Quick analysis: Do high housing prices cause longer commutes?

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.

Research question

Hypothesis: Higher housing costs are associated with longer commute times. I came across this idea in a newsletter I got this week that was looking at commute times in cities around the world. The author included the following sentence which got me thinking: “Transportation researchers often say a city's average commute tells you almost as much about its housing market as its roads.” 

The logic behind this hypothesis is that when housing costs are high, people have fewer options for housing in their price range. Less housing choices may lead to more people living farther away from where they work, which leads to longer commute times.

Methods

I’m going to use American Community Survey data for every metropolitan and micropolitan statistical area in the United States. I’ll collect their average commute time and their average housing cost and I will check the scatter plots to see if there are any obvious trends. Then, we can do a simple linear regression to see if those trends are statistically significant.

Because of the COVID-19 pandemic, I’ll do the regression on two samples: the 5-year ACS estimates from 2024 and from 2019. I suspect that if there is a trend, it is likely to be much stronger before the pandemic made working remotely significantly more common. 

Results

Both scatter plots appear to have a slight positive trend, though the effect size is likely quite small. Both linear regressions confirm this first glance and find that there is a statistically significant positive correlation between housing costs and commute times. The fact that both of these results are positive and statistically significant does provide some early evidence that our hypothesis is true. The regression using the most recent 2024 data suggests that commute times increase by an average of one minute for every $500 of increased monthly housing costs. Note that the median housing cost in the 2024 data was about $1,000 per month and the max was about $3,000, so there is not a ton of variation on average.

Figure 1: 2024 housing and commute times by metropolitan area

Looking at the 2019 data, we find that commute times increase by an average of one minute for every $250 of increased monthly housing costs. It is hard to tell because the two plots aren’t overlayed but the slope of that two lines is essentially twice as steep implies basically double the impact. This supports the prior hypothesis that the COVID-19 pandemic weakened the correlation between housing costs and commute times. An alternative hypothesis as to why the effect is weaker today than in 2019: rising housing prices nationwide has made the baseline cost for housing higher, so marginal increases in housing prices have less of an effect on commutes than they used to in a 2019 world of lower baseline housing costs. 

Figure 2: 2019 housing and commute times by metropolitan area

Limitations

This is not a causal analysis. This is not enough evidence to claim that higher housing costs cause people to have fewer options when it comes to where they live and therefore live farther away from their work. Notably we aren’t controlling for any other factors that might influence how long commutes are or how high housing costs are.

Another limitation is that I’m just looking at average housing costs and commute times across metropolitan areas. I don’t believe this is necessarily the best way to capture the relationship that I’m interested in, and maybe it would be better to look at individual level data. 

Preliminary conclusions and next steps

Based on these regressions, I would say there is definitely evidence to support this hypothesis and it would be worth exploring further in a more concrete way. A good next step would be to see if there are any historic examples of natural experiments that could be exploited to better identify a causal effect.

What will happen if we adopt permanent daylight savings time?

Last week, the U.S. House passed legislation that would make daylight savings time permanent across the country. If it passes in the Senate, then it would mean the annual tradition for most of the country of switching clocks back and forth twice a year would be officially over.

Back in 2023, we studied what the impacts of daylight savings time were in Ohio. The biggest takeaway from that study was that for Ohio, either permanent daylight savings or standard time was preferable to switching back and forth, but which of those you preferred depended on what the impacts were. Permanent daylight time led to lower crime thanks to more light in the evening, permanent standard time led to energy savings since the sun lines up better with human energy use.

While the average economic outcome might be the same for permanent daylight savings and permanent standard time, a closer look at the issue will tell you that not everyone has the “average” opinion, and a lot of people have reason to prefer one option over the other.

Why some people might oppose permanent daylight savings time

Although the evidence points to the fact that changing clocks twice a year causes a lot of problems to society as a whole, there are people who think that changing to permanent daylight savings time is a negative outcome. For instance, the American Academy of Sleep Medicine strictly prefers permanent standard time and its earlier sunrise. Their position is that across the country, standard time better lines up with our circadian rhythm which in turn leads to many health benefits. AASM has gone as far as to “condemn” this new legislation. 

Another major concern about switching to permanent daylight savings time is that in Northern states, the sun rising so late during the winter might present other challenges. In Minnesota for example, the sun won’t rise until 9:00 am during some times of the year. This is a difficult proposition for children and teachers who will start school in the pitch dark if school times stay the same.

The extent of many of these concerns depends on geography. States farther north experience shorter winter days, making the effects of permanent daylight savings time more noticeable than they would be in southern states. Similarly, the western portion of each time zone already experiences later sunrises than the eastern portions. A one-size-fits-all national policy inevitably produces different outcomes depending on where people live.

Other policy options

One thing local governments could do if they are forced to adopt permanent daylight savings time is move the start times of their schools. Schools could even go as far as twice a year changing the start time by an hour, moving back in winter and moving it forward again in spring. 

One alternative that isn’t readily available but might make the most sense is to just let states pick what timezones they want to be in. The timezone map is already full of oddities, and this would give states the flexibility to choose specifically when they want their sunlight to be. If you wanted to take this a step further, let states choose in half-hour increments! It’s only weird until people get used to it.

Any sort of significant restructuring to how we keep our time is an opportunity for creative changes to other parts of our daily schedules. Our society relies on clocks to tell us where to be when, and we’ve had to adjust our bodies to respond to changes in the numbers. The point of all of this is at least in theory to help line up our lives with the sun. Maybe settling on a fixed time for the whole year can be an opportunity to rethink when certain things start.

Should Ohio freeze utility rates for a year?

Earlier this month, Ohio state Rep. Desiree Tims introduced a bill to freeze utility rate increases in Ohio for the next year.

Tims’s bill comes in a time of increasing concern about affordability. According to the Energy Information Administration, the residential price of electricity per kilowatt hour in Ohio grew 29% from 2019 to 2024.

Utility prices are an element of family budgets where the state has a clear lever for controlling costs.

Since electricity companies are regulated entities, they negotiate with the state public utilities commission to set prices for their ratepayers. The state has the ability to freeze these prices for a year if it wishes to do so.

In a lot of ways, this sort of policy change is like a sales tax holiday, which the state of Ohio has embraced in the past, or a gas tax holiday, which legislators have bandied about recently to offset gas price increases endured by Ohio in the wake of the U.S. war with Iran.

It represents a temporary suspension of a cost for ratepayers to provide relief in the short-term.

In the short term, a utility freeze would help low-income households.

According to the United States Department of Energy, the average low-income household in Ohio spends about 11% of their income on utilities compared to 3% for a median-income household and 1% for a high-income household.

While a utility freeze would be less targeted than an earned income tax credit, it would likely disproportionately benefit low-income households.

Energy insecurity can lead to unsafe health outcomes for households.

As many as 1 in 10 U.S. households report keeping their household temperature at an “unsafe” level due to energy prices.

A utility freeze could help households stabilize their finances without forcing them to cut back on heating or cooling their homes.

Stability in energy prices could even lead to stabilization in home environments that could have spillovers into educational outcomes, like what we’ve seen with federal energy assistance programs.

While low-income households spend a larger percentage of their income on energy, evidence from the Energy Information Administration suggests the highest-income households consume energy at nearly twice the rate of the lowest-income households.

This means upper-income households could be saving twice as many dollars from a utility rate freeze as low-income households, reducing the inequality impact.

The biggest problem with a rate freeze is that it could lead utilities to defer maintenance of existing infrastructure, which could impact reliability of electricity service and grid modernization efforts.

This could lead to more blackouts, especially as growth in energy demand is leading to more users taxing the power system.

Ultimately, this policy represents a bet that utility increases over the next year will not be used for essential maintenance, efficient grid modernization, or reliability upgrades that outweigh the short-term benefit the freeze would have for household incomes, especially for low-income households.

If Rep. Tims is right, this could mean better outcomes for households at little cost to the public.

If she is wrong, the short-term benefits would come at the cost of a quicker degradation of the grid, which means more blackouts and could lead to higher costs in the long run. 

While there is part of me that wants to see if her bet is correct, there is another part of me that thinks a more targeted policy, like an earned income tax credit expansion, would achieve the same boost to household income without the potential costs to the grid.

We’ll see if the General Assembly has an appetite for either.

This commentary first appeared in the Ohio Capital Journal.

Where are Ohio’s data centers headed?

Data centers have been one of the biggest news stories of the past year. Last month, my colleague Rob wrote about this trend: everyone is talking about data centers. For community members, worries about the environment, electricity prices, and societal change have become widespread. For developers and investors, the exponential growth of data centers across the United States is generating excitement.

In Ohio, data center construction has received a lot of pushback from local residents. Many people don’t want to see more data centers being built near their homes. For instance, in Columbus, residents are urging local lawmakers to require large data centers to submit water conservation plans and create decommissioning plans. Some data center proposals have been rejected across the state and many communities are considering pausing data center development entirely (a proposal we asked Ohio economists about a few months ago). 

Earlier this week, the Ohio Environmental Protection Agency announced it would not move forward with a permit that would have allowed data centers to discharge wastewater into natural water sources after receiving more than 7,000 comments on the proposal from the public. With all of this talk about data centers, I wanted to take a deeper dive into the existing and proposed data centers across the state. What kind of facilities already exist, and where are they located?

Where are data centers in Ohio?

Using data from DataCenterMap and Cleanview, we compiled a list of all existing and proposed data centers across Ohio. We found that there are 151 existing data centers in Ohio and an additional 143 data centers that have been proposed or are in development. We count each building as a separate data center, meaning that for larger data center campuses that include multiple facilities, each building is being counted separately. Below is a map of where the existing and proposed data centers are across Ohio. Seven planned data centers are omitted from the map due to insufficient data about the locations of these facilities. 

Data centers are most concentrated in Licking County and Franklin County in Central Ohio, where there are 66 and 64 existing and planned data centers respectively. Franklin County has a lot of existing facilities, a total of 50, while 38 of the 66 facilities in Licking County are planned or proposed. 23 of the 38 planned facilities in Licking County are from Amazon Web Services alone. Between 2015 and 2024, Amazon invested $6.2 billion in its Amazon Web Services data center infrastructure in Licking County. 

The next largest county for data centers in Ohio is Fayette County, home to Washington Court House and Jeffersonville. Fayette County doesn’t have any existing data centers, but has 23 planned and proposed data centers. All 23 of these data centers are from Amazon Web Services, for a total of nearly 1,800 megawatts of power (which would power 1.44 million homes, 125 times the number of households in Fayette County). Early last year, Amazon purchased 590 acres of land in Fayette County for just over $100 million to build these 23 facilities. They are expected to be completed within the next one to four years.

Similar to Fayette County, Portage County, home to Kent, Ravenna, and Streetsboro, has no existing facilities but 14 planned data centers. All 14 facilities will be part of one campus in Shalersville run by Bitdeer, a cryptocurrency mining and artificial intelligence cloud infrastructure company. The Bitdeer campus will be at least 650 megawatts (which would power 520,000 homes, 8 times the number of households in Portage County) and 257 acres of land. Bitdeer is expected to open its first two facilities in 2029 and the remaining 12 facilities in 2031.

How fast are data centers growing in Ohio?

The first data center in Ohio opened in 1966. Through the twentieth and early twenty-first century, data center growth was steady. Starting in 2017, data center growth became exponential. For reference, the first large language model (the first GPT model from OpenAI) was released in 2018, and ChatGPT released publicly in 2022.

If anything, the next several years is likely an underestimate, as there may be more data centers that will be planned and developed in Ohio in the next several years that aren’t yet announced (unless Ohio residents vote for a ban on data center development).

In addition to the total number of data centers in Ohio, the size and energy capacity of data centers is increasing exponentially as well. The number of hyperscale facilities, defined as data centers that take up at least 100,000 square feet of physical space and an energy capacity of 100 megawatts or more, is growing rapidly.

By 2034, Ohio is projected to have at least 36,000 megawatts of data center energy capacity across the state. 36,000 megawatts could power 28.8 million homes, which is 6 times the number of households in Ohio, and 21.5% of the number of households across the entire United States.

As data center growth continues, development will be a balance of competing interests: demand for digital services and artificial intelligence, environmental concerns, and local community priorities. Navigating these trade-offs requires us to weigh immediate capital gains against broader societal external costs and long-term resource availability. Evidence-based policy analysis is the key to understanding these tradeoffs moving forward.

This fall, Scioto Analysis will be releasing a cost-benefit analysis of existing and planned data center projects in Ohio to give policymakers and communities a clearer picture of the economic and environmental impacts of data centers across the state.

Ohio Economists think university consolidation would threaten Ohio’s economic future

In a survey released this morning by Scioto Analysis, 11 of 14 economists agreed that reducing the number of universities in Ohio would reduce long-term economic output in the state.

Earlier this year, candidate for Ohio governor Vivek Ramaswamy proposed shutting down some public colleges and universities in Ohio if he believes they are subpar. Over the past decade, enrollment at Ohio’s public universities has declined by 14%. According to data from the United States Census Bureau, Ohio has an aging population, and more students are turning away from college due to financial reasons.

11 of 14 economists agreed that reducing the number of universities in Ohio would reduce long-run economic output in the state. According to Albert Sumell of Youngstown State University, “A number of economic studies have shown the biggest determinant of long term economic growth is the percentage of the population with a bachelor's degree - which is directly related to the number of universities.” The remaining 3 economists were unsure of the impact of public college closure on long-run economic output in Ohio.

Economists were perfectly split on whether reducing the number of universities in Ohio would generate savings for the state of Ohio, with 6 economists agreeing, 6 economists disagreeing, and 2 economists uncertain. According to David Brasington of the University of Cincinnati, “There would be direct cost savings, but only about 6% of the state's general revenue fund currently goes to higher ed anyway.” The economists who disagreed explained that closing public universities in Ohio is short-sighted, and while it may generate savings in the short-run, it doesn’t address the long-term issues with higher education in Ohio.

12 of 14 economists agreed that reducing the number of universities in Ohio would reduce Ohio’s bachelor degree attainment rate. Kevin Egan of the University of Toledo explains, “80% of the Univ. of Toledo students live off campus and commute to UToledo. This is similarly true for many of the Ohio public universities. Closing any one of them will decrease the local cheaper public university option and some will not attend college.” The remaining 2 economists were unsure of the impact of public college closure on bachelor degree attainment in Ohio.

The Ohio Economic Experts Panel is a panel of over 30 Ohio Economists from over 30 Ohio higher educational institutions conducted by Scioto Analysis. The goal of the Ohio Economic Experts Panel is to promote better policy outcomes by providing policymakers, policy influencers, and the public with the informed opinions of Ohio’s leading economists. Individual responses to all surveys can be found here.

What makes policy analysis credible?

Earlier this month, I wrote a blog post reflecting on Eugene Bardach’s A Practical Guide for Policy Analysis: The Eightfold Path to More Effective Problem Solving. I wrote about the “Grandma Bessie Test,” a strategy Bardach recommends to make sure we understand the conclusions from our work well. To conduct the “Grandma Bessie Test,” we try to explain our data, findings, and conclusions to an intelligent but non-expert audience (like our Grandma Bessie) in about one minute.

Bardach’s eightfold path is a book that we reference a lot at Scioto Analysis. In some ways, it’s the very bread and butter of our work as policy analysts. The eightfold path is an eight-step framework to help us understand public policy and conduct policy analysis more effectively. The eight steps in the framework are as follows:

1. Define the Problem

2. Assemble Some Evidence

3. Construct the Alternatives

4. Select the Criteria

5. Project the Outcomes

6. Confront the Trade-Offs

7. Decide!

8. Tell Your Story

As Bardach develops these eight steps throughout his book, an important theme of credibility emerges repeatedly. Today, I want to focus on an important theme that Bardach references throughout his book: what makes an analysis credible, and what makes one not credible?

What makes policy analysis credible?

When we’re doing policy analysis, we should be thinking about credibility multiple times throughout the analysis. Credibility means selecting trusted sources, creating well-researched methodologies, and fact-checking our analysis at the end. The following considerations can help ensure our work is credible.

Defending against Politically Inspired Criticism

Defending against politically inspired criticism is the first strategy Bardach recommends to protect credibility. If we think there might be parties who will oppose or attempt to undermine our work, then we should touch base with them before publishing our analysis. 

We can do this by having conversations or conducting interviews with parties who might be opposed. To bolster our analysis, we can include quotes from opposing parties. In addition, we should always be thorough with our analysis: we shouldn’t conduct analysis that heavily favors one side without considering the other.

Ultimately, conducting evidence-based analysis with data and statistics may be the best way to defend against politically inspired criticism. If our findings are grounded in data and statistics, then critics cannot so easily sweep these findings aside as mere personal bias.

Preparing for Premature Exposure

The second strategy Bardach recommends to protect credibility is to prepare for premature exposure. Research and policy analysis can be a slow process. Throughout our analysis, we should be prepared for unexpected deadlines or sudden windows of opportunity. For example, reporters might reach out suddenly looking for a quote on a certain policy. Alternatively, an event might pop up that our analysis would be useful for.

To help prepare for these opportunities without sacrificing credibility, Bardach recommends that we spend time early in our analysis figuring out the answers to the most basic, crude questions that people are bound to ask. That way, if opportunities pop up throughout our analysis, we can share information without overstating our analysis.

Sensitivity Analysis

One of my favorite parts of performing policy analysis is sensitivity analysis. Sensitivity analysis is the process of estimating how assumptions included in an analysis impact the uncertainty around the findings of the analysis. Conducting sensitivity analysis is an exercise in testing the precision of our results. 

Bardach recommends we ask ourselves the question, “how big a mistake can I afford in this assumption before this analysis is in really big trouble?” Sensitivity analysis helps answer that question by forcing us to look at our assumptions and determine where results might be vulnerable. By conducting sensitivity analysis, we can prove that our results aren’t just based on a couple of key assumptions.

What makes policy analysis not credible?

Understanding how to produce credible analysis is important, but so is recognizing the pitfalls that can undermine its credibility. Bardach points out that some of the biggest problems with a project’s credibility come from carelessness and letting our own opinions get in the way of the data.

The quickest way to destroy our credibility is by allowing our preferred solution to interfere with how we define the problem. For instance, if we define a problem as “there is too little funding for this program”, we have already assumed the answer to the problem before doing any research. To ensure credibility, we should leave the door open for all potential solutions.

We can also kill our credibility by letting optimism get in the way of our analysis. Bardach warns us not to count on one of our alternatives working out just because we expect it to. Instead of assuming results, we should rely on data to show us what the truth of a certain policy really is.

How do analysts decide what not to count in a cost-benefit analysis?

An underrated skill policy analysts need to have is understanding when something should not be included in an analysis. It often feels safer to always want to include more, but managing the balancing act of deciding when enough is enough can lead to better results.

How standing can limit results

One reason an analyst might choose to exclude something from their analysis is because they decide that it doesn’t have standing. Standing is the principle that a cost or benefit should only be included if it accrues to the group whose welfare is being evaluated. While this sounds straightforward, it can dramatically change the scope of an analysis. 

Suppose a state government is evaluating a transportation project using a state-level cost-benefit analysis. The primary question is whether the project makes the residents of that state better off. If the project causes travelers from another state to save time, those benefits are certainly real, but they are not benefits to the population being evaluated. Likewise, if a policy simply shifts economic activity from one state to another, the gains in one jurisdiction are offset by losses in another and may not represent a net benefit from the perspective of the analysis. 

For more information about standing, check out our State Handbook of Cost-Benefit Analysis.

When further analysis costs too much

Another reason an analyst might decide not to include something is because the marginal benefit of more research doesn’t meet the marginal cost. As with any project, there comes a point where additional effort produces very little additional value. Good policy analysis is not about measuring every conceivable impact, but about measuring the impacts that matter most. 

Analysts can hone their intuition when it comes to prioritizing which inputs to measure. For example, it is probably always worth exploring a mortality impact if the policy you are analyzing might have one since reductions in mortality are often extremely valuable. Conversely, if you think that an impact is probably not very valuable, then it’s worth using your limited time and resources to explore other impacts first.

If the data isn’t good enough

One final reason you might want to exclude something from an analysis is if the quality of the data isn’t good enough to support something quantitatively. Nearly every policy has effects that are difficult to measure, and there is often a temptation to assign numbers to those impacts simply because they seem important. Unfortunately, a quantitative estimate built on weak evidence can be more misleading than acknowledging that the effect cannot yet be measured. 

That does not mean these impacts should be ignored entirely. In many cases, the appropriate approach is to discuss them qualitatively while being transparent about the limitations of the available evidence. This allows decision-makers to consider potentially important effects without giving a false sense of precision. 

Note that poor data quality and high variance estimates are not the same thing. At the start of this year, the Trump administration stopped counting the health benefits of environmental policies, claiming that the monetization step was too uncertain. What they should have done if they were concerned was spend more time unpacking how that uncertainty might affect results through sensitivity analysis.

Ultimately, deciding what to leave out is just as important as deciding what to include. Every policy analysis involves limited time, limited data, and limited resources. The best analysts recognize these constraints and focus their efforts where they can provide the greatest insight. By carefully considering standing, weighing the costs and benefits of additional research, and being honest about the quality of the available evidence, analysts can produce work that is not only more efficient, but also more credible.

What is “social welfare?”

At Scioto Analysis, we perform policy analysis through an economic lens. While that is a true statement, it doesn’t do a great job of actually describing our work. You could use that exact sentence to describe what organizations like the World Bank or the Federal Reserve do in their research. I bring this up because I wanted to describe in more detail what it is we actually do at Scioto Analysis and what some of the theoretical underpinnings are.

A better description of our work would be something like this: we perform policy analysis at the state and local level using microeconomic tools rooted in the ideas of welfare economics. Policy analysis at the state and local level is fairly self explanatory, that is just an issue of scope. The more interesting part to discuss and what I want to explain in more detail today is what welfare economics is and why it has the tools that we like to use when it comes to policy analysis.

What is welfare economics?

Welfare economics is a branch of microeconomics that focuses on measuring how a society as a whole allocates its resources and how well-off that makes its local residents. Microeconomics gives us the toolbox to evaluate individual decision making, and welfare economics provides the framework to make assessments about how those individual decisions impact society as a whole.

One of the challenges in welfare economics is that there is no objective way to measure or compare one person's wellbeing against another's. There is the field of subjective wellbeing research and that data could be an extremely valuable input into a welfare calculation, but we’d need more and better data to really dig much deeper.

 Because of this, welfare economics relies on assumptions about how individual wellbeing should be aggregated into a measure of society's overall wellbeing. Different assumptions lead to different conclusions about which policies are desirable, which is why economists have developed a variety of social welfare functions that reflect different views about fairness, efficiency, and the tradeoffs between them. 

Types of social welfare functions

While you could technically have any social welfare function you wanted, in order to be useful it should probably use the individual welfare functions of everyone in a society as an input. Most commonly, we use some form of a utilitarian social welfare function as our starting point.

Under a simple utilitarian social welfare function, the total wellbeing of a society is calculated as the sum of all individual welfare functions. This has some useful features that lend themselves to cost-benefit analysis. If we are agnostic about who in society is benefiting from a change, then it makes sense that we’d prefer policy options so long as the net change in individual wellbeing is positive. 

A natural extension of the utilitarian model is to account for diminishing marginal returns of income. In other words, $100 given to a millionaire doesn’t increase their wellbeing by very much at all, while that same $100 given to a person in poverty could significantly improve their wellbeing. The end result is that taking $100 from a super wealthy person and giving it to a poor person increases total social welfare.

One of the most important contrasts to utilitarian social welfare is the Rawlsian social welfare function. This concept is based on the ideas of philosopher John Rawls whose work centered around ideas of social justice. The economic framework named after him is that a society is only as well off as its least well off member. Mathematically, this is represented as Social Utility = min(U1 ,  … , Un). Under this framework, a policy change would only be worth it if it benefits the least well off person.

Having a well defined social welfare function is an essential part of policy analysis. It is the heart of how we assess the tradeoffs of a policy. This is not to say that this is the only approach for policy analysts to use. Some evaluations might not explicitly rely on welfare economics, and reasonable people can disagree about which social welfare function best reflects society's values. At Scioto Analysis, however, welfare economics provides the foundation for much of our work because it offers a consistent framework for comparing the costs and benefits of different policy options and making those tradeoffs explicit.

Which states have the highest pre-k enrollment?

Last year, we released a cost-benefit analysis about universal prekindergarten in Ohio. We found that universal prekindergarten in Ohio could lead to between 9,300 and 29,000 new children enrolled in prekindergarten and between $220 million to $750 million in net economic returns. These benefits include increased lifetime earnings, reductions in criminal justice costs, and savings in special education.

Universal prekindergarten means that all families have the option of enrolling their children into a universal prekindergarten program that is publicly funded and accessible to all. Usually, universal prekindergarten is limited to four-year olds and sometimes three-year-olds. As of now, only four states (Colorado, Florida, Oklahoma, and Vermont) and Washington D.C. have universal prekindergarten. Another eight have universal eligibility programs that do not necessarily provide prekindergarten to every eligible child. Today, I want to look at prekindergarten enrollment by state as reported in the American Community Survey to see if they align with the states that offer universal prekindergarten programs.

Pre-K Enrollment Across the United States

The following table shows the number of children enrolled in prekindergarten by state. The table is ordered by the percentage of children enrolled in prekindergarten in each state.

Pre-K Enrollment by State
State	Number of Children Enrolled in Pre-K	Total Children Aged 3-4	Percentage of Children Enrolled in Pre-K
Vermont	8,646	12,164	71.08%
New Jersey	150,741	220,240	68.44%
New York	278,601	436,513	63.82%
Connecticut	45,537	72,133	63.13%
Mississippi	37,847	64,986	58.24%
Massachusetts	85,351	148,874	57.33%
New Hampshire	15,923	28,253	56.36%
Illinois	148,149	269,663	54.94%
Wyoming	6,232	11,382	54.75%
Colorado	65,653	122,308	53.68%
Louisiana	61,553	114,828	53.60%
Florida	254,189	493,696	51.49%
Kansas	34,515	67,543	51.10%
California	430,286	853,418	50.42%
Pennsylvania	139,085	277,992	50.03%
Rhode Island	12,039	24,217	49.71%
Maryland	69,906	143,001	48.88%
Michigan	107,189	220,343	48.65%
Georgia	121,366	249,671	48.61%
Virginia	96,086	197,701	48.60%
South Carolina	55,752	115,622	48.22%
Maine	12,350	25,669	48.11%
Texas	386,523	808,459	47.81%
Minnesota	61,667	129,083	47.77%
New Mexico	19,092	40,582	47.05%
Washington	78,144	167,498	46.65%
Delaware	10,018	21,504	46.59%
Oregon	37,732	82,534	45.72%
Missouri	64,800	141,853	45.68%
Alabama	51,516	114,660	44.93%
Iowa	34,497	77,108	44.74%
Arkansas	31,442	70,715	44.46%
Wisconsin	57,890	131,336	44.08%
Ohio	122,176	281,087	43.47%
Utah	42,343	97,649	43.36%
North Carolina	103,385	238,779	43.30%
Nebraska	21,386	49,735	43.00%
Montana	9,475	22,284	42.52%
Tennessee	68,684	168,220	40.83%
Hawaii	12,701	31,275	40.61%
Oklahoma	39,248	99,149	39.58%
Indiana	66,128	169,134	39.10%
Nevada	28,342	72,874	38.89%
Alaska	7,735	20,325	38.06%
Arizona	61,233	161,159	38.00%
Kentucky	41,187	108,996	37.79%
South Dakota	8,248	21,921	37.63%
West Virginia	13,611	36,536	37.25%
Idaho	17,215	46,570	36.97%
North Dakota	6,582	18,345	35.88%

The American Community Survey asks families if their children attended school at all in the last three months, so this data likely overestimates actual prekindergarten enrollment rates. The top five states for prekindergarten enrollment are Vermont, New Jersey, New York, Connecticut, and Mississippi, ranging from 58% to 71% prekindergarten enrollment. This means that only one state with universal prekindergarten, Vermont, ranks in the top five states for prekindergarten enrollment. The other three states with universal prekindergarten, Colorado, Florida, and Oklahoma, rank 10th (54% enrollment), 12th (51% enrollment), and 41st (40% enrollment) respectively. 

Colorado and Florida still rank relatively high for prekindergarten enrollment, while Oklahoma is in the bottom 20% of states. One reason could be that Oklahoma’s universal prekindergarten program is only for four-year olds, so prekindergarten enrollment for three-year olds is likely much lower than for four-year olds.

What about the top five states for prekindergarten enrollment? What are the factors driving enrollment among these states?

Vermont

About 8,600 of 12,100 children aged 3-4 years old in Vermont are enrolled in school, accounting for about 71% of the population. Vermont’s universal prekindergarten program requires at least 10 hours of prekindergarten to be available for all children aged 3 to 5 years old in Vermont for 35 weeks in a school year.

Vermont is also the second smallest state by population and the sixth smallest state by land area, so administering universal prekindergarten and boosting enrollment is more seamless than for larger states. Vermont’s high prekindergarten participation rate may contribute to future educational attainment as well, as Vermont is the third-highest state for bachelor’s degree attainment with a bachelor’s degree attainment rate of 45.1%, only ranking behind Massachusetts and Colorado.

New Jersey

About 151,000 of 220,000 children aged 3-4 years old in Vermont are enrolled in school, accounting for about 68% of the population. New Jersey does not have universal prekindergarten, but they have an expansive prekindergarten program available for vulnerable populations. New Jersey also heavily invests into prekindergarten programs. The state spends nearly $19,000 per child on preschool education, ranking second in the nation, only behind Washington, D.C. Alongside New York and California, New Jersey is just one of three states that spends more than $1 billion on preschool annually.

Similar to Vermont, New Jersey is the fourth-smallest state in terms of land area, but they are the eleventh largest state in terms of population. In fact, New Jersey is the most densely populated state in the nation. A denser population historically results in higher participation in public programs, which may partially explain why New Jersey’s prekindergarten participation rate is so high.

New York

About 279,000 of 437,000 children aged 3-4 years old in New York are enrolled in school, accounting for about 64% of the population. In New York City, which accounts for close to half of the entire state’s population, all four-year olds already have access to prekindergarten. Prekindergarten access and New York City’s dense population are both key reasons for its high prekindergarten enrollment rate.

Earlier this year, New York Governor Kathy Hochul announced that by the 2028-2029 school year, New York will begin offering full universal prekindergarten for all children under 5 years old. This policy change is likely to increase New York’s prekindergarten enrollment rates even higher.

Connecticut

About 46,000 of 72,000 children aged 3-4 years old are enrolled in school in Connecticut, or about 63% of the population. Connecticut joins the first three states as another state located in the Northeast with high prekindergarten enrollment. Similar to Vermont and New Jersey, Connecticut is the third smallest state in terms of land area with a medium-sized population. Connecticut’s small size might make it easier for families to enroll in prekindergarten. 

Connecticut doesn’t currently offer universal prekindergarten. However, Connecticut has recently been making strides to make prekindergarten more accessible for vulnerable populations, and some legislators have proposed plans for universal prekindergarten within the next several years. Connecticut also ranks ninth in median household income in the country at about $99,240. With more income, Connecticut households may be able to enroll their children in prekindergarten at higher rates.

Mississippi

About 38,000 of 65,000 children aged 3-4 years old in Mississippi are enrolled in school, accounting for about 58% of the population. Being located in the South, Mississippi is a deviation from the Northeast trend of the rest of the highest ranking states for prekindergarten enrollment. 

According to the National Institute for Early Education Research, Mississippi is one of six states in the country to meet 10 of 10 benchmarks for preschool quality. Mississippi doesn’t have universal prekindergarten, and state spending on prekindergarten is only $4,832 per child, ranking 35th in the nation. So, why does Mississippi have such high prekindergarten enrollment rates?

Prekindergarten in Mississippi is mostly administered through Early Learning Collaboratives. Early Learning Collaboratives are state-funded pre-K programs which include school districts, Head Starts, and licensed childcare providers. Communities can receive state funding for Early Learning Collaboratives by forming collaborative partnerships between school districts and childcare providers. This means that many prekindergarten programs in Mississippi are offered through a partnership between public school districts and private childcare providers. Mississippi also offers a 1:1 state tax credit for individuals or corporations who donate funds to Early Learning Collaborates for up to $1 million, which encourages private entities to increase access and quality for prekindergarten in their communities.

The future of pre-K in the states

As time passes, more states are beginning to offer universal or subsidized prekindergarten programs. Universal prekindergarten can increase prekindergarten participation rates and create future financial savings for the state and households alike. However, as Mississippi has shown, universal prekindergarten might not be the only answer for states and communities to improve prekindergarten quality and access.

Federal bill would open the door for Ohio housing reform

Last week, U.S. Congress sent the 21st Century ROAD to Housing Act to the White House for signature, the most significant piece of bipartisan legislation of President Trump’s second presidency and the largest housing affordability measure in a generation, though Trump is refusing to sign it.

The bill makes a number of policy changes to increase home supply like streamlining federal regulations, restricting institutional investors, and incentivizing local governments to lower housing barriers.

This follows an unprecedented rise in housing costs fueled by antiquated construction rules, underbuilding during the Great Recession, and rising inflation.

Though Ohio has enjoyed a relatively affordable cost of living for decades, the state has not been immune to these national housing trends.

The average Ohio house cost 65% more in January 2026 than it did in January 2020.

Without enough homes to meet the country’s growing demand, prices are going up and, with them, the median age of first-time homebuyers which now tops 40 years old. 

But Ohio doesn’t just have a housing supply issue. It also has a housing choice issue.

Increasingly, buyers do not have the freedom to choose the type of home that best fits their family’s unique wants and needs, constrained by local zoning regulations that often ban historic housing styles. 

In Ohio, only 3% of all homes are duplexes which Dr. Jason Reece, professor of City and Regional Planning at the Ohio State University, has called America’s “original workforce housing” because of their natural affordability and power to build family wealth through sweat equity. 

These so-called “missing middle” homes — duplexes, triplexes, and small townhomes — previously played a crucial role in expanding homeownership opportunities to middle-income families.

In the past, a young family that purchased a duplex could use their spare unit to generate extra rental income. They could also offer decent housing to family members at the start of their career or in their twilight years. This not only kept expenses low but also made family caregiving easier.

In theory, this path to the middle class is still possible.

In a recent analysis my firm Scioto Analysis conducted with the Affordable Housing Alliance of Central Ohio, we estimated a typical family can now earn or save nearly $1.3 million over a lifetime through duplex homeownership. And that doesn’t include the value fo the equity that the home itself builds over time. 

But Ohio families can only benefit from a duplex if they can find one.

We also mapped the zoning code rules in all eleven of central Ohio’s counties and estimated that duplexes were prohibited on 79% of all parcels.

In some communities, including those facing intense job growth pressures, they’re banned entirely. 

This isn’t just a problem in Ohio.

Across the country, duplexes and other traditional styles have gradually been removed from the marketplace.

Consumers have more freedom than ever to choose the kinds of food they eat or the places where they get their news, but they have fewer housing choices than their grandparents. 

From Maine to Montana to Arizona, and many places in between, cities and states are confronting this challenge with laws aimed at re-legalizing the duplex and restoring that tradition of choice. 

The 21st Century ROAD to Housing Act offers Ohio an opportunity to join them.

In particular, local communities will soon be able to learn zoning and land-use “best practices” from national research and case studies, compete for planning grants to localize these strategies, and seek a share of more than $200 million in competitive innovation funds to make these plans a reality. 

The state can accelerate this work, too, with its own reforms and supports, many of which were already laid out in the Ohio Senate Select Committee’s “Housing Reimagined” report and the Home Matters to Ohio plan.  

Each of these tools would make building housing in Ohio easier, which will improve prices and quality while expanding meaningful resident choice. And that will benefit everyone.

This commentary first appeared in the Ohio Capital Journal.