In December 2009, the education department head, Professor Kenneth K. Wong, another graduate student and myself were part of a three-person team consulting the Rhode Island Department of Education (RIDE) on how to establish a new state funding formula. We worked with finance and legal staff at the department to develop the legislation for the 2010 session that would establish a state funding formula for the first time in 15 years. 
 The Board of Regents had already passed a resolution with its policy priorities that they wanted enshrined in the formula. Additionally, there had been many attempts over the past 5-10 years to pass a new formula that failed for various reasons, chief among them that all previously proposed formulas were accompanied with a call to increase state funding for education 30-50%, with some even envisioning nearly doubling the state education funding. Our task was to research funding formulas, both in practice in other states and in the literature research on school finance, and achieve the goals of the Board of Regents without proposing a mammoth increase in state aid that would sink the entire endeavor 1. The general sense was that while more state aid had the potential to improve the progressiveness of education expenditures, the reality is the overall spending level in Rhode Island is high, and introducing new money was less important than redistributing state aid. I share this belief, particularly because I think adding money to the right places is simple once there is already a way to equitably distribute those funds. Tying up the increase in funding alongside a distribution method is a recipe for political horse-trading that can result in all kinds of distortions that prevent aid from flowing where needed.


My role in this process was primarily to create Excel simulators that would allow us to immediately track the impacts of changing different parts of the formula. I also helped RIDE staff interpret the meaning of changes to the math behind the funding formula and understand what levers existed to change the formula and how these changes impacted both the resulting distribution and policy.


We had three months.


There are a lot of people who are unhappy about the results of the formula that ultimately passed in June 2010. Because we are redistributing essentially the same amount of state aid, there are some districts that are losing money while others are gaining funds 2. Some dislike the fact that we used only a “weight” for free and reduced price lunch status. Alternative formulas (and formulas in other states) typically include numerous weights, from limited English proficiency and special education status, to gifted and talented and career and technical education 3. And yet others were displeased that many costs, including transportation and facilities maintenance, were excluded from the state-base of education aid. Then there are those who think the transition is all off– five years is too long to wait to get the increases the formula proposes, and ten years is far too fast to lose the money the formula proposes. 4


A Good State Aid System


In the end, I am proud of the formula produced for several reasons.
 First, it passed successfully and has been fully funded (and sometimes more than fully funded) each year of implementation throughout a period of massive structural budget deficits. This is no small accomplishment. The advocacy community rightfully pushes us to build ideal systems, but in our role of policy entrepreneurs we are faced with the reality that a policy that does not become law and is not supported as law may as well not exist. Producing a formula that has some of the other positive qualities discussed below, passing that formula, and implementing the formula with little fanfare is not a small accomplishment.


Second, the formula is highly progressive, sending as much as 20 times more aid to some of our poorest communities in Rhode Island compared to the wealthiest. I am not positive how this compares to other states– that’s a topic I certainly want to work on for a future post– but with just 39 cities and towns, it seems to show a high preference for vertical equity, treating different cities and towns differently. There are communities on both ends of the distribution who want substantially more state funding, and our state aid formula is not sufficient to effectively crowd out local capacity for education spending and ensure that our poorest communities are spending more than our wealthier ones 5, but it’s a very strong start.


Third, the formula is relatively simple. While I do not necessarily agree that it is a virtue to have fewer weights and a simple formula in perpetuity, the experience with other states and other formula-based programs show that weights and complexities are very easy to add and very hard to take away. Once a particular policy preference is enshrined in the distribution method, it had better be right because a community of advocacy will maintain that weight long into the future. Personally, I felt it critical to start with the very simple “core” formula that could be adjusted over time. I have some ideas on how I might modify/add to this core that I will be sharing in this post, but I firmly believe that starting with a simple core was the right move. It is also worth noting that because of the need to ensure the transition is smoothed out so that gains in some districts equal the total losses in the others meant that even a more progressive weighting scheme would not impact school funding until the far back end of the transition period (which we proposed as 7 years but was pushed to 5 years during the legislative process), since communities were already gaining funds as fast as we could move them. For this reason, not only was a simple core preferable from my technocratic perspective, but it also was not likely to have any immediate downside.


Fourth, we removed the long-term regionalization bonuses. Rhode Island had sought to reduce the ridiculous number of school districts by providing a bonus for regionalizing in the early 90s. Unfortunately, because of the timing of the abandonment of the previous state aid formula, the districts that did choose to regionalize had their base funding locked in at a level 6-8% higher than it should have been, because they were receiving a bonus that was meant to fade away over the course of several years. I could justify a small increase in state funding to pay some of the transition expenses of regionalizing districts, but long term funding increases? Part of the goal of regionalization is the reduction of overhead that allows for decreased costs (or increased services at the same costs). There is no ongoing need to supply a massive state bonus for regionalizing.


Now just because I am proud of this work does not mean that I think we have “solved” education funding in Rhode Island. Personally, I believe there are other defensible ways to distribute funding in Rhode Island, each of which represents slightly different policy preferences. There is no hard and fast “right” or “wrong” way to do this, within certain guidelines. As I see it, so long as the formula is progressive and moving toward a greater chance of seeing a day where Providence has the highest paid staff in the state 5, we are on the right path. I don’t believe that Rhode Island will have a truly “great” education finance climate without a substantial growth in the economy or a huge new tax that dramatically lowers the ability of municipalities to generate school funding while bolstering state aid. However, I think we have a great foundation and a “good” system.


For the remainder of this post, I would like to propose a few ideas that could help move Rhode Island from “good” to “very good” that I think are feasible within the next five years 6.


A Very Good State Aid Program


After a little over three years since its establishment, I think we are ready to tackle several additional aspects of state education funding in Rhode Island. One thing you may notice is that few of these ideas impact the original formula. Part of why that is comes from my aforementioned preference for a simple formula, and part is because these include some non-formula issues that were not pursued in 2010 in an effort to keep the focus on the main policy matters.
 First, and perhaps the most consequential change that can be made to state funding, is the teacher pension fund payments. Currently, the state and local districts split the cost of teacher pension contributions 60/40. This is a flat split, regardless of the wealth of the community. I think it’s absurd to ignore community wealth for such a large portion of state education expenditures. Using the Adjusted Equalized Weighted Assessment Values (AEWAV) to determine the reimbursement rates would be a big improvement on the progressiveness of school funding.


Second, I would make a slight change to the way that we fund charter schools. When we were developing the formula, there was broad agreement among policymakers that the “money should follow the child”. In one sense, this is the system we proposed since school district funding is based on enrollments. However, I think an irrational desire to not “double count” students, alongside the need to keep funding as flat as possible, pushed the formula a bit too far when it comes to charters. The old way of funding charter schools allowed districts to hold back 5% of the total per pupil expenditure from their charter school tuitions. This meant charter schools received 5% less funding than traditional public schools, but it also recognized that there are some fixed costs in districts that are not immediately recoverable when students leave on the margins. I think the state should return to this practice, however only if the state is willing to pay the withheld 5% to charters. I do think its fair to take into account some fixed costs, but I don’t believe it’s fair that charter schools received less funding as a result.


Third, we excluded all building maintenance costs from the base amount of state aid. This was largely because the formula was supposed to represent only the marginal instructional costs associated with each student. I don’t necessarily think that these costs have to be added into the base amount. However, I would like to see the state contribute to the maintenance of buildings more directly. I think the state should provide a flat dollar amount, say $100,000, per building in each district, provided that key criteria are met. The buildings should be at 90% occupancy/utilization, should have a minimum size set based on the research on efficiency (roughly 300 students at the elementary level and 600 students for high schools), and there should be some minimum standard for building systems quality and upkeep. These requirements are mostly about making sure this flat fund, which is really about the fixed costs of maintaining buildings, doesn’t create incentives to build more. It may seem inconsequential, but I think it’s important to state the preference for well-sized, occupied 7, and maintained buildings is worthwhile.


I think it’s wrong that the minimum reimbursement rate for school construction aid was raised to 40% during the funding formula debates in the General Assembly. This amounts to a massive subsidy for suburban schools and the previous 30% minimum is part of why we have such stark facilities inequities in the state. We should remove the minimums on construction reimbursement and simply use AEWAV to determine the reimbursement rate. Also, we need to establish a revolving facilities loan fund, much like the one used for sewers (and now roads and bridges). Access to lower interest bonds should not be dependent on city finances.


Fourth, one thing we did not include in the original funding formula that has come under considerably criticism is a special weight for students who are labeled English language learners. There are a few reasons we made this decision. The districts that have ELLs are the same districts that have high levels of poverty. In fact, the five communities that had more than 5% of their students classified as ELLs were, in order, also the top five districts with regards to free and reduced price lunch eligibility. Combined with a transition plan that was already increasing funding to these districts as rapidly as could be afforded, there were virtually no short-term consequences of not including an ELL weight. It’s worth noting that formula dollars are not categorical funds– there are no restrictions on how districts should spend this money, and there are no guarantees that an ELL weight would have any impact on ELL spending.


We were also concerned with incentivizing over-identification and failing to exit students who should no longer be classified as ELLs. I am also personally concerned with mistaking the additional supports we want to target as needed for English language acquisition; it would not only inspire the wrong policies and supports for these students, but it fails to recognize a host of needs that persist for these students well beyond English acquisition.


During the funding formula hearings at House and Senate Finance Committees we discussed the need for further study on this issue. I think that the next weight in the formula should be based on the Census and American Communities Survey. By using these data sets, classification of students who are eligible for the weight would not be dependent on the school district itself. Rather than focus on child language acquisition, I think we should broaden this weight to be applied based on the percentage of households that speak a language other than English in the home, where English is spoken at a level below “very well” 8. This would ensure that students who live in language minority households receive additional supports throughout their education, regardless of their language acquisition status. I would make this weight lower than some in the literature because it would apply to a broader set of students, probably somewhere around 40% like the poverty weight. For reference, the latest five-year estimate from the ACS data shows that 24.3% of households fit this definition in the city of Providence. With a 40% weight, at 22,500 students, with a foundation amount of around $9,000 per student, this weight would increase funding to Providence by a little over $16,000,000. Similar to other formula aid, these funds would be unrestricted.


Now, while I think that $16,000,000 is no small potatoes, and I am happy to express our policy preference to drive funding into communities where families are not using English in the home, some perspective is warranted. Providence will receive almost $240,000,000 in state aid when the formula is fully transitioned, compared to about $190,000,000 before. Adding this weight would only represent a 6% increase in state aid from the full formula amount. It’s an important increase, but I hope you’ll forgive me if I felt it was not grossly unfair to exclude it in the first iteration of the funding formula, especially considering we still have not fully transitioned to those higher dollar amounts sent to districts that would benefit from these funds.


It Takes Money


Each of these recommendations, in my view, would improve the way that Rhode Island distributes education aid. Some of the changes are technical, others address areas that are currently not considered, and some are purely about increasing the progressiveness of aid. All of these changes will require an even greater state contribution to education aid, but these increases would be an order of magnitude lower than what it would take to increase the state aid to covering 50-60% of all education expenditures. While I would support some pretty radical changes to drive more money into the state aid system, I think that each of these improvements are worth doing on the path to increased aid.


  1. I should note, that few people I spoke to were not in favor of raising the amount of state aid. We all want more money to come from the state because those dollars are far more progressive. However, Rhode Island was deep in its recession at this point in time and the dollar amounts to make a real dent in the state to local share in education are just staggering. Rhode Island currently funds just short of 40% of total school expenditures at the state level. To increase that to 60%, which is closer to the national average, they would have to contribute $500M more– a roughly 60% increase from the current level. Just for some context, the main tax fight of Rhode Island progressives has been to repeal tax cuts for higher income individuals that were instituted starting in 2006 in an attempt to move toward a flat income tax rate in Rhode Island. The impact of this repeal would be an increase in revenues that would cover roughly 10% of the increase in school funding required to move from 40% to 60% state aid. Of course, those dollars are supposed to pay for some portion of restoring pension benefits, so it’s already spoken for. ↩︎

  2. Hold harmless provisions, when introduced in other states, serve to dramatically distort the redistributive properties of state aid and almost always require a huge influx of funds. In fact, a hold harmless provision in Rhode Island would have required a doubling of state aid, which ultimately would have guaranteed that wealthy communities continue to receive too much state aid while less wealthy communities are stuck fighting year after year for tremendous revenue increases through taxation just to get their fair share. Essentially, hold harmless would ensure that you never reach formula-level spending and guarantee that state aid would not be very progressive. ↩︎

  3. One very popular progressive member of the Rhode Island General Assembly had been working for years to pass a new funding formula and had five or six such weights in her version. Interestingly, with the glaring exception of sending $0 to Newport in state aid, the difference in the overall distribution of funds by district in Rhode Island using this formula and our formula was tiny, almost always <5%. ↩︎

  4. Smoothing the “gains” and “losses” overtime was important to keep the formula as close to revenue neutral as possible. Of course, there are increases due to inflation and other factors each year as a part of the base, but our goal was to truly redistribute the funds such that not only is the end number not a big increase in total state aid but that getting through the transition period did not have huge costs. If it did, there is no way we could feel confident we would ever reach the point where the formula actually dictated state aid, much like the hold harmless provision prevents a full transition. Modeling various transition plans was a nightmare for me. ↩︎

  5. Many people forget that education spending is about competition within a single market. Overall spending matters less within this market than how you spend compared to others. The trick is that an urban school primarily working with traditionally under served families needs to be able to pay not just for more material supplies, but mostly for higher quality teachers and staff (and perhaps quantity). Because of compensating wage differentials, even hiring teachers and staff that are the same quality as wealthy communities costs more. ↩︎ ↩︎

  6. Perhaps I will write a future post on some ideas of how to push Rhode Island to “great”, even though I view all of those solutions as politically impossible. ↩︎

  7. I would include any leased space as occupied. We should encourage full utilization of the buildings, whether that includes charter schools, central office use, city government, or private companies. ↩︎

  8. This definition is clunky, but its how the ACS and Census track these things. This definition is clunky, but its how the ACS and Census track these things. We could verify the data using the data reported by districts about language spoken in the home. I would recommend using this data point to assist with whether or not to include these weights for charter schools. For example, approximately half of those families that do not speak English in the home also speak English very poorly. Therefore, I might apply half of the weight to each individual child whose family reports speaking a language other than English at home. Of course, the actual proportion of the weight should be specific to the ratio of speakers of language other than English to non-very well speakers of English by community. ↩︎