rate have generally found that we don't find any effect. But what we do find is that the patterns of openings and closings are steering the city towards less labor intensive restaurants. So, we're moving away from full service. We're moving towards order-at-the counter and other forms of serving people food that involves fewer labor hours.
Russ Roberts: So, does that show up directly? I mean, obviously--I don't know how hard that is to measure. I don't know if you have that in the data. So, even if you had the same number of restaurants, you could have a reduction in those labor hours, you are suggesting. Can you measure it?
Jacob Vigdor: Yeah. We actually do have the capacity to estimate effects on the number of hours worked in the restaurant industry. So, the biggest reduction in hours worked in the restaurant industry we found is amongst the lowest-paid workers. Which is [?] a restaurant that has done away with the bus boy, because now the customers are bussing dishes. You are getting rid of the lowest-paid worker in the establishment. And then the people who are earning higher wages--the people who are actually cooking the food and who do the customer service tasks like taking the orders at the counter--they are going to tend to be higher-paid because you need a slightly higher skill level to do something like, you know, cook food to order. You know, I think that the statistics that I've seen suggest that a lot of people with even a little bit of experience, it's quite easy to find work well above the minimum wage, as high as it is, in Seattle. It's just that there are these other categories of jobs that have fallen.
Russ Roberts: So that busboy is in your data set, and was earning $9 or $10 or whatever it was before this legislation. And now they are just--what? What do you see? Do you see them doing something else with fewer hours? Do you see them working only one part-time job instead of two? What are you seeing there?
Jacob Vigdor: Yeah. So, if you were a busboy in 2014--let's say, in February of 2014, so that minimum wage doesn't start going up until April--then you had at least some experience when the minimum wage starts going up. And so, what we're finding is that those people who are in entry level jobs at the time that minimum wage went up, they are more or less breaking even. So, there might be a tendency for them to witness a reduction in hours; but the fact that you were at least on the bottom rung at the time that the bottom rung was eliminated, you've got some possibility that you managed to stay with the organization and move up to the next rung. And what the data are suggesting to us is that the people who are really coming out behind--it's not the busboys of 2014, because the industry is such that if you were a busboy in 2014, it's very common for you to have moved up in the world by 2015, or by 2016. The people who are falling behind are the people who would have been hired as the busboy in 2016 or 2017. By 2017, someone who was hoping to be hired for a busboy-type position is going to discover that many restaurants have eliminated the job category altogether. And the people who had the job of busboy as of 2014, they by and large are at least breaking even, because it's fairly natural in this industry to move up from the position of busboy: once you learn the business a little bit better, once you learn a little bit more, to join the wait staff or to join the kitchen staff.
Russ Roberts: And you've got some familiarity with the boss, and they might know that you're reliable and all those things that are intangible and can't be measured.
Jacob Vigdor: So the natural moves up the ladder are the saving grace. So, the idea that busboy is not typically a terminal position is what helps out the people who had these lowest-paid jobs as of 2014 or early 2015.
Russ Roberts: Just to make it clear: You're not saying that every single restaurant went from table service to take-out at the counter. Obviously, some restaurants stayed with table service and kept busboys. And there are people who are, presumably, in 2019 able to be a bus boy, I would think, at $16 an hour--
Jacob Vigdor: Yes--
Russ Roberts: because that restaurant wants to keep a certain style: they've raised their prices, some. And some of them presumably will go out of business, because the customers aren't willing to pay that extra amount and competition with other restaurants. But, there are some out there, I assume.
Jacob Vigdor: Yes. Actually, there's a really interesting study. It's not ours. But it's a study that looked at restaurants on the basis of their Yelp reviews. And discovered that restaurants with low Yelp reviews--meaning that they weren't very popular, or people didn't like them very much--they were more likely to go out of business after the minimum wage increase. I think this was a study based in New York City; and it wasn't Seattle at all. But I think that that's sort of--if you are a popular restaurant, I tell ya', Seattle is an affluent town. There is a lot of money here. Amazon's been hiring tens of thousands of programmers, and they pay them really high salaries. So, there's a lot of people that have the income to be able to afford going out to eat at a full service restaurant. So, absolutely, we still have a pretty large number of full-service restaurants in the city. But, you're right: This is kind of a premium product. It's premium service. And you are going to pay a premium price for it. I actually--I went online a few months ago to just sort of look through the archives to see what online menu prices had been like just a few years ago. And, absolutely, you can see that menu prices have been going up at some of the--the more well-established restaurants that offer full service.
Russ Roberts: Which is what you'd expect.
Russ Roberts: Now, a number of people have studied Seattle and found no impact. I know there was a Berkeley study in 2016, or at least a working paper.
Jacob Vigdor: Yes.
Russ Roberts: And, I'd be curious what you think of that work. But, more generally, there are many, many studies done in the last 10 years that find very little impact on employment. Now, they generally don't have data on hours--as you have. But they find, I would say, among a certain group of economists--wouldn't include me, but a certain group of economists, there is a view that within the current range of minimum wage increases we've observed it's something close to a free lunch. That, the demand for labor--or that, or those wage levels, is--a technical term, relatively inelastic. That is, it doesn't respond much, the employment level. And so we can use this tool without much consequence. You've come and found something I would say is different. Much more disturbing. And, I'm curious, one, what you think of those other studies, and what they think of yours.
Jacob Vigdor: Well, the Berkeley study in particular, it follows a common pattern in the literature, which is to do a study of the restaurant industry. And not necessarily the low-wage labor market. You know, we've been talking a lot about restaurants today. But, most of the low-wage work in the economy is not in the restaurant industry. So--
Russ Roberts: Which is confusing. Because, a lot of employment in the restaurant business is low wage. But that does not follow, then, that most low-wage workers are in the restaurant business.
Jacob Vigdor: Yes--
Russ Roberts: It's a subtle distinction, but it's important.
Jacob Vigdor: There are just as many low-wage workers in the health care industry as there are in the restaurant industry. The difference is that--you're right. It's a higher proportion of restaurant workers are low-wage workers. Because in the health care industry you also have doctors and nurses and people who--you've also got custodial staff, cafeteria staff. You've got all sorts of employees in the health care sector that are low paid. Anyway, I think that the Berkeley study of the restaurant industry--it's reliable as a study of the restaurant industry, because they are finding the same result that we found when we did our analysis of restaurants in Seattle. Namely that, overall restaurant employment shows no negative impact. There are just as many jobs in Seattle restaurants as we would have expected without the minimum wage increase. Now, there's an asterisk there, which is, we're talking about all jobs in the restaurant industry. Not only low-wage jobs. So, the Berkeley study used a data set that didn't give them the capacity to study low-wage workers specifically. Our data set allows us to do that. And, what we found is that if you look at low-wage employment in the restaurant industry, rather than overall employment, and if you look specifically at hours instead of number of jobs, you do find these negative impacts. And so, I think that one of the things we're picking up from our data analysis is that there are quite a few people in the low-wage labor market in Seattle who have kept their jobs. And so, if you are just counting up the number of jobs, it might look like it hasn't changed very much. But the difference is that they are seeing reductions in their hours. So, a reduction in hours is something that [?] Berkeley's study can't.
Russ Roberts: But, how do you reconcile that employment staying constant with the story you told earlier that busboys can't--there won't be any busboys: restaurants have found those ways to reduce employment; they don't have somebody chopping? Shouldn't there be fewer jobs in the restaurant industry, if those effects are important?
Jacob Vigdor: Yeah: the difference is that it's a growing industry. The number of jobs overall has stayed relatively steady because the number of restaurants overall is expanding, because of the general expansion of Seattle relative to other areas. So, basically, here's the way you want to think about it. Imagine a world where there's 10 restaurants and all of them employ some relatively highly paid kitchen staffs and relatively highly paid wait staff, and a busboy. The minimum wage increases and some of those restaurants sort of get rid of the busboy and have the customers bus the tables. But at the same time, the number of restaurants expands. And so, what you could observe in that situation is that overall employment in the industry stays relatively constant, but the amount of low-wage employment, if the busboy was the only low-wage employee to begin with, declines.
Russ Roberts: Yeah. And the proportion of low-wage workers of the total has gone down, would be the other way to look at it.
Jacob Vigdor: That's right. Yeah. So, if you ask the question, 'Are there still jobs in the restaurant industry in Seattle?' the answer is, 'Yes.' But you want to be asking a more nuanced question, which is, 'Are there job opportunities available for people in Seattle with no prior experience?' And there, all the data is pointing us to the answer that that is what has dried up here: the number of opportunities available where you can be hired with no prior experience.
Russ Roberts: So, that's disturbing, to me. The media reaction to at least the first reports that you've done was, I would call it kind of pyrotechnic. A lot of fireworks. There were people who said, 'Hah! This proves what I knew all along.' On all sides of the issue people found different parts of the elephant to hold onto to confirm their priors. Some people said, you know, 'It's really okay, because, as you said, employment hasn't changed that much overall. We're not going to worry so much about the hours or whatever it is.' But it did make quite a splash. What's that been like?
Jacob Vigdor: Well, it was a little crazy. Particularly back in the summer of 2016, when we released our first major report on impacts, it was interesting, because there were a few studies that came out right away that were carefully written, where the reporters had taken the time to talk to the authors about the study. They had gone and they had talked to kind of third-party independent economists who had read the study and could comment on it. But then after that first round of careful studies, you got more media coverage where it was clear that the reporters and the columnists had not talked to anyone. They were writing things on the basis not of our paper, but on the media coverage of our paper. So it was like a game of telephone. So, and just like in the game of telephone you discover that the message that gets back to you is somewhat garbled compared to what you started with, I would see people writing about the study that were getting some of the fundamental details wrong. I saw one newspaper column that referred to our study as a Washington State University study--which, that is our rival over on the other side of Washington State--
Russ Roberts: shame on them--
Jacob Vigdor: It is not the U. of Washington. People were getting the details of our analysis exactly wrong. Our data limitations exactly wrong. And, I think that one of the things that we learned in working on this study and a sort of humbling life lesson, is that ultimately you can't control what other people say.
Russ Roberts: So, let's think about some meta-questions for a minute. And maybe you can provide some data on these questions. When we talked at the very beginning about what you did to try to control for what was going on in Seattle that was not related to the minimum wage, now that we've had the conversation for a while--people didn't hear and understand that, if the area is growing--and the number of restaurants would be growing a lot even if there hadn't been a minimum wage--there might be growing still after minimum wage but not as much. And the type of jobs available in those restaurants that are low-skilled would be maybe not growing nearly as much. And you are trying to tease that out by comparing the growth in Seattle, where the minimum wage was passed, to areas where there isn't this increase in the minimum wage but they are "like Seattle." And someone could argue--and I often do--that's just a wild attempt. That's the best you can do. I understand, you've got to do the best you can, one could argue. But the question then becomes: How reliable are these conclusions, even when they, in this case, confirm the way I'd like to look at the world? And I--I wonder--when I hear results like these, one of the things I always want to know is: How many regressions did you run? How many times did you tweak and fiddle with the variables that try to control for things? How robust, how sensitive were the results to these kind of changes? And, how many of the findings ended up on the floor that I don't get to see because they didn't make it into the final report, but you somehow convinced yourself that they were not reliable or not representative. Give us some of the flavor of that--the kitchen.
Jacob Vigdor: Yeah. So, there were a number of things that we did--because, of course, the first people that we have to convince are ourselves. So, I talked a little bit about the synthetic control methodology that we used, which is basically picking out other regions of Washington State that have a good track record of matching employment trends in Seattle. We used entirely different methodology called Interactive Fixed Effects, which showed largely the same results. I mentioned a little while ago that we had to sort of pick a threshold and say: A low-wage job is defined as under $19 an hour. We tried different versions of the analysis, trying anything from $15 to $25. The results were consistently pretty similar. The way that we tried to do things was to keep our hands above the table, and what ends up being in the paper are a lot of appendix tables and figures and footnotes that describe some of these alternate specifications that we tried. We tried looking at the restaurant industry specifically: there is more information in the paper about industries that had these funky patterns that we had to exclude from the analysis because we think there was something else going on. But, with this paper, it's really kind of a 'what you see is what you get' endeavor. We were trying to commit ourselves to writing down a specification before we conducted the analysis. Which is considered to be sort of a best practice in a lot of empirical research these days--is to not go fishing for a particular set of results, but sort of say, 'Okay: this is the regression, and I am not going to go and run a hundred different regressions and just pick the one that shows the results that I prefer.'
Russ Roberts: What's going to happen going forward? Are you continuing to analyze--because, I suggested that--you know, and a short period of time, to me, we can always debate what a short period of time is, but is it 6 weeks? is it 6 months? is it 6 years? But, as people have more time to respond to this situation in the restaurant industry and elsewhere--health care and so on--I'd expect the effects that are measured to be larger. Especially when the minimum wage is continuing to increase. Of course, if Seattle is growing faster than that increase because of other underlying factors, it would still be hard to measure and tease out. But I would expect some of these effects to get larger. Are you going to continue to measure these changes going forward?
Jacob Vigdor: You know, the long-run effects are really important, but the challenge with the long-run effects is that the methods that we use for this kind of analysis are really inadequate to try to pick them out. A large part of our confidence in the results really stems from timing. We see things change immediately when the minimum wage goes up. And that's what increases our confidence that what we are observing is the impact of the minimum wage. If something happens in Seattle 5 years from now--suppose that 5 years from now all of a sudden there's a collapse in low-wage employment--is that the delayed impact of the minimum wage? Or is that the result of something else? It just becomes much harder to tease that out further down the line.
Russ Roberts: Yeah. If Amazon were to not just open a second headquarters but close all of its headquarters, that would have an impact that would make it very hard to measure.
Jacob Vigdor: Yes. Yes. So, there's an argument that actually goes the other way. So, it's possible that in the long run, businesses will find more ways to adapt. But here's the counterargument. The counterargument is: Once the businesses in Seattle figure out ways to economize on labor, those methods will be transmitted to other locations. So, once a restaurant in Seattle figures out, 'Hey, I can save on labor costs by just having my customers bus their own dishes,' then, if that business practice filters out into other parts of the country, then our control groups start to see some reduction in employment as well. And this is actually the story--I mentioned a little while ago the conversation I had with the CEO of the Washington Restaurant Association. And this was his story. His story was, 'Minimum wage increases force businesses to think up new ways to economize on labor. Once they've thought them up, those business practices filter out into other parts of the country.' And so, it could very well be that in the long run, the treatment-versus-control difference in something like low-wage employment could actually dissipate. But the reason for the dissipation is that in the long run, we are going to see across the board the number of opportunity for low-wage workers declines, because all the businesses have adopted these changed practices to cut down on their use of low-wage labor.
Russ Roberts: Yeah. The only problem with that theory--I mean, I'm sympathetic to that point, obviously. But in general, people don't like to bus their own tables. And so, restaurants that try to "adopt this practice" as a way to save on their labor bill, even in areas where there isn't a minimum wage, are going to find they are going to struggle to attract clientele and customers because other restaurants won't do it. And so, for me, it reminds me of this--one area of change we haven't talked about is technology and automation, which is another way you can cope with an increase in labor cost. And I have always argued that--you know, the minimum wage speeds up the substitution of capital for labor. Speeds up the substitution of robots and artificial intelligence and other forms of economizing at the expense of labor. And that's--there's no reason we should want to do that. It's not necessarily--some people say, 'Well, that's a plus, because we can figure out all these new techniques.' Yeah, it's a plus, except it hurts people with low skills. It's a horrible--there's no--we don't want to encourage that. It's not a good thing. So, those are not free to the employer, those savings. They come at a cost, sometimes. Obviously, if I'm indifferent between bussing my own table and not; if I'm indifferent between being greeted by a person versus opening a mechanical door where some food is waiting for me, then those things will happen. I am indifferent between somebody automatically deep-frying--a robot deep-frying French fries and a human being, because I don't even usually see it. And it doesn't affect--it may even improve the quality of my French fries, which means those things will happen and spread out through the country. But, so I think that that kind of argument, that the long-run effects are going to be hard to measure: I take the general point. But I think in specifics it may not always be the case.
Jacob Vigdor: Yeah. Yeah. It's hard to know, for sure. One story that I like to tell the people, particularly when I'm talking to general audiences, about minimum wage: I ask people about the Starbucks App. And I will ask an audience, 'So, who has the Starbucks App on their phone?' And, especially in Seattle, there's always somebody in the crowd. And then I ask them, 'Well, what's your experience with the Starbucks App? How do you like it?'
Russ Roberts: Explain what it does, Jake, for people who don't have it.
Jacob Vigdor: The Starbucks App is, basically, you take your phone, and you have an app, and you can order whatever kind of fancy coffee beverage you want. And it will be available for pickup at a Starbucks of your choice. So, suppose that you are going to a meeting, and you want to grab a coffee on the way to the meeting, but you don't want to wait in line and you really don't want for your cappuccino to be brewed. What you do is you open up the app, you place your order, you pay for it right there in the app. And then there's just a cup there that has your drink inside. So, it saves you time. And it's--and I think that for most users, at least in my highly unscientific polling, people are really happy with the app. It provides the customer with a really good experience. And it is a big labor-saver. Because if there's an order placed with the app, the person at the cash register who would ordinarily have to take the order, have to take payment, have to return change or provide a receipt or something like that, that is all automated. And the best part of it, from Starbucks's perspective, is it's not as though there's a touch screen that Starbucks has to maintain. The touchscreen belongs to the consumer. And so, if we have innovations like this that, a). save labor, and b). are really popular amongst customers, then, you know, I think that you can envision a future where there is more general adoption of these kinds of technologies.
Russ Roberts: And the point being?
Jacob Vigdor: The point being that, if necessity is the mother of invention, then a higher minimum wage represents the necessity. But once you've got the invention, the invention can filter out to all parts of the country. Even the places with a low minimum wage. Because, if I've got the Starbucks App in Seattle, and it's helped the Starbucks here in Seattle cut back on their labor headcount, then when I go out to Idaho, where they've still got the Federal minimum of $7.25 an hour, I can use that same Starbucks App. And it's going to help the Starbucks in Idaho enjoy the same kind of labor savings.
Russ Roberts: And in theory, in theory, Pete's--different coffee chain--could also adopt an app--I'm sure they have an app--which would in theory mean that in Idaho, that labor savings through competition might have to have to passed on to the consumer, in the form of a lower price of a cup of coffee. And eventually will be also in Seattle. So, it's a--the brunt is--it doesn't mean these innovations are going to mean more profit necessarily for the corporation. But it will mean less employment for people of low skill. And I raised a similar question--I thought you were going to raise a moral question: whether it's moral to use the app. I always wonder if people who re-use their towels in the hotel or ask for no room service, housekeeping service, because they want to help the environment and not turn the towels and the sheets over, realize that they are reducing the employment for low-skilled, relatively low-skilled people who do that. Who make up the beds and pick up the towels and take them to the laundry. And, normally, we'd say, 'Well, that's how the standard of living improves,' in the case of the App, because it's going to make it cheaper to provide a cup of coffee. Which means we have more money to spend on other things. Which means that those other things can now expand. And those other things expanding will lead to some employment opportunity for people with low skills. And so, the long-run effect of this creative destruction is pretty good. And in the case of the towels, we get a cleaner environment. It's true there's lower-skilled workers, but at least they have, there's going to be something else going on out because the hotels now will lower their wages, their cost of the room, etc., etc., etc. The challenge, though, is that we are just making it harder for some people to get on that ladder. And I don't think that is a good thing.
Jacob Vigdor: Well, I think that the message--
Russ Roberts: I should say that the message--we are carefully--we are speeding up the rate at which people with low skills have to have a hard time. And it makes it harder for, say, the educational process to compensate for that. It's not doing a good job, anyway.
Jacob Vigdor: Yeah. So, I think--I've talked to a lot of people from across the spectrum. And I'd say that there are not a whole lot of people who express optimism about the future of low-wage employment. I had a political operative from the Seattle Mayor's office come visit me a couple of years ago. This particular staffer from the Mayor was wondering if I would be willing to sort of go out in public and advocate for a higher minimum wage. And, I responded by saying, 'Look, that's not my job. I'm not an advocate. I'm a researcher.' And I mentioned to him that our research was actually showing that there were some potentially adverse impacts of Seattle's minimum wage. And he responded to me by saying, 'Well, in the long run, aren't these jobs going to go away anyway?' And so, this is coming from someone whose job description is to be a minimum wage advocate. And so, I think that there's this sense that these jobs are not long for this world. And the real goal is to ensure that while they're still here, they pay enough that someone can make a living off of them. And, if the policy interventions that we use, sort of hasten the demise of the low-wage job, then that maybe doesn't bother people so much? Because I think you just have to say to yourself, 'Well, the technological innovation, we'll find a way to get these people into new lines of work.' Um, and we'll find--maybe we'll figure out how to better train people for a wider variety of high school jobs. And, you know, just sort of look through economic history and we find lots of examples of technologies that come along, and just completely eliminate low-skilled job categories. You know, mechanization of agriculture. That was a big deal. And it displaced a lot of people. But, these days, you don't find a lot of advocates out there saying, 'We need to send everybody back to rural areas to go engage in manual labor on farms.'
Russ Roberts: Yeah. As our standard of living climbs, there are not a lot of blacksmith jobs in America any more. That was a job that low-skilled people used to be able to do, or just a strong person who didn't have certain abilities elsewhere. And we've talked--
Jacob Vigdor: There's another interesting thing about the America low-skilled workforce, which is that it has a large number of immigrants.
Russ Roberts: Yeah.
Jacob Vigdor: Most of--I mentioned a little while ago that we've been following 50 families, trying to make ends meet with low-wage work in Seattle. Most of those families are headed by someone born in a foreign country. The long-run trends around immigration are such that the number of immigrants to the United States has kind of stabilized. And it's even maybe declining a little bit, when we look at low-skilled immigration. So, one of the factors that may help us transition to a world where there is less low-skilled work is the fact that, particularly in a country like the United States, our low-wage workforce, which for many years was fueled by immigration--as immigration starts to wane, the supply will be reduced around the same time the demand is reduced. So, maybe this becomes a natural transition.
Russ Roberts: Obviously people differ in how they view immigration. It's a hot topic. A lot of heat right now.
Jacob Vigdor: Yes.
Russ Roberts: But, from my perspective, while it's true that low-wage workers coming from elsewhere don't do great when they get here, they often do great eventually. And people risk their lives to be able to be poor in America. Or they have, historically. And if we close that off, either through our immigration policy or our wage policy, I think that would be a tragedy.
Jacob Vigdor: Yes. Well, the good news about immigration is that even as immigration has declined, I mean, one of the, the demography is a huge factor in determining immigration flows. And America's era of rapid immigration from Mexico coincided with demographic patterns in Mexico that involved high birth rates, low mortality rates, and big population growth. The birth rates in Mexico have declined to the point where their population is stable. And so I don't think that we are ever going to see the kind of immigration from Mexico that we had in this country prior to about 2007.