Introduction
The COVID-19 pandemic has had a severe toll on the United States. It has killed over 220,000 Americans since mid-October (New York Times, 2020) and damaged the economy by reducing consumer spending and necessitating business closures. This economic harm has created to deficits for state governments across the country due to declining tax revenues and increased expenses associated with virus mitigation and social services. These deficits will force many states to adopt policies to raise new revenue, cut their spending, or take other actions to balance their budgets.
The size of the budget deficits spurred by the pandemic vary by state depending on their previous fiscal conditions and major revenue sources. States may also differ in how they address their deficits. They may cut spending on different programs, raise revenue from different sources, or rely on borrowing or their existing reserves to differing extents. It is important to have reliable information about the fiscal impacts of the pandemic to better understand the challenges currently facing states. It is also essential to understand the spending cuts, borrowing, and other actions taken by states to address the crisis, and how those actions will impact residents.
This paper presents a methodology for collecting reliable information about the fiscal impacts of COVID-19 on states and state responses. This framework provides a replicable way to understand how the COVID-19 crisis has affected state governments. It also features an analysis of the collected data that provides an overview of how states have been impacted by the crisis and how they responded between March and July 2020.
First, we present a literature review of state actions in response to fiscal crises and existing information about their responses to COVID-19. Next, we outline the data collection and coding methodology used in this paper. Third, we analyze the data and present key findings. Finally, we end with a discussion about the results and their implications for states and their residents.
Literature Review
State Actions
States can take a variety of responses to address budget deficits, although they are limited by existing fiscal institutions. At the most basic level, states can respond by generating additional revenue or cutting expenditures. States may also draw down reserves, borrow money, or rely on federal aid to address their deficits. Most states have some combination of balanced budget requirements (BBRs), debt limits, and tax expenditure limits (TELs) that limit their ability to use certain strategies (Urban Institute, 2018). BBRs and debt limits constrains the ability of states to maintain deficits and borrow, while TELs restrict states from collecting more than a certain amount of revenue (cite).
About two-thirds of state revenue comes from their own taxes and fees, with the remainder largely coming from the federal government (Tax Policy Center, 2020). States can address deficits by increasing the own-source revenue they collect, although that is often difficult during economic downturns when residents are struggling already. The largest state revenue sources are sales taxes (23.1 percent of total revenue), individual income tax (17.8 percent), and charges and miscellaneous payments (18.8 percent). The latter category includes things like university tuition payments, hospital payments, and tolls. While most states heavily rely on sales and income taxes, some states do rely more on alternative revenue sources such as mineral extraction or property taxes (Cite).
States also commonly address deficits by cutting spending. The largest state expenditures across the nation are K-12 education (24.9%), Medicaid (16.4%), higher education (13.1%), transportation (7.9%), and corrections (4.4%), and public assistance (0.7%). The remaining spending includes the bulk of state agency spending that supports work such as health and welfare services, natural resources programs, and the legislative and judicial branches. Some state services experience increased demand during economic downturns, so it can be difficult to states cut them to make up deficits. During the COVID-19 pandemic, the same may be true for healthcare-related expenses.
States may also draw down their reserves to shrink deficits. Heading into the pandemic, states had collectively amassed the most money in their reserve accounts in at least 20 years (Pew, 2020). With this amount of funding, the median state would have been able to operate for about a month without taking in new revenue. About two-thirds of this money comes from rainy day funds, while the other third is from funding left over in state accounts at the end of a fiscal year (Pew, 2020). State legislatures must typically grant state governors to power to access these reserves. These funds are intended for use during deficits, so it is expected the states will draw on them during the COVID-19 pandemic, although there be reluctance to drain them completely.
Some states could also borrow funds to cover this expenses, but that is typically used for only infrastructure. Most states disallow long-term borrowing for operating expenses (CBPP, 2018) and may also have debt limits. Still, it is possible for some states to use borrowing to cover short-term expenses, such as through authorization legislation. States may also borrow in the short-term, such as a loan under over a period under a year (CBPP, 2020).
Additional federal aid is also an important tool for addressing state budget deficits. The Coronavirus Aid, Relief, and Economic Security (CARES) Act was signed into law in March 2020, and provided $150 billion to state, territorial, and tribal governments. Each state received a minimum allocation of $1.25 billion, but this money was only eligible to be spent on expenses directly related to COVID-19 (NCSL) such as virus mitigation efforts. States are unable to use it to close deficits created by revenue declines as a result of the pandemic.
Finally, an alternative option for some states is simply to maintain a budget deficit. However, Vermont is the only state with a balanced budget requirement (Investopedia), although other states such as New Jersey and Illinois routinely maintain deficits. –STILL NEED TO DO SOME RESEARCH ON THIS.
Historical Responses
Prior to COVID-19, states were arguably in their strongest collective fiscal position since before the Great Recession (Urban Institute, 2018). The Great Recession, from 2007-2009, was the most recent period that states faced a similar combination of falling revenues and rising expenses on social services and the actions taken during that time period are important for understanding how states may respond to this crisis. State budgets were more severely affected during the Great Recessions than previous ones dating back to the 1980s, and are still recovering in many ways in 2020.
Overall state own-source receipts by $100 billion between 2007 and 2009, which was a larger drop than in any previous recession (CBPP, 2010). In the first year, states took in $87 billion less in tax revenue than during the previous 12 months. This was the largest decline on record (11%), and was the result of lost jobs, reduced wages, and lowered economic activity. During this time, income tax collection fell 27 percent, sales tax fell 10 percent, and state taxes were down 17 percent total (Brookings, 2012). States faced over $500 billion in cumulative shortfalls during this time due to the combination of declining revenues and increasing service costs.
By September 2009, 33 states enacted changes that increased annual revenues. Twenty of those states enacted taxes that raised revenue by at least 1 percent, and 10 raised revenue by more than 5 percent. Income and sales taxes were responsible for much of the increases. Overall, these new taxes made up $32 billion of the $87 million lost over that year. In the same time period, 42 states cut spending even though they were facing increased need for state services. Compared to previous crises, states relied more heavily on cuts than new revenue (Urban Institute, 2018). Over the course of the recession, state cuts mostly affected their largest types of expenditures: 34 states cut K-12 spending, 43 cut spending on colleges and universities, 31 lowered health care spending, 29 cut services to the elderly and disabled, and 44 reduced employee compensation (cite). In total, general fund spending fell by 4 percent in 2009 and almost 6 in 2010, which was first decline since 1983 and unprecedented in size (IMF, 2012). Spending began to increase again in 2011.
States also received significant federal assistance from the American Reinvestment and Recovery Act (ARRA) in 2009. Funding from federal grants increased from 30 to 40 percent of state tax receipts in 2010, the highest point it reached during the crisis. Figure 1 shows the most common measures states take to close budget gaps between 2008 and 2012 (IMF, 2012).
The Great Recession also had significant long-term effects for many states. By 2019, many states had still not undone spending cuts and made up deferred payments undertaken during the Great Recession (Pew, 2019). In 2019, 15 states still had lower general fund spending than in 2008 and 40 states had lower higher education spending per student than 10 years earlier, adjusted for inflation. Pew also found that before the pandemic, state investment in roads, bridges, water and sewer, and other infrastructure was at its lowest point in more than 50 years. The lingering effects of the Great Recession will be relevant for states that now must make more difficult fiscal choices.
Existing Resources
Finally, there have already been some attempts to collect information about the effect of COVID-19 on states. One of strongest additional sources of information comes from the National Conference of State Legislatures (NCSL). They maintain three separate data sources for information on state spending to deal with COVID-19, the revenue impacts of the pandemic, and state actions to close budget shortfalls created by the virus. This information is collected from government sources and media reports, although there is no public methodology that describes how it is collected. The Center on Budget and Policy Priorities also collects some information on state revenue shortfalls. These resources provide some good information about the impact of COVID-19 on states, but the methodology for their information collected is not clear.
Methodology
The data used in this paper was collected using the Proquest database of U.S. Newspaper articles called ‘U.S. Newsstream’. Articles were found using the search string “state AND (COVID or coronavirus) AND (budget or revenue) AND (cut or shortfall or reduction or deficit) AND (million or billion)”. This string was developed testing a signifcan amount pf different search terms and weighing the size of their narrowing effect versus the number of relevant articles they excluded. Search results were further filtered in the Proquest database to exclude articles published by the Targeted News Service and restrict source and document types to ‘newspaper’ and ‘news’, respectively. Once this search methodology was used to narrow down possible articles, the research team coded the first 100 articles returned by relevance for each week between March 1st and June 27. Overall, the researchers read through 1,700 articles, and collected articles with information relevant to the impacts of COVID-19 on state finances and state reactions. Duplicate articles, either published by different papers or the same paper on multiple days, returned in the search process were only coded once.
Once articles were collected, they were uploaded to DeDoose for coding. 337 articles were collected for coding, or about 20 percent of the articles read, and about two pieces of information were coded from each article, on average. The researchers developed codes to capture information about COVID-19 impacts on state budget and state reactions to these fiscal challenges. Codes were to provide information about the type of each impact or reaction, whether it had actually been implemented, and its direction (positive or negative), among other things.
XXX Could add a figure here, either a map by state or a bar chart showing collection over time. XX
COVID-19 Impacts
The primary type of information collected about COVID-19 budgetary impacts was the type of revenue source that declined. Figure 1 shows the number of articles we collected with this information for each state. The number of articles collected for each states is not necessarily the same as the total number of actions found for each states because each article may contain information about more than one revenue impact. Articles with information about impacts were collected from 41 states and the states with the highest number of articles were New York (27), Pennsylvania (16), and Georgia (16).
Figure 1. Impact Articles by State
impact_bins<-c(0, 2, 6, 12, 27)
pal_impact <- colorBin(palette = "viridis", domain = theme_map_data$impact, bins=round(impact_bins, digits = 0), na.color = "#D0D0D0")
theme_map_data %>%
st_transform(crs = "+init=epsg:4326") %>%
leaflet(width = "100%", options = leafletOptions(crs = epsg2163, zoomControl=FALSE)) %>%
setView(-93, 37.8283, zoom=3) %>%
addPolygons(popup = paste("<strong>State:</strong>", theme_map_data$ID, "<br>",
"<strong>Impacts:</strong>", theme_map_data$impact),
stroke = TRUE,
smoothFactor = 1.5,
weight = 2.5,
fillOpacity = 0.9,
opacity = 0.7,
color = "black",
fillColor = ~ pal_impact(impact)) %>%
addLegend("bottomright",
pal = pal_impact,
values = ~ impact,
title = "Articles with Impacts",
opacity = 1) %>%
setMapWidgetStyle(list(background= "white"))Figure 2 shows the number of articles we collected for several types of revenue impacts, and the number of unique states the information was collected for. Most commonly, articles provided information about total declines across all revenue sources. We collected 124 articles with this information, across 31 unique states. This information is enough to establish that most states experienced total declines in revenue between March and July, unsurprising given the loss of business sales and personal income that are the largest funding sources for most states.
#this constructs the dataframe with article and state counts
article_impacts <- article_data %>%
filter(impact > 0) %>%
select(sales_tax:gas_tax, revenue_impact_other) %>%
summarise_all(sum) %>%
rename(
`Total Revenue`= total_taxes_fees,
`Sales Tax`= sales_tax,
`Income Tax`= income_tax,
`Other Tax`= revenue_impact_other,
`Gas Tax`= gas_tax,
`Tourism Tax`= tourism_tax) %>%
pivot_longer(`Sales Tax`:`Other Tax`, names_to= 'Impact', values_to= 'Articles') %>%
arrange(desc(Articles)) %>%
add_row(Impact= "Property Tax", Articles= 0)
total <- article_data %>%
filter(impact > 0) %>%
select(state, total_taxes_fees) %>%
filter(total_taxes_fees > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
sales <- article_data %>%
filter(impact > 0) %>%
select(state, sales_tax) %>%
filter(sales_tax > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
income <- article_data %>%
filter(impact > 0) %>%
select(state, income_tax) %>%
filter(income_tax > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
rev_other <- article_data %>%
filter(impact > 0) %>%
select(state, revenue_impact_other) %>%
filter(revenue_impact_other > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
gas <- article_data %>%
filter(impact > 0) %>%
select(state, gas_tax) %>%
filter(gas_tax > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
tourism <- article_data %>%
filter(impact > 0) %>%
select(state, tourism_tax) %>%
filter(tourism_tax > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
unique_impacts <- c(total$n, sales$n, income$n, rev_other$n, gas$n, tourism$n, 0) %>%
as.data.frame() %>%
magrittr::set_colnames("States")
article_impacts2 <- article_impacts %>%
cbind(unique_impacts) %>%
pivot_longer(!Impact, names_to= "Type", values_to= "Count") %>%
mutate(Type= factor(Type, levels= c('Articles', 'States'))) %>% #limit to certain categories
filter(Impact=='Total Revenue'|Impact=='Sales Tax'|Impact=='Income Tax'| Impact=='Property Tax')
other_articles <- article_data %>%
filter(impact > 0) %>%
select(title, state, tourism_tax:gas_tax, revenue_impact_other) %>%
filter(tourism_tax==1| gas_tax==1|revenue_impact_other==1) %>%
distinct(state) #titleA significant amount of information was also collected about declines in sales (28 articles) and income taxes (25 articles). This information was collected for 15 and 13 states respectively. It is expected that revenue from these sources declined for most states, so this indicates this approach to data collection likely missed this information. This may be because articles commonly focused on total declines rather than declines from individual sources, despite the importance of sales and income taxes.
Another important revenue category for some states is property taxes. As shown by Figure 1, we did not find any mention of property tax revenue being affected by COVID-19. This is not that surprising, both because property taxes are less important to most states than the previously discussed taxes and because property taxes reflect property values which would not be immediately impacted by the crisis, compared to income or business sales.
In addition to the revenue sources shown in the figure, data was also collected about changes to a variety of other revenue sources. Overall, 23 additional articles were collected about 16 states that had other types of revenue losses. Common types of declines were in gambling, tourism, and gas taxes.
pal <- c("#2d708eff", "#20a387ff","#2d708eff", "#20a387ff","#2d708eff", "#20a387ff","#2d708eff", "#20a387ff")
ggplot(article_impacts2, aes(x= reorder(Impact, Count), y= Count, fill= reorder(Type, Count), group= reorder(Type,
Count)))+
geom_col(position='dodge', fill= pal) +
geom_text(aes(label=Count), position=position_dodge(width=0.9), hjust=-0.25)+
scale_y_continuous(breaks= c(0, 25, 50, 75, 100, 125), expand = expand_scale(mult= c(0.02, 0.1)))+
theme_minimal()+
coord_flip()+
#scale_color_manual(labels= c("Articles", "States"), values = c("#482677ff", "#2d708eff"))+
labs(
title= "Figure 2. Impact Frequency",
x="",
fill="") +
theme(
panel.grid.minor = element_blank(),
panel.grid.major.y = element_blank(),
panel.grid.major = element_line(colour="light grey"),
legend.text = element_text(size= 11),
axis.title.y = element_text(size=13),
axis.text.y = element_text(size=14, color = "black"),
axis.text.x = element_text(size= 12, color = "black"),
plot.title = element_text(size=14, hjust=0.5),
legend.position = 'right')
State Reactions
We also collected a significant amount about how states responded to the fiscal challenges caused by COVID-19. Figure 3 shows the number of collected articles with this information that were collected for each state. The most articles were collected about Georgia (27) and New York (23). About the same number of articles were collected about state reactions as for impacts, and they provided data about 40 states over the studied time period.
Figure 3. Reactions by State
react_bins<-c(0, 2, 6, 13, 27) #max(themes_by_state$reaction, na.rm=TRUE)
pal_react <- colorBin(palette = "viridis", domain = theme_map_data$reaction, bins=round(react_bins, digits = 0), na.color = "#D0D0D0")
theme_map_data %>%
st_transform(crs = "+init=epsg:4326") %>%
leaflet(width = "100%", options = leafletOptions(crs = epsg2163, zoomControl=FALSE)) %>%
setView(-93, 37.8283, zoom=3) %>%
addPolygons(popup = paste("<strong>State:</strong>", theme_map_data$ID, "<br>",
"<strong>Actions:</strong>", theme_map_data$reaction, "<br>"),
stroke = TRUE,
smoothFactor = 1.5,
weight = 2.5,
fillOpacity = 0.9,
opacity = 0.7,
color = "black",
fillColor = ~ pal_react(reaction)) %>%
addLegend("bottomright",
pal = pal_react,
values = ~ reaction,
title = "Articles with Reactions",
opacity = 1) %>%
setMapWidgetStyle(list(background= "white"))The data collected about states’ most common types of reactions is very important for understanding how COVID-19 will impact the services state are able to provide. Figure 4 shows a number of different reaction types we identified. By far the most common type of state reaction was to cut spending. The most common types of spending cuts were spending on state employee salaries and positions (77 articles), across the board cuts to state agency budgets (59 articles), and cuts the K-12 education spending (57). This data was collected for 26, 21, and 20 states, respectively. Less commonly coded types of expenditure cuts included cuts to higher education, healthcare, and infrastructure spending. The prevalence of state expenditure cuts compared to alternative methods of balancing budgets early in the crisis may indicate that similarly to the Great Recession, states are relying on spending cuts more than revenue increases and other methods. However, it may also be the case that states will enact new methods of revenue collection after the immediate economic downturn caused by the crisis has passed.
** Note: I think it would be best to break this up into two figures. The first would show expenditure, new revenue, calls for federal aid, borrowing, deferment, and reserves. The second would show the break down of different types of expenditure changes.
#this constructs the data frame with article and state counts
article_reactions <- article_data %>%
filter(reaction > 0) %>%
select(reserves:state_employees, eliminated_tax_breaks, expend_action_other, borrowing, infrastructure, deferment,
federal_aid) %>%
summarise_all(sum, na.rm= TRUE) %>%
pivot_longer(reserves:federal_aid, names_to= 'Reaction', values_to= 'Articles') %>%
arrange(desc(Articles))
state_emp <- article_data %>%
filter(reaction > 0) %>%
select(state, state_employees) %>%
filter(state_employees > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
atb <- article_data %>%
filter(reaction > 0) %>%
select(state, across_the_board) %>%
filter(across_the_board > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
k12 <- article_data %>%
filter(reaction > 0) %>%
select(state, k12_education) %>%
filter(k12_education > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
reserves <- article_data %>%
filter(reaction > 0) %>%
select(state, reserves) %>%
filter(reserves > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
expend_action_other <- article_data %>%
filter(reaction > 0) %>%
select(state, expend_action_other) %>%
filter(expend_action_other > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
higher_ed <- article_data %>%
filter(reaction > 0) %>%
select(state, higher_education) %>%
filter(higher_education > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
healthcare <- article_data %>%
filter(reaction > 0) %>%
select(state, healthcare) %>%
filter(healthcare > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
tax_breaks <- article_data %>%
filter(reaction > 0) %>%
select(state, eliminated_tax_breaks) %>%
filter(eliminated_tax_breaks > 0) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
infra <- article_data %>%
filter(infrastructure > 0) %>%
select(state, infrastructure) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
borrow <- article_data %>%
filter(borrowing > 0) %>%
select(state, borrowing) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
defer <- article_data %>%
filter(deferment > 0) %>%
select(state, deferment) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
fed_aid <- article_data %>%
filter(federal_aid > 0) %>%
select(state, federal_aid) %>%
distinct() %>%
count() %>%
mutate(n= as.double(n))
unique_reactions <- c(state_emp$n, atb$n, k12$n, reserves$n, expend_action_other$n, higher_ed$n, fed_aid$n, borrow$n, healthcare$n, tax_breaks$n, infra$n, defer$n) %>%
as.data.frame() %>%
magrittr::set_colnames("States")
article_reactions2 <- article_reactions %>%
cbind(unique_reactions) %>%
pivot_longer(!Reaction, names_to= "Type", values_to= "Count")pal2 <- c("#2d708eff", "#20a387ff","#2d708eff", "#20a387ff","#2d708eff", "#20a387ff","#2d708eff", "#20a387ff","#2d708eff", "#20a387ff","#2d708eff", "#20a387ff","#2d708eff", "#20a387ff","#2d708eff", "#20a387ff","#2d708eff", "#20a387ff", "#2d708eff", "#20a387ff", "#2d708eff", "#20a387ff", "#2d708eff", "#20a387ff")
ggplot(article_reactions2, aes(x= reorder(Reaction, Count), y= Count, fill= reorder(Type, Count), group= reorder(Type, Count))) +
geom_col(position='dodge', fill= pal2) +
geom_text(aes(label=Count), position=position_dodge(width=0.9), hjust=-0.25)+
scale_y_continuous(breaks= c(0, 25, 50, 75, 100), expand = expand_scale(mult= c(0.02, 0.1)))+
theme_minimal()+
coord_flip()+
labs(
title= "Figure 4. Reaction Frequency",
x="",
fill="") +
theme(
panel.grid.minor = element_blank(),
panel.grid.major.y = element_blank(),
panel.grid.major = element_line(colour="light grey"),
legend.text = element_text(size= 11),
axis.title.y = element_text(size=14),
axis.text.y = element_text(size=14, color = "black"),
axis.text.x = element_text(size= 12, color = "black"),
plot.title = element_text(size=14, hjust=0.5),
legend.position = 'bottom')
The most common action taken by states besides spending cuts was to draw down their existing reserves. This is not surprising, given that many states had record levels of reserves prior to the crisis. Another action commonly taken by states was to call for additional aid from the federal government. The collected data shows that officials from 19 states made similar calls for increased funding, which shows the importance many states place in receiving federal aid to reduce their deficits. Borrowing was less common, but the data did indicate that 11 states were pursuing options to borrow funds to reduce their deficits. Finally, a smaller amount of states took action to eliminate existing tax breaks to raise new revenue or defer state payments to local governments and other organizations for providing services such as education.
The common spending cuts made by states give some insight into the process by which the they were made. Cuts to state employee expenditures likely do not require legislative approval and can be done quickly, so it makes sense that these are some of the most common actions so far. However, employee payroll is a much smaller portion of most state budgets than spending on things like education and health and human services. As states continue to deal with the fiscal effects of the virus they may begin to make more substantial cuts to state programs. It is clear that many states are still waiting to understand the long-term effects of the virus and will be forced to make larger cuts in the future. For states that we found across the board cuts, it is also likely that this strategy included cuts to many areas of spending that if specified would have been otherwise coded.
Another factor to consider when interpreting the data collected about state reactions is that we collected data about both actual and potential reactions. Actual reactions are pretty straightforward, they are policy actions that were signed into law or otherwise actually implemented (for example, by a governor’s executive order). Potential reactions represent a range of actions that had not been actually implemented at the time of the article. It could include bills that have been passed by the legislature but not signed by the governor, a governor’s budget outlines, or a legislator’s idea for spending cuts.
Overall, there were 136 actual reactions coded and 285 potential ones. This number is not the number of articles, but total number of reactions coded. We collected data about the source of these claims, so we can further explore where these potential reactions originate from. The difference between a bill that has been passed by the legislature but has not yet been signed by the governor and a comment from a legislator can be quite large. If necessary, we could use these potential reactions as a starting point to search for whether they were actually implemented.
It is also true that many states are still in the middle of responding to the crisis, so not all actual reactions we found may be permanent. In fact, a number of states have passed budgets but plan to make further changes in coming months as the impact of COVID becomes more clear. As a result, even some of the actual reactions we found could be changed at a later date.
Conclusions
The collected information about the economic impacts indicates that most states are facing drops in revenue that have led to deficits. Most commonly, the drops have been driven by falls in important sales and income tax revenues. States have most commonly responded to these challenges through funding cuts. By the end of June, the most frequent types of cuts occurred to state agencies and employees. As states continue to deal with the effects of COVID-19, they may be forced to make additional cuts to larger expenditure categories.
Currently, states have implemented far more spending cuts than revenue increases, a similar pattern to the Great Recession. Given that some of the spending cuts implemented during the previous crisis were never rescinded, this could further limit some types of services provided by the state. It is possible that as the immediate economic affects of COVID-19 decline, state legislators will become more willing to raise taxes.
Overall, the methodology showcased in this paper is a fairly reliable way to collect information about the fiscal effects of COVID-19 on state governments. Unlike similar sources, it provides a replaceable way to collect the necessary information. However, there still may be room for improvement, as evidenced by the number of states for which no information was collected. For instance, data about loss of revenue was only collected for 41, even though it can be assumed to all of them had revenue losses of some sort. Information about the declines of specific sources, such as sales and income tax, were collected for an even smaller subset of states. In the future, the value of this data could be expanded on by quantifying the size of the impacts and comparing them across states.
Finally, most states are still adopting measures to deal with the deficits caused by the COVID-19 pandemic. This paper presents a strong survey of the initial actions taken by states, but research covering a longer time period will be important to fully capture state responses to the coronavirus. Especially given the rise in confirmed cases across the U.S. in the fall of 2020, states are continuing to respond to the challenges created by the virus. Future research can expand the time period analyzed.