Moving the blog to the new website

We have moved our blog to our new website. Please follow the blog there, since this blog will no longer be regularly updated.


How The Poor Got Framed

People respond not the world as it is, but to the world as they perceive it to be. This simple but powerful truth about perspective is particularly important when it comes to setting poverty policy in the United States. Are the poor ‘cheating’ and ‘lazy,’ or is poverty the product of a system that throws up barriers to work and opportunity leading to the neglect and misery of part of the population?

In an excellent new study in Policy Studies Journal, Max Rose and Frank Baumgartner look at the data on media framing of poverty and government generosity from 1960-2008. Even for someone like me who has long thought media matters, the results were stunning. A full 82 percent of the changes in government generosity to the poor can be explained by the tone of media coverage over the previous ten years.

Rose and Baumgartner identified five frames in poverty stories, three positive frames and two negative ones. The chart below is a stacked line graph showing the distribution of stories by framing. The lines always total to 100%, so it is the spaces between the lines that show the percentage of news stories that apply each frame. The negative frames are shown at the top of the chart, above the bold line. Since the 1960s there has been a huge growth in the percentage of poverty stories that use the frame “Lazy” and a major decline in those that frame poverty in terms of ‘Misery and Neglect.’

Rose Baumgartner media 1

This overall decrease in positive frames has been matched by a decrease in the generosity of government. The media tone for the previous ten years is an excellent predictor of the current level of government generosity.

Rose Baumgartner media and generosity


The fit between media coverage and policy appears almost too good to be credible, a point Rose and Baumgartner address in their paper.

…the story appears too simple. However, recall that our measure of generosity incorporates the number of poor, the depth of their poverty, and the percentage of all government spending on alleviating poverty. Similarly, our framing indicator combines the level of attention (e.g., how many stories are printed) with the tone of that attention. One way to understand these surprisingly simple results is to focus on how they summarize and put into context what many qualitative and quantitative studies have shown us over the decades: after the War on Poverty, the discussion turned toward a more negative view of the poor and the policies that supported the poor, making them easy targets when looking for spending cuts.

It is important to note that this is not simply a partisan issue, but instead represents a major shift in the way the United States has looked at poverty. From the paper:

…the new elite discourse on the poor is not simply conservative or ideologically right wing….it has shifted from an abstract ideological stance to one more focused on more operational issues of ‘what works’ and on a long-standing unease at the idea of recipients not working for the benefits they receive. Our data suggest that this focus on the individual, as opposed to the system, may be one of the most important elements of the general ideological ascendance of neoliberalism in American Politics since the 1970s.

I encourage you to read the full paper if you’re interested in the details of the research, but Rose and Baumgartner pull together their main points in this concluding paragraph:

Policymakers, members of the public, and journalists once focused on aspects of poverty that are beyond the control of those who find themselves with dire economic prospects or which focus on the collective costs to all Americans from having large numbers of poor. This resulted in a large decrease in the amount of poverty in this country. From this initial focus, associated with optimistic efforts to alleviate poverty and which justified massive interventions and spending, the public has given up, tired, frustrated, discouraged. Collectively, attention now focuses on what we have called the ‘stingy’ frames: The poor are individually responsible for their problems, and government efforts to help them may do more harm than good. We have shifted from an overwhelming focus on one side of the coin to an equally disproportionate focus on the other side, and policy has followed the framing.

Shifting the dominant frame back to one that focuses on systems and the role they play in producing outcomes is a difficult task, but a necessary one if we are to return to a public policy that treats poverty as a problem worth solving.


CBO finds increasing the minimum wage does not have a statistically significant impact on employment

That’s probably not the headline you were expecting if you’ve turned on the news in the last few days. The Congressional Budget Office (CBO) says raising the wage to $10.10 will cost 500,000 jobs! (We posted a full analysis of the CBO report yesterday). Thats certainly seems significant. But statistical significance is not a measure of magnitude, it’s a measure of certainty. How sure are we that we’ve got the correct number? In this case, the CBO is not very certain at all. In fact, by the standards normally used in social science research, they aren’t justified in rejecting the hypothesis that the minimum wage has no impact at all on employment.

But let’s back up a little to talk about how we can evaluate certainty. In most cases, when we test for statistical significance, we’re testing if our result is different from zero. The most intuitive way to test this is to look at a confidence interval. Social scientists usually report a 95% confidence interval, indicating a range that they would expect to observe 19 times out of 20 if we were enact a policy (in this case, raising the minimum wage).  A wide confidence interval indicates a high range of uncertainty. If the confidence interval includes zero, then we say that result is not statistically significant. (1)

Now, based on the CBO’s own numbers, the confidence interval for the impact of raising the minimum wage on employment includes zero. So, why doesn’t the CBO report clearly say this? Well, instead of reporting the 95% confidence interval, they chose to report a 67% confidence interval. Now, I’ve never seen anyone report a 67% confidence interval. So I asked a few economists and an econometrician (someone who deals with advanced statistics in economics), and they’ve never seen a 67% confidence interval either. They only reason someone would choose such a low confidence interval is to make their result appear statistically significant.

Choosing a confidence interval after running the analysis in order report a significant result is a form of manipulating data. Chances are, even people who routinely deal with confidence intervals skipped the footnote in which the CBO noted that they were not reporting the standard 95% confidence interval, but instead a 67% confidence interval.

The CBO’s analysis showed, correctly, that they really have no idea what the impact of a minimum wage increase would be on employment. However, they then chose to report that analysis in a misleading fashion by intentionally choosing a confidence interval that allowed them to report a significant result. The correct thing to say would have been along the lines of, “we really have no idea, but we think there might be a slight negative impact.” Or, in slightly more technical terms, “we found a negative coefficient, but it wasn’t statistically significant.”  It’s a shame the CBO didn’t choose to report the data in a more straightforward manner.


(1) I’m assuming that we’re testing for a statistically significant difference from zero. This is the most common, but one could also test if the result is significantly different from 1 or 2 or any value one cares to test. However, in the case of raising the minimum wage, the pressing question is if the employment impact is zero, or not.

What Could Cut the Poverty Rate Right Now

What does the CBO report actually tell us about increasing the minimum wage?

Yesterday’s Congressional Budget Office (CBO) report on the impact of raising the minimum wage has been all over the news today. The headlines from the report (16.5 million workers with increased wages, 500,000 jobs lost, and 900,000 people lifted out of poverty) come as mixed news for both sides of the debate. It’s worth taking a little bit of time to dig into the report and its likely impact on minimum wage legislation.

First, let’s talk about why the CBO report matters. Reports on the minimum wage are a dime a dozen, and anyone with an internet connection can quickly conjure up a study that tells them whatever they want to hear.  The CBO is the closest thing there is to a referee within policy circles. Although it is occasionally challenged, it is generally respected as a source of good, impartial analysis.

In this report, the CBO analyzes the impacts of raising the minimum wage to $10.10 and of raising it to $9.00. Because the current push is to raise the wage to $10.10 I will only be talking about that part of the report. Let’s start with the headline numbers presented in Table 1.

CBO min wage table 1

First, we have the estimated job losses, 500,000. As the CBO notes, this is highly uncertain. There is a 67% chance that job losses will be between 0 and 1,000,000. That leaves a 33% change, which is not insignificant, that the minimum wage could have an even stronger effect in either  direction. In other words, the CBO is really not sure how many jobs it costs, to the extent that there’s a 17% chance the impact is actually positive! Their estimate is based on empirical studies of the employment impacts on teenagers, and then an attempt to place those studies in the context of other studies on adults. There are decent reasons to think that their resulting estimate is too high. (See, in particular, this review of empirical studies on the minimum wage from the Center for Economic and Policy Research).  The easiest summary to see is this visual one provided by Jared Bernstien in the NY Times this morning, which shows a very clear clustering around no impact on employment:

min wage funnel graph

That was the bad news from the minimum wage report. The rest is highly positive. Raising the wage would:

  • Increase wages for 16.5 billion people.
  • Lift 900,000 people out of poverty
  • Of the people impacted, only 12% are teenagers and over half are full-time workers.
  • Increase overall incomes by $2 billion.

For some reason, the myth that the minimum wage mainly benefits part-time teenage workers never dies, despite clear evidence that most people impacted by the wage increase are over 20. Hopefully the CBO report can put that myth to rest even as it rekindles debate over the employment impacts.

The benefits of the minimum wage are not evenly spread. Figure 3 shows the impact on income groups classified based on their distance from the poverty line.

CBO min wage fig 3

It suddenly becomes very clear that a minimum wage increase is good news for most people making less than 6 times the poverty line.

Suppose, for the sake of argument, we accept the CBO’s analysis that the minimum wage will lead to job loss for 500,000 workers. It also helps 16.5 million workers directly (and others indirectly, but since CBO does not attempt to quantify those workers we’ll ignore them for now). That leaves us with a policy that helps 97% of low-wage workers. If we could distribute gains evenly, this would be the equivalent of taking a 3% cut in pay in exchange for a 39% increase in hourly wages. (1) That’s a really good deal, and that’s why, in aggregate, an increase in the minimum wage increases earnings for low-income families and lifts 900,000 families over the poverty line.

If someone offered me a 3% reduction in my hours and a 39% increase in my hourly wages I’d jump at the chance to take that deal. Now, unfortunately, businesses are likely to cut people instead of cutting hours. I think the CBO has overstated the potential employment losses, but supposing they have it right, we’re still looking at a policy with an undeniable upside. It’s not only quantity of jobs that counts, it’s quality as well.

(1) The 3% cut is the cut in hours moving from 17 to 16.5 million jobs. The 39% increase is the increase in the minimum wage from $7.25 to $10.10.

A Visual Essay on the Wage-Productivity Gap

In the late 1970s a gap opened up between the productivity of workers, and the amount they were being paid. The not-so-creative name for this trend is the wage-productivity gap. The causes are hotly debated, but the consequences are straightforward. The economy continues to grow through increases in productivity, but the benefits of increased productivity are not shared by workers. This means real wages stagnate, and inequality increases. It also calls into question the justice of a system in which the production of workers is divorced from the wages of workers.  The Bureau of Labor Statistics has put together a visual essay on wage-productivity gap.

BLS wage prod 1 BLS wage prod 1a

In chart 1 we see the rate of change in productivity growth compared to the rate of change in real (inflation-adjusted) hourly compensation (1). Chart 2 breaks down the trend by selected time periods. We can see that from 1947 to 1979, wages and productivity grew together. They started to separate in the 1970s, and came completely unglued in the 1980s, through the early 1990s.  The late 1990s was the only time in the past 30 years that real (inflation-adjusted) wages grew with productivity, and then in the 2000s wages remained stagnant while productivity continued to increase.

The growing gap is also reflected in the decrease in labor’s share of output.

BLS wage prod 3



For a long time the labor share was thought to be constant. Here again, the causes are both complex and disputed, but the consequences are clear, less of the work being done by workers is reflected in their wages.

Without really figuring out what’s going on, it’s tough to reverse these trends. Chances are technology, globalization, and loss of union power all play roles in driving down labor’s share of income and opening up the wage-productivity gap. In some cases (technology and globalization) those processes have benefits as well as costs, so the trick is to capture the benefits while reducing or eliminating the costs. This is not an easy task, but there are a few clear implications to these trends.

First, we should support a strong safety net. Stagnant real wages and unemployment are being caused by macroeconomic changes, not individual’s suddenly becoming less productive or less deserving. Second, we can raise the minimum wage to help insure that workers share in the gains of productivity.


(1) Technical note: Taking the log of both indexes allows us to see the rate of change presented visually. Instead of each tick mark representing an increase of X  in a meaningless index, each tick mark represents a percentage increase (in this case, each tick mark is a 22% increase). The gap actually looks much larger if you don’t log-transform the data first, as later changes are bigger in absolute terms than they are in percentage terms.

The Minimum Wage and Employment: The Case of Two Terrible, Horrible, No Good, Very Bad Assumptions

In any debate about the minimum wage, inevitably someone will dig out their knowledge of economics 101 and argue that raising the minimum wage is self-defeating because increasing the price of labor will decrease the demand of labor, and so unemployment will rise. Like most econ 101 arguments, this rest on some terrible, horrible, no good, very bad assumptions.(1) For the sake of space, let’s just talk about two of them.

  1. The labor market works like other markets.
  2. Raising the minimum wage will not affect anything else in the economy. (Economists call this ceteris paribus, Latin for “other things equal.” Because we all know that if you say it in Latin it must be true.)

Let’s think about the labor market compared to other markets. If I were to ask you the ideal price of any normal good or service, like a new computer or a haircut, you would probably say free. After all, in our ideal world, all of our material needs are satisfied at zero or at least very little costs. But what about the price of labor? Almost certainly you would not want labor to be free, because labor, on a fundamental level, is not the same as a commodity. (Thank you to blogger Squarely Rooted for this general idea).

The labor market is also characterized by large information asymmetries and differences in bargaining power. Potential employees have less information than their employers, and tend to have fewer options. This inequality of bargaining power only gets worse when there is high unemployment. You don’t have to be a radical to think this is a problem, you can find an analysis of uneven bargaining power between workers and employers in capitalism’s foundational text, Adam Smith’s Wealth of Nations. This inequality is a driving force behind unions, labor market regulations and yes, the minimum wage. Any time one party to a contract has significantly more power than the other there is always the opportunity for exploitation. (Fun Fact: Urban Sinclair’s The Jungle was not primarily about unsanitary food, the focus was actually on the terrible working conditions of the people preparing the food. As one example, he wrote about someone falling in a meat grinder thinking that people might care that unsafe factories were killing people, but instead the audience just freaked out about the possibility of there being human in their food). If you do not believe that power differentials lead to exploitation, please spend some time reviewing the history of the industrial revolution. Labor market laws were a direct response to historical abuses of power.

Second, raising the minimum wage will inevitably impact other parts of the economy. The Economic Policy Institute has argued that by increasing the earnings of people who are likely to spend all their extra money we can boost aggregate demand and therefore help stimulate employment. John Schmitt at the Center for Economic and Policy Research suggests 11 theoretical reasons a minimum wage increase might not increase unemployment. 

Fortunately, we don’t have to restrain ourselves to competing theories. The minimum wage has been empirically studied. A lot. So much that there are now many studies of studies (meta-studies) that try to see if there is a consensus about the minimum wage.  John Schmitt has a review of those, but the most compelling way to see the results of the meta-studies is this graphic on teenage employment and the minimum wage.

min wage funnel graph


The results are weighted by the accuracy of the estimate, with higher points being considered more accurate. So of the 1,492 estimates available, we see a clear convergence around zero. Now this is just teenage unemployment, but there’s widespread agreement that if the minimum wage has a dis-employment affect it will be strongest among teenagers. So if we can’t see it in looking at the data on teenagers, we won’t be able to see it with adults either. So at some point we may be justified in concluding that there just isn’t a strong effect of the minimum wage on employment.

Of course, if minimum wage doesn’t increase unemployment and does increase wages, we have a clear win for low-wage workers. One final note, although I criticized ‘econ 101’ logic at the beginning of this post, I’m not saying anything here that hasn’t been said by a lot of economists. In fact, a survey of leading economists found them ambivalent on the question of unemployment effects, but fairly decisively in favor of raising the minimum wage. Even among those economists who think there may be a small impact on employment, there’s largely agreement that the benefits of higher wages outweigh the modest effect on employment. Here are their responses asked if the benefits of raising the minimum wage to $9 an hour outweigh the costs:

minimum wage survey


So why do all these economists disagree with econ 101? Econ 101 makes assumptions in order to make things easier to understand as a starting point for learning about economics. Good economists understand that these are assumptions, look at empirical evidence, and then adjust by making increasingly realistic theoretical models. So next time someone says “It’s basic Econ 101” remember that’s not the end of the story. Economics only gets more complicated from there, so making policy decisions based on “simple supply and demand” is never a good idea.


(1) I realize this is redundant, but I want to (a) make a point about just how bad these assumptions are, and (b) make a reference to one of my favorite children’s books, “Alexander and the Terrible, Horrible, No Good, Very Bad Day”