Earmarks and State Appropriations for Higher Education
This study considers the relationship between federal academic earmarks and state appropriations for higher education. Often referred to as "pork," federal academic earmarks are both controversial and understudied. Using a unique panel dataset which spans 1990-2006, this study conducts a panel analysis with two-way state and year-fixed effects. It finds a positive, significant relationship between federal spending on academic earmarks and state spending on higher education appropriations. The effect is large in magnitude. For every dollar increase in federal earmarks received by institutions within a state, state appropriations for higher education increase by $1.98 to $4.75, depending on the measure of appropriations used. Additional analyses show that the relationship is a recent phenomenon, appearing most strongly after 1997.
Introduction
Federal academic earmarks are funds that are given by the United States Congress to a college or university. Often referred to as "pork," earmarks are usually inserted into spending bills by individual members of Congress for use in their home district. These "pet projects" differ from regular Congressional appropriations because they have not been requested by the president and skirt normal procedures for budgeting, appropriations, competitive bidding, and review. Federal earmarks to higher education represent a substantial amount of [End Page 3] money. In 2008 alone, 2,300 earmarks were granted to 920 institutions totaling at least $2.25 billion. A decade ago, spending on earmarks totaled $528 million (Brainard and Hermes 2008). Earmarks are not just large, but also growing at a rapid rate.
Earmarks are controversial because they are unregulated. Any project a member of congress deems fundable has the potential to become an earmark. Earmarks are not subject to scrutiny in regard to public need, affordability, efficiency, or effectiveness. On the front end, there is no review process or standard that earmarks must meet. Earmarks can be inserted into bills late in the lawmaking process, which leaves little or no time for debate—especially public debate. Once funds are appropriated, there is often no oversight of how funds are spent or evaluation of outcomes. The lack of oversight of earmarks makes them ripe for concern over graft, corruption, or conflicts of interest, since there is no check on these funds.
The unregulated nature of earmarks stands in sharp contrast to the federal peer-review funding process. Most federal departments and agencies have some sort of competitive process for allocating federal resources in support of academic research. In the peer-reviewed process, applications for government funds are carefully reviewed by experts. Once awarded, grants are monitored and evaluated.
Although not granted on the basis of peer-review, earmarks are allocated via identifiable means—the political process. Instead of the decision process for funding being controlled by experts in a peer-reviewed process, the power of allocation rests with elected officials, who, arguably, request and grant earmarks for the betterment of their constituents. Earmarks also offer a way for institutions—and thereby states—that are generally not winners of peer-reviewed research to gain a chance at receiving federal support for their academic enterprise.
While recognizing the controversy surrounding academic earmarks, this study will not weigh in on the relative merits and drawbacks of federal academic earmarks, but instead will treat earmarks for what they are—a funding source for colleges and universities. Whether institutions of higher education receive federal funds through peer-reviewed research competitions, student aid programs, or Congressional earmarks, these funds are important to the functioning and operation of institutions and therefore deserve careful study.
This article focuses on the distribution and allocation of Congressional earmarks at the state level. In the U.S. federalist system, states have primary responsibility for supporting higher education. However, federal earmarks flow, sometimes unexpectedly, to institutions within states each year. This study explores the research question: What is the relationship between the granting of federal earmarks to institutions within a state and state spending on higher education? [End Page 4]
Federal academic earmarks are ripe for study because they are an understudied funding stream for universities. Earmarks are a large funding stream, which has directed billions of dollars to institutions of higher education. Over the past decade, both the total amount of earmarks and the number of institutions receiving earmarks have grown considerably. However, scholarship in this area remains sparse. In addition, researchers need to better understand how states budget for higher education. This study contributes to knowledge in this area by exploring the interaction of spending by different levels of government for higher education.
Theoretical Framework
At the outset, one can imagine three possible relationships between federal academic earmarks and state appropriations for higher education:
1. States might perceive federal earmarks as being an alternative revenue source and reduce funds to institutions.
2. States might increase spending to amplify the effect of the federal earmarks.
3. There might be no relationship between federal earmarks and state spending on higher education.
The first possible relationship describes a substitution effect. If the federal government pays for a research project through an academic earmark, then the state funds that would have supported that research are freed and can be used for another purpose. Substituting funds from one level of government for another is common in the U.S. system of federalism. In current news sources, numerous stories exist regarding federal stimulus funds that are being substituted for state funds to support core state functions, like K-12 education. Given the prevalence of substituting funds at one level of government for another, it would not be surprising to find that states substitute Congressional earmarks for regular state appropriations for higher education.
The second alternative is that states might actually spend money to help entice earmarks to be granted within their borders. One can easily imagine a scenario in which a state supports the salary of a researcher who will work in a research lab that was built by a federal academic earmark. It is also conceivable that states could match funds with an earmark to amplify the effect, and the proposed benefit, of the earmark. The final alternative is that there is no relationship between federal academic earmarks and state spending on higher education.
If academic earmarks change the price of research conducted within a state, then the receipt of earmarks would reduce the price of a given research project, [End Page 5] and the state would be more likely to invest in it. This would be a change in the price that leads to a movement along the demand curve. Therefore, price theory would predict that this study should find evidence of the aforementioned second option.
Review of the Literature
This review of literature covers two bodies of knowledge—the first is related to academic earmarks and the second is related to state appropriations for higher education. Although earmarks are often cited in the popular press or are debated rhetorically, there is not a large scholarly literature on academic earmarks. Ehrenberg (2004) in his review of econometric studies of higher education cites only two studies related to academic earmarks (De Figueiredo and Silverman; Payne and Siow). Although this review includes more studies than the two reviewed by Ehrenberg, the scholarly literature on academic earmarks remains sparse.
Academic Earmarks Literature
Savage's 1999 book focuses on the political processes and debates surrounding the granting of these funds. Theoretically, Savage (1999) views earmarks as a collective action problem, where the benefits to individual institutions (and presidents) outweigh collective concerns about using a political process by which to allocate scarce federal academic research funds. Savage also calls federal academic earmarks a "small revolution" within the dominant policy regime of peer-reviewed funding. His work discusses the politics and history of federal academic earmarks, particularly among Association of American Universities (AAU) institutions. He also provides descriptive analysis which shows an increase in earmarking activity over time and an inequitable distribution of earmarks by state and institution. In addition, he explores the relationship between academic earmarks and academic competitiveness by looking to see if National Science Foundation (NSF) rankings change after an institution receives one million dollars or more in earmarks. There is inherent endogeneity in this approach since NSF rankings are influenced by earmark funds. Savage's results are mixed and he posits that changes in NSF rankings might relate to the type of earmark received.
Like Savage, De Figueiredo and Silverman (2006) use a Congressional perspective when considering academic earmarks. Their study focuses on higher education lobbying in Washington and uses academic earmarks as a measure of lobbying success. They also explore positional power within the House and Senate and find that for universities that are represented on important Congressional [End Page 6] committees a 10% increase in lobbying leads to a 2.8-3.5% increase in earmarks received. Providing further evidence of the influence of positional power on earmarks, Savage (1991) finds that individual members who serve as chairs of appropriations subcommittees can prevent or limit earmarks in appropriations bills.
Payne (2002) tests the impact of earmarks on research output for 120 research universities. Empirically, she uses university and year fixed effects, and an instrumental variable approach, which is based on a constructed instrument that captures the number of articles published in social science, engineering, life science, and agriculture in similar institutions located outside the geographical region of the institution of interest. She finds that for a one million dollar increase in federal earmarks, the number of articles increases by either 21 or 42—an increase between 8% and 14%, depending on the measure of earmarks used. However, she also finds that a one million dollar increase in federal earmarks decreases the number of citations per article by 0.25 or 1.3—a decline between 9% and 57%, depending on the measure of earmarks used. Payne argues that this shows evidence that earmarks increase the quantity of research, but not the quality. However, these results should be interpreted with caution as the instrument used is open to criticism due to its elaborate construction, the lack of similarity in research productivity across regions, and the fact that Harvard and John Hopkins, two large "winners" of federal earmarks, were excluded from the analysis.
In both a book chapter and a journal article (Payne 2006, 2003b), Payne compares the distribution, quantity, and quality of university research based on funding from federal earmarks and the NSF's Experimental Program to Stimulate Competitive Research (EPSCoR). 2 Payne finds that increasing the stock of federal earmarks increases the quantity of research (a one million dollar increase in federal earmarks leads to 22 additional publications per institution, on average), but decreases the quality (an extra one million dollars in federal earmarks results in an average reduction of 0.74 citations per article). EPSCoR reduces the quantity of research (qualifying for the program reduces publications by 31 articles per institution on average), but increases the quality after 1992 when the program was expanded (qualifying for the program increases citations per article by 1.1, on average). These results should be used carefully as the limitations of using publications and a citation index as measures of research productivity raise some concerns. In addition, Payne's measure of federal earmarks includes funds that are intended for non-research purposes such as [End Page 7] residence hall construction. The two programs compared in the analysis are not distinct, since Payne's sample shows that all states with an EPSCoR designation have institutions that also received federal earmarks.
Payne has also written on the impact of district representation or alumnus status of Congressional appropriations committee members on the distribution of federal research funding on a sample of 120 research universities (Payne, 2003a). In addition, a co-authored piece (Payne and Siow, 2003) uses an instrumental variable approach, which is based on the alumni status of committee members, to test the relationship between research output and federal research funding for a sample of 68 research universities. Aghion, et al. (2009) also use an instrumental variable approach, based on the probability of membership on federal appropriations committees, in a model that uses the process of granting federal academic earmarks in a multi-step design to identify the causal impact of higher education on growth. Although all of these works rely on earmarks (or instruments derived from academic earmarks) in their analysis, earmarks are not the main focus of the studies.
Balla et al. (2002) incorporates partisanship among Congressional members to better understand how pork is distributed. The authors first hypothesize that majority party control matters for gaining the "best" pork. Second, they hypothesize that because the majority party wants to avoid blame, they allow some minority pork. If the minority party also receives some level of pork they are less able to critique the majority party for engaging in pork and the distribution of pork can no longer be framed as a partisan issue. Offering evidence supporting their first hypothesis, party membership is not found to be a significant indicator of the probability that a higher education institution within a single Congressional district will receive an earmark. Balla et al. also find evidence supporting their second hypothesis; majority party membership matters for the amount of earmarks received by higher education institutions within a particular House district if the earmarks are lagged by one Congress (or two years). However, they were not able to find evidence of a similar effect in the Senate. Consistent with other literature, they also find that committee membership, holding a party leadership position, and seniority matters for the distribution of academic earmarks.
Considering academic earmarks from an institutional perspective, Cook (1998) shows evidence that earmarks pit individual universities against higher education associations in Washington. Cook (1998) argues that academic earmarks contributed to an erosion of confidence in higher education. She finds that institutional self interest in earmarks override the collective view that allocating funds on the basis of peer-review is preferable. Cook (1998) also finds that some institutions use high-priced Washington lobbyists to secure earmark [End Page 8] funding and that institutions are best able to garner earmarks if they have connections to powerful people in Congress. From a journalistic point of view, Greenberg (2001) also writes about how institutions seek to direct resources for science to their campuses by using lobbying and academic earmarks.
State Appropriations for Higher Education Literature
There is a growing literature on how states appropriate funds for higher education. This strand of literature began with articles that focused on the features of states that would change the level of funds, which states appropriated to higher education. In his 1976 article, Clotfelter found a positive effect of per capita enrollment on appropriations and a negative effect of out-migration on appropriations. Peterson (1976) included an analysis of the political characteristics of the states and determined that, when comparing 1960 to 1969, higher education environment, socioeconomic influences, a more professional legislature, and a factor that measured interparty competition and voter turnout, were important factors in predicting higher education appropriations. McLendon, Hearn, and Mokher (2009) also find evidence of political factors that influence state appropriations levels (partisanship, legislative professionalism, term limits, governor's powers, and interest groups). Leslie and Ramey (1986) ran a time series analysis on the relationship between enrollments and state funding for higher education. They found this link to be most strong in the 1960s, but that it had declined over time. Betts and McFarland (1995) found that the enrollment relationship was particularly strong for community colleges especially with increases in unemployment rates. Using panel data from 1982 to 1996, Toutkoushian and Hollis (1998) show that the percent of the population under the age of 18 and average faculty salaries at public institutions matter for levels of appropriations. Humphreys (2004) links spending on higher education to the business cycle. Okunade (2004) shows that a state's previous level of indebtedness predicts the share of state appropriations allocated to higher education, between 1993 and 1995.
A final strand of literature considers the relationship of funding for higher education to other state budget categories. Some studies consider higher education in relation to one other aspect of state budgets. For example, Kane, Orszag, and Gunter (2003) look at the relationship between spending on Medicaid and spending on higher education showing that Medicaid "squeezes" funding for higher education. By contrast, Hovey (1999) suggests that higher education serves as a balance wheel for the entire state budget. Delaney and Doyle (2007) formalized this idea by empirically testing the balance wheel model. They found evidence that higher education does serve as a balance wheel for state budgets and that other state budget categories do not conform to the balance wheel functional form. [End Page 9]
In an article that rhetorically links state appropriations and federal academic earmarks, Feller (2004) argues that variation in state spending creates an environment in which there is increased pressure on institutions to seek federal earmarks for research and development expenditures.
Despite these two prior lines of research, there is still a large gap in the scholarly literature exploring the interaction between state funding for higher education and federal academic earmarks. This study seeks to fill that hole by empirically exploring if academic earmarks interact with state funding for higher education, in what magnitude, and if this relationship has changed over time. It also seeks to update prior literature on federal academic earmarks by including more recent years, which have ushered in the largest increases in earmark funding and most extreme changes in earmark regulation to date.
Data and Methods
A unique dataset was constructed for this study. The panel dataset spans the period from 1990 to 2006. Most of the earmark data used in this study were graciously shared by the Chronicle of Higher Education. 3 During the time period of the study, the Chronicle published data on Congressional earmarks from 1990 to 2003. From 2004 to 2006, the Chronicle did not publish data on academic earmarks. For these years, data are used from Citizens Against Government Waste (CAGW)—a watchdog group that provides an online forum for the public reporting of all earmarks made by Congress in a given year. Since the CAGW includes all earmarks made by Congress regardless of type of recipient, the data from CAGW were cleaned to only include earmarks that were received by a higher education institution.
This study will focus on federal academic earmarks, or federal earmarks that are granted to colleges and universities. It includes all types of institutions and all types of earmarks, from those that directly support research to those that provide funds for buildings on campuses. This study does not include Congressional earmarks that are granted to non-higher education recipients, such as funds that are awarded to municipalities. Only earmarks that are granted within the 50 U.S. states are included. Earmarks that were granted to higher education institutions in non-U.S. states (like D.C., Guam, or the U.S. Virgin Islands) have been excluded from the dataset. [End Page 10]
Only non-shared earmarks are counted for the purposes of this study. 4 It is unclear in the data how earmark funds are divided if they are shared between two or more institutions. To ensure that earmarks are not double counted, all earmarks that were granted to more than one recipient (or are shared) were dropped from the analysis.
The source data for this study are identified at the earmark level, since it is possible for the same institution to receive more than one earmark in a given year. For purposes of analysis, the earmark level data are aggregated up to the state level. This is done to match the level of analysis with the research question, which relates to state, not institutional, behavior. 5
I use data for the variable of interest, state appropriations for higher education, from two sources: Grapevine and the U.S. Census bureau. There are some differences between the two measures of state appropriations for higher education. The Grapevine data report appropriations, not actual expenditures, and are focused on annual operating funds. 6 The U.S. Census data are focused on total expenditures on higher education, not just operating expenditures. The Census data define higher education more broadly than Grapevine, and consider any degree-granting institution which provides academic training above grade 12. 7 From the Census data, I use a measure of direct expenditures for higher education, which excludes capital outlays, construction, and intergovernmental transfers to either state or local governments. Because of the different foci of the two measures of state appropriations for higher education, both data sources are used in this study.
In addition to the data on federal academic earmarks and state appropriations for higher education, a number of control variables are used in this analysis. Data on total state revenues and voter participation come from the U.S. Census Bureau. Total state revenue measures the size of each state's budget. Voter participation rates from presidential elections are used since they historically have higher voter turnout than off-year, Congressional elections, and would [End Page 11] capture the highest level of political participation in a state. 8 Per capita income is used as a measure of the income of the residents of a state and data come from the U.S. Bureau of Economic Analysis. Unemployment rates are included as an indication of the health of a state's economy, and because these rates have been shown to be related to postsecondary enrollment level—especially in two year institutions. Data on state unemployment are from the U.S. Bureau of Labor Statistics. I include two measures of the nature of the higher education infrastructure within a state. The first is a measure of the number of public institutions in a state. The second is the share of enrollment in different sectors within a state. The enrollment share variables are broken out by type of control (public/private) as well as level (two-year/four-year). In my analyses, the percent enrollment share in public four-year institutions is the excluded category. The data on enrollment by higher education sector and the number of public institutions 9 in a state come from the U.S. Department of Education (USDE). All dollar amounts are adjusted for the consumer price index (CPI) using data on the CPI from the U.S. Bureau of Labor Statistics. The estimating equation for this study can be found in Equation 1.
Results
Descriptive information about the amount and distribution of federal academic earmarks by state can be found in Figures 1 and 2. Figure 1 displays earmark data from 1990 and Figure 2 shows the amount of academic earmarks received by states in 2006. Table 1 relates to the data displayed in Figures 1 and 2. It shows the number of states within each band of spending on academic earmarks for 1990 and 2006. Table 1 also displays the total amount of federal earmarks granted for both 1990 and 2006. These raw data show a growth in federal academic earmarks over time. Comparing 1990 to 2006, there is an 11 million dollar real increase in the total amount of federal earmarks received by all states. There is also a change in the distribution of states receiving earmarks. In 1990, seven states received no federal academic earmarks. In 2006, all states received some federal academic earmarks. In addition, the upper limit of earmarks by state appears to have grown over time. In 1990, three states received more than $20 million in earmarks (Michigan = $20 million, Mississippi = $20.1 million, and Iowa = $26.4 million) for a combined total of $66.5 million. In 2006, the number of states receiving more than $20 million in earmarks grew to four. The amount [End Page 12] of earmarks received by these four states grew even larger (Nevada = $22.6 million, Alabama = $28.9 million, West Virginia = $36.9 million, Mississippi = $45.9 million) for a combined total of $134.3 million.
Total Academic Earmarks by State, 1990
Total Earmarks by State, 2006
In addition to these descriptive statistics and, in order to explore the relationship between federal earmarks and state appropriations for higher education, this study also ran a series of regressions. The results of these regression analyses are discussed below. [End Page 13]
Number of States in Each Spending Band of Federal Earmarks and Total Earmarks Granted, 1990 and 2006
Table 2 displays summary statistics for the data used in the analysis. The total amount of federal earmarks received by all institutions in a state ranges from zero dollars to a maximum of $169.5 million. The mean amount of federal academic earmarks received by all institutions in a state in a year is $14.6 million. Mean state appropriations for higher education as measured by Grapevine is $1.17 billion, with a range between $64.4 million to $9.95 billion. Because the U.S. Census measure includes more institutions and captures more than just operating expenditures, state appropriations for higher education with this measure are larger than the Grapevine data. With the Census measure, mean state appropriations are $2.34 billion with a range between $204 million to $15.1 billion.
Table 3 presents regression results when pooling the panel data to run it as though it is a cross-section. The columns 1 and 2 present the results of the pooled analysis without any control variables for the Grapevine and Census data respectively. Regressions using each measure of appropriations find a positive, significant relationship between federal academic earmarks and state appropriations for higher education (p<0.01). The columns 3 and 4 improve upon the barebones results by including a full set of control variables using both the Grapevine and Census measures of appropriations. These results also show a positive, significant relationship between federal academic earmarks and state appropriations for higher education (p<0.01).
Since unobserved heterogeneity is likely to be a problem in the pooled analyses, these results can be more highly specified by using the data as a panel. Table 4 presents regression results using the data as a panel. The first model (columns 1 and 2) shows results for a state fixed effects model with the full set of control variables. State fixed effects can be thought of as including a dummy variable for each state. This forces the analysis to vary within each state and controls [End Page 14] for any unobserved heterogeneity in the data. The state fixed effects model in both column 1 with the Grapevine measure of appropriations and column 2 with the Census measure of appropriations again shows a positive, significant relationship between federal academic earmarks and state appropriations for higher education (p<0.01).
Descriptive Statistics, 1990–2006
Another approach to remedying the problems inherent in pooled analysis is to run the data as a panel and to create a model that controls for common time trends in the data by implementing year fixed effects. These were implemented by including a dummy variable for each year. The model shown in columns 3 and 4 is a random effects model, which contains a full set of controls and year fixed effects. The results using the Grapevine measure of appropriations for higher education show a positive, significant relationship between federal academic earmarks and state appropriations for higher education (p<0.01). Likewise the results using the Census measure show a positive, significant relationship (p<0.05).
The models shown in columns 5 and 6 even more strictly specify the models by including a full set of controls and two way fixed effects by both state and [End Page 15] year. State fixed effects control for unobserved heterogeneity in the data. Year fixed effects control for any year-specific common time trends in the data. Hence, this model considers variation within each state, controls for unobserved heterogeneity in the data, and removes common time trends. The results of this last panel model are discussed in the greatest detail since it is the most rigorously specified model of all of the models tested.
Earmarks and State Appropriations for Higher Education: Pooled Anaylsis, 1990–2006
In the two-way fixed effects panel analysis, the results show a positive, significant relationship between earmarks and state appropriations for higher education using the Grapevine and the Census measures of appropriations (p<0.1). In this highly specified model, the magnitude of the effect is remarkably large. An increase of a dollar in federal academic earmarks leads to approximately a $1.98 increase in state appropriations for higher education with the Grapevine [End Page 16] measure and approximately a $4.75 increase in state appropriations for higher education with the Census measure. 10
Earmarks and State Appropriations for Higher Education: Panel Anaylsis, State and Year Fixed Effects, 1990–2006
When considering the control variables in the two-way fixed effects model, the results show a positive, significant relationship between total state revenues and earmarks in the test using data from both Grapevine and the Census (p<0.01). This indicates, when a state's total revenues are higher, state appropriations for [End Page 17] higher education are more likely to be higher—all else equal. The number of public institutions in a state has a positive, significant relationship to earmarks with the Grapevine data (p<0.1). This can be interpreted to mean that when a state increases its number of public institutions, then it is more likely to increase state appropriations—all else equal. The percent of students enrolled in private four-year institutions has a negative, significant effect on state appropriations using both the Grapevine measure (p<0.01) and the Census measure of appropriations (p<0.1). When the share of enrollment in private four-year institutions increases, as compared to the enrollment share in public four-year institutions, state appropriations are more likely to fall—all else equal. State per capita personal income, state unemployment rate, the enrollment share in private and public two-year institutions, and voter participation rates in presidential elections were insignificant.
Is There Variation in the State Response to Federal Earmarks Over Time?
The next set of results expands upon the main finding of a positive, significant relationship between federal earmarks and state appropriations and considers if this result has always held or if it has changed over time. To more formally test this idea, I conducted an analysis that looks at the relationship between federal academic earmarks and state appropriations and includes an interaction term for years following 1997. The year of 1997 was chosen as the cut-off since there is an uptick in the mean amount of earmarks received by states following 1997—as shown in Figure 3.
The results of my analysis using the interaction term are shown in Table 5. With the interaction term (total federal academic earmarks received post-1997), the results show a positive, significant relationship for the years 1998-2006 for the Census measure (p<0.01). The Grapevine measure also yields a positive
Mean Total Earmarks Received by States 1990-2006
[End Page 18]
result (p<0.109), which is close to borderline significance. These results provide evidence that the relationship between federal academic earmarks and state appropriations appears most strongly after the year 1997 and is a fairly recent phenomenon. In other words, states appear to have reacted more strongly to federal earmarks in later years than they did in earlier years.
Conclusion
This study explored the relationship between federal academic earmarks and
Earmarks and State Appropriations for Higher Education: Panel Analysis, State and Year Fixed Effects, Pre- and Post-1997
[End Page 19]
state appropriations for higher education. It found that there is a positive, significant relationship between federal spending on academic earmarks and state spending on higher education appropriations. Either by allocating funds in the hopes of receiving an earmark, or by matching state funds to amplify the impact of earmark funds within their borders, states appear to respond in a positive, significant way to earmarks. Of course, this does not in itself establish a causal relationship between earmarks and state appropriations. However, given the absence of prior research on this question, finding a strong and robust relationship between these variables is an important first step towards further research on this topic.
The magnitude of the effect, as found in the two-way fixed effects panel model, is that for every dollar increase in federal earmarks received by institutions within a state, state appropriations for higher education will increase by $1.98 to $4.75, depending on the measure of appropriations used. This is a large and substantial finding and the direction of the effect was shown to be robust to numerous model specifications. As was shown in the analyses that considered this effect over time, the relationship between federal academic earmarks and state spending on higher education appropriations appears to be a recent phenomenon, since the results are significant when interacted with the years 1998-2006.
These results offer support for the outcome predicted by price theory. When the cost of research is reduced through the receipt of federal earmarked funds, state demand for that research appears to increase. This study offers new knowledge about the effect of revenue streams from different levels of government on higher education funding. It also provides an expanded understanding of federal academic earmarks as a funding source for colleges and universities.
Policy Implications
From the perspective of a single institution, the finding that there is no substitution effect between federal earmarked funds and state appropriations is good news. Although in practice, earmarks tend to reduce federal agency budgets for peer-reviewed funding, there is no rule or mechanism to stop a single institution from requesting funds, even for the same project, from both peer-reviewed sources and earmarks. In addition, there appears to be no penalty at the state level for receiving a federal academic earmark. An entrepreneurial president acting in his or her institution's self interest could conceivably put together a funding package for a research project that contains earmarks, peer-reviewed, and state funds. (A truly enterprising president could also add in private foundation or alumni raised funds to further multiply the scale of the endeavor.) It appears that the [End Page 20] end result would be an overall increase in funding for the research project, since there is unlikely to be a substitution effect among the different funding sources.
Although this article treated earmarks simply as a revenue stream for colleges and universities, these funds are controversial and inherently political. As earmarks become increasingly common and grow in size, frequency, and influence, the academy ought to revisit conversations about the role earmarks should play in funding academic research and how they should interact with state and peer-reviewed funding sources. Simply leaving the issue to be resolved by the self interest of individual institutions and Congressional representatives is an unsatisfactory approach that could hinder collective and national interests in academic research.
|
|
| The estimating equation used for the panel analysis in this study is represented as follows: |
| yst = a + ß1xst + ß2ast + ß3bst + ß4cst + ß5dst + ß6fst + ß7gst + ß8hst + ß9jst + ds + µt + est |
| where, |
| y = state appropriations for higher education |
| s = state |
| t = year |
| a = constant |
| x = total dollar amount of federal academic earmarks received |
| a = total revenue |
| b = per capita income |
| c = unemployment rate |
| d = number of public institutions |
| f = percent enrollment in 2-year private institutions |
| g = percent enrollment in 4-year private institutions |
| h = percent enrollment in 2-year public institutions |
| j = voter participation rate in presidential elections |
| d = state effect |
| µ = year effect |
| e = error term |
|
|
[End Page 21]
Jennifer A. Delaney, PhD, is an Assistant Professor in the Department of Education Policy, Organization, and Leadership in the College of Education at the University of Illinois at Urbana-Champaign.
References
Footnotes
1. Jennifer would like to thank the Chronicle of Higher Education, especially Jeffrey Brainard, for graciously sharing their data on federal academic earmarks, which made this work possible. She also expresses her gratitude to Erik Larson, Patricia Yu, and Mai Vang for their research assistance on this project, WISCAPE for supporting some of the data collection for this project, and Brian Noland for his helpful discussant comments on this article from the 2009 ASHE conference.
2. Like earmarks, EPSCoR is a program that circumvents the traditional peer-review process. EPSCoR was founded to make funding more accessible for institutions in states that were not historically successful in traditional peer-review processes. The program sets aside funds that can only be awarded to groups of institutions within states with EPSCoR designations. These funds are then awarded through a special peer-review process.
3. An alternative source of data for congressional earmarks was constructed by Savage (1999). It is used in Savage's work and Payne (2003b, 2002, and 2006) uses both the Savage and the Chronicle datasets in her analyses. The Savage dataset use Congressional records to capture the earmarks as inserted into legislation. The Chronicle dataset is derived by contacting agencies to capture the amount spent on earmarks by the federal agencies. Because I am most interested in actual spending levels, not intended spending, I use the Chronicle dataset throughout this study.
4. In their study, Balla, et al. (2002) also take this approach and drop all shared earmarks from analysis.
5. In aggregating the earmark level data to the state level, some of the counted earmarks are awarded to private institutions within a state. These earmarks to private institutions are included in the dataset, since it seems reasonable that a state will benefit from a federal academic earmark and from the resulting research that is conducted—regardless of the type of institution where that research is conducted. All states utilize their public and private sector differently. Some states like Michigan provide direct appropriations to private institutions. Other states like Massachusetts provide indirect support to private institutions through loan programs and student aid. A number of other states provide no support to their private institutions. Some states have very few students enrolled in private institutions, whereas others are dependent on private institutions to provide a college educated workforce for the state. Given this diversity in role of private institutions by state, I contend that the analysis is more complete when federal earmarks that are awarded to private institutions are included.
6. More information about the Grapevine survey and data limitations can be found at: https://www.grapevine.ilstu.edu/datalimitations.htm
7. For more information about the definitions and data collection techniques used by the Census bureau see: https://www.census.gov/govs/state/definitions.html#h
8. For years in which there was no presidential election, voter participation rates from the preceding election year are used.
9. For the year (1999) in which this data was not reported at the state level, data from the preceding year (1998) are used.
10. I find results of a similar magnitude with both measures of appropriations when using random effects models (both with and without year fixed effects). In addition, there is no change in sign or level of significance in the random effects models. However, I have reported the state fixed effects models in this article because the random effects model imposes the assumption that earmarks and the control variables are not related to unobserved state characteristics. This assumption is unlikely to be true in this context.
