This morning I gave oral testimony to the Vermont Senate Economic Committee on their proposal to raise the state minimum wage to $15 an hour by 2024. As part of my written evidence, I explored in detail the rationale for minimum wage hikes from the “Fight for $15” campaigners and other think-tanks. Below is a slightly edited version of that section of my testimony, which has wider applicability.
Theoretically, a minimum wage hike can improve the functioning of a labor market when employers are “monopsonistic.” When firms have significant labor market power over employees, raising a wage floor can increase both pay for workers and employment levels. Some economists have argued that most firms do have a slight degree of monopsony power over their employees. This suggests a modest minimum wage can, in some cases, actually improve economic efficiency, ending so-called “exploitative” low wage rates without reducing employment.
But those advocating minimum wage hikes across America today do not use this economic line of argument. Bureau of Labor Statistics data shows 84 percent of employees paid at or below the federal minimum wage work for businesses in retail, leisure and hospitality, and education and health services. These industries do not tend to be characterized by powerful companies which dominate local labor markets. The theoretical economic case for minimum wage hikes to solve “market failures” is therefore weak.
Instead, proponents of higher minimum wages assert that legislation should deliver some target level for minimum hourly wages to fulfill other objectives. Their rationale is really about giving minimum wage earners more money. And they use a host of metrics to show that, currently, minimum wage earners are just not being paid enough.
But what should determine where the minimum wage level is set? Ideas seemingly differ across $15 wage proponents. Some implicitly argue that minimum wages should be set to cover certain essential living costs. Democratic presidential primary candidate Elizabeth Warren, for example, suggested this when she claimed:
When I was a kid, a minimum wage job in America would support a family of three. It would pay a mortgage, keep the utilities on and put food on the table. Today, a minimum-wage job in America will not keep a mama and a baby out of poverty.
Others prefer different metrics. In testimony to this committee, David Cooper of the Economic Policy Institute documented extensively how current minimum wage rates had not “kept up” with average wages or economy-wide productivity levels, and did not provide incomes for full-time workers to cross certain poverty thresholds.
None of these claims are disputable as facts. But I want to suggest that those metrics are not appropriate for judging what the Vermont minimum wage should be. Devoid of broader context, they are misleading and might lead to damaging policy conclusions.
Cost of living: It’s important to remember that employers pay employees for the perceived value of the work undertaken, not to compensate workers for their rent, food, energy, transport, clothes or child-care bills (which will differ hugely by family and are beyond employers’ control).
High living costs are a very under-discussed cause of economic hardship. My own research has found typical poor American households face higher prices on essential goods and services due to misguided interventions and regulations that cost them between $800 and $3,500 in total per year.
Rather than tackle the causes of these high prices, minimum wage hike campaigners want businesses to bear the cost of compensating workers for their living expenses. This is not economically sensible. Putting the full burden of the cost of living on shareholders and customers of low wage employers, divorcing pay rates from the work employees undertake, market conditions, and firms’ ability to pay, could risk a significant diminution in low wage job opportunities.
Productivity: Economy-wide labor productivity has risen faster than minimum wage rates over the last fifty years. Given productivity changes should affect worker total compensation levels, $15 wage campaigners imply that minimum wage workers are being paid below what their productivity commands.
But comparing productivity gains of all workers to speculate what hourly wage rates should be for minimum wage workers alone is misguided. After all, different industries experience different productivity growth rates over time.
Sadly, a productivity growth series solely for minimum wage workers is not available. But long-term data for the food services industry might be a reasonable proxy. Bureau of Labor Statistics data show that from 1987 to 2017, labor productivity in the food service sector rose by an average of just 0.4 percent per year (with unit labor costs increasing by 3.2 percent per year). If the minimum wage had been pegged to this productivity measure, it would have increased by 13 percent in real terms – from $7.16 in 1987 (2017 dollars) to $8.06 in 2017.
The actual 2017 federal minimum was, of course, $7.25 in 2017 and the Vermont minimum wage was $10. Using this productivity series and start date then, the Vermont minimum wage is now higher than justified by food service productivity improvements since 1987.
This does not prove that Vermont minimum wage rate increases have exceeded the productivity growth of all minimum wage employees, nor does it tell us what the minimum wage level should be according to this measure (given the necessary arbitrary start date). But it does show the danger of making spurious comparisons between economy-wide productivity and minimum wage rates. Pegging minimum wage rates to aggregate productivity trends might lead us to deliver much higher wage floors than justified by the productivity of workers in certain sectors, causing significant job losses or other adjustments.
Poverty: A stated ambition of $15 minimum wage advocates is to lift households above poverty thresholds. Minimum wage hikes can achieve this by raising incomes for minimum wage earners. Yet poverty is measured at the household level. The very reason why minimum wages were not used as a primary tool to reduce poverty for households in recent decades is that they were not considered particularly well targeted or effective.
First, people who earn around the minimum wage are often not from poor households. In 2014, the economist Joseph Sabia estimated just 13 percent of workers on hourly rates between $7.25 and $10.10 lived in households below the poverty line. Many people earning around the current minimum wage rate are second earners (particularly part-time) or young people who live in households with parents not in poverty. Second, minimum wage hikes could have adverse consequences on employment prospects by raising business costs.
That’s why economists have long concluded that in-work income transfers to families with children through programs such as the earned income tax credit (EITC) are better targeted at reducing poverty. Many believe they are preferable to minimum wages as a poverty reduction policy, because they encourage work while boosting incomes of poor households, rather than discouraging employers from hiring (though they come with other problems).
Yet poverty thresholds themselves ignore the EITC and other transfers that have expanded in recent decades. Comparing the income for a full-time minimum wage earner to poverty thresholds is therefore very misleading on the living standards many households actually enjoy.
In short, the metrics that $15 minimum wage advocates use to make the case for substantial hikes are not economically sensible benchmarks. For all the noise and economics-y language of their arguments, economists generally oppose manipulating prices to obtain social policy objectives.
Your focus in considering this policy proposal should center instead on traditional labor market economics.
Changing the legislated minimum wage affects both:
1) the incomes of minimum wage workers who maintain their jobs and hours
2) the number of jobs and hours of work employers demand, and broader working conditions, due to the change in business costs.
Standard economics tells us that, in most cases, raising the minimum wage will boost 1) and reduce 2). What becomes crucial then is considering the potential size of these two effects and judging their importance to achieving your objectives. The first requires drawing on economic evidence. The latter is more a question of philosophy.
Germany introduced a new economy-wide minimum wage for the first time in 2015, at a relatively high rate of €8.50 ($9.67 today). This rose to €8.84 in 2017. For reference: between 10 and 14 percent of eligible workers were thought to earn less than €8.50 before the policy was introduced.
This is interesting from a research perspective. Most minimum wage studies examine the impact of minimum wages at low levels or assess small changes to their rate. But here we have a case study of a whole regime change with a high rate introduced for the first time.
A new paper by IZA Institute of Labor Economics provides a clear literature review on the effects so far. Studies have exploited three different strategies to assess the impact: utilizing regional variation of the “bite” of the minimum wage, using treatment and control groups, and assessing the impact on firms. As the table below shows, a broad consensus is emerging, which sits well within the existing minimum wage literature:
- Unsurprisingly, hourly wages have increased at the bottom of the income distribution, though there is little evidence of a ripple effect further up.
- Most studies find a small but negative effect on overall employment (up to 260,000 fewer jobs), driven by reduced hiring (not layoffs) and a reduction of casual and atypical employment.
- All studies that assessed it find a negative effect on contractual hours.
- As a result, although hourly wages increased, the reduction in hours meant gross monthly earnings does not appear to have increased much for low-paid employees.
- Since gross monthly earnings have not substantially increased, and those earning minimum wage are often not from the poorest households, the policy hasn’t seemingly reduced the risk of being in poverty.
For more on the state of the academic debate on minimum wages, read my Regulation article.
One story about poverty in the United States goes like this: Poverty is simple to escape. Finish high school. Get a job, even a menial one. Do not have kids until you're married. And if you do all these things, you're pretty unlikely to be poor.
Conservatives like this story because it suggests that no significant social changes are needed to end poverty. On this view, poverty may even be just a personal choice. It's largely up to you whether you follow the so-called "success sequence" or not.
Critics, though, are quick to point out that the success sequence is much easier described than followed, and that following it is much easier for some people than for others. Failing or dangerous schools offer little reason for students to remain. Getting a job is easier in some places than others, and easier for some types of people than others. In some communities, marriage partners are all too few. And avoiding having children is a lot to ask, because it's a natural human desire to want to have them.
If the success sequence doesn't hold up so well, what do we do about it? And what specifically libertarian steps remain to be done to fight poverty?
This month at Cato Unbound, we're debating the usefulness of the success sequence as a tool for thinking about American poverty. Cato Senior Fellow Michael Tanner has written the lead essay, which I encourage you to read. Comments are open, and we welcome readers' feedback. Discussion with a panel of diverse outside experts will continue through the end of the month.
An overwhelming majority—95 percent—of Americans are confused about the state of global poverty. A survey from the late Hans Rosling’s web project Gapminder assessed the public’s knowledge on that subject. The survey asked twelve thousand people in fourteen countries if, over the last two decades, the proportion of the world's population living in extreme poverty has a) almost doubled, b) stayed the same, or c) almost halved.
The correct answer, as frequent visitors of HumanProgress.org know, is c. Extreme poverty has halved. But a staggering 19 in 20 Americans got the answer wrong. In fact, most people in all fourteen countries surveyed got it wrong. How is it possible that so many people are unaware of the extraordinary and unprecedented decline in world poverty that has been achieved in the last twenty years?
In some cases, they’re following the headlines instead of the trend lines: people in the news know that pessimistic narratives attract more clicks than heartening long-term trends. As the saying goes, “If it bleeds, it leads.” And, of course, many in the media share the broader public’s ignorance of the progress that humanity has made in its fight against world poverty. We at HumanProgress.org will continue to do our part to correct mistaken perceptions about the state of humanity and advocate for a realistic, empirically-based view of the world.
Paul Krugman’s column yesterday lamented Republican policy towards the poor. He has particular gripes with Ben Carson’s changes to housing subsidies, increased work requirements for those seeking food stamps, and waivers granted to states to enable new work requirements for Medicaid.
I’m not going to get into these specific policy changes here. But let’s take Krugman’s analysis of the changes and the motivation for them at face value, and pose a question: how robust is an anti-poverty agenda that depends so much on political and societal attitudes to the poor?
As I outlined in a recent blog, it’s a mistake to think of policy towards the poor being merely about government transfers, services and benefits-in-kind. In fact, this focus on income and services has blinded the debate about poverty from the truth that there are lots and lots of state, local and federal policies that increase the price of goods and services the poor spend a disproportionate amount on.
Zoning laws and urban growth boundaries raise house prices. Regulations make childcare more expensive. Sugar, milk programs and the ethanol mandate increase food costs. Tariffs on clothes and footwear have particularly regressive effects. Energy regulations which seek to subsidize renewables rather than being “technology neutral” can raise prices. CAFE standards, constraints against ride sharing, and some regulations on gas taxes raise some transport prices too. Not to mention the broader effects of protectionism and occupational licensing in both raising prices and reducing efficiency across the economy.
Some of these things affect families by orders of magnitude greater than the changes Krugman is concerned about. Combined, they would have a huge impact for many households. What’s more, most of the status quo interventions make the economy less efficient too, reducing market-wages and, in the case of housing and childcare, deterring the mobility of labor over different dimensions. One cannot talk about a “war on the poor” without acknowledging these fronts and the armies which battle on them, not least because these bad policies in part drive significant demands for redistributive transfers in the first place.
In my view, it would be far more fruitful for liberals concerned with the well-being of the poor to focus on all these issues as part of a “first do no harm” poverty agenda. Why?
1. There’s evidence that fiscal transfers may have hit diminishing returns in terms of their role in poverty alleviation.
2. The fiscal environment is not conducive to huge new expenditures on programs, and evidence from other countries (not least Britain) suggests working age welfare is the first port of call for cuts when a fiscal crisis hits.
3. There are clear economic trade-offs where transfers are concerned. As this accompanying Twitter thread by Paul Krugman acknowledges, even increasing the availability and generosity of transfers to more people disincentivizes people from earning more income.
4. And crucially for Krugman’s column, attitudes to redistribution are volatile, and support can be replaced by narratives about “moochers” or “welfare queens” relatively quickly.
In contrast, a pro-market agenda seeking to undo existing damaging regulations at the local, state and federal levels could: reduce poverty, reduce the demands for redistributive activity, would not undermine work incentives and would be harder to undo given its dispersed nature. Those in favor of extensive redistribution should see this too: you do not have to believe existing anti-poverty programs have failed to acknowledge they can have negative unintended consequences, hit diminishing returns, or that their effectiveness is undermined by bad policies which drive up living costs.
A pro-market cost of living agenda would not "solve" poverty, of course. And there are major vested interests in each of these areas who would resist reform. But there are clearly lots of different wars on the poor being raged, even if inadvertently. As long as the poverty debate focuses on just income transfers and government services, the more bountiful battles against vested interests who drive up the poor’s living costs go unfought.
Following Trump’s electoral success in rustbelt states, the spotlight has been on white, rural, post-industrial poverty. J.D. Vance, author of the now-famous memoir Hillbilly Elegy, discussed some possible explanations for rural poverty yesterday in a podcast. In the interview, he suggests that geographic (im)mobility is partly to blame for the erosion of areas like Appalachia: the poor simply aren’t migrating to jobs.
Vance is right that Americans have limited interest in relocating, and are relocating less than before. According to calculations using University of Chicago data, the proportion of individuals unwilling to relocate for work is high: 42% of Americans say they will not move within the United States for work, and 68% of Americans will not move outside the country for work. A full quarter (25%) of Americans would not consider traveling further for a job, even if the decision resulted in unemployment. Meanwhile, Census data suggests that relocation—whether inter-state, inter-county, or intra-county—is down (Figure 1). 2016 had the lowest relocation rate in seventy years (Figure 2).
Figure 1. Type of Move, 1948–2016
Figure 2: Number of Movers and Mover Rate, 1948–2016
Why are people relocating less than ever? One explanation Vance misses is that government policy gets in the way. For example, research provides evidence that land use regulations put pressure on housing prices in high-opportunity areas, which in turn eliminates the fiscal rewards of relocation for the poor and unskilled. Tragically, this means low-income Americans are trapped in job deserts with little in the way of opportunities, amenities, or hope.
You can listen to more of Vance’s interview regarding the causes of white, rural, post-industrial poverty, here.
 Author’s own.
Earlier this month, James Kwak penned an extensive critique for The Atlantic of the Econ 101 view that government-imposed minimum wage rates lead to job losses. There is a lot of muddled thinking in the article, not least that it constantly conflates poverty and inequality. But for the sake of brevity, here are 9 observations:
- The theoretical and empirical literature does suggest the minimum wage question is more complex than first imagined, and it should not surprise us that reality lies somewhere between the perfectly competitive model of the labor market and the monopsonistic one, depending on the time and sector analyzed.
- However, the bulk of the detailed empirical literature still supports Econ 101’s prediction that raising wages by government dictat reduces labor demand, particularly for certain groups (the young and lowest skilled). Of course, the “bite” of the minimum wage is significant. That an increase from $7.25 to $8 may not have a large effect on the labor market does not mean that a rise to $15 wouldn’t have a much bigger effect.
- Reductions in labor demand need not mean higher unemployment per se (and so tracking time series of unemployment against minimum wage rates is unhelpful.) It may affect hours offered, lower the quality of jobs as firms cut back on financing training, or lead to cuts to other employee benefits. More recent evidence also stresses the dynamism of the labor market – minimum wage hikes may not manifest themselves through immediate job cuts (not least because of redundancy costs and stickiness of contracts/orders), but can lower the propensity to hire in future and hence slow job growth.
- Kwak’s explanations for why Econ 101 might be wrong seem like motivated reasoning. One potential explanation is that firms in sectors affected by the minimum wage have significant market power. If a firm is a monopsonist, then theoretically you can increase both wages and employment by raising the minimum wage. In reality the days of the so-called ‘company town’ are over. Looking at the industries in which most minimum wage workers actually operate: food preparation, serving, sales, personal care and office support and administration – does anyone think these aren’t very competitive industries?
- Kwak uses Keynesian logic to claim that low paid workers’ higher marginal propensity to consume means minimum wages can boost demand in the economy. Of course, this ignores any contractionary effects arising from lower profits, higher prices or reduced employment. It’s worth noting Keynesian theories are also predicated on the idea that market wage rigidities are precisely what leads to unemployment when an economy contracts (so it’s difficult to see why that would be any different with state-imposed rigidities).
- Another argument outlined is that minimum wage hikes can improve productivity. It is said that firms paying more can reduce employee turnover and attract higher skilled workers. This is a variant of the ‘efficiency wage hypothesis.’ But the efficiency wage hypothesis is precisely a theory of persistent unemployment! Even in that theory, much of the productivity gains that firms can theoretically get from paying higher wages occur from being able to hire the best workers in an industry. If *every* firm is mandated to pay a higher wage, then these gains are lost, and so the line of argument cannot be generalized to the whole economy.
- The hurdle in assuming there is some form of free lunch that firms are not exploiting is that a government setting a minimum wage must know more about the potential productivity of individual firms’ workers than the firms themselves. Even if this were true, which is highly doubtful, it need not be the case that setting a minimum wage to reduce turnover would be good for economy-wide productivity. A higher wage in certain sectors might reduce the incentive for workers to move to higher-skilled, higher-paying sectors over time by building or investing in human capital – actually lowering the productive potential of the economy in the long-term.
- Kwak writes: “Even if a higher minimum wage does cause some people to lose their jobs, that cost has to be balanced against the benefit of greater earnings for other low-income workers.” This is undoubtedly true, but in the rest of the article he seems to really judge the morals of those who seek to protect the rights and opportunities of low-skilled individuals. If he and other advocates want to advocate for significant minimum wage increases, he also has to own the estimated 500,000 low skilled workers on the scrap heap from raising the minimum wage to $10.10. This might be uncomfortable, but then the minimum wage has uncomfortable roots – with UK Fabians desiring it precisely to exclude the low-skilled, disabled and migrants from working.
- Finally, despite repeated allusion to the idea that raising the minimum wage would be a good way of reducing poverty, Kwak produces scant evidence to that effect. Joseph Sabia’s work – reported in a Cato bulletin – actually suggests that minimum wages are a badly targeted anti-poverty policy. He found workers earning between $7.25 and $10.10 per hour overwhelmingly live in non-poor households (the unit by which we measure poverty). In fact, only 13 percent of those affected lived in poor households, while nearly two-thirds lived in households with incomes over twice the poverty line. In other words, many people earning slightly above the current minimum wage may be second earners (particularly part-time) or young people who live in households where total household income is above the poverty line.