Does the One Big Beautiful Bill leave SNAP policy up to chance?

Beginning in October 2027, the Supplemental Nutrition Assistance Program (otherwise known as SNAP and  formerly “food stamps”) is going to be significantly more expensive for states. Currently, the program is fully guaranteed by the federal government, so aside from some administrative costs this amounts to free money for the states. 

Now with the passage of the One Big Beautiful Bill Act, states will have to pay a portion of their SNAP benefits themselves. This change alone is potentially devastating to state governments. SNAP is a massive entitlement program which means that anyone who is eligible for benefits may claim them. This means that 1) some states might not be able to afford paying for any portion of their total SNAP benefits (especially in the face of balanced budget requirements) and 2) states wont know how much spending they are on the hook for until after the fact. 

Those two problems are significant enough that it might be enough to cause states to restrict eligibility or cancel the program altogether, but the actual implementation that states have to contend with stands to make matters even more complicated.

According to the new legislation, a state’s SNAP error rate will determine what percentage of the total costs they end up paying for each year.

What is the SNAP error rate?

The SNAP error rate is the percentage of total SNAP benefit payments that are estimated to have been incorrectly overpaid or underpaid to participating households. This is calculated by taking a sample of households that receive benefits, and auditing them to see if they received the correct amount of benefits each month.

There are two points about this calculation that are crucial to understand. First, this is about total error at the household level. Overpaying one household and underpaying another don’t cancel out in the aggregate estimate, they both count towards a higher error rate. The second point is that this is calculated on a random sample of households, and is subject to sampling variation. 

That second point is especially important because the new law uses this error rate to determine how much states will have to pay. Beginning in 2027, states with an error rate below 6 percent will continue to not have to contribute state funds toward SNAP benefits. States with error rates between 6 and 8 percent will pay 5 percent of benefits, states between 8 and 10 percent will pay 10 percent, and states with error rates of 10 percent or higher will generally pay 15 percent.*

How will this impact states?

A Brookings study published last month looked into state SNAP error rates and measured how much they vary from year to year. In short, they found that these cost-sharing thresholds are far too small to account for the natural variability of the error rate estimate. This is a clear example of why it is foolish to build policies around point estimates. Uncertainty that is inherent whenever we look at a sample makes it so states have basically no way to plan for how much their SNAP program will cost year to year.

Brookings estimates that the average standard error of a state’s annual SNAP error rate is about 1.1 percentage points. That means a typical 95 percent confidence interval spans roughly 4.4 percentage points. But the cost-sharing thresholds between 6, 8, and 10 percent are only two percentage points apart. 

To put that into perspective, imagine that a state has an estimated error rate of 8.5 percent. Under the new policy, that state would have to pay 10 percent of its SNAP benefits. But the statistical uncertainty surrounding that estimate means the state cannot reliably be distinguished from one with an error rate of 7.1 percent or 10.9 percent. Those three estimates would result in the state paying 5, 10, or 15 percent of its SNAP benefits, respectively, which could amount to hundreds of millions of dollars of differences in state required spending. 

When looking at the historical data, Brookings found that 43 states had error rates that moved between at least two of the cost-sharing categories between 2022 and 2025. Fourteen states moved between three categories, and North Carolina moved between all four. Only five states (Idaho, South Dakota, Vermont, Wisconsin, and Wyoming) had error rates consistently below 6 percent over that period. 

The potential dollar amounts involved are enormous. Using 2023 data, Brookings estimates that California, for example, could face anywhere from no state contribution to more than $2 billion in SNAP costs simply because of the statistical uncertainty surrounding its estimated error rate. 

This creates a strange situation where states can make substantial investments in improving their SNAP administration without knowing whether those improvements will actually reduce their financial responsibility. A state’s error rate can change substantially from one year to the next simply because a different set of households happened to be selected for review. 

The underlying issue is not that statistical sampling is inherently bad. Sampling is an essential tool for estimating characteristics of large populations, and every estimate based on a sample comes with some degree of uncertainty. The problem is using a noisy estimate to make a high-stakes financial decision without accounting for that uncertainty. 

States are required to balance their budgets, which makes unpredictable expenses particularly difficult to manage. A state cannot simply assume that its SNAP costs will be the same next year, because its required contribution could change by hundreds of millions of dollars based largely on statistical noise. 

If the goal of this policy is to encourage states to administer SNAP more efficiently, there are better ways to accomplish that goal. Making state governments responsible for billions of dollars in benefits based on an estimate that can change substantially because of random sampling variation does not create a particularly strong incentive for better administration. Instead, it creates uncertainty for state budgets and, potentially, for the people who rely on SNAP. 

*  There is a temporary exception for states with particularly high error rates during the first two years of the policy.

Could lower birth rates actually help the economy?

Birth rates across the globe have fallen sharply over the past several decades. This trend is talked about frequently among policymakers—most people suggest that we need to have more children to continue to support the economy. The key assumption that this relies on is that having fewer babies is bad for future economic growth. The reasoning is intuitive: a shrinking population means fewer workers, more retirees depending on a smaller workforce, and therefore an economy that slows down.

A recent working paper from the National Bureau of Economic Research challenges this conventional wisdom. To better understand the relationship between birth rates and economic growth, researchers looked at what has actually happened in national and state economies over the past 75 years as birth rates have declined. Today, I wanted to discuss some of the study’s key findings.

Does a shrinking population actually slow down the economy?

According to researchers from the National Bureau of Economic Research, the global birth rate fell by more than half between 1950 and 2025. In 1950, the global birth rate was 3.78 per 100 people, and in 2025, the global birth rate was 1.71 per 100 people. This means that in 2025, 1.71 people were born for every 100 people across the world.

To see how this demographic shift affected economies, researchers looked at data from 1970 to 2020 across all countries with populations exceeding one million in 2019 and across 722 commuting zones across the United States. Tax havens, such as Luxembourg and Singapore, and crisis-affected states, such as Yemen and Venezuela were omitted due to unusual trends around economic growth and population growth.

The authors of the paper find that lower birth rates are associated with higher growth rates in income per worker. Across the countries in the study, a one percentage-point lower birth rate was associated with 26.8% higher gross domestic product per worker. At the same time, total economic output did not shrink. In other words, higher incomes per worker didn’t only occur due to there being fewer workers–the total size of the economy stayed steady too.

When labor becomes scarce, industries and firms adapt by innovating labor-saving technologies. Firms in places with falling birthrates have more incentive to automate tasks and reorganize work to compensate for a smaller workforce. This explains why per capita gross domestic product rises while birth rates fall: each worker becomes marginally more efficient due to innovation.

In countries with lower birth rates, researchers found higher total factor productivity, which measures how efficiently an economy generates output from a fixed amount of capital and labor. These countries also saw a higher share of high-tech exports and higher rates of patenting in labor-saving industries like information technologies and automation.

But isn’t it possible that there are some other factors that are associated with low birth rates in countries that would also be associated with innovation of these kinds of technologies and industries? What if there is some underlying characteristic among countries with low birth rates that better explains such economic growth?

How can we establish causality between low birth rates and economic growth?

The researchers use two main strategies to determine whether birth rates actually cause per capita economic growth. First, the researchers lag birth rates by 20 years rather than using current population growth. This reflects the idea that newborns take around two decades to enter the workforce. Researchers found no impact on gross domestic product before the 20 year mark.

Second, researchers isolated the demographic changes associated with lower birth rates. Lower birth rates cause two simultaneous demographic changes: total population size decreases, as death rates begin to outpace birth rates, and the average age of the workforce increases. To simulate which of these demographic changes is responsible for economic growth, the researchers analyzed casualties from World War II across countries. 

Civilian war deaths reduced population across all age brackets without significantly altering the age distribution, whereas military deaths were heavily concentrated among young men. Civilian deaths, which reduce a country’s population among all age groups, were associated with lower gross domestic product per capita. However, military deaths, which were concentrated among young people, were associated with higher gross domestic product per worker. In other words, higher economic output per worker is specifically associated with a shrinking and aging workforce, not just a smaller population.

Falling birth rates still present a challenge for policymakers. While it appears that per capita income and production actually benefit from lower birth rates, there are other problems associated with a smaller workforce, such as a smaller tax base for social programs that serve older residents at disproportionately higher rates than younger residents. However, this does suggest falling birth rates might not be as bad for the overall economy as we once thought.

Will foster care payment reform save money or harm kids?

Last month, two members of the Ohio House of Representatives introduced legislation to fix payment rates for foster care statewide.

As covered previously in the Ohio Capital Journal, the bill is largely a response to rising costs in Ohio’s foster care system. According to Scioto County Commissioner Scottie Powell, the number of Ohio children placed in foster care has dropped 9% since 2020, but costs have increased by 68%, a $160 million increase over the same time period.

Scioto County Job and Family Services Director Tamela Moore Morton says that this is due to foster care providers increasing prices to take advantage of the short supply of providers. She says providers are able to ask for whatever they want to place children and the state and counties have little they can do about it since these children need to be placed.

Ohio House Bill 984 is essentially a price-control bill for publicly funded foster care placements. The bill will create a statewide rate-setting system for care provided by certified foster homes and residential facilities and will prohibit counties from paying more than that rate for services. The hope is that this bill will rein in excess costs of foster care placements in places where a limited number of providers are able to keep rates artificially high.

The market power story is a plausible one and the fact that at least one Job and Family Services director sees this happening on the ground makes this bill seem like a good idea. But there is a counterweight consideration that also matters when it comes to foster care payments: how payments impact supply.

Foster care is expensive. Caring for children, especially children with serious behavioral or mental health needs, requires staffing, supervision, specialized training, transportation, facilities, and crisis management. If reimbursement rates fall below the cost of providing that care, providers may stop accepting high-needs children, reduce capacity, or stop providing care altogether.

If current rates are higher than the cost of providing care because providers are exercising market power, H.B. 984 could save taxpayers a lot of money without reducing the quantity or quality of care. But if these rates reflect the actual cost of serving children with increasingly complex needs, capping rates could discourage new providers from providing care and stop existing providers from doing it.

The best way to find out which way the bill will behave would be to build evaluation into the proposal. The bill could require the Department of Children and Youth to establish a pre-implementation baseline and publicly report key measures six months, one year, and two years after the new rates take effect. By tracking the cost per child-day, provider supply, placement difficulty, displacement, placement stability, and number of exceptions granted to the bill, policymakers should be able to get a good idea after a few years whether the bill is reducing costs at no harm to the system or if it is restricting supply of care.

If costs go down without any changes to provider participation and placement waiting times and disruptions, then we would have good evidence that this bill is saving money at no costs to quality care for children. If, on the other hand, the number of providers plummets while refusals, placement times, out-of-state placements, and disruptions increase, then we know that supply is being constricted.

The fact is, we don’t know if H.B. 984 will save money or harm kids. But legislators can write the bill in a way so they can find out.

This commentary first appeared in the Ohio Capital Journal.

Are we wrong about how dangerous heat can be?

Last week, I came across a Morning Edition Story talking about the results of a study they conducted in partnership with Boston University about heat deaths across the country. They wondered whether the Centers for Disease Control was undercounting heat-related deaths in their published mortality data.

The reason they suspected this might be true is because heat can be a factor for other causes of death that get reported. For example, an abnormally hot day might induce a heart attack in someone. While their cause of death was the heart attack, we can say that had that person had better access to cooling resources, they would not have died from that heart attack. 

It is similar to how we talk about the avoidable deaths from smoking. When we say that people die because of their cigarette smoking, we often attribute some other cause (cancer, heart disease, etc.) to their smoking behavior. Had that person not smoked, we suspect they would not have otherwise developed these fatal diseases. 

This new research suggests that across the country as a whole, actual heat related deaths are as high as five times the reported figure by the CDC. 

To come to this conclusion, the researchers developed a model which looked at weekly temperature data along with weekly mortality data at the county level. They measured the impact that unusually hot weeks (relative to a county’s average temperature to account for differences in acclimatization) impacted mortality rates at the county level. The article is light on specific modeling details, but they note that they intend to publish more detailed results in an academic journal in the near future. 

This kind of approach was also used to measure the mortality impact of the COVID-19 pandemic. Researchers measured the number of excess deaths that occurred because of the pandemic relative to the number we would have expected under normal conditions.

One detail about this model that really highlights the scope of the problem is that there are some counties which actually report higher heat-related death counts than the model estimates. One example they highlight is Maricopa County Arizona, where Phoenix is. 

Health officials in Maricopa County have made a concerted effort to better understand the impact that temperature has on local mortality rates. The result of their efforts is that they report more than double the number of heat related deaths than the model estimates. 

This highlights the fact that these estimates are averages across the whole country. I’ve written before about the problems with point estimates, and while I expect the final results of this work to go into more detail about the variation in these estimates policymakers should be aware that their jurisdiction might not look exactly like these results. 

Counties that have experienced the effects of climate change more harshly or those with more at-risk populations might experience even worse effects than those estimated in this report, while the inverse may be true for more stable or healthier counties.

Understanding the full range of heat-related impacts is important for policymakers and analysts because it will change the way we value certain impacts. If we believe that more people are dying because of their exposure to heat, then policies that work to reduce that exposure may end up saving more lives than we might otherwise expect. It can also impact our valuation of climate impacts. If heat is more deadly than we previously believed, then a hotter world is a more dangerous world for its residents.

Are electric vehicles a problem for the gas tax?

Last week, my colleague Rob wrote a blog post analyzing various taxes and how distortionary each one is. One type of tax is a user fee, which is designed to recover the cost of offering a good or service. 

A classic example of a user fee is a gas tax. The only people who directly pay a gas tax are people who buy gas, and those people use the roads, bridges, and other infrastructure intended for cars. Since gas taxes are a major source of funding for this kind of infrastructure, more gas taxes paid translates directly into more funding for roads.

In theory, gas taxes aren’t distortionary at all. However, as electric vehicles become more prominent, there are more and more people who are using roads who aren’t paying into gas taxes. This presents an issue for policymakers: how do we ensure that electric vehicle owners are contributing to infrastructure for cars as much as consumers of gasoline?

What do gas taxes fund in Ohio?

The state of Ohio charges a 38.5-cent tax per gallon on gasoline and 47-cent tax per gallon on diesel and other fuel types. If we view gas taxes in Ohio as user fees, then the 8.5-cent premium for diesel and other fuel types reflects that diesel-powered vehicles, such as trucks, buses, and other heavy vehicles, cause greater wear on roads than passenger vehicles. In 2025, Ohio collected nearly $2.7 billion in tax revenue from gas taxes.

Of that $2.7 billion, about $1.7 billion (63%) went to the Gasoline Excise Tax Fund, about $950 million (36%) went to the Highway Operating Fund, and about $34 million (1%) went to other funding sources.

The Gasoline Excise Tax Fund is sent to municipalities, counties, and townships across Ohio for the purpose of local road projects, while the Highway Operating Fund is used for state level roads, bridges, and other statewide infrastructure projects.

Are electric vehicles becoming more common?

Between 2019 and 2025, electric vehicle registration increased nearly eight-fold, from 170,000 registered electric vehicles in 2019 to over 1.3 million registered electric vehicles in 2025.

As more people drive electric vehicles, relatively fewer people are purchasing gasoline. According to data from the Federal Highway Administration, the average car uses about 447 gallons of gasoline per year across the United States. That means that for each fewer gasoline car on the road every year, the state of Ohio loses about $170 in tax revenue. 

Assuming each newly registered electric vehicle in Ohio between 2019 and 2025 replaced a gasoline car, and each hybrid uses roughly half the fuel of a gasoline car, Ohio loses approximately $170 million per year in gas tax revenue from the adoption of electric vehicles. 

So how does Ohio recoup this money? For one, the state raised the gasoline tax by 10.5 cents and the diesel tax by 19 cents in 2019, which might have offset some of the growth in electric vehicles. However, simply raising the fuel tax places a disproportionate amount of the tax burden on drivers who still use gasoline cars.

Another way that many states are dealing with this problem is by implementing electric vehicle registration fees. In Ohio, it costs an additional $100 for hybrid vehicles, $150 for plug-in hybrid vehicles, and $200 for fully electric vehicles to renew registration each year. These prices are about in line with the amount of gas tax revenue that the state loses per electric vehicle, so these registration fees are fairly equal gas tax replacements for electric vehicle drivers. 

However, a flat-rate fee paid each year doesn’t fully replace the functionality of a user fee. With a gas fee, drivers are charged according to how much gas they purchase, which roughly reflects their annual usage of roads. Registration fees for electric vehicles don’t reflect the amount that drivers use roads, so it is a less precise user fee. Some electric vehicle charging stations have taxes per kilowatt hour of energy used, which better reflects the usage of electric vehicle drivers, though private residences are exempt.

Ideally, this problem can be fixed with a vehicle miles traveled fee, which some states are currently implementing. One of Scioto Analysis’s first studies was on how new technologies are making vehicle miles traveled fees more possible. If states like Ohio can successfully implement vehicle miles traveled fees, they will be able to bring user fees closer to an efficient mimicry of a market.

How does social media impact happiness?

The World Happiness Report is released every March alongside the United Nations International Day of Happiness. One of the biggest takeaways from this year’s study is that young people are less happy than they used to be, and that this may have something to do with social media use. Today, I wanted to explore some of the findings from this year’s report and look into what the effects of a social media ban could be.

Is social media making us miserable?

According to the 2026 World Happiness Report, most people indicated that they are willing to pay less money to use social media platforms than they would need to be paid to stop using them. In other words, people would pay a relatively low price to access social media, but they would expect a relatively high price to give it up. 

This finding could point to something deeper about social media use: perhaps free access to social media generates a disproportionate amount of value that makes it hard to give up despite its negative effects, such as algorithmically curated content that keeps people hooked. Alternatively, this gap could stem from behavioral biases. People are used to a pre-existing zero-dollar price tag, and giving up an established habit feels costly.

Another interesting finding around how people value access to social media is that researchers found people demand a significant amount of money to stop using social media platforms like Instagram and TikTok, but they would also be willing to pay to eliminate those platforms from their community entirely. This suggests that people’s lives are so entrenched with social media that it's challenging to give it up voluntarily, but people are far more receptive if they know their friends and family won’t be using it.

There are a few interesting takeaways from this finding. First, people recognize that platforms like Instagram and TikTok  harm themselves and people they care about and that they would be better off without social media. Second, individuals still feel that their lives would be worse off without social media if everyone else in their community still has access to it, despite the harms they recognize. If everyone else is on the platform and you’re not, you start to feel like an outsider, even if you and your friends would be better off without social media.

A broader finding from the 2026 World Happiness Report is that life satisfaction is highest for people with low rates of social media use and lower for people with high rates of social media use. Internet use is associated with higher levels of life satisfaction for activities like communication, news, learning, and content creation, but only if they’re consumed in moderation. At high levels of use, all internet activities were found to be correlated with lower life satisfaction.

Among adolescents, there are some interesting trends. Adolescent girls tend to have the highest life satisfaction when they use social media for less than an hour per day, but adolescent boys tend to have both very high and very low levels of life satisfaction concentrated at high levels of social media use (7+ hours per day). 

This is a great example of why economists are generally against outright bans: some amount of economic value gets destroyed (like the life satisfaction that some people are getting from 7+ hours of social media use per day). There are, of course, other factors that might be bad about extreme social media use outside of life satisfaction which could warrant a ban.

Would a social media ban increase happiness?

According to a survey conducted by the Pew Research Center in May 2026, nearly 6 in 10 Americans support a social media ban for children under 16 years old. The demographic group that most strongly supports a social media ban for children under 16 years old is parents, 65% of whom support a social media ban. Since 2023, social media policies aimed at minors, such as requiring parental consent, age verification, or time limits, have all gained support. In general, support for a social media ban is nonpartisan and strong across most demographics.

Research from the American Psychological Association shows that limiting daily social media use improves mental health outcomes for young adults. Young adults who limited social media usage to 30 minutes per day reported significantly lower levels of anxiety, depression, loneliness, and fear of missing out. Other research from Liu et al. has found that doing “social media detoxes” has small but positive effects on well-being.

But many of these studies are flawed. For example, in a meta-analysis conducted by researchers at the University of California, Irvine, researchers could not find a single study on social media bans that included children or adolescents. Children and adolescents have different social dynamics than adults where losing access to social media could be harmful, so findings from studies about adults might not be applicable to children and adolescents. Additionally, most research is centered around brief social media intervention, and literature about long-term social media bans is scarce.

Nevertheless, various countries across the world are starting to implement social media bans for children under 16. In Australia, a ban took effect in December of last year. Both Denmark and the United Kingdom have signaled potential plans for banning social media for young children as well.

As of now, there is little evidence that social media bans are an effective solution, and Australia’s social media ban isn’t exactly promising. Take Business Insider’s article as an example, which is literally titled, “Australia’s social media ban for teens is going, uh, pretty badly”. As of last month, 26% of children aged 13-15 in Australia were on TikTok, just one percentage point lower than the rate prior to the ban. And the social media use rate for children between the ages of 10 and 12 is higher now than it was before the ban.

Extreme social media use is correlated with higher rates of unhappiness, especially among children, but social media bans have so far been ineffective. So what's the solution? Age verification tools from social media companies are easy to circumvent, and some families might have concerns about privacy. There doesn’t seem to be a clear path forward on how to solve the problem of social media use and well-being, but public policy analysis around some other alternatives could be a place to start.

Quick analysis: Is there a relationship between income and physical exercise?

This is going to be a quick analysis where I attempt to answer a question in what basically looks like the first pass I’d do for a larger research project. The goal is not to conclusively answer the question I pose, but rather to show how you could go start exploring a question like this and get at least some evidence and see if this is worth exploring further. I’ve previously done a similar analysis looking at commute times and housing prices.

Research question: how do income and exercise relate to one another?

Hypothesis: There is a non-linear relationship between income and physical activity levels. As with my previous quick research blog, I got this idea from a newsletter that was looking at physical activity in different countries. One thing that jumped out at me when I saw the rankings was how there didn’t seem to be any relationship between how rich a country was and how active it was. My intuition based on living in the U.S. would be that income and exercise would be much more closely related.

The question that follows for me is whether this trend holds at the national level. Do we see a similar non-linear trend between income and physical activity if we just look at data from one country, or are there indeed broader cultural trends that make some low-income countries more active than some wealthy countries? 

Methods

I will be looking at the Center for Disease Control’s Behavioral Risk Factor Surveillance System (BRFSS). This is the largest health survey conducted in the U.S. each year, and in particular it asks respondents some basic demographic questions (including their household income) as well as a few relevant questions about the amount of physical activity they report.

There are two outcome questions we will look at: a binary variable for whether or not a respondent reported exercising enough to meet the recommendation for the amount of aerobic exercise they should be doing and a categorical variable for whether a respondent is classified as highly active, active, insufficiently active, or inactive. Both of our dependent variables are calculated based on respondents' answers to other questions about their type and frequency of their exercise. These are not self-reported measures, so we don’t need to worry about how people interpret things like “highly active.”

We will use an ANOVA (analysis of variance) model to test whether there is some impact. The reason we are going to use ANOVA to start is because I am curious if there is a non-linear relationship between income and exercise. By treating each income category separately rather than assuming that physical activity increases or decreases at a constant rate with income, ANOVA allows us to identify differences between income groups that a simple linear regression might miss.  

Results

The first thing we see when we look at the responses is that if there is an effect, it is almost certainly linear. Table 1 reports the percentage of respondents in each income group that meet their aerobic exercise recommendation, as well as the average response category for the categorical variable about total activity level. This variable is coded such that one is associated with highly active and four is associated with inactive, so smaller numbers mean more physical activity on average.

Table 1: Higher income respondents are more active than lower income respondents

The ANOVA results back for both variables both conclusively suggest that there is a relationship between these income categories and both of the response variables, with both p-values coming out to essentially zero. We can in both cases reject the null hypothesis that there is no relationship between income and exercise level. 

Limitations

I have not really gone far enough to fully understand the relationship between income and amount of physical activity. We found by looking at these data that there is some relationship, and the clear order of these averages suggests that it is quite likely linear. More work is needed to demonstrate more strongly that the relationship is indeed linear, and what the strength of the relationship is. 

Preliminary conclusions and next steps

There certainly is some relationship between income and the amount of physical activity respondents participate in, and we have plenty of evidence to dispute the original hypothesis that it is non-linear. The connection between income and exercise appears to be positively correlated, meaning higher income is correlated with higher likelihood of being physically active, though the exact strength of this relationship is still open to further research. The next steps to fully understand this relationship would be to look for potential confounders and account for those as well.

Data center debate exposes Ohio’s uneven energy landscape

In the battle over data center siting in Ohio, one of the flashpoints is how data centers impact electricity prices.

Data centers are indeed energy-intensive and the worry about their impact on local energy prices has moved most developers to work to get their energy generated in behind-the-meter projects that don’t draw from the electrical grid.

So if you are a data center developer, what kind of power are you going to put behind the meter?

Looking at the top sources of power in the United States, you can whittle the options down pretty quickly.

Coal, once the heavyweight for energy in the United States, has seen its economics turn against itself and the United States has only begun construction on one coal-fired power plant in the past 13 years.

For all the talk of small modular reactors, nuclear power is still not viable in Ohio due to the massive up-front costs and decades of regulatory hurdles to clear.

Hydropower and geothermal power demand specific topographic conditions, and biomass has economies of scale that don’t make it competitive with other technologies.

This leaves developers with three choices: solar, wind, and natural gas.

The state has put its thumb on the scale when it comes to choosing between these technologies. A range of state decisions have made siting solar, wind, and natural gas projects very different from one another.

Solar and natural gas projects over 50 megawatts must be approved by the Ohio Power Siting Board. For reference, this would be large enough to power all the homes in Canton with a little bit of energy left over.

Wind projects, on the other hand, only need to be 5 megawatts to face Ohio Power Siting Board scrutiny. That is only enough to power about half the homes in Athens.

The state has also given considerable latitude to county governments to ban solar and wind projects in unincorporated areas, a barrier gas-powered plants do not have to overcome.

For projects that are not outright banned, two local representatives get a vote on Siting Board decisions for wind and solar projects, a requirement not faced by natural gas projects.

Solar and wind projects also face a regime of siting rules that do not apply to gas plants.

Solar facilities face specified setbacks, landscaping, fencing, stormwater, noise, and vegetation requirements. Wind facilities face turbine setbacks and shadow-flicker, ice-throw, blade-failure, communications-interference, noise, and aviation requirements.

Solar and wind projects also face decommissioning planning requirements that gas-powered plants are not subject to.

Gas-powered plants do have one requirement that solar and wind projects do not: they must submit a range of operational air-quality analyses. This makes sense to a certain extent given solar and wind generation is emissions-free.

In an ideal world, technologies can compete against each other on a level playing field.

If there are specific costs associated with outcomes like public health, environmental sustainability, or even aesthetics, these can be captured through fees and taxes specifically designed to internalize these costs into the market.

Creating separate regulatory regimes for different technologies, on the other hand, makes legislators the arbiters of technological superiority rather than the market.

This commentary first appeared in the Ohio Capital Journal.

Taxes are necessary. Which ones do the least harm?

When I first enrolled in graduate school, I was pretty high on taxes.

Maybe it was the “contrarian” in me. Maybe it was the excitement of Bernie Sanders suddenly making “socialism” a mainstream word in the political lexicon. Or maybe it was just the fact that I was going to policy school and I knew programs were funded by taxes so I liked them. Suffice it to say, I was quick to dismiss anything nasty people had to say about taxes as political posturing.

This made it hard for me to accept the assurances from my economics professors that taxes did indeed destroy social value.

This does not mean that government necessarily destroys social value, but many taxes do destroy social value in one place, even if it creates it in another place.

The main way taxes do this is through distorting markets. By tagging an extra price on transactions, consumers and producers self-select out of markets they would otherwise be competitive in. This increases prices and reduces quantities traded at the same time.

Not all taxes are created equal, though. Some taxes are incredibly distortionary. Some barely distort markets at all. Some are negatively distortionary–they actually make markets work better.

How do we know which taxes are more distortionary and which are less? We mainly find this through empirical study. That being said, there is a general hierarchy of taxes that can help you have a reasonable idea of which are most or least distortionary.

Most distortionary: Narrow-based taxes on elastic goods

The taxes that distort the economy most are narrow taxes on single goods. These end up being distortionary because they shift spending decisions so wildly. To understand the intuition behind this, imagine a tax on hamburger buns. This would lead to a reduction in purchases of hamburger buns because they are more expensive, but would also lead to less consumption of hamburger patties, more consumption of hotdog buns, and more consumption of hotdogs, the latter two of which serve as substitutes for hamburger buns and hamburgers.

Analyses of federal taxes in Australia lend support to this claim. House sale taxes have tax burdens that exceed their value of revenue raised. New car taxes destroy nearly a dollar in value for every dollar they raise in revenue. Since purchasers of new cars can substitute to purchasing used cars and buyers of property can instead become renters of property, these taxes end up shifting preferences from ideal purchases to secondary purchases and reshuffling markets.

Less distortionary: broad-based taxes

So if assessing taxes on narrow goods leads to distortion, shouldn’t we instead levy taxes on broad sets of goods? Well this has been the logic of economists for years now: levy taxes on a broad range of goods so they have less of an impact on the broader economy. The logic here is that if you tax something broad like labor, purchases, or property, which people cannot avoid, then their purchasing decisions will not change as much as if you tax something narrow.

The federal Office of Management and Budget estimates marginal excess tax burden at 25 cents of value destroyed for every dollar of revenue raised, much lower than the above estimates for the cost of narrow taxes. Researchers have put the cost of a broad-based sales tax at 13 cents on the dollar and a broad-based property tax at 14 cents on the dollar. Estimates on value added taxes, which spread a tax throughout the stages of production in an economy, should theoretically be even lower but land in the same territory as these estimates in the empirical evidence.

Not distortionary: user fees

Some taxes and fees like gas taxes are designed to mimic market mechanisms. While roads are public goods constructed and maintained with tax dollars, a major source of their funding is through gas taxes. Levying a gas tax and earmarking these funds for road maintenance means that people and companies who buy more gas (and presumably use roads more) pay more taxes. So the people who are paying for the roads are the people who are using the roads.

Ideally, a perfect user fee would have a social cost that equals its social benefit, netting a marginal excess tax burden of zero. This is not always the case, however. The rise of electric vehicles means that many people who use roads are not paying gas taxes, which makes gas taxes a less perfect user fee than they would be otherwise.

You can imagine other user fees that would have no distortion, like fees for water, sewage, and waste disposal conducted at the city level. If these are made in proportion to the volume of service, then these fees provide no distortion to the economy: their social benefits can theoretically equal their social costs.

Negatively distortionary: Pigouvian taxes

A final category of taxes are taxes that actually enhance market efficiency: Pigouvian taxes. Named after the father of welfare economics Arthur Pigou, this is a category of taxes that are deployed when a market has total social costs that exceed the private benefits realized from the transaction. Examples of these are carbon taxes and cigarette taxes. Since future generations impacted by climate change and breathers of secondhand smoke are not willing participants in markets in carbon and cigarettes respectively, the marginal social costs of these transactions exceed the marginal private costs, representing a failure in the current market. Taxes on carbon and cigarettes bring private costs in line with social costs, making the markets more efficient than they would be otherwise. 

These are all just rules of thumb: there can be some narrow-based taxes that have low elasticities that are not very distortionary. There are also some user fees that do not operate efficiently and cause large distortions. And distortion is not the only policy-relevant dimension of tax policy. Often we are willing to trade off efficiency for equity outcomes, like we would do with a graduated income tax or a corporate tax. But I hope this framework at least gives you an idea of the logic behind marginal excess tax burden and a framework to approach tax policy and design.

Is money really fungible?

Recently, I came across a new working paper that looks at some of the impacts that came about as a result of recent changes to the Supplemental Nutrition Assistance Program, otherwise known as SNAP (formerly “food stamps”). In some states, SNAP benefits became more limited and are no longer able to be used on the purchase of certain unhealthy items including sugary drinks. 

The whole paper is extremely interesting and definitely worth a read, but today I wanted to dive in more closely to one of the key questions the paper raises, is money fungible?

What does it mean for money to be fungible?

If something is fungible, that means it can be interchanged with something equivalent. It shouldn’t matter whether someone gets paid in cash, a check, or via direct deposit. As long as the value is the same, (aside from some minor inconvenience) those differences don’t matter.

Two of the core functions of money are that it is a unit of account and a medium of exchange. In practice, these two characteristics should suggest that all money is equal. In a broad sense, money is certainly fungible (i.e. when considering how people trade money).* However, it is more interesting to think about this in terms of how people make decisions about their personal spending. 

Is money fungible across a budget?

This is where the paper comes in. The researchers looked at changes to the SNAP program that banned the purchase of sugary drinks and food. This is an interesting change, because according to the paper over 80% of SNAP households spend more on food each month than their SNAP benefits cover. According to economic theory, this implies that SNAP essentially acts as a cash transfer, since we’d expect monthly food budgets to remain constant with or without these benefits. 

If this is true, then we should expect these SNAP changes to have little to no effect on specific spending decisions. The benefit size isn’t changing, these are just new restrictions on how it can be spent. If these households are already spending more than their SNAP benefits on food, then they could just switch around what items they buy with their SNAP cards and what items they buy with cash after. 

The main takeaway from the paper is that money isn’t necessarily fungible across budgets. People who receive SNAP benefits do not appear to be changing their non-SNAP grocery spending to compensate for the fact that their SNAP-supported grocery spending changed. While there may be situations where a one-to-one change isn’t possible for some reason, economic theory would suggest that the fungibility of money should make it so this transition was much smoother than it ended up being. 

What does this mean for policymakers

While many of us will find it inherently interesting that we have another concrete counter-example to a basic principle of economic theory, we should ask ourselves how this is actually relevant to policymakers in the real world. 

One takeaway is simply that policies such as these SNAP restrictions can actually reduce consumption of certain goods. This can be both a blessing and a curse, as it gives policymakers another tool for changing behaviors which can be difficult, but it also highlights how some policies might have unintended consequences.

Another takeaway is that household budgets play an important role in how people make spending decisions. In this SNAP example, nothing about the market for sugary drinks was impacted by this policy change. There were no new taxes or subsidies, nothing impacted a substitute or complimentary good, from a theoretical perspective the equilibrium price and quantity should have remained constant. The fact that we saw an observable change in the amount consumed suggests that policies such as these can have indirect effects that shape markets.

One important difference that makes SNAP different from other tax-and-transfer programs is that SNAP benefits come preloaded on an electronic benefits transfer card (essentially a debit card). This is functionally quite different from programs like Social Security or the Earned Income Tax Credit that are distributed as cash. Those programs are the most similar to SNAP given their size and scope, but it may not be possible to achieve a similar impact because their administration doesn’t create a separate pool of resources that can be budgeted separately.

One argument against this kind of policy intervention is that it is an overly paternalistic decision made by the government, and that people should have the ability to make decisions about their food intake more freely. That is a tradeoff that policymakers should be wary of, since overly restrictive policies can reduce the overall economic benefit that comes from improved health outcomes. Whether these policies end up becoming more widespread, it is fascinating to see how people react to changes like these.

* There are plenty of examples where money can have different values. Someone who wants to buy something from a vending machine might have a higher value for five $1 dollar bills relative to a single $5 bill. Similarly, most people I know don’t like carrying around $100 bills, preferring the easier to use $20 denomination.