When Equity Is Optional: Does Accountability Drive School Spending? 

For over two decades, states have run accountability systems, reported on student outcomes, assigned school ratings, and identified lists of schools needing improvement. These lists, and the names of schools on them, may be published in the newspaper, discussed at school board meetings, and passed around in parent groups. But naming schools that need extra help is just the first step. The benefits of school accountability hinge on states using these systems to identify student needs, provide services and interventions, and address resource equity gaps to better support students.  

The Every Student Succeeds Act (ESSA) gave states flexibility in designing school accountability systems, while retaining a set of “equity guardrails” to ensure states focus resources on low-performing schools and marginalized students. ESSA requires states to identify at least the lowest-performing 5% of Title I schools and high schools where fewer than two-thirds of students graduate on time to receive comprehensive support. States must also identify schools with one or more underperforming groups of students to receive targeted support.  

Improving these schools and implementing evidence-based interventions requires extra funding. In previous work in our When Equity Is Optional series, we found that low-performing schools in different states had very different odds of being identified—and that students in those schools, therefore, had very different odds of receiving the additional services and resources they needed. For example, only identified schools are eligible to receive additional federal Title I funds set aside for school improvement (also known as the “7% Title I set-aside”). 

In this analysis, we analyzed per-pupil spending in identified and non-identified schools across nine states—at the time of school identification and in the school year following that—to help determine the extent to which low-performing schools have access to resources to help them improve. We believe accountability results should be reflected in policymakers’ funding decisions and identified schools should receive additional resources. 

Findings

This isn’t always the case. While identified schools in the states we analyzed tended to spend more per pupil than non-identified schools at the time they were identified, we found that education leaders did not consistently follow up identification of schools for additional support with those additional resources. Although the majority of identified schools spent more per pupil after being identified for support, our data reveals compelling evidence that funding for school improvement is inadequate to support all identified schools. 

• In every state but one, at least 25% of identified schools spent less per-pupil following their identification for extra support, not more. And despite dedicated federal funding for school improvement in the “7% Title I set-aside,” an even higher share of identified schools spent less per pupil when only looking at spending from federal sources at the time of, compared to after, identification. 

• The reason some identified schools spent less overall per pupil after identification stemmed more often from declines in spending from state and local sources than federal ones. This is because state and local funds make up a much larger share of K-12 education spending. When identified schools spent less per pupil after identification, across the states analyzed, between 68% and 97% of them spent less from state and local sources (compared to a range from 43% to 90% for federal sources). 

• The declines in overall per-pupil spending at identified schools were often significant. Among identified schools spending less per pupil following identification from both state and local sources and from federal sources, the average decrease in spending ranged from $705 to $2,819 per pupil. In most states, the decline in per-pupil spending neared or exceeded $1,000. 

Based on these findings, we believe policymakers should direct more money to identified schools and make sure all identified schools have additional resources to improve student outcomes. State and local leaders should review their funding systems, shield identified schools from budget cuts, and potentially set aside funds that exclusively support school improvement. Meanwhile, federal leaders should consider increasing the “7% Title I set-aside” to prevent so many identified schools from spending less per pupil from federal sources following their identification. 

About This Analysis

About Our Data 

Was per-pupil spending lower in identified schools relative to non-identified schools at the time they were identified? 

We did not find evidence that average spending per pupil was lower in identified schools relative to non-identified schools at the time they were identified. In fact, in most states, identified schools tended to spend more per pupil than non-identified schools.¹Connecticut’s expenditure data is disaggregated by three funding sources (federal, state/local, and private/other) in the National Education Resource Database on Schools. Analyses related to total per-pupil expenditures in Connecticut only include federal and state/local sources.

• Seven out of nine states (Indiana, Florida, New Mexico, Louisiana, Michigan, Ohio, and Washington) reported higher average total per-pupil expenditure in identified schools than in non-identified schools in FY 2018-19 (i.e., the 2018-19 school year). 
  
  
• The difference between total per-pupil spending in identified schools and non-identified schools was especially substantial in some states, but not others.  
  
  
  • For example, on average, Florida, Michigan, and Ohio spent at least 12% more per pupil in identified schools than in non-identified schools. In contrast, on average, Washington and New Mexico spent 2% and 3% more per pupil in identified schools than in non-identified schools, respectively. 
    
    
• Only Mississippi and Connecticut reported lower total per-pupil spending in identified schools than in non-identified ones at the time of identification. In these states, total per-pupil spending was 6% and 5% less in identified schools than in non-identified schools, respectively.  

How to read this figure: In Indiana, identified schools spent 7% more, on average, ($9,831 per pupil) than non-identified schools ($9,200 per pupil) in the 2018-19 school year (FY 2018-19). 

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 from the National Education Resource Database on Schools.

How to read this figure: In Indiana, identified schools spent 4% more from state and local funding sources, on average, ($9,004 per pupil) than non-identified schools ($8,685 per pupil) in the 2018-19 school year (FY 2018-19). 

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 from the National Education Resource Database on Schools. 

Public elementary and secondary schools are funded through a combination of federal, state, and local money but predominantly rely on state and local dollars. For example, in the 2018-19 school year, the federal government provided 8% of K-12 public education revenues, compared to roughly 47% and 45% from state and local sources, respectively.

Consequently, per-pupil spending from state and local sources in identified and non-identified schools largely mirrored spending patterns overall. 

• In five of the eight states with available data,²Ohio is excluded from analyses related to funding sources, because expenditures by state and local sources and by federal sources for the state were not available in the National Education Resource Database on Schools. the average per-pupil expenditure from state and local sources in identified schools was higher than in non-identified schools at the time they were identified: Indiana, Florida, Michigan, Louisiana, and Washington. 

However, when only considering state and local sources, the difference in per-pupil spending between identified and non-identified schools changed—to the benefit of non-identified schools. 

• In every state with higher total per-pupil spending in identified schools, the spending differential between identified and non-identified schools narrowed when only examining state and local funding sources. And in one state (New Mexico), the trend was reversed, with non-identified schools spending more per pupil on average than identified schools from state and local sources. 
  
  
• In the two states with lower total per-pupil spending in identified schools (Connecticut and Mississippi), the spending gap grew even bigger when only state and local sources were included. 

Unlike spending from state and local sources, per-pupil spending from federal sources was greater in identified schools than in non-identified ones across every state analyzed.

This is, perhaps, expected because federal education funding (particularly Title I funding) is designed to be targeted to districts with the highest concentrations of poverty; for example, an Urban Institute analysis found that 18 states with regressive state and local school funding became progressive when also including federal dollars.  

• The percentage differences between per-pupil spending in identified and non-identified schools based on federal funding alone also tended to be larger. In Indiana, Florida, Connecticut, and Michigan, for example, per-pupil spending from federal sources in identified schools was at least 60% higher than in non-identified ones.  

How to read this figure: In Indiana, identified schools spent 61% more from federal funding sources, on average, ($827 per pupil) than non-identified schools ($515 per pupil) in the 2018-19 school year (FY 2018-19). 

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 from the National Education Resource Database on Schools. 

How to read this figure: In Connecticut, identified schools spent 10% less from state and local funding sources, on average, ($15,490 per pupil) than non-identified schools ($17,158 per pupil). Yet despite identified schools spending 116% more from federal funding sources ($1,321 per pupil) than non-identified schools ($612 per pupil), they still spent 5% less overall ($16,811 per pupil) than non-identified schools ($17,770 per pupil) in the 2018-19 school year (FY 2018-19).

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 from the National Education Resource Database on Schools. 

Despite the gaps in per-pupil spending from federal sources in favor of identified schools, federal funding did not have a large impact on total per-pupil spending due to its comparatively small dollar amount. 

For example, in Connecticut, average per-pupil spending from federal sources in identified schools was more than twice as much as in non-identified schools. Meanwhile, average per-pupil spending from state and local sources in identified schools was 10% less than in non-identified schools. However, because federal dollars make up such a small fraction of public education funding, average total per-pupil spending in identified schools was 5% less than in non-identified schools. 

Did school identification lead to greater spending in identified schools? 

Under ESSA, a primary reason why states are required to identify schools for support and improvement is to ensure those schools receive extra assistance and resources to implement evidence-based strategies so that they can improve student outcomes and school quality. If spending in identified schools declined following their identification, it would be a troubling sign that policymakers may not be providing the resources schools need to improve. 

Many Identified Schools Spent Less after Identification 

The good news? In all nine states, most identified schools experienced an increase in total per-pupil expenditure in the year following their identification (i.e., comparing spending from FY 2018-19 to FY 2019-20). This means school identification, in most cases, led to greater spending in schools that states deemed as needing extra help. These identified schools likely spent more per pupil to improve student outcomes and school quality.  

However, not all identified schools were able to spend more. In every state except New Mexico, at least one quarter of identified schools spent less per pupil after they were identified as needing extra support. 

How to read this figure: In Florida, 27% of identified schools spent less per pupil overall after they were identified for support (comparing spending in FY 2019-20 to FY 2018-19). 

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 and FY 2019-20 from the National Education Resource Database on Schools. 

How to read this figure: In Florida, 45% of identified schools spent less per pupil from federal sources after they were identified for support (comparing spending in FY 2019-20 to FY 2018-19). 

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 and FY 2019-20 from the National Education Resource Database on Schools. 

Despite the availability of extra federal funding just to support identified schools (i.e., the “7% Title I set-aside”), lower per-pupil spending in identified schools following school identification was even more common when looking at spending from federal sources only.  

• In every state, at least a quarter of identified schools spent less per pupil from federal sources in the 2019-20 school year than in 2018-19.  
  
  
• And in five states, at least one out of three identified schools spent less per pupil from federal sources following their identification: Florida, Indiana, New Mexico, Connecticut, and Louisiana.

Keep in mind, because federal funding is a much smaller source of schools’ revenue, the decline in per-pupil spending from federal sources in identified schools could be offset by relatively small increases in spending from state and local sources. For example, even though 70% of identified schools in Louisiana spent less per pupil from federal sources in FY 2019-20 than in FY 2018-19, only 36% of identified schools spent less per pupil from all sources.   

Still, the fact remains that many schools spent less per pupil from federal sources following their identification for support, revealing inadequacies in the “7% Title I set-aside” to fund robust school improvement activities in all identified schools.  

To better understand patterns among identified schools spending less overall per pupil following their identification, we examined the frequency of three scenarios among these schools: 

1. Spending less per pupil from state and local sources and from federal sources after identification. 

2. Spending less per pupil from state and local sources while spending more from federal sources after identification. 

3. Spending less per pupil from federal sources while spending more from state and local sources after identification. 

In five of eight states (Florida³In Florida, 22 identified schools that spent less overall per pupil spent less from state and local sources and the same amount from federal sources. These schools were included in the same group as schools spending less per pupil from state and local sources but more from federal sources. One Florida school that spent less overall per pupil spent less from federal sources but spent the same amount from state and local sources. This school was included in the same group as schools spending less per pupil from federal sources but more from state and local sources. , New Mexico, Indiana, Connecticut, and Louisiana), the most common scenario among identified schools that spent less overall was to spend less per pupil from state and local sources as well as from federal sources, resulting in a decrease in overall per-pupil spending after they were identified for extra support. 

In these states, the next most common pattern was the second scenario: identified schools spent more per pupil from federal sources but less from state and local sources following their identification. This could have occurred because more federal funds became available to them through the “7% Title I set-aside.” Yet, any increase in per-pupil spending from federal sources was not significant enough to offset decreases from state and local sources. This is not entirely unexpected given that state and local funds make up a much larger share of school spending. 

• In two of the remaining states (Mississippi and Washington), the second scenario was the most common, observed in over half of identified schools spending less per pupil overall in FY 2019-20 than in FY 2018-19. 
  
  
• In Michigan⁴In Michigan, one school that spent less overall per pupil spent less from state and local sources but spent the same amount from federal sources. This school was included in the same group as schools spending less per pupil from state and local sources but more per pupil from federal sources. , the first and second scenario were equally as common, observed in 49% of the identified schools spending less per pupil overall after identification. 

In every state, at least some identified schools experienced the third scenario: spending less per pupil overall because they spent more per pupil from state and local sources but less from federal sources after identification. In these identified schools, decreases in per-pupil spending from federal sources were too large to be compensated by higher spending from state and local sources.  

How to read this figure: In Mississippi, among identified schools that spent less per pupil after they were identified (comparing spending in FY 2019-20 to FY 2018-19), 34% of those schools spent less per pupil from both state and local sources and federal sources; 56% spent less per pupil from state and local sources but spent more from federal sources; and 10% spent less per pupil from federal sources but spent more from state and local sources. 

Note: In New Mexico, the “Other” category includes two schools for which the state did not report per-pupil expenditure disaggregated by source of funding. Percentages may not add up to 100% due to rounding.

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 and FY 2019-20 from the National Education Resource Database on Schools. 

How to read this figure: Among identified schools that spent less per pupil after they were identified (comparing spending in FY 2019-20 to FY 2018-19), across the eight states in this analysis, between 43% and 90% of those schools spent less per pupil from federal sources (regardless of whether they also spent less from state and local sources). 

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 and FY 2019-20 from the National Education Resource Database on Schools. 

Although the third scenario was the least common, these findings again reinforce that the “7% Title I set-aside” is inadequate to support all identified schools. Across all the states we examined, between 43% and 90% of identified schools that spent less per pupil overall after identification spent less from federal sources (whether or not they also spent less from state and local sources).  

The findings also suggest, however, that state and local leaders need to do a better job allocating resources to low-performing schools as well. When identified schools spent less per pupil following their identification, an even higher share – between 68% and 97% – spent less from state and local sources. 

While it is troubling that so many low-performing schools spent less per pupil following their identification, it would cause less concern if the decrease in per-pupil expenditure in these schools was trivial in terms of the overall dollar amount spent. However, that does not appear to be the case.  

Among identified schools spending less per pupil following identification from both state and local and from federal sources (the most common scenario in most of the states we examined), the average decrease in total per-pupil spending ranged from $705 to $2,819 per pupil. In most states, the typical decline in per-pupil spending neared or exceeded $1,000, and in two states, average declines exceeded $2,500 per pupil. 

Keep in mind, similar decreases in per-pupil spending have a different impact depending on the state. For example, the average decrease in per-pupil spending among identified schools spending less from all sources in Louisiana ($705 per pupil) and Indiana ($769 per pupil) was comparable. However, average spending across all identified schools in Louisiana in FY 2018-19 was $12,445 per pupil, compared to $9,831 per pupil in Indiana. This means the negative impact of a roughly $700 decrease in per-pupil spending may feel more significant for schools and students in Indiana than in Louisiana. That said, a loss of $700, let alone $1,500 or $2,500 per pupil, is meaningful in and of itself, even in higher-spending states.

How to read this figure: In Florida, identified schools that spent less per pupil overall after identification (comparing spending in FY 2019-20 to FY 2018-19) due to declines in spending from all funding sources spent $968 less, on average, per pupil.

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 and FY 2019-20 from the National Education Resource Database on Schools. 

How to read this figure: In Florida, identified schools that spent less per pupil overall after identification (comparing spending in FY 2019-20 to FY 2018-19) due to declines in spending from all funding sources spent $968 less, on average, per pupil. In contrast, identified schools that spent less due to declines in spending from state and local sources only spent $866 less per pupil, while those that spent less due to declines in spending from federal sources only spent $172 less per pupil, on average.

Data source: All4Ed analysis of per-pupil expenditure data from FY 2018-19 and FY 2019-20 from the National Education Resource Database on Schools. 

Among identified schools that spent less per pupil solely due to decreases in state and local spending, the average decrease in spending was, unsurprisingly, smaller than for identified schools spending less from all sources. However, the average decrease in per-pupil spending in these schools was still substantial in many states, ranging from $363 to $1,301 per pupil.  

On the other hand, because federal funding pales in comparison to state and local funding for education, the average decrease in per-pupil spending was much smaller among the group of identified schools that spent less only from federal sources, staying below roughly $200 per pupil in every state except New Mexico. 

Given that the impact of declines in per-pupil spending from state and local sources is much more substantial than declines in spending from federal sources, it is imperative for state and local policymakers to examine how they allocate funding to ensure identified schools are resourced appropriately following their identification, shielded from budget cuts, and – ideally – given supplemental state and local funding for the supports, services, and other evidence-based interventions students need. 

Conclusion

The purpose of ESSA’s accountability requirements is not to create a list of struggling schools and stop there. It is to guide education leaders’ decisions, determine school-level interventions, and address inequities in resource allocation to drive positive changes in those schools and help them meet students’ needs.  

While identified schools generally, but not always, spent more per pupil than non-identified schools at the time they were identified (helped by federal funding targeted to high-need communities), our analysis reveals many identified schools did not receive the extra resources they needed to improve. By considering per-pupil spending at the time of schools’ identification and just after, we found that resources were inadequate to support all schools identified for improvement. In all but one state, at least 25% of identified schools spent less per pupil overall in the school year after identification than at the time they were identified. In several states, nearly half of identified schools spent less.  

Worse, identified schools that spent less per pupil following their identification often spent significantly less. Among identified schools with reduced spending from all funding sources in the 2019-20 school year compared to 2018-19, the average decrease ranged from $705 to $2,819 per pupil—with the average decline nearing or exceeding $1,000 per pupil in most states. 

Identified schools that spent less per pupil after identification more often did so because of declines in spending from state and local sources, not federal sources, likely because state and local funds make up a far greater share of K-12 spending. Across the states in our analysis, when identified schools spent less after identification, between 68% and 97% of them spent less from state and local sources. On the other hand, between 43% and 90% of identified schools spending less per pupil overall after identification spent less from federal sources. 

That said, when we did only consider federal spending, where additional funds are available just for school improvement and where one might, therefore, expect to see fewer decreases in per-pupil spending, we saw evidence of inadequate funding for low-performing schools. In most states in our analysis, the share of identified schools spending less per pupil was even higher when only examining changes in spending from federal sources. And in five of the eight states, nearly half, or more, of schools identified for support and improvement spent less per pupil from federal sources after being identified.  

Although our analysis is focused on school-level spending, our findings have implications for how funds are distributed, because, at the end of the day, a school can only spend its allocated share of funds.  

• State and local leaders. To address the inadequacy of school improvement funding, state and local governments should review their funding systems to ensure they are directing relatively more resources to identified schools and shielding them from budget cuts. They should also consider adding categorical funding mechanisms that exclusively support school improvement in identified schools.  
  
  
• Federal leaders. Likewise, even though federal spending was more targeted to identified schools than state and local funds, federal policymakers should consider increasing funding for the “7% Title I set-aside.” With at least one quarter of identified schools in every state spending less per pupil from federal sources following identification, current funding levels for school improvement are apparently inadequate to support the needs of all identified schools. 

Accountability systems cannot deliver on their promise to lead to meaningful school improvement if those schools do not receive additional resources after they are identified. Policymakers should consider school identification results when making school funding decisions and proactively address the inadequacy of school improvement funding. 

Appendix 

This analysis includes nine of the ten states in our When Equity Is Optional series on ESSA accountability systems: Connecticut, Florida, Indiana, Louisiana, Michigan, Mississippi, New Mexico, Ohio, and Washington. In each state, we built a data set including all schools that received ratings in the state’s ESSA accountability system, as well as schools identified for support and improvement, during the first year of ESSA implementation (which is typically the 2018-19 school year). We obtained school-level PPE data for FY 2018-19 and FY 2019-20 from the National Education Resource Database on Schools and merged their PPE data with our accountability data set for the nine states. 

The following tables show (1) data on schools included in this analysis and (2) raw data corresponding to all figures. 

Data Tables

Notes

• 1
  Connecticut’s expenditure data is disaggregated by three funding sources (federal, state/local, and private/other) in the National Education Resource Database on Schools. Analyses related to total per-pupil expenditures in Connecticut only include federal and state/local sources.
• 2
  Ohio is excluded from analyses related to funding sources, because expenditures by state and local sources and by federal sources for the state were not available in the National Education Resource Database on Schools.
• 3
  In Florida, 22 identified schools that spent less overall per pupil spent less from state and local sources and the same amount from federal sources. These schools were included in the same group as schools spending less per pupil from state and local sources but more from federal sources. One Florida school that spent less overall per pupil spent less from federal sources but spent the same amount from state and local sources. This school was included in the same group as schools spending less per pupil from federal sources but more from state and local sources. 
• 4
  In Michigan, one school that spent less overall per pupil spent less from state and local sources but spent the same amount from federal sources. This school was included in the same group as schools spending less per pupil from state and local sources but more per pupil from federal sources. 

Acknowledgments:
This series would not have been possible without the hard work, collaboration, and important contributions of many on the All4Ed team, especially Ally McCraw for her work with the design and dissemination of this report. We also thank the Edunomics Lab at Georgetown University, particularly Ash Dhammani, for sharing his insights and expertise regarding their school per-pupil expenditure data. Finally, All4Ed thanks the W.K. Kellogg Foundation for its generous support of this work.