Mask mandates save lives but are a political thornbush

COVID-19 is bigger than ever in Ohio, with the number of new daily cases reaching its highest points of the pandemic in July. As the disease continues to spread throughout the state, Gov. Mike DeWine has struggled to balance public health needs with a growing sentiment against public health measures. 

Before last week’s statewide mask mandate, the main strategy from the administration has leaned on a county-by-county approach that mandates masks in high-spread counties with less resistance to such measures and preserves local authority in areas of lower spread and more antipathy to mandates. This strategy led to a mask mandate for 60% of Ohioans and lighter regulation in places with less spread and less taste for mandates.

Whatever your personal view is on masks mandates, however, available evidence on their use during the COVID-19 pandemic seems to show they slow the spread of disease. Much of the news coverage of the use of masks has focused on very technical questions: What kind of mask needs to be worn to spread disease? How should a mask be worn? Are masks more protective for people wearing them or people around those who wear them?

Whatever the answers are to these questions, the available evidence seems to show us one thing: that mask mandates do slow the spread of disease. A study published in Health Affairs last month by two University of Iowa public health researchers is possibly the most rigorous and applicable study on COVID-19 mask mandates to date. The researchers compared the infection trajectory of 15 states that implemented mask mandates in April and May to the other 35 without mandates to see if mandates showed any measurable impact on infection rates.

The “natural experiment” of variable mask mandates suggests that mask mandates reduced infection rates. The researchers found that growth rates in COVID-19 cases declined by one percentage point in the first couple of weeks after a mandate compared to states without mandates and that case decline rose to two percentage points in the weeks after.

One or two percentage points of growth rate decline might not seem like a lot, but the researchers estimate that the mask mandates prevented at least 230,000-450,000 new infections through the end of May, implying they saved a combined 9,000-17,000 lives over that time period assuming the current 3.7% national COVID fatality rate.

While this study seems to suggest mask mandates can save lives, reduced growth rates might be the effect of other factors. For instance, states that have a larger cultural commitment to reducing the spread might be likely to both adhere to social distancing guidelines and pass mask mandates as well. While the study tries to control for these factors, only so much control can be made in a natural experiment such as this.

These problems present a sticky challenge for DeWine. On the one hand, the evidence available suggests that mask mandates slow spread of disease and save lives. On the other hand, Ohio is a hotbed for anti-mask sentiment and communities need to be on board with masks for mandates to work. DeWine has a difficult path to tread as he tries to convince people to take steps that will slow the spread of COVID-19 and save lives in the process.

This commentary first appeared in the Ohio Capital Journal.

On Balance: New Editor, Same Mission

What a year it has been. Between the unfolding of a global pandemic and nationwide protests around the topic of police brutality, 2020 has already been a jam-packed year for public policy, and we haven’t even made it into election season.

I was appointed the new editor of On Balance in January, taking over for our founding editor Fran Sussman. I had big shoes to fill so figured I’d get a few entries out before posting my own introductory blog post. Then the world turned upside-down in March. I came back from a vacation to see the SBCA board voting to cancel its March conference and then the next few months became about COVID. So forgive this late introduction: there were more important things to attend to.
 
To give you some background about myself, I have a public policy analysis practice based in Columbus, Ohio called Scioto Analysis. Along with doing work for clients, we frequently engage in social impact work focused on improving policy analysis at the state and local level. We have a publication called The Ohio Handbook of Cost-Benefit Analysis and last week we released a best-practices benefit-cost analysis on COVID school closures.
 
I first came across economic analysis of public policy as a lobbyist in the Nebraska state legislature, when I was amazed to see a bill that would have expanded contraceptive coverage unexpectedly gain support from a number of conservative legislators.  The support arose in reaction to a legislative fiscal office report showing the bill would save the state health and safety dollars in the long run. From that experience, I learned the lesson that rigorous analysis presented by a nonpartisan, credible agency can have an impact on the policy conversation.
 
I was also lucky enough to take a class in benefit-cost analysis in graduate school at the University of California, Berkeley, taught by current SBCA board member Dan Acland. It was in that class that I learned that benefit-cost analysis is less about finding that “magic number” of a net present value greater than zero and more about taking apart a policy problem and understanding the real consequences of public action. I am glad I have had the opportunity to stay in touch with Dan because he has consistently given me good advice and thought-provoking commentary on benefit-cost analysis and first introduced me to SBCA in graduate school.
 
At this  moment, benefit-cost analysis is as relevant as ever. In our blog so far this year, we have had posts on COVID-19 policy, response to the opioid crisis, public investment in scientific research, and even the value of a statistical dog life. Benefit-cost analysis forces us to take a step back and understand the economic mechanisms that undergird every policy decision. Whether the problem is environmental, health-related, or even social, public policy is largely about resource allocation and benefit-cost analysis gives us a framework for understanding how scarcity impacts public decision making.
 
I would love to see your words in this blog. Below are some examples of contributions you could make to On Balance.

• Have you published an article in a journal recently with a significant benefit-cost component? Publish an article tie-in summarizing it and putting your research in a broader context.
   

• Read a good book on benefit-cost analysis? Publish a book review with us that tells readers about the content and how it fits into the literature.
   

• Have a perspective on a current event informed by benefit-cost principles? Share it with us in a commentary.
   

• Have a take on how benefit-cost analysis can be better carried out? Submit a practical perspectives piece to us aimed at scholars and practitioners.

These are just a few examples of content you can submit to On Balance: if you have an idea for something creative, I am absolutely open to talking about it. We’re only as good as the content you provide us, so don’t be afraid to step up and submit something so your voice can be heard. As crazy as the world is right now, there is no shortage of need for evidence-based policy decisions. I look forward to seeing this blog continue to be a place for the advancement, exchange of ideas, and dissemination of research related to benefit-cost analysis, all based on submissions from scholars and practitioners.

This blog post first appeared in On Balance, the blog for the Society for Benefit-Cost Analysis. Scioto Analysis Principal Rob Moore is currently serving as the editor for On Balance.

How should we measure poverty in the United States?

By Rachel Hammond

In 2018, 12.9% of Ohioans lived below the poverty line according to the Official Poverty Measure.  Under the Supplemental Poverty Measure, 10.4% of Ohioans lived in poverty.  Why are these numbers different and what do they tell us about people’s abilities to meet their needs?

The Official Poverty Measure was first constructed in the sixties and has been reported annually since 1969.  The poverty line was defined as three times the cost of an economy food plan because at that time, most families spent approximately a third of their income on food.  This threshold has been adjusted for inflation each year since.  Over the past fifty years, spending patterns have changed.  Food purchases now account for only one eighth of a household’s spending and the cost of living varies dramatically across the country. 

Over time, people have recognized the changes in spending patterns and shortcomings of the Official Poverty Measure.  In 1995, Constance Citro and Robert Michaels of the National Academy of Sciences published Measuring Poverty: A New Approach.  In the report, they recommended a new official poverty measure for the country that adjusts for geographic differences in cost of living and includes a more complete assessment of a household’s resources.  The Census Bureau used these recommendations to create the Supplemental Poverty Measure, and sixteen years later in 2011, they began publishing Supplemental Poverty Measure data alongside Official Poverty Measure data.

The first change in the Supplemental Poverty Measure is the definition of the poverty threshold.  The Supplemental Poverty Measure sets the poverty threshold as the average of food, clothing, shelter, and utility expenditures at the 33rd-36th percentiles, multiplied by 1.2 for “a little extra.”  This extra 20% added on to the threshold accounts for other necessary expenditures such as household supplies and transportation not related to work.  Both measures adjust the threshold for family size.  Additionally, the Supplemental Poverty Measure adjusts the threshold geographically by state and metro/non-metro areas within states.  It also includes an adjustment for three housing statuses (own with mortgage, own without mortgage, or rent) since people who own their house have less of a resource burden than those who rent or have a mortgage.

The next major change is how household resources are defined.  When assessing a household’s resources, the Official Poverty Measure only counts cash income.  The Supplemental Poverty Measure, on the other hand, attempts to capture a more complete picture of resources by including the value of safety net features such as SNAP (formerly known as Food Stamps) benefits, housing subsidies, WIC (a supplemental nutritional program for mothers and their children), TANF (cash welfare), tax credits, and LIHEAP (heating and cooling subsidies).  The SPM also subtracts from resources medical out-of-pocket expenses, childcare expenses, commuting costs, and taxes paid.  

The Census Bureau constructs the Official Poverty Measure and Supplemental Poverty Measure using data from the Current Population Survey.  The Current Population Survey contains information on enough households that the Census Bureau can report an accurate measure of poverty at the state level but not at the county level.  Since the National Academy of Sciences report was published, New York City, Wisconsin, California, Virginia, and Oregon have published their own supplemental poverty measures using data from the American Community Survey, which has a larger sample size, to provide a much more detailed description of poverty in their states.  It is worth noting that New York City released their first report in 2008, three years before the Census Bureau released the Supplemental Poverty Measure.   

Using American Community Survey data, however, comes at a cost.  While it contains data on more households, each household is asked fewer questions than those included in the Current Population Survey.  For example, the Current Population Survey includes information on the number of people in a household who received SNAP benefits in the past year, the number of months they received benefits, and the monetary value of the benefits.  The American Community Survey, on the other hand, only indicates if someone in the household received SNAP benefits.  Given this, researchers must impute the value of the resources and expenses included in the Supplemental Poverty Measure using regression models.  

Scioto Analysis is currently working on developing a state specific supplemental poverty measure for Ohio with the goal of providing a more complete picture of what poverty looks like across the state.  The Ohio Poverty Measure will help policymakers understand if and how the safety net is reducing poverty in the state and geographic differences in poverty across the state. More accurate measurement of poverty is the first step towards understanding how to effectively alleviate it.

Scioto Analysis covered in Wall Street Journal

This week, Scioto Analysis’s cost-benefit analysis on school closures for COVID-19 was featured in the Wall Street Journal (article paywalled). Scioto Analysis’s research was highlighted in an opinion piece on virtual learning at Harvard University due to COVID-19. Below is an excerpt that covered Scioto’s analysis.

In the Pac-12, the Goldman School of Public Policy at the University of California, Berkeley is pointing to a recent cost-benefit analysis of school closures co-authored by its graduate Rob Moore. The study focuses on the possibility of closing K-12 schools in Ohio for four months this fall. According to Mr. Moore’s firm, Scioto Analysis, the shutdown would have a disproportionate impact on kids’ future earning potential:

Overall, the cost-benefit analysis found that total costs in lost wages outweigh benefits measured in risk of death reduction by a factor of 14 to 1. The paper also touches on distributional impacts since further school closings would amount to a relatively small cost exacted on a large number of school-age children in exchange for a large benefit for a small number of elderly residents.

“More than nine out of ten COVID deaths in Ohio are among people age sixty and up and we have yet to record a COVID death in Ohio among school-age children,” said Moore. “Meanwhile, the average student loses out on $12,000-27,000 in lifetime earnings by losing four months of schooling. School closings are in essence an intergenerational transfer.”

Overall, the analysts estimate that further school closings would exact $22-37 billion in net social costs to the state when balancing wage losses with risk of death reduction benefits.

Mr. Moore’s study was released June 22. The state of Ohio is currently reporting a cumulative total of two Covid deaths among people under the age of 20. The median age of those dying with the virus is 80.

The cost-benefit analysis on school closings was the second best-practice cost-benefit analysis in Ohio in over a decade. The previous one was Scioto Analysis’s 2019 study on the state earned income tax credit. To learn more about cost-benefit analysis, check out the Ohio Handbook of Cost-Benefit Analysis, a free resource available for download on the Scioto Analysis website.

State advisory system may finally bring COVID clarity

Last week, the Ohio Department of Health rolled out a public health advisory system aimed at providing residents of the state of Ohio with county-level information on the spread of COVID-19 in their communities.

Advisory systems like this are a key tool for state governments. As the state of Ohio has eased restrictions on movement and work from its stay-at-home order implemented in March, it has devolved control of public health from the state level to the individual level, hoping that individuals, families, and businesses will be able to make decisions that will slow spread of the virus while allowing people to work, socialize, dine, and shop as needed.

Individuals, families, and businesses can’t make decisions about when to socialize and shop responsibly without good information. The purpose of a public health advisory system is to allow individuals to assess the severity of local spread and make social, family, and work plans based off this information. Ideally, a public health advisory system not only provides people with information about spread, but also with guidance on how to react to local conditions.

While Ohio’s response to COVID-19 has been serious, it has also been overwhelmingly voluntary. Ohio is not fining residents for not wearing masks or punishing the homeless for not staying indoors: It has relied on a system of voluntary compliance that lead to pretty extreme reductions in movement during the height of Ohio’s stay-at-home order.

As Ohio has eased off public health restrictions, though, guidance has not been clear. The last iteration of Ohio’s “stay-at-home” order allowed retail businesses to open without allowing people to shop there, indicating the unraveling of the system of orders.

The guidance since the lifting of stay-at-home has been less than optimally clear. Public officials have gone back and forth on the importance of wearing masks. The definition of “congregating” has been bent to the point of weddings being allowed with hundreds of people. Maybe most concerning, though, is the confusion that has caused for well-meaning people. Opening restaurants, bars, and stores for business implies safety, but COVID is still spreading and people are still dying. How is someone to know what to do and not do under these circumstances?

The public health advisory system has the possibility to be that guide that people need right now. Luckily, DeWine has eased off his previous insistence on statewide guidance only, acknowledging that the threat of COVID-19 has a geographic component to it. This should pay dividends not only in helping slow activity in areas with worse outbreaks, but also in easing political tensions in low-threat areas with businesses and politicians who wants to encourage economic activity.

The system as it stands now, however, is still not quite where it needs to be as a tool for action. Risk levels are determined by a hodgepodge of community spread, clinical diagnosis, and health care system capacity indicators that provide an educated guess at the danger of engaging in activities in the community. Risk level guidance is vague, marrying standard messaging about 6-foot rules, hand washing, and mask wearing with suggestions to “decrease interactions” at medium and high threat levels, as if we have any standard level of interaction these days to use as a baseline.

Most frustratingly, the guidance skirts the issue of what types of activities to avoid, a question that has been central to past guidance but which would open the system for criticisms by interest groups representing the entertainment industry and other industries that have an interest in promoting high-risk events.

Despite these drawbacks, the Public Health Advisory System could be a big step forward for people trying to figure out how to balance safety with social interaction. Let’s hope it is used and that future iterations provide more clear guidance to users.

This commentary first appeared in the Ohio Capital Journal.

Don’t forget about the children when closing schools

This fall, 1.7 million Ohio children will return to school for the first time in nearly half a year. Coronavirus brought about a second, unexpected summertime for Ohio’s K-12 students, surprising parents across the country with a new role as full-time educators and forcing teachers across the state to become remote educators. In the meantime, Ohio’s students suffered.

The goal of school closings is to slow the spread of COVID-19. Schools are locations of high traffic and high contact in close quarters, so the potential for spread of infectious disease in these locations is high. This is why the influenza protocol has placed school closures high on the list for social distancing measures to slow the spread of disease.

COVID-19, however, has had an uneven impact on people of different ages in the state of Ohio. While not a single school-age person in the state of Ohio has died of COVID-19, 90% of Ohio’s COVID-19 deaths have been suffered by those age 60 and up.

This means school closures are only a moderately effective tool for reducing COVID-19 deaths. According to a recent study in The Lancet that surveyed 616 studies and conducted meta-analysis on 16 high-quality studies, school closures alone would only prevent about 2-4% of deaths caused by COVID-19. A recent analysis conducted by my firm found this means we should only expect school closures in the fall to save at most a couple hundred lives.

Saving lives is important. But it is also something that we pay heavily to do. A recent analysis by the Brookings Institute suggests that learning loss by students over the period of COVID-19 school closures could cost students tens of thousands of dollars in future lost wages. Using more conservative assumptions than the Brookings study, we find that further school closures could cost students more than $22 billion in future discounted earnings due to learning loss over that time period.

This implies that school closures act as an intergenerational transfer. By closing schools, we extend the lives of a couple hundred elderly from anywhere from a few months to a couple decades by taking out a loan to be paid for with lost future earnings of children. Thus, a few mostly elderly people receive very large benefits paid for with relatively smaller costs (on the order of $18,000 or so) on a per-person basis for 1.7 million children.

When using standard valuation techniques, the lost future wages for today’s children, even when discounting to factor in present bias, vastly outweigh the present risk of death reduction brought about by school closures. This isn’t to say school closures are bad policy. School closures could be favored by policymakers who heavily prefer current lives being saved to future benefits for today’s children.

School closures could also be justified by a policymaker who wants to spare a small number of people of a large cost by spreading that cost (albeit inefficiently) across a large number of other people. When push comes to shove, however, the reality is that school closures for COVID-19 are strategy to reduce current risk of death for mostly elderly people financed with the future wages of today’s children.

This commentary first appeared in the Ohio Capital Journal.

Scioto Analysis Releases Cost-Benefit Analysis on COVID School Closures

Scioto Analysis released a cost-benefit analysis on school closures for COVID-19 Monday morning. Co-authors Rob Moore and Noah Stein project that an additional four-month closing of schools in the fall of 2020 would likely save 100-210 lives but would come at significant cost to K-12 students in the form of future earnings.

“We know from the literature on ‘summer slide’ that months off of school hurt students’ future labor market earnings potential,” said Stein. “On a statewide scale, this adds up to tens of billions of dollars of discounted future wage losses for Ohio’s 1.7 million K-12 students.”

Overall, the cost-benefit analysis found that total costs in lost wages outweigh benefits measured in risk of death reduction by a factor of 14 to 1. The paper also touches on distributional impacts since further school closings would amount to a relatively small cost exacted on a large number of school-age children in exchange for a large benefit for a small number of elderly residents.

“More than nine out of ten COVID deaths in Ohio are among people age sixty and up and we have yet to record a COVID death in Ohio among school-age children,” said Moore. “Meanwhile, the average student loses out on $12,000-27,000 in lifetime earnings by losing four months of schooling. School closings are in essence an intergenerational transfer.”

Overall, the analysts estimate that further school closings would exact $22-37 billion in net social costs to the state when balancing wage losses with risk of death reduction benefits.

This is the second best-practices cost-benefit analysis conducted on a state policy in the state of Ohio in the past decade. The first, Ohio Earned Income Tax Credit Refundability: A Cost-Benefit Analysis, was released by Scioto Analysis in August 2019.

Cap-And-Trade for Carbon Emissions

On Monday, Scioto Analysis released a brief on what a cap-and-trade system for carbon emissions would do for the state economy.

“Cap-and-trade is an efficient way to abate carbon emissions,” said Scioto Analysis Principal Rob Moore. “By allowing firms that can efficiently reduce emissions to sell emissions allowances and those that cannot to incur costs, a cap-and-trade system is able to promote economic efficiency and improve the environment simultaneously.”

Scioto Analysis also estimated that the auction in carbon emissions could raise $2.3-4.1 billion in annual revenue by 2030.

This analysis was conducted by Cruz Eduardo Flores Vera, MPP candidate at the University of California, Berkeley’s Goldman School of Public Policy. Writing support was provided by Rob Moore of Scioto Analysis and additional analytical support was provided by Isabel Clayter, Masashi Hamano, and Ashwin MB of the University of California, Berkeley.

Moore Elected President of Gross National Happiness USA

On Wednesday, Scioto Analysis Principal Rob Moore was elected president of Gross National Happiness USA, a non-profit organization with a mission to increase personal happiness and our collective wellbeing by changing how we measure progress and success.

“Scioto Analysis’s mission is Gross National Happiness USA’s mission—helping improve public policy by helping policymakers understand the impacts of their decisions,” said Moore. “This is a natural fit.”

Moore has been a board member for Gross National Happiness USA since 2018 and Scioto Analysis published its first genuine progress indicator study, Ohio’s Economy: 2009-2016, in conjunction with Gross National Happiness USA.

Moore will lead GNHUSA as it executes its strategic plan, which comprises of gathering data that contributes to measures of wellbeing, advocating for the use of holistic measures of wellbeing by promoting data-driven approaches, working with decision-makers to facilitate policies based on wellbeing, and engaging with individuals and organizations to build a nationwide effort for better metrics.

His term will be begin July 1st, 2020.

Race is a big factor in Ohio police killings

According to CNN, police shot and killed about 1,000 people in the United States in 2018. Compare this to Germany, where police shot and killed 11 people, Sweden, where police shot and killed six people, the U.K., where police shot and killed three people, and New Zealand, where police shot and killed a whopping one person.

These numbers aren’t just because the U.S. has more people, either. Even on a per-capita basis, U.S. citizens are shot and killed by police anywhere from five times as much as in Sweden to over thirty times as much as in the U.K.

It’s not just the prevalence of police killings that is prompting hundreds of thousands of people across the country to take to the streets, though. It’s the disparity in victimization rates between black and white Americans. About a quarter of people killed by police in America are black, much higher than the 13% of total Americans who are black.

Ohio is an especially bad offender when it comes to racial disparities in police killings. From 2013 to 2019, 215 Ohioans were killed by police, 80 of whom were black. That means over a third (37%) of people killed by police over that time period were black, despite the fact that only 12% of Ohioans as a whole are black according to Census Bureau estimates. This 25-percentage-point disparity makes Ohio a top 10 worst state for racial disparities in police killings and constitutes a bigger disparity than all of its neighboring states.

When people across the country think about large-scale activism, they might not think Columbus, Ohio, but Ohio’s capital city has exploded with protests over the past couple of weeks in the wake of the killing of Minneapolis resident George Floyd. Columbus is no stranger to police killings, though: Officers of the Columbus Division of Police killed 40 people from 2013 to 2019, more than three times as many as any other agency in the state of Ohio. Of those 40 people killed by Columbus Division of Police officers, 27 were black, a number more than three times as high as any other agency in the state.

In case you were following along, this means that over two-thirds of people killed by Columbus Division of Police officers were black. This is in a city that is, according to the US Census Bureau, only 29% black as a whole. That means that black people are being killed by police at a rate 39 percentage points higher than would be expected by their share of the Columbus population. If Columbus was its own state, it would only be behind Rhode Island for how disproportionately its police officers kill black residents.

So this means that Columbus police are killing black people at higher rates than police across Ohio are, who are killing black people at higher rates than police across the United States are, who are killing people as a whole at higher rates than police in other developed countries are.

Despite these sobering statistics, police in the Ohio kill people on average less than the country as a whole, killing 26 people per 10 million citizens per year compared the nationwide average of 34, a lower rate than 36 other states. Hopefully after this nationwide wakeup call, cities across the state will figure out why they can’t extend the same courtesy to their black residents.

This commentary first appeared in the Ohio Capital Journal.