Friday, July 20, 2018

Early Childhood Education Fails Another Randomized Trial

Public programs for pre-K education have a worthy goal: reducing the gaps in educational achievement that manifest themselves in early grades. I find myself rooting for such programs to succeed. But there are now two major randomized control trial studies looking at the the results of publicly provided pre-K programs, and neither one finds lasting success. Mark W. Lipsey, Dale C. Farran, and Kelley Durkin report the results of the most recent study in "Effects of the Tennessee Prekindergarten Program on children’s achievement and behavior through third grade" (Early Childhood Research Quarterly, 2018, online version April 21, 2018). 

The Tennessee Voluntary Pre-K Program offers support for families with low-income levels. In 2008-2009, the program only had sufficient funding to cover 42% of applicants. Admission to the program was thus decided by a lottery. This selection procedure makes the eyes of academic researchers light up, because it means that there was random selection of who was in the program (the "treatment" group) and those who were not (the "control" group). As these students continue into Tennessee public schools, there is follow-up information on how these two groups performed. The result was that that students in the pre-K program has a short-term boost in performance, which quickly faded. The abstract of the study says:
"Low-income children (N = 2990) applying to oversubscribed programs were randomly assigned to receive offers of admission or remain on a waiting list. Data from pre-k through 3rd grade were obtained from state education records; additional data were collected for a subset of children with parental consent (N = 1076). At the end of pre-k, pre-k participants in the consented subsample performed better than control children on a battery of achievement tests, with non-native English speakers and children scoring lowest at base-line showing the greatest gains. During the kindergarten year and thereafter, the control children caught up with the pre-k participants on those tests and generally surpassed them. Similar results appeared on the 3rd grade state achievement tests for the full randomized sample – pre-k participants did not perform as well as the control children. Teacher ratings of classroom behavior did not favor either group overall, though some negative treatment effects were seen in 1st and 2nd grade."
The previous major randomized control trial study of an early childhood education program was done on the federal Head start program. I discussed it here. Lipsey, Farran and Durkin summarize the finding in this way:
"This study [of Head Start] began in 2002 with a national sample of 5000 children who applied to 84 programs expected to have more applicants than spaces. Children were randomly selected for offers of admission with those not selected providing the control group.The 4-year-old children admitted to Head Start made greater gains across the pre-k year than nonparticipating children on measures of language and literacy, although not on math. However, by the  end of kindergarten the control children had caught up on most achievement outcomes; subsequent positive effects for Head Start participants were found on only one achievement measure at the end of 1st grade and another at the end of 3rd grade. There were no statistically significant effects on social–emotional measures at the end of the pre-k or kindergarten years. A few positive effects appeared in parent reports at the end of the 1st and 3rd grade years, but teacher and child reports in those years showed either null or negative effects."
Of course, the fact that the two major studies of publicly provided pre-K find near-zero results by third grade doesn't prove that such programs never or can't work. There are many studies of early childhood education programs. The results of the Tennessee and Head Start studies don't rule out the possibility of benefits from narrower programs specifically targeted at children with certain needs. Also, some studies with longer-term follow-up found that although measures of educational performance didn't move much. pre-K programs had longer-term effects on outcomes like high school graduation rates.

But the case for believing that publicly provided  pre-K programs will boost long-term educational outcomes for the disadvantaged is not very strong. If positive results were clear-cut and of meaningful size, it seems as if they should have shown up in these major studies.

Some researchers in this area have suggested that interventions earlier than pre-K might be needed to close achievement gaps. For example, some evidence suggests that the black-white educational achievement gap is apparent by age 2. It may be, although the evidence on this point isn't clear, that some of the funding now being spent on pre-K programs would have a bigger payoff if spent on home visits to the parents of very young children. Greater attention to the health status of pregnant women--including both personal health and exposure to environmental risks--might have a substantial payoff for the eventual educational performance of their children, too.

Thursday, July 19, 2018

Conflict Minerals and Unexpected Tradeoffs

The cause seemed worthy, and the policy mild. Militia groups in the Democratic Republic of the Congo (DRC) were taxing and extorting revenue from those who mined like tin, tungsten, and tantalum. Thus, Section 1502 of the Dodd-Frank Act of 2010 required companies to disclose the source of their purchases of such minerals. The hope was to reduce funding for the militias, and thus to benefit people in the area. Human rights advocacy groups supported the idea. The Good Intentions Paving Company was up and running.

But tradeoffs are no respecters of good intentions. Dominic Parker describes some research on the tradeoffs that occurred in  "Conflict Minerals or Conflict Policies? New research on the unintended consequences of conflict-mineral regulation" (PERC Reports, Summer 2018, 37:1, pp. 36-40). Parker writes:
"First, Section 1502 initially caused a widespread, de facto boycott on minerals from the eastern DRC. Rather than engaging in costly due diligence to identify the sources of minerals—and risking being considered a supporter of rebel violence—some U.S. companies simply stopped buying minerals from the region. This de facto boycott had the intended effect of reducing funding to militias, but its unintended effect was to undercut families who depended on mining for income and access to health care. The decreases in mineral production rocked an artisanal mining sector that had supported an estimated 785,000 miners prior to Dodd-Frank, with spillovers from their economic activity thought to affect millions.
"Second, the legislation changed the relative value of controlling certain mining areas from the perspective of militias, who changed their tactics accordingly. Before the boycott, the militias could maximize revenue by taxing tin, tungsten, and tantalum at or near mining sites. They therefore had an interest in keeping mining areas productive and relatively safe for miners. After the legislation, the militias sought to make up for reduced revenue in other ways. According to the evidence, they started to loot civilians who were not necessarily involved in mining. They also started to fight for control over other commodities, including gold, which was in effect exempt from the regulation."
One result of the economic losses in the area was a sharp rise in infant mortality rates: "The combined evidence suggests that Dodd-Frank increased the probability of infant deaths (that is, babies who died before reaching their first birthday) from 2010 to 2013 for children who lived part of or all of their first year in villages targeted by the legislation and mining ban. The most conservative estimate is that the legislation increased infant mortality from a baseline average of 60 deaths per 1,000 births to 146 deaths per 1,000 births over this period—a 143 percent increase."

The level of violent conflict affecting civilians actually seemed to rise, rather that fall: "At the end of 2010, after the passage of Dodd-Frank, looting in the territories targeted by the mining policies became more common and remained that way through much of 2011 and 2012, when our study period ended. ... The incidence of violence against civilians also increased in the policy regions after the legislation ..."

One economic insight here is the "stationary bandit" theory that when a bandit remains in one location, there are incentives for the bandit to keep local workers and companies safe and productive.

The political insights are fuzzier. One can't rule out that if the Dodd-Frank provisions had been better thought-out or better targeted, maybe the effects would have been better, too. Or maybe this is a case where long-run benefits of these provisions will outweigh short-run costs. But it's also possible that an alternative strategy for bolstering the economy and human rights in the area might have worked better. And it's quite clear that those who supported this particular conflict mineral policy did not predict or acknowledge that their good intentions could have these adverse consequences.

Tuesday, July 17, 2018

On Preferring A to B, While Also Preferring B to A

"In the last quarter-century, one of the most intriguing findings in behavioral science goes under the unlovely name of `preference reversals between joint and separate evaluations of options.' The basic idea is that when people evaluate options A and B separately, they prefer A to B, but when they
evaluate the two jointly, they prefer B to A." Thus, Cass R. Sunstein begins his interesting and readable paper "On preferring A to B, while also preferring B to A" (Rationality and Society 2018,  first published July 11, 2018, subscription required)

Here is one such problem that has been studied: 

Dictionary A: 20,000 entries, torn cover but otherwise like new
Dictionary B: 10,000 entries, like new

"When the two options are assessed separately, people are willing to pay more for B; when they are assessed jointly, they are willing to pay more for A." A common explanation is that when assessed separately, people have no basis for knowing if 10,000 or 20,000 words is a medium or large number for a dictionary, so they tend to focus on "new" or "torn cover." But when comparing the two, people focus on the number of words.

Here's another example, which (as Sunstein notes is "involving an admittedly
outdated technology":

CD Changer A: Can hold 5 CDs; Total Harmonic Distortion = 0.003%
CD Changer B: Can hold 20 CDs; Total Harmonic Distortion = 0.01%

"Subjects were informed that the smaller the Total Harmonic Distortion, the better the sound quality. In separate evaluation, they were willing to pay more for CD Changer B. In joint evaluation, they were willing to pay more for CD Changer A." When looking at them separately, holding 20 CDs seems more more important. When comparing them, the sound quality in Total Harmonic Distortion seems more important--although most people have no basis for knowing if this difference ins sound quality would be meaningful to their ears or not.

And one more example:

Baseball Card Package A: 10 valuable baseball cards, 3 not-so-valuable baseball cards
Baseball Card Package B: 10 valuable baseball cards

"In separate evaluation, inexperienced baseball card traders would pay more for Package B than for Package A. In joint evaluation, they would pay more for Package A (naturally enough). Intriguingly, experienced traders also show a reversal, though it is less stark." When comparing them, choosing A is obvious. But without comparing them, there is something about getting all valuable cards, with no less valuable cards mixed in, which seems attractive.

And yet another example:

Congressional Candidate A: Would create 5000 jobs; has been convicted of a misdemeanor
Congressional Candidate B: Would create 1000 jobs; has no criminal convictions

"In separate evaluation, people rated Candidate B more favorably, but in joint evaluation they preferred candidate A." When looking at them separately, the focus is on criminal history; when looking at them together, the focus is on jobs.
And one more: 

Cause A: Program to improve detection of skin cancer in farm workers
Cause B: Fund to clean up and protect dolphin breeding locations

When people see the two in isolation, they show a higher satisfaction rating from giving to Cause B, and they are willing to pay about the same. But when they evaluate them jointly, they show a much higher satisfaction rating from A, and they want to pay far more for it." The explanation here seems to be a form of category-bound thinking, where just thinking about the dolphins generates a stronger visceral response, but when comparing directly, the humans weigh more heavily. 

One temptation in these and many other examples given by Sunstein is that joint evaluation must be more meaningful, because there is more context for comparison. But he argues strongly that this conclusion is unwarranted. He writes: 
"In cases subject to preference reversals, the problem is that in separate evaluation, some characteristic of an option is difficult or impossible to evaluate—which means that it will not receive the attention that it may deserve. The risk, then, is that a characteristic that is important to welfare or actual experience will be ignored. In joint evaluation, the problem is that the characteristic that is evaluable may receive undue attention. The risk, then, is that a characteristic that is unimportant to welfare or to actual experience will be given excessive weight."
In addition, life does not usually give us a random selection of choices and characteristics for our limited attention spans to consider. Instead, choices are defined and described by sellers of products, or by politicians selling policies. They choose how to frame issues. Sunstein writes: 
"Sellers can manipulate choosers in either separate evaluation or joint evaluation, and the design of the manipulation should now be clear. In separate evaluation, the challenge is to show choosers a characteristic that they can evaluate, if it is good (intact cover), and to show them a characteristic  that they cannot evaluate, if it is not so good (0.01 Total Harmonic Distortion). In joint evaluation, the challenge is to allow an easy comparison along a dimension that seems self-evidently important, whether or not the difference along that dimension matters to experience or to people’s lives. ... Sellers (and others) can choose to display a range of easily evaluable characteristics (appealing ones) and also display a range of others that are difficult or impossible to assess (not so appealing ones). It is well known that some product attributes are “shrouded,” in the sense that they are hidden from view, either because of selective attention on the part of choosers or because of deliberative action on the part of sellers." 
We often think of ourselves as having a set of personal preferences that are fundamental to who we are--part of our personality and self. But in many contexts, people (including me and you) can be influenced by the framing and presentation of choices. Whether the choice is between products or politicians, beware.

Monday, July 16, 2018

Carbon Dioxide Emissions: Global and US

US emissions of carbon have been falling, while nations in the Asia-Pacific region have already become the main contributors to the rise in atmospheric carbon dioxide. These and other conclusions are apparent from the BP Statistical Review of World Energy (June 2018), a useful annual compilation of global trends in energy production, consumption, and prices. 

Here's a table from the report on carbon emissions (I clipped out columns showing annual data for the years from 2008-2016). The report is careful to note: "The carbon emissions above reflect only those through consumption of oil, gas and coal for combustion related activities ... This does not allow for any carbon that is sequestered, for other sources of carbon emissions, or for emissions of other greenhouse gases. Our data is therefore not comparable to official national emissions data." But the data does show some central plot-lines in the carbon emissions story.

A few thoughts: 

1) The US has often had the biggest declines in the world in carbon emissions in absolute magnitudes in recent years. Granted, this is in part because the quantity of US carbon emissions is so large that even a small percentage drop is large in absolute size. Still, better down than up. The BP report notes: "This is the ninth time in this century that the US has had the largest decline in emissions in the world. This also was the third consecutive year that emissions in the US declined, though the fall was the smallest over the last three years. ... Carbon emissions from energy use from the US are the lowest since 1992, the year that the UNFCCC came into existence.:

2) Anyone who follows this topic at all knows that China leads the world in carbon emissions. Still, it's striking to me that China accounts for 27.6% of world carbon emissions, compared to 15.2% for the US. On a regional basis, the Asia Pacific region--led by China, India, and Japan, but also with substantial contributions from Indonesia, South Korea, and Australia--by itself accounts for nearly half of global carbon emissions. If you're concerned about carbon emissions, you need to think about proposals that would have strong effects on China and this region. 

3) Total carbon emissions from the three regions of South and Central America, the Middle East, and Africa total 13.8% of the global total, and thus their combined total is less than either the United States or the European/Eurasian economies. However, if the carbon emissions for this group of three regions keeps growing at about 3% per year, while the carbon emissions for the US economy keeps falling at 1% per year, their carbon emissions will outstrip the US in a few years.

4) In an interconnected global economy, it's worth remembering that the country where energy is used doesn't always reflect where the final product is consumed. If China produces something through an energy-intensive process that is later consumed in the US, it counts as energy use in China--but both countries play a role.

For some more US-specific data, here's some data from the Monthly Energy Review (June 2018) published by the US Energy Information Administration. This table shows total carbon emissions for the US, emissions per capita, and emissions relative to GDP, going back to 1950.

A few comments: 

1) US carbon emissions on this measure peaked around 2007, and have generally declined since then.  An underlying pattern here is a reduction in the use of coal and rise in the use of natural gas, along with greater use of renewables. US emissions are now back to the levels from the late 1980s and early 1990s. 

2) Carbon emissions per capita in the US economy have fallen back to the level of the early 1950s. 

3) Carbon emissions relative to GDP produced have been falling pretty steadily for the almost 70 years shown in this data. 

Friday, July 13, 2018

Time to Reform Unemployment Insurance

The best time to fix your roof is when the weather is sunny and warm, not when it's rainy, cold--and actually leaking. In a similar spirit, the best time to fix unemployment insurance is when the unemployment rate is low. Conor McKay, Ethan Pollack, andAlastair Fitzpayne offer some ideas in "Modernizing Unemployment Insurance for the Changing Nature of Work" (Aspen Institute, January 2018). They write:

The UI [Unemployment Insurance] program — which is overseen by the U.S. Department of Labor and administered by the states — collects payroll taxes from employers to insure workers against unexpected job loss. Eligible workers who become unemployed through no fault of their own can receive temporary income support while they search for reemployment. In 2016, the program paid $32 billion to 6.2 million out-of-work individuals. UI is one of America’s most important anti-poverty programs for individuals and families, serving as a key counter-cyclical stabilizer for the broader economy. In 2009, the worst year of the Great Recession, the UI program kept 5 million Americans out of poverty, and prevented an estimated 1.4 million foreclosures between 2008 and 2012."

So what's wrong with UI as it stands? The system was designed for full-time workers, who have been with an employer for some time, losing full-time jobs. It doesn't do a good job of covering independent contractors, freelancers, short-timers, and part-timers. However, if you are receiving unemployment insurance and you take a freelance or part-time or self-employed job, your benefits usually stop. The share of unemployed workers who are actually covered by unemployment insurance is falling over time.

Many of the reforms mentioned here have been kicked around before, but it's still useful to have them compiled in one place.

The main conceptual difficulty is that unemployment insurance needs someone to pay into the system on a regular basis, but only to withdraw money from the system when it's really needed. Some legislative creativity may be needed  here. But for example, independent and freelance workers could pay unemployment insurance premiums while employed, and if they did so for some period of time (maybe a year or more), they could become eligible for some level of unemployment insurance payouts.  Figure out ways to offer some protection to those who hold multiple jobs, but are not currently eligible for unemployment insurance from any single employer. Figure out how to offer at least some protection to self-employed and temporary workers.

Alternatively, nontraditional workers could be allowed to set up tax-free savings accounts that they would only use if they became unemployed. Such an account could be combined with a retirement account: basically, a worker with a short-term financial need could withdraw some money from the account, but only up to a certain maximum--while the rest stayed in the retirement account.

Finally, it seems wise not to be too quick to cut off unemployment benefits for those who try to work their way back with a part-time job or by starting their own company. Or unemployment insurance could be designed to encourage workers to conduct a long-distance job search and consider moving to another city or state. 

The OECD Employment Outlook 2018 that was just published includes a chapter on unemployment insurance issues across high-income countries. The problem of limited coverage of unemployment insurance is common.
"Across 24 OECD countries, fewer than one-in-three unemployed, and fewer than one-in-four jobseekers, receive unemployment benefits on average. Coverage rates for jobseekers are below 15% in Greece, Italy, Poland, Slovak Republic, Slovenia and the United States. Austria, Belgium and Finland show the highest coverage rates in 2016, ranging between approximately 45% and 60%: In countries with the highest coverage in the OECD, at least four-in-ten jobseekers still report not receiving an unemployment benefit."
Unemployment insurance systems differ quite a bit across countries: qualifications to receive benefits (like what kind of job you previously had, for how long, and how long you have been unemployed from that job), level of benefits, time limits on benefits, whether you are required to get training or some kind of job search assistance while unemployed--and how all of these factors were adjusted by political systems during and after the rise in unemployment during the Great Recession.

But the OECD report emphasizes that unemployment insurance doesn't just help those who are unemployed--it also provides a mechanisms for government to focus on what kinds of assistance might help the unemployed get jobs again. The chapter in the OECD report ends (citations omitted):
"[U]nemployment benefits provide the principal instrument for linking jobless people to employment services and active labour market programmes to improve their job prospects. In the absence of accessible unemployment benefits, it can be difficult to reach out to those facing multiple barriers to employment, who therefore risk being left behind. In these cases, achieving good benefit coverage can be essential to make an activation strategy effective and sustainable. For this reason the new OECD Jobs Strategy calls for clear policy action to extend access to unemployment benefit within a rigorously-enforced `mutual obligation' framework, in which governments have the duty to provide jobseekers with benefits and effective services to enable them to find work and, in turn, beneficiaries have to take active steps to find work or improve their employability ..." 

Thursday, July 12, 2018

China Stops Importing Waste Plastic

For a few decades now, the US and Europe have been managing their plastic waste by shipping it to China and other countries in east Asia for recycling and reuse. But in the last few years, China has been tightening up what it was willing to import, wanting only plastic waste that is uncontaminated. In 2017, China announced that in the future it was banning the import of nonindustrial plastic waste--that is the plastic waste generated by households.

Amy L. Brooks, Shunli Wang, Jenna R. Jambeck look at some consequences in "The Chinese import ban and its impact on global plastic waste trade," published in Science Advances (June 20, 2018). Here's a figure showing patterns of exports and imports of plastic waste, in quantities and values, based on UN data. In theory, of course, the lines for imports and exports should match exactly, but the data is collected from different countries and errors of classification and inclusion do creep in. Still, the overall pattern of a dramatic rise, leveling off in the last few years as China imposed additional restrictions, is clear.
Brooks, Wang, and Jambeck summarize in this way:
"The rapid growth of the use and disposal of plastic materials has proved to be a challenge for solid waste management systems with impacts on our environment and ocean. While recycling and the circular economy have been touted as potential solutions, upward of half of the plastic waste intended for recycling has been exported to hundreds of countries around the world. China, which has imported a cumulative 45% of plastic waste since 1992, recently implemented a new policy banning the importation of most plastic waste, begging the question of where the plastic waste will go now. We use commodity trade data for mass and value, region, and income level to illustrate that higher-income countries in the Organization for Economic Cooperation have been exporting plastic waste (70% in 2016) to lower-income countries in the East Asia and Pacific for decades. An estimated 111 million  metric tons of plastic waste will be displaced with the new Chinese policy by 2030. As 89% of historical exports consist of polymer groups often used in single-use plastic food packaging (polyethylene, polypropylene, and polyethylene terephthalate), bold global ideas and actions for reducing quantities of nonrecyclable materials, redesigning products, and funding domestic plastic waste management are needed."
The pattern of high-income countries sending their recycling to lower- and middle income countries is common. The share of plastic waste going to China seems to actually be greater than the 45% mentioned above. The authors write: "China has imported 106 million MT of plastic waste, making up 45.1% of all cumulative imports. Collectively, China and Hong Kong have imported 72.4% of all plastic waste. However, Hong Kong acts as an entry port into China, with most of the plastic waste imported to Hong Kong (63%) going directly to China as an export in 2016." 

I don't have a solution here. The authors write: "Suggestions from the recycling industry demonstrate that, if no adjustments are made in solid waste management, and plastic waste management in particular, then much of the waste originally diverted from landfills by consumers paying for a recycling service will ultimately be landfilled." Other nations of east Asia don't have the capacity to absorb this flow of plastic waste, at least not right now. There doesn't seem to be much market for this type of plastic waste in the US or Europe, at least not right now. Substitutes for these plastics that either degrade or recycle more easily do not seem to be immediately available. But there is a mountain of plastic waste coming, so we will have a chance to see how the forces of supply, demand, and regulation deal with it. 

For short readable surveys of the study, I can recommend Ellen Airhart, "China Won't Solve the World's Plastics Problem Anymore," in Wired (June 20, 2018) and Jason Daley, "China’s Plastic Ban Will Flood Us With Trash," in Smithsonian (June 21, 2018).

Wednesday, July 11, 2018

When Growth of US Education Attainment Went Flat

Human capital in general, and educational background in particular, are one of the key ingredients for economic growth. But the US had a period of about 20 years, for those born through most of the 1950s and 1960s, where educational attainment barely budged.  Urvi Neelakantan and Jessie Romero provide an overview in "Slowing Growth in Educational Attainment," an Economic Brief written for the Federal Reserve Bank of Richmond (July 2018, EB18-07).

Here's a figure showing years of schooling for Americans going back to the 1870s. You can see the steady rise for both men and women up until the birth cohorts, when the educational gains for women slow down and those for men go flat.

In their essay, Neelakantan and Romero argue that this strengthens the case for improving K-12 education, and offer some thoughts. Here are a few related points I would emphasize.

1) Lots of factors affect productivity growth for an economy. But rapid US education growth starting back in the 19th century has been tied to later US economic growth. And it's probably not just a coincidence that when those born around 1950 were entering the workforce in the 1970s, there is a sustained slump in productivity that lasts about 20 years--into the early 1990s.

2) One reason for the rise in inequality of incomes that started in the late 1970s is that the demand for high-skilled workers was growing faster than the supply. For example, the wage gap between college-educates worker and workers with no more than a high school education increased substantially. As Neelakantan and Romero write: "This slowdown in skill acquisition, combined with growing demand for high-skill workers, contributed to a large increase in the `college premium' — the higher wages and earnings of college graduates relative to workers with only high school degrees." When educational attainment went flat, it also helped to create the conditions for US inequality to rise.

3) When a society has a period of a couple of decades where educational attainment doesn't rise, there's no way to go back later and "fix" it. The consequences like slower growth and higher inequality just march on through time. Similarly, the current generation of students--all of them, K-12, college and university--will be the next generation of US workers.