The State of Ohio's Investments in the 2010s in Five Graphs

In June, the federal Congressional Budget Office released its periodic report on federal investment—a report that catalogs federal spending on physical capital, education, and research and development over time and compares it to the growth of the overall economy. After seeing this report, I was interested in seeing how Ohio’s state investments compared.

The Ohio Legislative Service Commission provides historical data on Ohio’s revenues and expenditures going back to 1975 for operating funds and 2008 for capital funds. In order to mirror a previous study we did on the state economy, I looked at data from 2009 to the present to look at state investments throughout Ohio’s recovery from the Great Recession.

Unfortunately, the Legislative Service Commission does not provide readily-accessible data on state research and development spending. That being said, the agency does a good job of making physical capital and education spending available. Below are some of the takeaways I had looking at this data.

1. Total physical capital and education spending is down 11% in the past decade. In inflation-adjusted terms (using the Bureau of Labor Statistics’ CPI-U measure), the state spent 11% less on physical and human capital in 2019 than in 2009. This is driven by a $600 million reduction in education spending and a $700 million reduction in physical capital spending over that time period. Spending on physical and human capital is up $1.1 billion in 2019 from its nadir in 2012, but has not reached 2009 levels in real terms.

Data from Ohio Legislative Service Commission

2. Reductions in education spending over the past decade were driven by higher education cuts. While K-12 real expenditures were flat from 2009 to 2019, higher education spending was cut by 23% over the time period. Reductions in higher education expenditures made up almost half of the total reduction in combined human and physical capital over the past decade.

Data from Ohio Legislative Service Commission

3. Reductions in physical capital spending were driven by school facility investment cuts. The state of Ohio spend almost $900 million less on school facilities in 2019 than in 2009, representing two-thirds of the net reduction in human and physical capital spending over the decade. This may have been due to a “catch-up” period in K-12 facility building in the late 2000s or may represent a policy decision to reduce facility spending.

While K-12 facility cuts cuts were partially offset by increases in investment in natural resources, corrections, and transportation capital, they weigh down total physical capital spending, leading to an over 40% reduction in physical capital spending from 2009 to 2019. While physical capital investment is up $175 million from its low point in 2013, it was lower in 2019 than any year since 2014.

Data from Ohio Legislative Service Commission

4. State human and physical capital investment has declined as a percentage of Gross State Product (GSP) over the past decade. Taken as a percentage of GSP, or total economic output for the state as reported by the Bureau of Economic Analysis, state human and physical capital investment is down almost a full percentage point from 2009 to 2018 (the most recent year we have BEA GSP data for), a reduction of one third of its previous level. Spending as a percentage of GSP has not had the recovery that human or even physical capital has, reaching 1.7%, its lowest point over the time period, in 2018. This could mean that more investments are being made privately, but also means the state is investing less in capital that could have economic spillovers or equity benefits for the population.

Data from Ohio Legislative Service Commission and the Bureau of Economic Analysis

5. State human and physical capital investment has declined as a percentage of Genuine Progress Indicator (GPI) over the past decade. GPI is an economic growth indicator that adjusts for factors not included in GSP that have known economic impacts like environmental damage, the value of unpaid housework and child care, and the cost of family breakdown. GPI in Ohio is currently calculated by Scioto Analysis and data is available through 2016. Using this measure as the calculation for the size of the economy, human and physical capital investment is down by a percentage point from 2009 to 2016, which only amounts to a reduction of a fifth of state human and physical capital investment as a share of the economy as a whole. Much of this, though, owes to slower GPI growth over the period fueled by rising inequality, culminating in a GPI “recession” in 2016. The slower growth in GPI thus makes the reduction in human and physical capital investment by the state look smaller as well, though investment is still lower in 2016 than in 2009.

Data from Ohio Legislative Service Commission and Scioto Analysis

Overall, the Legislative Service Commission data suggests that the state of Ohio has reduced human and physical capital investment over the past decade. Total real human and physical capital investment, human and physical capital investment as a share of the economy, human capital investment, physical capital investment, and K-12 facility spending are all down in the past decade, while K-12 non-facility investment is flat and there have been smaller increases in natural resource, corrections, and transportation investments. This signals a belief among current policymakers that investments are not the most effective strategy for reducing market failures or achieving equity goals over the period of recovery from the Great Recession.

Poverty and Car Ownership in Ohio's Cities

In 2016, one in six households in Ohio’s six largest cities did not have a car. This number varied between cities in the state, though: in Columbus the number of households without a car was less than one in ten while in Cleveland the number approached one in four.

Part of this is a story about poverty. Higher-poverty cities tend to also have lower levels of car ownership nationally, and this is a trend that we see among Ohio cities as well.

Data from the US Census Bureau as reported by Governing Magazine and the Ohio Development Services Agency.

This data should not be especially surprising: it is well established that higher incomes lead to greater levels of car ownership. Something to take away from this data is that car ownership is still rather prevalent among poor households. Even if all houses without cars were poor, anywhere from a quarter to over half of all poor households in each city own a car.

On implication of this data is for financing of mass transit. One of the justifications given for public financing of mass transit is redistributionary: that mass transit is disproportionately used by the poor and thus financing for mass transit serves as a mechanism for promoting equity. This data, though, shows how imprecise this method is for achieving equity. Add the fact that taxes used to pay for mass transit tend to be regressive sales taxes, and we can see that much of our mass transit system is financed by poor families with cars.

Another implication is that gas taxes may not be as regressive as they are often pitched. While a lot of poor households own cars, many also do not, which means that a gas tax to capture infrastructure, congestion, fatality, and emission costs with a low-income rebate could be more progressive than we often assume.

Scioto Analysis Releases Ohio Handbook of Cost-Benefit Analysis

On Tuesday, Scioto Analysis released the Ohio Handbook of Cost-Benefit Analysis, a handbook for policy analysts in the state of Ohio interested in adding cost-benefit analysis to their policy analytic toolkit.

The Handbook provides an overview of the theory behind cost-benefit analysis, standard formulas used in cost-benefit analysis, and a checklist for those carrying out a cost-benefit analysis.

“Analysts in Ohio interested in cost-benefit analysis will be able to use this guide as a quick and dirty introduction to the practice,” said Rob Moore, principal of Scioto Analysis. “Cost-benefit analysis is a key tool for bringing numbers to the discussion on budgetary and regulatory decisions policymakers need to make.”

Last Spring, Scioto Analysis conducted a study of the use of cost-benefit analysis in the state of Ohio, finding 27 studies from 2012 to 2018 that measured costs and projected some sort of outcomes of state-level policies. None of these studies, though, followed all the best practices of cost-benefit analysis. Scioto Analysis’s cost-benefit analysis on the state earned income tax credit released last month was the first best-practices cost-benefit analysis carried out on a state program in Ohio in over a decade.

“More resources like these will make it easier for analysts to conduct cost-benefit analyses according to best practice: monetizing costs and benefits against a baseline, discounting, conducting sensitivity analysis, and disclosing assumptions,” said Moore. “We look forward to seeing the improved analysis that this handbook makes possible.”

How do College Students Impact County-Level Poverty Rates in Ohio?

When college students live off campus, poverty rates go up. This is because the incomes of full-time college students are counted in poverty counts alongside the incomes of other local residents.

There is plenty of reason to be worried about college student poverty rates. That being said, college students are more likely to have higher family income and social capital than those who do not attend college. Also, students from high-income families who decide to not generate income during college years in order to increase wages later may necessitate a different policy intervention than local residents experiencing involuntary chronic or intermittent poverty.

In December 2017, researchers at the Census Bureau estimated the impact of presence of off-campus college students on poverty rates on a county-by-county basis. Nine Ohio Counties, Athens, Butler, Cuyahoga, Franklin, Greene, Hamilton, Lucas, Portage, and Wood, were found to have college student populations that significantly shifted the poverty rate in the counties. These counties were also home to 12 of Ohio’s 15 largest colleges and universities.

Of these counties, the three where college students had the largest impact on county-level poverty rates were Athens (9 percentage points), Wood (4.6 percentage points), and Portage (2.5 percentage points) counties, rural and suburban counties that are home to large state universities Ohio University, Bowling Green State University, and Kent State University.

The presence of college students also impacts the ranking of counties relative to one another when it comes to poverty. As can be seen in table 1, Wood and Portage Counties falls from a slightly low-poverty county to one of the lowest-poverty counties in the state when adjusting for college students. Even Franklin and Hamilton County, two of the most populous urban counties in the state, drop from relatively high poverty to more middling poverty numbers when adjusting for college students.

Also notable is Athens County, which still maintains an extremely high poverty rate after adjusting for college students, but no longer holds the top spot after the adjustment.

College student poverty is an issue that policymakers should pay attention to, but it also looks very different than poverty among the general population. Adjusting for college student prevalence can help policymakers understand the different types of poverty that live in university-dominated counties versus those with smaller student populations.

Moore named editor of Society for Benefit-Cost Analysis's Official Blog

On Wednesday, Scioto Analysis Principal Rob Moore was named editor of On Balance, the official blog of the Society for Benefit-Cost Analysis.

As editor, Moore will solicit, review, and edit submission to the Society’s blog and will serve as an ex-officio member of the board of the Society for Benefit-Cost Analysis . He will begin duties as editor in January.

“It is an honor to be selected to take on this role with the leading association for practitioners of benefit-cost analysis worldwide,” said Moore. “I look forward to working to ensure the blog presents a range of material for members and people interested in learning more about benefit-cost analysis.”

Moore had previously served as the editor of the blog for the Policy Matters, now the Berkeley Public Policy Journal, the student journal for the University of California, Berkeley’s Goldman School of Public Policy as a graduate student. He has also contributed to On Balance before, writing about Scioto Analysis’s study on the use of cost-benefit analysis in Ohio.

What Drives Poverty in Ohio?

As part of our double bottom line mission, Scioto is carrying out a multi-year research project on five key aspects of well-being. With the conclusion of this summer’s cost-benefit analysis on the earned income tax credit, we are shifting from a focus on economic growth to a focus on poverty.

As we started our 2018-2019 research on economic growth, we carried out a study of Ohio’s economy using the Genuine Progress Indicator (GPI) framework, a “GDP+” framework that incorporates environmental and social costs along with traditional economic indicators to gauge economic growth. This study served as a baseline by which to gauge the future economic growth and policy interventions to improve the economy.

As we start working on poverty, we already have two detailed poverty reports at our disposal, the Development Services Agency’s Ohio Poverty Report and the Ohio Association of Community Action Agency’s (OACAA) State of Poverty in Ohio, both published in the past six months. These two poverty reports provide a plethora of data on poverty, food insecurity, housing cost burdens, and other income-related measures disaggregated by race and geography. Between these two reports, we can find most measures of poverty addressed, with the notable exception of relative poverty.

One of the most valuable parts of the OACAA report is a large table with detailed information on the state of poverty in Ohio’s 88 counties. This set of 88 datapoints can help us explore a litany of questions about poverty in Ohio. For this post, let us make just a few surface-level observations about the state of poverty in Ohio.

1. Poverty is not primarily urban in Ohio.

Using a standard linear regression, there is little relationship between population size and poverty level in Ohio counties. Ohio’s largest counties fall fairly close to the average poverty rates and Ohio’s large number of small- and medium-size counties generally range between 5 and 20% like the larger counties.

Above, we highlight three reference counties, Franklin County, a large, urban, medium-high poverty county, Delaware County, a medium-large, suburban, low poverty county, and Athens County, a small, rural, high poverty county.

2. Poverty is not primarily minority-driven in Ohio.

Just as we found little relationship between population size and poverty rates, we also find little relationship between what percentage minority a county is and its poverty rate. Poverty rates show more stability with more minority-heavy counties and are slightly higher in those counties, but hardly significantly higher and lower than many small and medium-size counties.

3. Poverty is weakly related to population growth since 2012

About 15% of the variation in poverty rates in the state can be explained by population growth since 2012. It should be noted that causation could run either way in this relationship: while population growth could mean a stronger economy and more opportunities for residents, it also could be a result of lower poverty levels as people move to opportunity.

Also notable here, though, is outlier counties. Athens County is an outlier in almost every graph we plot because its poverty rates are so incredibly high. A contributor to this is Ohio University: the Census Bureau estimates the Athens County poverty rate would have been 5-10 percentage points lower from 2012 to 2016 if off-campus college students were excluded from the poverty calculation. That being said, even factoring in this adjustment leaves Athens as the highest at worst or one of the highest at best in poverty rates statewide. This is especially surprising considering Athens County is one of the fastest-growing counties in the state.

Franklin County is also impacted by large numbers of college students, though the Census Bureau estimates the impact is only worth 0 to 2.5 percentage points. Even factoring this adjustment in, Franklin County’s poverty rate is much higher than its population growth rate would suggest it should be.

These three tests give us a small peek into what drives county-level poverty in the state of Ohio. Over the next year, we at Scioto Analysis will be working to uncover more of what drives poverty in the state as well as exploring alternative measures of poverty such as supplemental measures and relative poverty measures. If you are interested in talking about state-level poverty and approaches to reduce it, please email me. I’m always looking to connect with folks who are working to understand and reduce poverty at the state and local level.

Thank you to Tong Zhou for data manipulation support that contributed to this post.

Akron Beacon Journal Endorses Refundable Earned Income Tax Credit

On Sunday, the Akron Beacon-Journal and Ohio.com’s editorial staff endorsed making Ohio’s state earned income tax credit refundable.

In their editorial, the newspaper’s editors cited Scioto Analysis’s cost-benefit analysis released earlier this month, particularly citing the public health and education benefits brought to light by the report.

A refundable credit would benefit Ohio overall, a conclusion amplified in a report released by Scioto Analysis last week. For instance, the study finds that a refundable credit at 30 percent of the federal version would result in an average payment of $750. Among other things, the additional resources would lead to more children born at normal weights, reducing the state’s dismaying level of infant mortality. Such a refundable credit would result in more young people succeeding in higher education.

The cost-benefit analysis was a culmination of a summer-long analysis done by Scioto Analysis on the state earned income tax credit. The study was the first best-practices cost-benefit analysis on a statewide program carried out in over a decade.

“Policymakers have a thirst for solid numbers that can help them understand the impacts of the policy options they have,” said Scioto Analysis Principal Rob Moore. “This the first of many new cost-benefit analyses that will be providing vital information to decision makers in the state.”

What will the impact of Ohio’s Recent EITC Expansion Be?

By Rob Moore and Tong Zhou

The federal earned income tax credit (EITC) is one of the most successful antipoverty programs in the United States, lifting 6 million Americans out of poverty each year by giving low-income families a break on their taxes. Seeing the success of the EITC at the national level, 29 state governments have instituted their own earned income tax credits, including Ohio.

In wrangling over this year’s high-profile gas tax increase, some opponents of an increase worried about the impact of a regressive tax increase on low-income families. To offset these impacts, the Ohio General Assembly expanded the earned income tax credit, increasing the federal match from 10% to 30% and nearly doubling the expected size of the state earned income tax credit.

One feature of the current state earned income tax credit in Ohio is that it is nonrefundable, meaning the value of the credit cannot exceed the taxpayer’s tax liability. While this keeps costs of the program down, it also minimizes the impact of the program. Because of this, some have advocated for a refundable earned income tax credit. On top of this, when Scioto Analysis surveyed the Ohio legislature earlier this year, legislators asked about refundable tax credits more than any other topic.

In order to assess the impact of a refundable earned income tax credit, Scioto Analysis built a model this summer, which resulted in the cost-benefit analysis on refundable credits released last week. In this paper, we model the budgetary and wage impacts of the program, then project and monetize four key economic impacts: tax distortions, labor market distortions, birthweight benefits, and college enrollment benefits.

Using LSC estimates for the cost of the EITC expansion in the transportation bill, we can use our model to predict the impacts the change that was made this year. Tables 1 and 2 below show the monetized impacts of the separate policies. Table 3 shows the impact of the change from 10% nonrefundable to 30% nonrefundable.

Basically, by using past studies on the EITC and its impacts on public health and education, we can estimate that the expansion will lead to a $45 increase in wages per low-income worker, about 2,000 new workers entering the workforce, about 5 less instances of low birthweight per year, and 11 new college enrollees per year. This means the change in the EITC in the transportation budget will generate a net benefit of $4 million per year over the status quo 10% nonrefundable policy.

While the recent policy change likely had net benefits, refundability would yield net benefits anywhere from 3 to 15 times higher than the policy change this year. They also would cost the state much more, creating political feasibility issues for such a change in the future.

Tax Credit Refundability Would Yield Economic, Education, and Health Benefits

(Columbus, OH) – On Friday, Scioto Analysis released a cost-benefit analysis of proposals to make Ohio’s earned income tax credit refundable.

The cost-benefit analysis incorporated findings from policy research studying past earned income tax credit expansions and elements of its design to measure human development impacts of the policy.

“In this analysis, we found that refundability reforms would put an extra $150 to $900 in the pocket of the average low-income Ohioan,” said Rob Moore, principal for Scioto Analysis. "It would also bring anywhere from 3,000 to 60,000 new workers into the workforce, would prevent 20 to 120 cases of low infant birthweight per year, and would lead to 40-230 new college enrollments every year, generating anywhere from a $5 to $130 million in new economic activity,”

This analysis sheds new light on the state earned income tax credit, showing for the first time that human development benefits tied to education and health outweigh the tax distortion costs and potential labor market costs levied by the earned income tax credit.

“In all 10,000 simulations we ran of the change from a 30% nonrefundable to a 10% refundable tax credit, we found the change in policy would have net economic benefits,” said Tong Zhou, co-author of the analysis. “This was also the case in all but 7 of the 10,000 simulations we ran of the change to a more robust 40% refundable tax credit.”

This paper represents the first best-practices cost-benefit analysis on a state policy in Ohio in over a decade. Scioto Analysis looks forward to promoting the use of more cost-benefit analysis in the upcoming years.

For more information, please contact Rob Moore, principal, Scioto Analysis, (614) 743-1840, rob@sciotoanalysis.com

Tips for Conducting Cost-Benefit Analysis

By Tong Zhou

Politicians always argue that their policies will benefit citizens and society. However, we often have little information to assess how right they are about these claims. But how do we know when these claims are correct or not?

To create a standardized approach for objectively assess the economic efficiency of public policies, economists created cost-benefit analysis, a tool for weighing the economic costs and benefits of proposed policies. In this post, I will provide some tips for beginners doing cost benefit analysis.

Firstly, let’s look at the cost benefit analysis as laid out in Boardman et. al’s Cost-Benefit Analysis: Concepts and Practice. The following nine steps summarize the process of cost benefit analysis:

1.     Define the policy that we are analyzing and the set of alternative policies that could be adopted in place of the policy. 

2.     Define who has standing in the study, or whose benefits and costs count for the analysis. Sometimes the standing can be the citizens of a city, sometimes the standing can be broader like all human beings on Earth. Standing depends on the specific policy project.

3.     Then, we must determine all the impacts caused by the policy. Any policy will have benefit categories and cost categories just as a coin has two sides. We want to be as objective as possible in this process as biased determination of impacts can make your analysis meaningless. Sometimes we can be guardians and spenders unconsciously. Peer review is a good way to avoid this problem.

4.     Now we can quantify the impacts found in the last step. For example, in the study we are doing this summer, we found that the Earned Income Tax Credit (EITC) will increase the number of workers in the labor market. In this step, we want to determine how many more workers will enter the job market. Sometimes it can be hard to translate impact categories into numbers as translation requires serious analysis and literature review to find an accurate number. You will read a lot of literature to find the best sources of data and effect sizes. Reliable sources play a significant role here: we must stand on the shoulders of giants in cost-benefit work.  

5.     Monetize all impacts. Some people may argue that not everything has a price: things like time and life are often considered “priceless.” However, people make tradeoffs of their time and risks of life all the time. We can survey people to ask them their willingness to pay for these goods. Monetization is a key step in any cost-benefit analysis because it allows us to compare seemingly incomparable outcomes.

6.     Even though we have the monetized all impacts, not all benefit or cost are realized today. In order to get an accurate estimation of present value, or how we value these future outcomes now, we will discount all impacts occurring in the future. The discount rate usually ranges from 3% to 7%, though this is a point of contention in the world of cost-benefit analysis. 

7.     Finally, we have determined the cost and benefits of the policy. Simply subtract costs from benefit to calculate the net benefit. A positive net benefit suggests the policy is economically efficient, while a negative net benefit suggests the policy is economically inefficient.  

8.     Even though we have determined the results, we still want to know how accurate our results are. To do this, we perform sensitivity analysis such as a Monte Carlo Simulation, the best practice in determining the accuracy of cost-benefit analytical outcomes. Sensitivity analysis is a tool that will tell us the most likely net benefit.

9.     The final step is to report your findings in a format that provides guidance to policymakers. 

After carrying out all the steps above, we want to double-check that the policy impacts we found that are directly caused by the policy. We do not want to count correlated impacts that would not bear out as causal in a final report. Even though researchers may find that Nobel Prize winners usually eat more chocolate than normal people, eating more chocolates will not necessarily increase your likelihood to win a Nobel Prize. 

Although the steps listed above seem pretty straightforward, it is very easy to still encounter uncertainties in the model which may undermine the accuracy of the final results. For example, some impacts are more important than others so we might need to weight the impact for policymakers. Prices may fluctuate over time so we need to use the current price. My suggestion is to spend some time reading how others do cost-benefit analysis. Finding other similar examples can not only teach us innovative ways to deal with uncertainties but also let us compare different ways to find the best one.

My final suggestion for beginners is to keep practicing cost-benefit analysis. We cannot truly master it unless we practice. Practice goes a long way and only by refining our craft will we improve at it.

 Tong is a Data Analysis Intern at Scioto Analysis and a data science student at Denison University.