Crop Insurance Should Die, Yet It Lives On

As I post this, Congress is debating the farm bill renewal. In a rational world, it would eliminate or greatly scale back our absurd system of crop insurance, but it appears that once again, the program will live on.

Insurance plays an essential role in any market economy. By spreading losses among members of a group with similar exposure, insurance encourages people to take prudent risks while protecting them from financial ruin in case they are the unlucky ones. But not all insurance is equal. Sometimes, rather than representing the public interest, a public insurance program can come to represent a special interest subsidy in disguise. Crop insurance, the multi-billion-dollar government subsidy that lies at the heart of the farm bill that is now working its way through Congress, is a case in point.

Let’s take a look at the strange economics of crop insurance, and what can be done to fix it.

Crop losses are not an insurable risk

The problems of crop “insurance” begin with the fact that crop losses are not really an insurable risk. Crop insurance violates three of the most important rules that economists have developed to identify which risks are insurable and which are not.

First, to be insurable, losses should be fortuitous — the result of events that are outside the control of the insured party. Some crop losses (for example, damage from hail or tornados) fit this definition. They strike randomly. However, many other risks depend on the choice of farming practices, which, in turn, are influenced by the presence of insurance.

The tendency for people who are protected from loss by insurance to take greater risks is known as moral hazard. Commercial insurers guard against moral hazard as best they can by encouraging practices that reduce losses. For example, you can get a discount on your home insurance if you have a smoke alarm and security system. However, as the National Resources Defense Council explains in a recent report, rules of the federal crop insurance program encourage risky practices, such as planting crops on lands that are poorly suited for them. Rather than encouraging loss mitigation practices such as diversification and planting cover crops, crop insurance discourages them. When crops are planted that are likely to fail, resources are wasted and program costs soar.

Second, insurance is normally limited to situations in which people face a pure risk; that is, a risk of loss that is not offset by a hope of gain. For example, if I insure my house against fire, I either experience a fire, in which case I suffer a loss, or I do not, in which case I have neither a loss nor a gain. In contrast, speculative risks carry a chance of gain as well as loss. If I plant a field of corn, I may suffer a loss if the weather is bad or prices are low, but if the harvest and market conditions are good, I expect to make a profit. Insurers have traditionally been unwilling to cover speculative risks. Other financial mechanisms, such as futures- and options-markets, are more suited to that job.

Third, in order for a risk to be insurable, it must be possible to establish a premium that is affordable to the party seeking coverage, yet high enough to cover claims and the administrative expenses of the insurer. That is not possible for crop insurance. Due to the prevalence of moral hazard and the speculative nature of the risks, an actuarially fair premium for crop insurance would be unaffordable.
Together, moral hazard, speculative risks, and lack of affordability make crop “insurance” a sham. It is able to exist only because the government subsidizes more than 60 percent of the cost of the program. It is the subsidy, not the insurance that makes the program so popular with farmers.

Benefits are skewed toward the largest farms

People seem to like farmers. A recent industry-commissioned public opinion poll found that 86 percent of those questioned had a favorable opinion of farmers. More than 90 percent thought that it was important for the federal government to spend money to support them. Favorable responses varied hardly at all among Republicans, Democrats, and independents.

Why the popularity? One item in the poll gives us a clue: More than two-thirds of those polled disagreed with the statement, “Most of our nation’s food is grown by farming corporations that can easily afford to pay for their own crop insurance premiums without taxpayer money. It is the responsibility of any business to protect itself with insurance, and farmers should not get special treatment that the federal government simply cannot afford.” It seems likely that the disagreement reflects a widespread perception that the principal beneficiaries of crop insurance are small family-farms that would face a high risk of failure without subsidized crop insurance. However, the numbers, as reported for 2016 by the U.S. Department of Agriculture, do not support that view.

The first source of confusion arises from the notion that “family farms” and “corporate farms” are two different things. It is true that the USDA classifies 99 percent of all farms as “family farms.” However, it applies that term broadly to “any farm where the majority of the business is owned by the principal operator — the person most responsible for running the farm — and individuals related to the principal operator.” In practice, many such farms, including most large family-farms, are legally organized as limited-liability corporations or partnerships. Contrary to what farm lobbyists would like us to think, “corporate farms” and “family farms” are not separate categories.

According to USDA data, less than half of all farm output is produced by farms that are organized as traditional sole proprietorships. What’s more, most small farm-proprietorships are not the primary source of incomes for their owners. Some 18 percent of small family-farms are operated by people who report that they are retired, and another 42 percent report a nonfarm occupation as their main source of income.

As the data in the following table show, small family-farms — those with gross sales less than $350,000 — are not, on average, very profitable. Even when we exclude the 60 percent of small family-farms run by retirees or people with other principal jobs, the average farm income earned by farm households is less than $50,000. Some 80 percent of such farms have an average operating profit margin (OPM) — defined as revenue less operating costs — below 10 percent, putting them at high financial risk by USDA standards. Yet, despite facing the highest financial risk of any farms, they receive just 17 percent of all insurance payouts for crop losses.

Midsize farms are more prosperous, averaging some $120,000 of household income from farming. Only about 40 percent of them are in the high-risk category. Another 20 percent face moderate financial risks, with OPMs between 10 and 25 percent. These midsize farms, which account for 6 percent of all farms and 21 percent of all acreage, receive 32 percent of crop insurance payouts.

The 2.9 percent of family farms that are classified as large or very large exist in a different world. Large farms each have more than a million dollars in sales, and average $357,000 in annual household income. Those that are very large each have more than $5 million in sales and average $1.7 million in household income. More than half of large farms and more than 40 percent of very large farms fall in the low-risk category, with OPMs of 25 percent or more. Together, these high-income, low-risk family farms received 46 percent of all crop insurance payouts. If we add in non-family farms, which are also very large, on average, we find that just 4.3 percent of all farms receive more than half of all crop insurance payouts.

Finally, it is worth noting that crop insurance benefits are not only skewed toward the largest farms, but are also skewed toward just a few crops. In principle, insurance is available for over 100 different crops, but according to data from the Congressional Research Service, between federal crop insurance and other subsidies provided by the Commodity Credit Corporation, corn, wheat, and soybeans accounted for 77 percent of all support payments. Adding cotton, rice, and peanuts brought the total to 94 percent. Yet these six crops accounted for just 28 percent of total farm output in that year. The millions of farmers, large and small, who produce the milk, beef, eggs, tomatoes, lettuce, and other things you eat receive little or no support from crop insurance and the like.

If we add all these facts together, it becomes much harder to disagree with the proposition that “most of our nation’s food is grown by farming corporations that can easily afford to pay for their own crop insurance premiums without taxpayer money.” The data speak for themselves: The great majority of American farm output is produced by farmers who are, or should be, willing to accept “the responsibility of any business to protect itself with insurance” without special subsidies from the federal government.

What can be done?

The obvious thing would be to phase out the federal crop insurance program altogether, giving farmers reasonable time to adjust their mix of crops, farming practices, and financial arrangements. Short of that, critics of crop insurance offer a number of reforms that would make the program less costly to taxpayers and less disruptive to markets. Here just a few ideas from some of the most prominent liberal and conservative critics:

Eliminate yield exclusion. Premiums for crop insurance are adjusted according to each farm’s past yields. The idea is a reasonable one — farmers that have a record of consistently good yields and less frequent crop failures should pay lower rates for insurance than those with consistently poor yields and high losses. However, in calculating average yields, the government allows farmers to exclude as many as 15 low-yielding years from the average. The exclusions exaggerate past performance, artificially lower premiums, encourage farmers to plant risky crops, and increase payouts. Recommendation: Base premiums on true average-yields without exclusions. (Natural Resources Defense Council)

Stricter caps for subsidies. The 2014 farm bill, which the new version will replace, attempted to place caps on the subsidies that could go to the largest farms. However, it left many loopholes. One of the largest loopholes was the ability of large farms to form “pass-through entities” that allowed them to split total farm income among family members. Far from fixing this, the 2018 bill would widen the loophole by adding nephews, nieces, and cousins to the siblings and adult children who can already share farm income.

Recommendation: Make the caps realistic. (Environmental Working Group)
As the small-farmer advocacy group Center for Rural Affairs puts it,

Unlimited crop-insurance-premium subsidies are a loophole that allows the largest farmers to reap the greatest benefits from government subsidy. That is why we support capping crop-insurance-premium subsidies at $50,000. We believe that crop insurance should be a safety net, not a government subsidy to finance unlimited expansion.

Limit protection to yields, not revenue. If the government is to offer any protection against farming risk, it should focus on risks to crop yields from droughts, tornadoes, pests, and other natural hazards, as opposed to risks to the profitability of a given yield arising from market conditions. That would mean eliminating programs like Agricultural Risk Coverage, which protects farms from even “shallow” losses in revenue. “Any myth that commodity programs are supposed to be a safety net as opposed to an income guarantee gets quickly dispelled by this program.” Recommendation: Limit coverage to loss of crop yield. (Heritage Foundation)

Eliminate the Harvest-Price Option. Under standard insurance, a farmer would be compensated for crop losses at the price prevailing at the time of planting. Presumably, if that price had not been high enough to offer a profit, the farmer would not have planted, or would have planted something else. Under the Harvest-Price Option included in subsidized policies under the current farm bill, compensation is based on the higher of the price at time of planting or the price at the time of harvest. That is not only costly to the government, but it duplicates protection that could be obtained through futures- and options-markets without government assistance. Recommendation: Eliminate the Harvest-Price Option. (American Enterprise Institute)

The Bottom Line

Crop “insurance” is not true insurance. It is sham insurance that could not exist without government subsidies. Although the iconic small family-farm is often invoked in the program’s defense, its primary beneficiaries are, in practice, the largest and wealthiest farms — huge limited-liability corporations and other tax-advantaged pass-through entities that are “family farms” only in name. Outside the favored sectors of corn, wheat, and soybeans, the vast majority of American farmers and ranchers manage the risks of farming without subsidized insurance. As passed out of committee, the 2018 farm bill is no better than the old one, but there is still a chance to fix it while it is open for debate on the floor. Seize that chance!

Previously posted at NiskanenCenter.org

Why We Should be Skeptical About Guaranteed Jobs

The idea of a national job guarantee (JG) is about to go mainstream. The concept is far from new, but for the first time in decades, it is being endorsed by politicians with national stature. Sen. Bernie Sanders has promised to submit a legislative proposal. Other Democratic presidential hopefuls are showing interest. Academics, including Levy Economic Institute’s L. Randall Wray and Pavlina Tcherneva, have provided detailed blueprints for a national JG program.

No one denies that it would be nice if everyone who wanted to work could find a job, but before we start to beat the drum for a full-bore national job guarantee, we need a reality check. In a recent post, my  Niskanen Center colleague  Samuel Hammond outlined three reasons to be skeptical:

  1.     The private sector is better at allocating labor than public bureaucracies.
  2.     A JG program would be too easily politicized.
  3.     Other active labor-market policies, including wage subsidies, would work better than a JG.

These are valid points. Let me add three more reasons to be cautious about a national job guarantee.

4. Don’t exaggerate the pool of eligible candidates


As of April 2018, some 6 million people were officially unemployed, that is, counted as not working but actively looking for work. However, not all of those would be candidates for public-service jobs. Both in good times and bad, many of the unemployed are merely on temporary layoff or engaged in short spells of unemployment between jobs. At present, 33 percent of unemployed workers have been out of work for 5 weeks or less and another 31 percent for 5 to 14 weeks. Even in a bad year like 2010, nearly 40 percent of the unemployed were out of work for 14 weeks or less. Providing short-term in-and-out jobs for the temporarily unemployed is not the purpose of a JG. Even if offered such jobs, most of the short-term unemployed would probably prefer to keep looking for something more suited to their skills and interests.

But the officially unemployed are not the only targets of a job guarantee. Advocates claim there is a huge pool of other potential candidates who, for one reason or another, are not actively participating in the labor force, but might be drawn into it. As evidence, they point out that the employed share of the population has been falling since the beginning of the century, after a 50-year upward trend. As the following chart shows, that pattern holds both for the adult population as a whole and for prime-aged workers.

Yet the pool of potential workers is not really as large as it seems. Of the 95 million adults who were not in the labor force as of April 2018, only 5 million said they wanted a job now. Many of the rest were retired, in school, or medically incapacitated. Others were financially secure due to savings or earnings of other family members and preferred to spend their time  on activities other than paid work.

The 5 million who were out of the labor force but said they did want a job cited various reasons for not working, many of which would make it hard to draw them into public-service jobs. Based on annual data for 2017, 20 percent of them reported that school, family responsibilities, ill health, lack of transportation or other reasons kept them from looking for jobs. Only 9 percent reported that their reasons for not job-hunting were discouragement about the availability of work, fear that private employers would discriminate against them, or similar reasons. In some cases, these barriers to work could be overcome by a JG program that offered extensive support services, but it would be unrealistic to think that all 5 million could easily be slotted into public-service jobs.

Realistically, a JG program might absorb half of the long-term unemployed and half of those who want a job but are not currently in the labor force — fewer than 4 million people, of whom 3 million or so might come from the prime-age group. That would raise the share of the prime-age population who are employed to about 62 percent from its current level of 60.3, leaving it still well below peak rates of the past.

Some JG proponents envision a much larger program. Wray’s version aims for as many as 15 million public-service jobs. Except during periods of deep recession, however, a JG program could achieve that size only by offering wage and benefit premiums high enough to draw in millions of workers from the private sector. Doing so would do nothing to raise the employment-population ratio.

5. Don’t underestimate administrative costs


Critics of job guarantees worry about potentially high administrative costs. JG proponent Wray brushes that problem aside, urging that “federal spending should not subsidize administrative expenses.” However, the experience of other federal programs suggests that skimping on administrative support would sharply reduce the chances that a JG program would reach its goals.

For example, consider programs that attempt to increase employment by means of work requirements. Work requirements use a stick to move people into the labor force, rather than the carrot used by JG, but despite that difference, their target demographic is the same:  people who are able to work but are currently jobless.

The welfare reforms of the 1990s remain the largest and best-studied example of the effects of work requirements. At the time they were implemented, researchers conducted rigorous controlled experiments to measure how variations in program design influenced effectiveness. Each experiment compared the employment experience of a group of welfare recipients subject to work requirements to that of a group who received the same welfare benefits without the work requirements. Findings were published in 2001 in a document called National Evaluation of Welfare-to-Work Strategies (NEWWS). (For a summary, see my article for Milken Institute Review earlier this year.)

The experiments found that work requirements had a surprisingly small effect on employment. In one of the most successful trials, in Portland, OR, the number of welfare recipients who worked during at least one calendar quarter of the 20 quarters of the experiment rose to 85.7 in the group subject to work requirements compared to 81.5 in the control group. In the least successful experiment, in Oklahoma City, the number who worked in the group facing work requirements was actually lower than in the control group. In five of eleven experiments, the difference was statistically insignificant.

One key take-away from the experiments of the 1990s, and from more recent experience with work requirements for food stamps, is that there are no bright lines between “unwilling to work,” “willing and able to work but not working,” and “unable to work.” Many willing welfare recipients face practical barriers to work, such as child care or lack of transportation. Other barriers to work include physical or mental health problems that fall short of full disability; emotional issues; criminal records; substance abuse; and low skills.

With or without work requirements, people with such problems tend to move in and out of work frequently, even when jobs are available. They do not move through welfare-to-work programs in a simple, linear fashion, from unemployed to trainee to permanent job holder. Frequent failures and backsliding undermine the best-intentioned policies.

Another thing the experiments of the 1990s made clear was that success requires adequately funded administrative support and well-trained staff. The best results were obtained when case workers and other administrators did more than simply monitor eligibility, participation, and compliance with program rules. Instead, they needed to be proactive and to stress self-sufficiency. However, such efforts come at a cost. Results were disappointing in cities like Oklahoma City and Detroit where administrative support was underfunded.

Today’s nonworking population faces the same spectrum of barriers to employment that prevailed in the 1990s. Any realistic JG program would have to deliver the support people need to arrange child care or elder care, transportation, and the like. Such a program would need to create jobs where people can contribute productively despite difficulty walking, back pain, poor vision, headaches, and other health problems. Its administrators would need to work one-on-one with job candidates who suffer from anxiety, depression, substance abuse, violent domestic situations, and other problems that make it hard even to show up for work each day, let alone perform on the job. None of these barriers is insurmountable, but they cannot be overcome on the cheap.

6. Beware of illusory additions to GDP

Optimists contend that a JG program would provide a huge boost to the economy. For example, Wray estimates that a national job guarantee with 15 million participants would add $560 billion dollars to GDP, about 3 percent, with little impact on inflation. But, even if he is right about the numbers, how much of the measured addition to GDP would represent a real increase in the economy’s output of goods and services?

It would be easier to know if, for example, the JG workers were put to work sewing shirts. Their output would then be measured by multiplying the number of shirts times the price at which they were sold. If that came up to $560 billion, you would have added an honest increment to GDP. If customers didn’t like the shirts, or too many were produced, their price would fall and the factory’s contribution to GDP would decrease accordingly.

However, JG advocates don’t have clothing factories in mind. Instead, almost all JG participants would be given public-service jobs. For example, in her working paper for the Levy Economics Institute, Tcherneva describes what she calls a “National Care Act.” The jobs it created would focus on three areas:

  • Care for the environment, including soil-erosion- and flood-control; environmental surveys; species monitoring; park maintenance; removal of invasive species; support for local fisheries; community supported agriculture (CSA) farms; rooftop gardens; tree planting; fire- and other disaster-prevention measures; weatherization of homes; and composting.
  • Care for the community, including cleanup of vacant properties; reclamation of materials; restoration of public spaces; establishment of school gardens; solar arrays; tool-lending libraries; community theaters; restoration of historical sites; organization of carpooling programs; recycling; water-collection initiatives; food waste programs; and oral history projects.
  • Care for people, including elderly care; after-school programs; programs for children, new mothers, at-risk youth, veterans, former inmates, and people with disabilities; organizing after-school activities in schools or local libraries; shadowing teachers, coaches, hospice workers and librarians to learn new skills and assist them in their duties; organizing nutrition surveys in schools; and coordinating health-awareness programs for young mothers.

All of these are worthy activities. Millions of public employees and volunteers engage in them every day. But how can we measure their contribution to GDP when we can’t put dollar value on removal of invasive species or accurately measure the quantity and quality of output of an oral history project?

The answer is that national-income accountants don’t even try to measure the output of public-service workers. Instead, they measure the cost of their contributions to GDP by adding up the costs of the labor and other inputs that go into producing the services. If city workers are assigned to composting food waste from school cafeterias, their contribution to GDP is measured by the salary they are paid. It does not matter whether the compost turns out to be useless, or whether it turns out to be black gold that enormously boosts the output of lettuce in a community vegetable garden.

Following standard accounting procedures, then, paying each of 15 million people $30,000 per year to perform tasks from Tcherneva’s list would produce a $450 billion bump to GDP regardless of the value of the services the workers produced.

Accounting by cost rather than output greatly increases the possibility of overstatement of a JG program’s contribution to GDP. To see why, let’s return to the example of the shirt factory. Suppose a private shirt factory has 10 workers who produce $300,000 worth of shirts a year. The owners then decide to hire another worker, at $30,000 a year, to run a day care center for workers’ children. If lower absenteeism and better employee morale increase productivity, extra shirts are produced, and the factory’s contribution to GDP goes up. If the day care experiment fails, and no extra shirts are produced, the factory’s contribution to GDP does not change and the cost of the day care project comes out of profits.

In contrast, if a community arts group employs 10 JG workers at $30,000 a year, its contribution to GDP would be measured as $300,000. If the arts group adds an eleventh worker to take care of the children of the first 10, its measured contribution to GDP goes up by $30,000, even if the amount of art the group produces is unchanged. If it adds a twelfth employee for administrative duties, its contribution to GDP goes up again, whether or not it produces more or better art.

The likelihood of exaggeration is even greater if we consider that some of the services that would be produced by a JG program are already being produced, but not for pay, by the same people that the program would employ. For example, we noted earlier that many of those who are out of the labor force, but want a job, are unavailable for work because they are caregivers for children or other family members. Imagine two parents, in two separate households. Each parent has been staying home to care for two young children. If a JG program hires one of them at $30,000 a year to work in a community tool-lending library, and hires the other as a day care worker to care for all four children, the measured amount of new GDP produced will be $60,000. However, the only new services produced are those of the tool-lending library. The day care services were already being produced by the parents themselves, but not for pay.

Child care services are not the only example. The United States has a vast volunteer sector. A Bureau of Labor Statistics report on Volunteering in America indicates that more than 62 million people aged 16 and older participated in volunteer work in 2015. Of those, 5.9 percent each spent more than 500 hours a year volunteering — a vast amount of work. Much of that was spent on exactly the kind jobs that would be targeted by JG programs. In some cases, people who are now out of the labor force and volunteer as wildlife monitors would happily take paid jobs as JG wildlife monitors. In other cases, people who are already employed but volunteer on weekends would find they are no longer needed to pick up trash in the local park, since a JG worker is already doing the job. They would go jogging, instead. Any way you look at it, at least part of the work done by JG participants would displace something already being done by volunteers. The program’s addition to GDP would, to that extent, be illusory.

Caution is the bottom line

I do not mean to be completely dismissive about public-service jobs. Volunteers are great, but they are not always enough to monitor water quality or maintain historical sites. The idea that local governments should keep a list of “shovel-ready” jobs in reserve to serve as fiscal stimulus in hard times is nothing new. But a full-scale national job guarantee — one that employs tens of millions during downturns and does not go to zero even in the best of times — is something else again.

The bottom line is that a national job guarantee sounds great until you actually think about it. When you do the numbers, you find out that many of the people you want most to help are not good candidates for public-service jobs. When you look at past efforts, you see that welfare-to-work programs of any kind are unlikely to succeed without expensive investments in staffing and administration. And when you do the accounting, you find that many of the supposed benefits of putting millions of people to work in new public-service jobs are illusory.

In short, we should think twice, three times, or more before we let ourselves get carried away with enthusiasm for a national job guarantee. Meanwhile, we are far from exhausting the alternatives to JG. Based on international best practices, we could do more with job placement, training, and other active labor-market policies. We could reform and consolidate existing anti-poverty programs in order to reduce disincentives to work. We could make tax policy more work-friendly by expanding earned income tax credits and easing regressive payroll taxes. And we could recognize that cash assistance makes more sense than make-work jobs for many who need help most.

Reposted with permission from NiskanenCenter.org

How Framing Affects Attitudes Toward the Social Safety Net

An article by Caitlin Dewey in the Washington Post led me to an interesting new piece of research by Rachel Wetts of U.C. Berkeley and Rob Willer of Stanford (unrestricted draft version here). The research, which is relevant for anyone in the area of public policy advocacy, shows how strongly framing—and the framing of charts, in particular—can affect attitudes toward pubic policies.

The specific issue that the authors investigate is the influence of information related to race on attitudes toward pubic assistance. As they put it in their abstract,

We argue that when whites perceive threats to their relative advantage in the racial status hierarchy, their resentment of minorities increases. This increased resentment in turn leads whites to withdraw support for welfare programs when they perceive these programs to primarily benefit minorities.

In one of several experiments, Wetts and Willer presented a questionnaire to a sample of participants recruited from Amazon Mechanical Turk. Before answering questions about welfare policy, participants were shown one of two variants of a chart depicting U.S population trends by race and ethnicity. Both charts were based on the same data from the Census Bureau, but Chart A shows a short time period, over which the trends appear relatively weak, while Chart B shows a longer time period, over which the trends appear stronger. Chart B also adds a line for “all nonwhite” which more dramatically shows that the white population is trending toward minority status.

After viewing the information, participants were presented with scenarios designed to reveal attitudes toward welfare. In one scenario,

participants were told to imagine that they were on a Congressional committee charged with cutting $500 million from the federal budget.  They were given a list of nine spending areas including “Temporary Assistance for Needy Families (Welfare)” and asked to indicate how much they would cut from each area.

White participants said they would cut TANF by 28 percent if they had viewed Chart A and cut by 51 percent if they had viewed Chart B—a statistically significant difference.

Although there were too few minority participants to produce a statistically significant results, their answers provide some food for thought. The nonwhites who had viewed Chart A said they would cut TANF by 53 percent, while those who had viewed Chart B said they would cut by 57 percent. About three-quarters of the recipients of cash assistance programs are nonwhite. Does TANF look better from a distance than from up close, or is this a statistical fluke?

And while we are on the subject of framing—would the results have been significantly different if the parenthetical word “Welfare,” which is perceived pejoratively by many people, had not been inserted after “Temporary Assistance for Needy Families,” which implies that the people who get it actually need it?

A previous version of this post appeared on Niskanen Notes

Inflation Increasingly Erodes Wage Gains Even as Unemployment Falls

According to the latest Employment Situation Summary from the Bureau of Labor Statistics, average nominal hourly earnings for all employees on private nonfarm payrolls rose at a compound annual rate of 3.6 percent in May, 2016. That rate is well above the 2.6 percent average for the preceding 12 months, and also above the average CPI inflation rate of 2.5 percent for the same period. Monthly observations are shown by the dotted lines in the chart, while the solid lines show 12-month moving averages.

Monthly data include a lot of statistical noise and are subject to revisions, so policymakers will be paying more attention to trends than to individual data points. The trend lines show that over the past three years, CPI inflation has accelerated more rapidly than has the rate of nominal wage gains. CPI data for May will not be released until June 12, but by April, the last month for which full data are available, the 12-month moving average for wages exceeded that for inflation by just 0.1 percent (2.6 percent vs. 2.5 percent).

The trends of the moving averages contain both not-so-good news and better news. For workers, the news is not so good, inasmuch as the wage gains for May, which are hopeful taken in isolation, may turn out to be a statistical fluke. From a macroeconomic point of view, however, the news is better. With CPI and wage trends still holding at or close to 2.5 percent, there is little sign of overheating yet.

Reposted from Niskanen Notes

Could We Afford Universal Catastrophic Health Care Coverage?

Universal catastrophic coverage (UCC) is a health care plan that aims to protect all Americans against financially ruinous medical expenses, while preserving the principle that those who can afford it should contribute toward the cost of their own care. It offers a potentially attractive compromise between the current system, which leaves millions of people uninsured or underinsured, and more expensive, “first dollar” proposals that would cover all health care costs for everyone.

Skeptics often ask whether such a plan is affordable. The short answer is “Yes,” but I would prefer to frame the question differently. Rather than asking how much any given health care plan would cost, it is more useful to ask, “What is the best plan we could design for what we are politically willing to spend?” If we set that amount somewhere close to what the government now spends on health care, universal catastrophic coverage looks rather good. This post explains why.

The parameters of universal catastrophic coverage

First, we need to review the basic parameters that define any UCC plan. The simplest version of UCC would have just two parameters, a low-income threshold and, for those above the threshold, a deductible that varies with household income.

The low-income threshold is a level of income below which any medical expenditures at all threaten serious financial distress. Under UCC, people whose incomes are at or below the threshold would need to be covered in full. The dollar value of the low-income threshold would vary with family size. Some versions of UCC simply set it equal to the federal poverty level (FPL), which, as of 2018, stands (in round numbers) at $12,000 for an individual and $25,000 for a family of four.

Given a low-income threshold, the deductible can be set as a percentage of eligible income, that is, the amount by which household income exceeds the threshold. For example, if the deductible is 10 percent of eligible income and the low-income threshold is the FPL, a family of four with total income of $50,000 would have eligible income of $25,000 and a deductible of $2,500. A family with total income of $100,000, which would put them at the upper limit of eligibility for premium subsidies on the ACA exchanges, would have a deductible of $7,500 — a little less than their deductible under a typical ACA silver plan. A family with total income of $1 million would have a deductible of $97,500, and so on. For administrative purposes, household income could be defined as adjusted gross income reported on the previous year’s tax return, perhaps with some form of averaging for those with highly variable incomes.

Some versions of UCC include copayments as well as deductibles. For example, we could set the low-income threshold at the FPL and the deductible at 5 percent of eligible income, and then add a copay of 20 percent on health care expenses from 5 percent to 30 percent of eligible income. A family of four with eligible income of $50,000 would then face a deductible of $2,500 and a 20 percent copay on expenditures from $2,500 to $15,000. Its out-of-pocket maximum, consisting of the $2,500 deductible plus $2,500 in copays, would be the same as under a UCC plan with a 10 percent deductible and no copay.

Some UCC plans also include a premium, which would be paid whether or not the family had any health care expenses in a given year. Typically, the premium is zero for people below the low-income threshold and follows a sliding scale for people with higher incomes. A family’s maximum total contribution to its health care expenses would then be the sum of its deductible, its copays, and its premium.

Finally, many versions of UCC specify a package of preventive services that would be exempted from deductibles and copays. If the package were limited to highly cost-effective services, it could reduce total out-of-pocket costs to insured households with little impact on the cost of the program to the government budget, since the cost of vaccinations, screenings, and the like would be largely, or even wholly, recouped through reductions in the cost of treatment of preventable conditions.

Trade-offs among the parameters

Each of the five parameters of a UCC plan — the low-income threshold, the deductible, the copay, the premium, and the preventive package — serves a particular objective. Tightening or loosening any one parameter lowers or raises the cost of the plan. By trading off the tightening of one parameter against the loosening of another, it is possible to emphasize one objective and de-emphasize others while holding the overall cost of the plan constant. With that in mind, let’s look at each of the parameters in turn.

The low-income threshold. Other things being equal, a higher threshold increases the budgetary cost of a UCC plan. That, in turn, creates pressure to control costs elsewhere, for example, by imposing higher deductibles and copays on middle- and upper-income households. For budgetary purposes, then, it would make sense to set the threshold at the lowest level that is consistent with the basic purpose of UCC, which is to ensure that no one is exposed to financially ruinous medical expenses.

For simplicity, many treatments of UCC set the low-income threshold equal to the FPL. However, the FPL is not especially well-suited for this purpose. Historically, the FPL, which dates from the 1950s, was based on the assumption that a family with income equal to three times the minimum needed for food would be able to provide adequate nutrition with enough left over for other necessities. At that time, however, health care prices were much lower, relative to those of food, shelter, and clothing, than they are now. The higher cost of health care services in today’s economy would argue for setting the low-income threshold for a UCC plan somewhat higher than the FPL, perhaps at the 138-percent-of-FPL that is used to determine Medicaid eligibility in states that chose to expand the program under the ACA.

Deductibles. Under UCC, deductibles serve a threefold purpose. First, they limit the budgetary cost of the plan in comparison with alternatives that would provide first-dollar coverage for everyone. Second, they uphold the principle that people who can afford to do so should assume responsibility for their own non-catastrophic health care expenses. And third, they give health care consumers “skin in the game,” which, according to advocates of market-based health care, should make them better shoppers who carefully compare the cost and quality of services and budget prudently to meet foreseeable medical expenses.

These three considerations lie behind the growing popularity, not just of UCC proposals, but of high-deductible health insurance of all kinds. Unfortunately, though, there is a downside to high deductibles: In practice, people are not always as prudent in their health care choices as they might be.

For example, a recent study by Jeffrey T. Kullgren and colleagues looked at the behavior of a large sample of health care consumers with high-deductible policies. Their findings, summarized in the following table, indicated that only 40 percent of respondents had saved any money toward the cost of future services, even though some 58 percent of respondents had set up health savings accounts. Only a small minority of those surveyed attempted to compare price or quality of services. And, although a quarter of respondents discussed price with providers, only six percent actually tried to negotiate prices.

Attempts by consumers to learn more about price and quality or to negotiate prices, when they were made, produced mixed results. Among consumers who engaged in at least one of these behaviors, 26 percent decided to put off the service until they could afford it, 10 percent decided that the service was not worth the cost, and 22 percent managed to obtain the service at a lower price. Only the last of these outcomes is unambiguously good. Postponing or forgoing a service because of its price is a good outcome if it deters the consumer from using a service that is not cost-effective, but not if the service has benefits that exceed its costs. Other research has found that consumers are often poor judges of cost-effectiveness and forgo both appropriate and inappropriate services for reasons of affordability.

Copayments. In principle, copays serve the same three purposes in a UCC plan as do deductibles. However, the two are not fully interchangeable, since copays have special advantages and disadvantages of their own.

On the positive side, copays extend the range of expenses over which consumers have an incentive to shop for the best health care values. Compare, for example, two versions of UCC, each of which imposes the same $5,000 maximum out-of-pocket cost on individuals with eligible income of $50,000. Plan A has a straight $5,000 deductible and no copays, while Plan B has a $2,500 deductible and a 20 percent copay on expenses from $2,501 to $15,000.

If I have Plan A, then once my medical expenses pass $5,000, I pay nothing for further services. According to the “skin in the game” theory, I will have no incentive to avoid services that are not cost-effective or to look for the lowest prices for those that are. In contrast, under Plan B, I would have an incentive to be a good health care shopper until the cost of my care rose all the way to $15,000, even though my own maximum exposure would be no higher than under Plan A.

However, copays also have their drawbacks. One is that for a given out-of-pocket maximum, a UCC plan with copays would be more expensive for the government to operate. To see why, suppose that our Plans A and B were applied to 100 individuals, whose health care spending was distributed according to the pattern shown in the following chart. As in the actual distribution of health care expenses for the U.S. population, this stylized distribution of spending is highly skewed. The average is $10,000 per person, but a third of the population spend less than $100 each, while the top 20 percent of individuals account for more than 80 percent of the total.

Those who spend less than $2,500 (60 percent of the population but only 3.5 percent of total spending) would not meet their deductibles under either Plan A or Plan B, and all those with spending of more than $15,000 would hit their out-of-pocket maximums. For both of those groups, out-of-pocket expenses would equal to total expenses. However, people with expenditures between $2,500 and $15,000 (22 percent of the population and 25.9 percent of all spending) would have lower out-of-pocket costs under Plan B, where they are in the copay range, than under Plan A.

When you do the numbers, it turns out that the budgetary cost of Plan B would be 7 percent greater than that of Plan A. To bring the government’s cost for the plan with copays down to that of the plan without them, we would have to allow a larger out-of-pocket maximum for Plan B than for Plan A, either by raising the deductible for Plan B, or extending the copay range, or both.

True, this comparison assumes that people’s actual health care expenses remain the same under Plan A as under Plan B. That would not be the case if, as intended, both deductibles and copays induce consumers to shop for better prices, or to forgo services that they perceive as unaffordable or not cost-effective. A more complete comparison would have to take into account elasticities of demand for services with respect to deductibles and copays. It is not clear a priori which of these two plans would have the stronger effect on health care spending or which would produce better health outcomes. Case studies of the effects of specific policy changes result in disparate findings.

A full comparison should also consider the relative administrative burdens of copays and deductibles. In a plan with deductibles only, any given service is either billed entirely to the patient or entirely to the insurer. With copays, there is an extended range of expenditure over which bills must be split, with the patient and the insurer each paying part. Based on the distribution of expenditures in the chart, that range covers more than a quarter of all health care spending. Within it, the extra paperwork and collection expense raises billing-and insurance-related-costs for doctors, insurers, and patients.

Preventive care. As explained in an earlier post, the risk that people will forgo cost-effective preventive services can be reduced by exempting those services from deductibles and copays. In some cases, such as childhood immunizations, doing so will actually decrease total national health care costs. Even when it does not, there are many cases in which the payoff to preventive care is large compared to its cost. As a starting point, a UCC could simply adopt the package of preventive services currently offered free of charge under the ACA. However, as Mark Pauly and colleagues explain in an article in Health Affairs, there is a lot of room for improvement in the current system for measuring cost-effectiveness.

Premiums. The final major parameter of a UCC plan is the premium, if any, to be paid by policyholders. Like deductibles and copays, premiums are motivated by the idea that people should pay a fair share of health care costs when they can afford to do so. Within a given cost to the government budget, adding premiums to a UCC plan would make it possible to reduce deductibles or copays. However, the distributional and incentive effects of premiums are quite different from those of deductibles and copays.

To be consistent with the goal of making coverage affordable to everyone, premiums, like deductibles and copays, would need to vary with income and to be waived for households below the low-income threshold. As a result, premiums would not have much effect on the distribution of health care costs according to income. Instead, the main effect of adding a premium to a UCC plan would be to redistribute health care costs toward people who are healthy and away from those who are ill within specific middle- and upper-income groups. If we view taxpayer-funded health care as a form of social insurance, that makes sense.

The downside is that introducing premiums might improve perceived fairness at the expense of incentives for more careful choice by health care consumers. A premium, once paid, is a sunk cost that has no effect on people’s decisions to visit their doctors or to wait to see if they feel better in the morning, nor do premiums provide any incentive to shop for the best price for a knee replacement rather than simply visiting the closest hospital.

Premiums also raise difficult issues regarding universality of coverage. With no premiums, there would be no problem with automatically enrolling everyone for UCC coverage. Some people might decline to use their policies even if they developed health problems, perhaps on religious grounds or perhaps because they are just too stubborn to spend anything on health care even when they need it. That would be up to them. But, introducing a premium requires deciding what to do about people who fail to pay it. Are you going to turn them away at the emergency room door? Are you going to let them play the free rider, and later pay up and join the system when they become ill? Are you going to seek court orders to garnish their wages?

If you kick out people who don’t pay their premiums, then coverage is no longer universal. If you don’t kick them out, you open the door to free riders. If you collect the premium by force of law, then it is not a premium, but a tax. And to the extent you are going to finance UCC with taxes, it is arguably better to do so out of general revenue than through a special health care tax that you euphemistically call a “premium.”

Estimating the cost of universal catastrophic coverage: A case study

So far, we have looked at the economics of UCC in the abstract. This section turns to the challenging issue of estimating the actual dollar cost of a specific UCC plan.

To date, the most ambitious attempt to do so is one by Jodi L. Liu of the RAND Corporation. Liu uses a detailed simulation model that includes population data, estimates of demand elasticities, and estimates of additional cost-saving measures, all drawn from reviews of the literature. She applies the model to two different national health care plans: a 2013 version of Sen. Bernie Sanders’ Medicare for All proposal and a UCC plan outlined in 2012 in National Affairs by Kip Hagopian and Dana Goldman.

For purposes of estimation, Liu sets the parameters of the catastrophic plan as follows: The low-income threshold is 100 percent of the FPL. The deductible is 10 percent of eligible income. Copays are 5 percent, subject to an out-of-pocket maximum of 14.5 percent of eligible income. Given these parameters, households thus hit their out-of-pocket maximum at the point where health care expenses reach 100 percent of eligible income. Liu also assumes a fixed charge, waived for incomes up to FPL and assessed on a sliding linear scale up to a maximum of $3,350 at 300 percent of the FPL. She calls this charge a “tax,” although Hagopian and Goldman themselves, writing in Forbes, call it a “premium.” Liu does not model the cost of a package of free preventive services, even though the original Hagopian-Goldman plan that she draws on recommends such a feature.

The basic version of the UCC plan that Liu considers leaves Medicaid and Medicare intact, and covers everyone who does not participate in either of those programs. As such, it completely replaces all employer-sponsored insurance. She also considers variants that preserve employer-sponsored plans as optional alternatives.

Liu estimates that for 2027, the basic UCC plan would reduce total national health care expenditures by $211 billion, or about 8.7 percent. She estimates that total federal expenditures on health care would increase by $648 billion compared with expenditures under the ACA for that year. Of that increase, $524 billion would be covered by revenue from the dedicated tax/premium. The remainder would be slightly more than offset by increased revenue from income and payroll taxes due to elimination of the deduction for employer-sponsored insurance. As a result, the net impact of the basic UCC plan on the national budget would be a saving of $40 billion.

For comparison, Liu estimates that Sanders-style first-dollar coverage would increase total national health care expenditures by 18 percent and federal health care expenditures by 60 percent. Most of the additional federal spending for the Sanders plan would come from new taxes.

Next, Liu estimates potential savings in administrative costs for insurers and providers, together with further savings from negotiation of better prices for prescription drugs, hospital services, and other provider services. The net effect, with the further cost savings, would be a reduction of total national health care expenditures by $767 billion dollars, or 35 percent.

Although some of the $767 billion of further savings in national health care expenditures would accrue to individuals, Liu estimates that $556 billion of those savings would accrue to the federal budget. Including the further cost savings, then, total impact of the basic UCC plan on the federal budget would be a net saving of $596 billion, rather than the $40 billion estimated for UCC without further cost saving. Note also that the federal share of further cost savings of $556 billion is  slightly greater than the estimated $524 billion of revenue that would be raised by the tax/premium feature of the basic UCC plan. Putting this all together, then, total federal expenditures on health care would be $72 billion less than under the ACA even if the tax/premium were dropped from the plan.

Liu’s estimates are carefully constructed and draw on the best available data. Nonetheless, they should be considered as illustrative, not as definitive. Further research might reach different conclusions regarding the responsiveness of health care consumers to system changes, the success of cost control efforts, and changes in tax revenues. Other investigators would doubtless want to explore the effects of changes in various UCC parameters, and to examine a broader UCC plan that replaced Medicaid and/or Medicare. Still, UCC proponents will find Liu’s estimates encouraging, since they are consistent with the expectation that a reasonable UCC plan could be implemented without new taxes or large increases federal health care spending.

Conclusion: Optimizing for health

This brings us back to our original question: Could we afford a national system of universal catastrophic health coverage? I think the answer is yes. Within the limits of what the government now spends on health care, it appears reasonable to suppose that we could afford a UCC plan with a low-income threshold at or a little above the federal poverty level; an out-of-pocket maximum (including deductibles and copays) in the range of 10 to 15 percent of eligible income; a package of exempt preventive services similar to that offered under the ACA; and no individual premium or new dedicated taxes.

That general outline of a UCC plan leaves many questions open. To list just a few, what combination of deductibles and copays would produce the best balance between incentives to shop carefully and the tendency of participants to forgo appropriate, cost-effective care? What kinds of reform could provide better information about health care prices and quality in order to facilitate consumer choice? What kinds of savings vehicles or supplementary insurance products might be useful to help middle- and upper-income consumers with out-of-pocket costs? What would constitute an appropriate package of exempt preventive services? What mix of public and private institutional mechanisms would minimize administrative costs?

The more we know about the answers to questions like these, the better we can optimize an affordable UCC plan to deliver superior health care outcomes.

Reposted from NiskanenCenter.com