Division of Research REVISED May 1987 Graduate School of Business Administration THE INFLUENCE OF POLITICAL COMPETITION ON STATE GOVERNMENT ACCOUNTING PRACTICES Working Paper #497 R Vivian L. Carpenter University of Michigan FOR DISCUSSION PURPOSES ONLY None of this material is to be quoted or reproduced without the expressed permission of the Division of Research. Copyright 1987 University of Michigan School of Business Administration Ann Arbor Michigan 48109

I

This is a working draft and is subject to change. Please do not quote or copy without the permission of the author. Helpful comments were received from Tom Lechner, Ehsan Feroz, Paul Danos, Waida Wallace, Robert Ingram, Sam Peltzman, and William R. Kinney. Earlier versions of this paper also benefited from the comments of workshop participants at The University of Michigan and the University of California-Berkeley. The financial support of the Provost Office of Wayne State University and the Ford Foundation is gratefully acknowledged, as is also the research assistance of Marti Wing and Kathy Ruby Petroni. Working Draft

ABSTRACT This paper uses the Economic Interest Group Theory of Government (EIG Theory) to investigate the incentives of state governments to produce GAAP annual financial reports. Use of the EIG Theory suggests that political competition among interest groups will affect the incentives of state governments to produce GAAP financial reports. Political competition affects a state's decision to produce GAAP financial reports because it affects the incentives of competing interests to accept the additional deadweight costs associated with producing GAAP disclosures. Evidence offered in this study implies that state government GAAP compliance is influenced by the political environment of each state.

1. Introduction Many recent studies have examined the economic determinants of financial disclosure practices of governments. These studies propose various economic reasons to explain why we observe such diverse accounting disclosure practices in the government sector. An example of these studies is Ingram [1984] who provides some possible explanations for financial disclosure practices of state governments in order to predict associations in his data, but he does not develop these explanations into an explicit theory. Governmental accounting research is currently being conducted without an explicit theoretical framework to base predictions of the associations between economic factors and cross-sectional variation in generally accepted accounting principles (GAAP). Without an explicit theory of state government disclosure practices, the validity of predictions made cannot be determined. In this paper, I use the economic interest group theory of government (hereafter referred to as the EIG theory) to develop a positive theory of state government financial disclosure practice. The positive theory developed in this paper posits that the level of state government GAAP disclosure is significantly influenced by the nature of the political competition among interest groups. The EIG theory will enable us to predict state government disclosure practices by predicting the behavior of dominant interest groups. In the EIG theory, dominant interest groups control the political budgetary process and, therefore, control the funding of costly GAAP disclosures. The funding of GAAP disclosures requires the support of some coalition of dominant interest groups. Dominant interest groups will use their political influence to promote or oppose the production of GAAP disclosures to further their own interests. Whether or not dominant interest groups support funding for GAAP disclosures depends on the level of political competition between taxed and subsidized groups.

-2 - In the next section, I extend the EIG theory to explain why the level of political competition between taxed and subsidized groups affects compliance with GAAP disclosure standards by state governments. In section 3, I explain how the possible determinants of state governments' GAAP compliance provided by Ingram relates to and can be explained by the EIG theory and in that section develop the primary hypothesis for this study. Section 4 describes the selection of the data and the methodology. In section 5, I present the empirical analysis, while section 6 contains the conclusion. 2. Political Competition and Governmental Accounting Practices: An Extension of the EIG Theory Understanding how most government agencies choose accounting disclosures requires a model of how government information is funded. This paper uses a model of government that is based on the EIG theory. The EIG theory suggests that interest groups are relevant in understanding the incentives of governments to produce GAAP information. A model of government based on the EIG theory can be used to explain why political competition between taxed and subsidized groups influences the variation in state government accounting practice. The EIG theory is drawn from the work of Downs [1957], Stigler [1971], Peltzman [1976, 1980], and Becker [1983, 1985]. Downs began the development of the EIG theory by showing that high information costs cause both politicians and the public to rely on interest groups to gather and to process government information. Stigler's contribution to the EIG theory was to explain why the optimal size of effective political coalitions (interest groups) is small in relation to the total voting population. Peltzman argued that the level and the allocation of government spending is determined by the political competition between these interest groups, specifically, the

-3 - competition between taxed and subsidized groups. Becker developed the EIG theory further, detailing the factors that determine the political effectiveness of interest groups. Becker [1983] finds that the determinants of political influence for a given interest group are: number of persons in different groups, additional pressure from competing interests, efficiency in producing political pressure, and deadweight costs. In the EIG theory, deadweight costs are defined as distortions in the use of resources that are caused by the use of taxes and subsidies. Deadweight costs include administrative costs that do not flow directly to subsidized groups. Hence, deadweight costs also include the costs of producing GAAP disclosures. In the EIG theory, interest groups exert political pressure to affect the taxes and subsidies of their members. Total taxes and total subsidies will not be equal because of deadweight costs. Both taxed groups and subsidized groups have incentives to reduce deadweight costs because more tax revenue is required to provide a given amount of subsidy and an increase in deadweight costs reduces the equilibrium subsidy (Becker [1983]). Therefore, the EIG theory suggests that both taxed and subsidized groups may not demand GAAP disclosures because the production of GAAP disclosures increases deadweight costs. However, we observe that some states do produce GAAP disclosures. GAAP disclosures are produced when some dominant coalition has economic incentives to accept the higher deadweight costs. The EIG theory maintains that government budgets are determined by the competition among interest groups and assumes that the state will implement the desires of dominant coalitions. Therefore, it must be that the decision to fund and produce GAAP disclosures is supported (or at least not opposed) by the interest groups or coalitions

-4 - that dominate the political competition. The EIG theory suggests that whether or not a particular state accepts the higher deadweight costs associated with producing GAAP disclosures depends on the competitive environment between taxed and subsidized groups. 2.1 Low Level of Political Competition When taxes are low, the political competition between taxed and subsidized groups will be low. Low political competition between taxed and subsidized groups signals the dominance of subsidized groups in the political process. Low political competition may not provide sufficient incentives for the dominant coalition (subsidized groups) to support funding for GAAP disclosures. This is because it is generally in the best interest of subsidized groups to minimize deadweight costs and, therefore, accounting costs. Consequently, in states where subsidized groups dominate the political process, GAAP disclosures will tend not to be funded. In states where subsidized groups dominate the political process, politicians can be pressured to raise the tax revenues necessary to fund subsidies. Hence, in states where subsidized groups are dominant, resources are available to fund subsidies. Subsidized groups do not have incentives to increase deadweight costs by funding GAAP disclosures when the level of political competition between taxed and subsidized groups is low. When competition is low, subsidized groups will tend to dominate the political process because taxed groups will have low incentives to invest in lobbying activities. 2.2 High Level of Political Competition Political competition will tend to be high when taxed groups have sufficient incentives to be active in the political process i.e., when tax rates are high. When taxes are high, the political competition between taxed and

-5 -subidized groups will be high, because the marginal benefits of taxpayers participating in the political process will exceed the marginal costs (Peltzman [1976, 1980]). If taxed groups dominate the political process, GAAP disclosures will tend to be funded. Government production of accounting information may improve the efficiency of taxed groups in producing political pressure because it reduces the information costs to taxed groups. As taxed groups improve their efficiency, the level of political competition between taxed and subsidized groups should increase. In a more competitive environment, subsidized groups may be forced to increase their efforts in producing political pressure to maintain their current level of subsidy. Increased political activity by taxed and subsidized groups does not insure positive effects for either group because political influence is zero-sum (Becker [1983]). However, increased political activity by taxed groups will tend to result in the production of more information to facilitate the assessment of overall financial condition. When the level of political competition between taxed and subsidized groups is high, the political environment is favorable for the funding of GAAP disclosures. 2.3 Intense Political Competition When political competition is intense with neither taxed nor subsidized group dominating, fiscal strain may result. Intense political competition between taxed and subsidized groups is evidenced by fiscal strain. When fiscal capacity is strained, both taxed and subsidized groups have incentives to be more active in the political process. However, neither taxed groups nor subsidized groups may dominate the political process when the level of political competition is high; taxed groups may have enough political influence to resist additional taxes, but not enough influence to significantly cut subsi dies. In such a case, the issuance of notes and bonds to finance subsidies

-6 - and deadweight costs may result. When neither taxed nor subsidized groups dominate the political process, a third major interest group enters the picture: the financial community. An intense degree of political competition between taxed and subsidized groups may lead to debt financing of government operations. The financial community requires governments to enter into legally binding contracts which may be penalized for non-GAAP disclosures. Therefore, the demands of the financial community for GAAP disclosures are only relevant incentives for producing GAAP disclosures when an intense level of political competition between taxed and subsidized groups leads to debt financing. 2.4 Other Effects of High Competition When taxed groups dominate the political process, taxes cannot be raised and may be lowered. The dominance of taxed groups increases the level of political competition within both taxed and subsidized groups. Increased within-group political competition may lead to an increased need for politicians to have GAAP information to determine if "equitable" allocations and tax burdens have been distributed by the bureaucrats. GAAP information to assess "equitable" allocations and tax burdens will be strongly supported by those interest groups expected to gain from its production. The supporters of GAAP disclosures, in this case, will tend to be taxed groups and politically weak subsidized groups. Politically strong subsidized groups will not have incentives to support funding of GAAP disclosures regardless of the political climate. Politically strong subsidized groups will be able to use their political influence to insure their maximum subsidy. However because of logrolling, politically weak subsidized groups may join coalitions with taxed groups to support funding for GAAP disclosures. Politically weak interests could gain political influence from the production of GAAP information with the support of taxed groups.

-7 - The increase in deadweight costs due to the production of GAAP disclosures will affect the nature of the political competition between taxed and subsidized groups. This is because deadweight costs affect the incentives of interest groups to produce political pressure. The EIG theory maintains that deadweight costs discourage the efforts of subsidized groups to raise subsidies, while stimulating the efforts of taxed groups to lower taxes (Becker [1985]). Since accounting costs increase deadweight costs, while decreasing the information costs, the efforts of taxed groups to reduce taxes will be stimulated. This stimulation of taxed group efforts will contribute to political competition remaining high in'states with GAAP disclosures, resulting in conditions that are favorable to the continued funding of GAAP disclosures.3 The EIG theory is important to government accounting research because government accounting and auditing costs are deadweight costs. Hence, the EIG theory provides us with a tool that can be used to develop a positive theory of state government accounting practice. Such a positive theory can be employed to aid in our understanding and our ability to predict variation in GAAP disclosure practices across state governments. 3. Political Competition Between Taxed and Subsidized Groups: An EIG Theory Interpretation of Ingram Ingram [1984] examined the association between economic factors and cross-sectional variations in the GAAP disclosure practices of state governments. Ingram posits that interest groups demand GAAP disclosures to determine resource availability and use; however, as previously explained,'this demand may not be sufficient to result in funding for GAAP disclosures if subsidized groups dominate the political process. Yet, Ingram's view is consistent with the EIG theory view that interest groups demands are important in understanding the demand for GAAP disclosures at the state level.

-8 - In addition to interest group demands, Ingram treats the following as possible independent determinants of compliance with GAAP disclosures: (1) voter demands to reduce the costs of monitoring; (2) bureaucrat demands to maintain their professional status; and (3) press demands to satisfy their information needs. Ingram does not attempt to structure his suggested determinants of GAAP disclosures for state governments into a positive theory of state government information production. The EIG theory, however, can both explain these determinants and structure them into a logical conceptual framework. The EIG theory can explain the influence of the voters, bureaucrats and the press on state government GAAP disclosure practices in terms of the political competition between taxed and subsidized groups for economic benefits. In the EIG theory, voters use interest groups to reduce the vast quantities of government information required to make informed election choices. Voter reliance on interest groups to reduce the high costs of gathering and processing government information results in the election decisions of voters being significantly influenced by the interest groups to which they belong. The EIG theory assumes that voters delegate their information processing and decision making responsibilities to interest groups to reduce the high costs of monitoring governments. Hence, voter demands for accounting information to reduce the costs of monitoring state governments are conveyed by interest group demands. The EIG theory further assumes that interest groups use politicians, bureaucrats, and voters to transmit their power. "Voter 'preferences' can be manipulated and created through the information and misinformation provided by interested pressure groups, who raise their political influence partly by changing the revealed 'preferences' of enough voters and politicians (Becker

-9 - [1985, p. 392])." Interest groups use statutes and appropriation bills to legalize their negotiated contracts with politicians (Carpenter [1987]). The incentives of politicians to monitor the execution of statutory provisions and to hire professionally trained accountants will be determined by the desires of dominant interest groups. Statutes and appropriation bills are also used by interest groups to direct bureaucrats. Bureaucrats are motivated by their own self-interest and are hired to carry out the negotiated contracts between politicians and interest groups. Therefore, interest groups can determine the professional standards on which accounting bureaucrats will be evaluated. The professional pressures for bureaucrats to select GAAP disclosures will be determined by the professional standards imposed by dominant interests. The EIG theory may also be used to examine the importance of the Press documented in Ingram's analysis by focusing on the affects that the Press has 4 on the information costs to taxed and subsidized groups. The Press affects the information costs to taxed and subsidized groups in two ways. First, the Press reduces the costs to interest groups of disseminating information to their members. Since interest groups often engage in public relations activities and issue press releases, they influence the information provided to their members and the public by the Press. Use of the Press by interest groups can reduce the costs of interest groups communicating with their members, as well as influencing the information available to the public on a given government issue. If the Press reduces information dissemination costs to interest groups, increased newspaper circulation would favor coalition formation, which should lead to increased competition between taxed and subsidized groups. Increased competition would tend to be associated with greater GAAP disclosure by states.

-10 - Second, the Press reduces the information costs of monitoring state governments by interest groups. The effect of the Press on monitoring costs would tend to be more positive with respect to the ability of taxed groups to produce political pressure than on subsidized groups. This is because of a bias in press coverage (Zimmerman [1977]). Government program scandals provide more newsworthy stories than stories with favorable effects; therefore, taxed groups have more opportunities to employ press information in producing political pressure. Since newspaper articles tend to favor the political needs of taxed groups, increased newspaper circulation can be expected to increase the influence of taxed groups in the political process. Therefore, increased newspaper circulation is generally expected to have negative effects on the ability of subsidized groups to produce political pressure and positive effects on the ability of taxed groups. Hence, the reduction of monitoring costs should be associated with greater GAAP disclosures since reduced monitoring costs are expected to increase the political influence of taxed groups. Increased political competition between taxed and subsidized groups, and within taxed and subsidized groups should make conditions more favorable for the passage of funding to support the production of GAAP disclosures. Dominant interests will support GAAP disclosures when they expect GAAP disclosures to reduce their information costs and produce information useful to them in producing political pressure. Since fiscal strain indicates a high or intense level of political competition, fiscal strain is a signal that conditions are favorable for the funding of GAAP disclosures. GAAP disclosures will be funded when taxed groups are active in the political process. High or intense levels of political competition result when taxed groups become overburdened with taxes, which creates incentives for

-11 - them to organize to reduce their taxes. Taxed groups may support an increase in deadweight costs related to the state producing GAAP disclosures, because GAAP disclosures will reduce information costs. Reducing information costs on the overall financial condition of the state creates a political environment more supportive of taxed groups. Figure 1 contains a summary of the effect of the political environment on GAAP disclosure practices of states. Use of the EIG theory leads to the primary hypothesis of this study: The level of state government financial disclosure in conformance with GAAP will tend to be positively associated with the level of political competition between taxed and subsidized groups. 4. Selection of Data This section describes the data used in this paper to test the hypothesis that state government conformance with GAAP disclosures is positively associated with the level of political competition among interest groups. In this section, I describe the GAAP disclosure indexes employed, the selection of political competition proxy variables, and the statistical methods employed. 4.1 The GAAP Disclosure Indexes Statewide financial reports were requested for the fiscal year ended 1984 from all fifty states. An annual financial report was received from all states that produce a statewide financial report. For states that did not send a statewide financial report, written or verbal confirmation was obtained to insure that the state did not, in fact, produce a statewide audited financial report. The financial reports obtained were analyzed twice. Once to determine conformance with the GAAP disclosures-used in Ingram [1984], while the second analysis was made using a GAAP compliance checklist based on National Council

-12 - on Governmental Accounting (NCGA) and American Institute of Certified Public Accountants (AICPA) pronouncements which were the authoritative sources of GAAP for governmental units in 1984. Only those items believed to be relevant and/or material for all state government entities were included in the checklist. Hence, two GAAP-based disclosure indexes are used in this study. States without a statewide audit report were given the same scores reported in Ingram [1984] for the INGRAM index and a zero for the NCGA index. The Ingram [1984] scores were verified over the telephone. INGRAM is the variable name for the updated Ingram [1984] index, while NCGA is the variable name for the 1984 GAAP compliance index. The selection of 1984 as the study year is potentially significant because in October of 1984, Congress passed the Single Audit Act of 1984. The Single Audit Act now requires all state governments to have an annual audit conducted in accordance with generally accepted auditing standards (GAAS). This requirement will result in the reporting of weaknesses in internal control, compliance with major grant regulations, and material departures from GAAP. A state government may elect to have a Single Audit performed on a departmental basis, therefore, passage of the Single Audit Act does not insure that states will produce statewide financial reports in conformance with GAAP. However, since departures from GAAP will become public information, states may now be under more pressure to fund the production of GAAP disclosures. The General Accounting Office (GAO) of the federal government had attempted to implement the Single Audit Concept without the force of a federal statute. However, the attempts of the GAO to implement the Single Audit Concept prior to 1984 were not very successful (McBride [1980] and Reynolds [1981]). Therefore, it is assumed that the incentives/for states to produce GAAP disclosures prior to 1984 were not primarily driven by federal audit

-13 - requirements,-but were significantly driven by the nature of political competition in each state. In addition to the two GAAP indexes previously described, a third variable on state GAAP compliance was based on a recent report published by the National Association of State Comptrollers. This report contained a list of states that self-reported if they produced financial reports in conformance with GAAP. This report contained responses from forty-four states. The six states that did not respond to the state comptroller study were classified as GAAP/NON-GAAP states based on review of their annual financial report or verification that they do not produce a statewide GAAP report. The classification of states into GAAP vs. NON-GAAP states is contained in Table 2. The GAAP compliance dummy variable used in this study is GAAPCODE. 4.2 Selection of Political Competition Proxy Variables Several variables were selected as proxy variables for the level of political competition between interest groups based on the premise that factors which facilitate the formation of political coalitions will also increase the level of political competition. Dye [1969] suggests that urbanization, personal income, industrialization, and education may lead to coalition formation. Peltzman [1980] clearly articulates why education and personal income would lead to more political competition among interest groups. An urbanization index based on percentage of the population in urban areas (URBAN) was employed to measure the extent of urbanization in each state. Per Capita income in 1979 (MONY7930) was used to measure personal income. The percentage of the population with four years of college (COLLEGE) was used to proxy for the level of education. In addition to urbanization (URBAN), personal income (MONY7930), and education (COLLEGE), I posit that those factors which decrease information

-14 - costs to the public will lead to increased formation of political coalitions and, hence, increased political competition among interest groups. This assertion leads to the selection of population per square mile (POPG8003) and Per Capita Newspaper consumption (NEWSPR) as additional proxies for the level of political competition. Since states may become more aggressive in seeking federal grants when the level of political competition is high (i.e., high taxes), total federal grants (GRANTS) may also proxy for the level of political competition. The Salary of the Accounting Administration may also proxy for the level of political competition because it signals that conditions are favorable for funding additional deadweight costs for information production. All of the proxies identified above capture some attributes related to the level of political competition, but none is a direct measure. To overcome this problem, a political competition index (CINDEX) was calculated. CINDEX was calculated by ranking the following variables: urbanization index (URBAN), Per Capita Income (MONY7930), percentage of the population with a college education (COLLEGE), population per square mile (POPG8003), Per Capita newspaper consumption (NEWSPR), Salary of the Accounting Administration (SALARY), and Per Capita federal grants (GRANTS). The ranking procedure would assign a value of 1 to the state with the lowest value and 50 to the state with the highest value. For example, the Per Capita Income ranking would assign a 1 to the state with the lowest Per Capita Income and 50 to the state with the highest. After all ranks were obtained they were summed for each state to obtain the value of CINDEX for each state. CINDEX is used as a proxy for the level of political competition.- Another proxy for political competition is based on deleting Grants from CINDEX. This variable is named GTINDEX and is used when the actual amount of federal grants is used as a control. The predicted association between the political competition indexes and the

-15 - GAAP disclosure indexes (CINDEX and GTINDEX) is positive. That is, those states with conditions which favor coalition formation will tend to be the states that produce GAAP financial reports. In addition to CINDEX and GTINDEX, two other political proxy variables were used, RANNEY and RAN84. RANNEY was used in Baber [1983] to proxy for political competition. RANNEY measures the degree of inter-party competition between the Democrats and the Republicans. The RANNEY index measures the inter-party competition for the period from 1962 to 1972. The RANNEY index is based on the average percent of the vote won by the Democratic gubernatorial candidates, the average percentage of the seats won in the state senate by the Democrats, the average percentage of the seats won in the state house by Democrats, and the percentage of all terms for governor, senate, and house in which the Democrats had control. Since, the RANNEY index measures the political competition over a period more than a decade earlier than the 1984 period that I am interested in, I recalculated the index for the period 1974 to 1984. The new RANNEY index is named RAN84. If Democrats are viewed as the party which primarily represents the interests of subsidized groups, while Republicans are viewed as the party which primarily represents the interests of taxed groups then we would expect low values of the RANNEY and RAN84 indexes to be associated with states that adopt GAAP disclosures, while high values would be associated with states that do not adopt GAAP disclosures. States with values in the middle ranges for RANNEY and RAN84 indicate intense political competition, since neither Republicans (taxed groups) nor Democrats (subsidized groups) dominate the state's politics. Therefore, middle range values of RANNEY and RAN84 would tend to be associated with GAAP disclosure. Given the associations for high and low values of RANNEY and RAN84, a negative association with the GAAP disclosures is predicted.

-16 - A description of the variables and their sources is contained in Table 2. Summary descriptive statistics are contained in Table 3. 4.3. Methodology This study uses univariate and multivariate methods for examining the association between political competition and the cross-sectional variation in the level of GAAP disclosures by state government entities. The univariate analysis includes the Pearson Product-Moment analysis for parametric tests and the Spearman Rank Order analysis for the nonparametric tests. The multivariate analysis uses OLS regression with the NCGA and INGRAM GAAP disclosure indexes as the dependent variables, and logistic regression for the dummy GAAPCODE dependent variable. The OLS models assumes a multivariate normal distribution, while the logistic model does not. The base OLS model is of the form: Yi = + B + B2X2i + B3X + B4X4 where Y. = INGRAM or NCGA, 1 Xi = Population, a proxy to control for size of the state, X = General Obligation Debt, X3 = CPACODE, X. = Proxies for the level of political competition. 4i The base logistic regression model relates the probability that a state will produce GAAP disclosures, conditional on the logistic cdf given below: P (Y. = 1)= 1/[1 + exp (-XiB)] 1 1 where Y. = GAAPCODE X = 1, so B if the intercept, li ' X2i = Population, X = General Obligation Debt, X4 = CPACODE, 4i X = Proxies for the level of political competition. 5i An interpretation of the logistic model is as follows. Whether or not a state produces GAAP disclosures depends on the size of the state, professional skills of the accounting administrators, influence of the capital market, and

-17 - the level of political competition. This, in turn, may depend on population, level of general obligation debt outstanding, professional background of the accounting administrator, factors that influence coalition formation, and the nature of the political competition. 5. The Empirical Analysis 5.1 Univariate Analysis This section presents and discusses results for the univariate tests of the cross-sectional relationship between state government GAAP disclosures and political competition among interest groups. The Pearson and Spearman correlation coefficients between the GAAP dependent variables and the political competition proxy variables are presented in Table 4. The results based on the Pearson Analysis are comparable to the Spearman analysis, however there are a few significant differences. All explanatory variables except for RANNEY, RAN84, and CPACODE are significantly correlated with the NCGA, INGRAM, and GAAPCODE Indexes. All significant correlation coefficients are consistent with predictions based on the EIG theory. There is a very high degree of significant positive correlation among several of the explanatory variables as indicated in Table 5. Positive correlation among several of the explanatory variables is expected. 5.2 OLS Regressions The results of the OLS regression analysis for the NCGA and INGRAM Indices are presented in Tables 6 and 7. Table 6 contains the analysis for the NCGA index controlling for the effects of population and general obligation debt for all regressions except for MODEL3 which controls for population and federal grants. In all regressions where general obligation debt is included the CINDEX and GTINDEX are significant. GRANTS, which is intended to

-18 - control for the effects of federal grants audit regulations on GAAP reporting practices, is not significant but GTINDEX is significant. When INGRAM is used (Table 7) as the dependent variable, some summary statistics improve, however, still only G.O. Debt, CINDEX and GTINDEX are significant. Of the eighteen regressions reported in Tables 6 and 7, MODEL7 in Table 7 (INGRAM INDEX) has the greatest explanatory power as indicated by the adjusted R summary statistic. MODEL7 in Table 7 is statistically significant and is able to explain 31.5% of the variation. MODEL2 in Table 7 has the smallest estimated standard error and explains 30.6% of the variation. Inferences that can be drawn from the OLS models are limited because a-ll OLS assumptions have not been satisfied. Logistic regression, which has fewer assumptions than OLS, is employed in the next section. 5.3 Logistic Regressions Table 8 contains the logistic regressions which use GAAPCODE as the dependent variable. Table 8 uses a similar format to that used in Tables 6 and 7. As is the case in the OLS analysis, the more parsimonious models that only control for POPULATION and G.O. DEBT have, in general, better summary statistics. Of the nine estimated versions of the GAAPCODE logistic model reported in Table 8, MODEL7 has the greatest explanatory power as measured by its R of 51.1. The percentage of concordant pairs for MODEL7 is.88. Using a one-tail test, GTINDEX, CINDEX, and RAN84 are significant at the 10% level in the predicted directions. All models reported in Table 8 are statistically significant with R values ranging from.394 to.511. 5.4 Other Evidence Finally, a more general approach to looking at the data is employed in this section. Table 9 divides the States into High Debt versus Low Debt

-19 - States based-on the median value of General Obligation Long Term Debt Outstanding. Table 10 divides the States into High Competition versus Low Competition States based on the value of the RAN84 Index. States with RAN84 values greater than.45 are classified as high competition states. Then the data in Table 1 (GAAP versus NONGAAP compliance), Table 9, and Table 10 are combined into a 2x2 classification table which gives the proportion of states that are GAAP given their level of political competition and outstanding debt. These data are presented in Table 11. The evidence presented in Table 11 supports the primary hypothesis that the level of political competition affects the incentives of states to adopt GAAP disclosures. Only one of the seven states that have low debt and a low level of political competition produces a GAAP annual financial report. While seven out of eighteen states with low debt and high competition produce a GAAP financial report. Clearly, from looking at Table 11, the effect of debt on a state's decision to produce GAAP disclosures is greater than the level of political competition, but the level of political competition still appears to be important. 5.5 Limitations The empirical results generally support the premise that cross-sectional variation in state government GAAP reporting practices is associated with the level of political competition. However, there are some methodological problems that limit the generalizations that can be made from the statistical evidence presented. The regression analyses attempted to control for factors other than the level of political competition that would influence the decision of a state government to produce GAAP disclosures. Addition of such control variables assumes that the variables can be exogenously specified, when actually these

-21 - between taxed and subsidized groups and/or within subsidized groups will increase, also creating conditions favorable for the funding of GAAP disclosures. Future research should involve a combined effort among economists, political scientists, and accountants to: (1) explore the relationships that influence the nature and the level of political competition among interest groups, (2) further investigate the influence of interest groups on 'the production of information by government entities, and (3) more rigorously test the premise that the level of state GAAP disclosure is positively associated with the level of political competition between taxed and subsidized groups. Clearly, all factors which influence the decision of states to adopt GAAP disclosures have not been identified in this study. Political competition appears to be an important factor, but additional work is needed to identify other significant factors that influence government's decision to conform to generally accepted accounting principles for external financial reporting. Finally, future research should explore the relationship between political competition and GAAP reporting practices of local units of government.

-20 - independent variables may be endogenous. Since the models employed did not specify the causal relationships among the variables nor use a simultaneous method of estimation the results could be misleading. In addition, the proxy variables selected may measure the variable of interest (level of political competition between taxed and subsidized groups) with error. The extent of any measurement error is unknown, however, if present it could bias the statistical tests. In spite of these limitations, this study does support the notion developed in Baber [1983] that political competition affects the incentives of state governments to supply information to the public. Political competition appears to have a significant influence on a state's decision to provide GAAP disclosures., 6. Conclusion This paper provides an extension to the EIG theory of government to explain the cross-sectional variation in GAAP reporting practices of state governments. Use of the EIG theory identifies accounting and auditing costs as deadweight costs which is important in understanding the incentives of state governments to produce GAAP disclosures. The variation in GAAP reporting practices of states is found to be positively associated with political competition indexes which measure conditions that favor coalition formation and negatively associated with political enivornments where subsidized groups dominant the political process. These findings are consistent with the overall hypothesis that the level of political competition is positively associated with the level of conformance with GAAP. When taxed groups dominant the political process, political activity within subsidized groups will increase, creating conditions favorable for the funding of GAAP disclosures. When political competition is intense with neither taxed groups nor subsidized groups dominating the political process, the level of political activity

-22 -FOOTNOTES These studies include Zimmerman [1977], Maher and Keller [1978], Evans and Patton [1983] and Ingram [1984]. Once GAAP disclosures are funded, it is unlikely that subsidized groups will be able to produce enough political pressure to eliminate them. This is because GAAP disclosures are reducing the monitoring costs to taxed groups and would signal a need to taxed groups to increase their political activities. 3Funding of GAAP disclosures is not a trival issue since it may cost a state in excess of $1 million dollars to implement an accounting system to produce all required GAAP disclosures. The State of New York has spent in excess of $10 million dollars improving its accounting system to produce GAAP disclosures. Ingram documents an insignificant, weak, and negative correlation coefficient between newspaper circulation and his GAAP indexes; in his multivariate analysis he finds a significant negative regression coefficient for newspaper circulation. I recalculated the Spearman correlation coefficient between newspaper circulation and Ingram's GAAP indexes using the SAS statistical package and the public data sources cited in Ingram. I found the correlation coefficients to be insignificant, weak, and positive. There are differences in the data reported by Ingram and the data that I obtained from the cited sources. The nature of this discrepancy is unknown. A high value for the RANNEY and RAN84 indexes occurs when Democrats are dominant (i.e., are winning the elections), while a low value for RANNEY and RAN84 occurs when the Republicans have dominated the elections in a state.

-23 - BIBLIOGRAPHY Becker, G. A. "Public Policies, Pressure Groups, and Deadweight Costs." Journal of Public Economics, Vol. 28, pp. 329-347.. "A Theory of Competition Among Pressure Groups for Political Influence," Quarterly Journal of Economic, Vol. XCVIII, August 1983, pp. 371-400. Carpenter, V. L. "The Effects of Interest Group Competition on Funding Government Information Systems." Research in Government and Non-Profit Accounting, Vol. 3, 1987. Congressional Quarterly Inc. Congressional Quarterly Weekly Report, Nov. 17, 1984, p. 2943. Congressional Quarterly Weekly Report, Nov. 13, 1982, p. 2847. Congressional Quarterly Weekly Report, Nov. 15, 1980, p. 3374. Congressional Quarterly Weekly Report, Nov. 18, 1978, p. 3301. Congressional Quarterly Weekly Report, Nov. 13, 1976, p. 3157.. Congressional Quarterly Weekly Report, Nov. 16, 1974, p. 3124. Council of State Governments. Preferred Accounting Practices for State Governments. Exposure Draft for a Research Report, 1982. Cox, D. R. The Analysis of Binary Data. London: Metheun, 1983. Downs, A. An Economic Theory of Democracy. New York: Harper, 1957.. Inside Bureaucracy. Boston: Little, Brown and Company, 1967. Dye, T. Politics in States and Communities. Englewood Cliffs, N.J.: Prentice-Hall, 1969. Engstrom, J. H. "The Governmental Report Entity." Journal of Accounting, Auditing, and Finance, Vol. 8, No. 4, Summer 1985, pp. 305-318. Evans III, J. H. & Patton. "An Economic Analysis of Participation in the Municipal Finance Offices Association Certificate of Conformance Program." Journal of Accounting and Economics, August 1983, pp. 151-175. Ingram, R. W. Accounting in the Public Sector: The Changing Environment. Salt Lake City: Brighton Publishing Company, 1980. ______ "Economic Incentives and the Choice of State Accounting Practices." Journal of Accounting Research, Spring 1984, pp. 126-4. _"The Importance of State Accounting Practices for Credit Decisions." Journal of Accounting and Public Policy, 2, 1 Spring 1983, pp. 5-19.

-24 - Maddala, G. S-. Limited Dependent and Qualitative Variables in Econometrics. Cambridge University Press, 1983. Madison, J. The Federalist No. 10 - The Source and Control of Factions, in the Federalist Papers, ed. R. P. Fairfield. Baltimore, Maryland: John Hopkins University Press, 1981, pp. 16-23 Maher, M., and E. Keller. "An Analysis of Reporting Practices in American Cities," Proceedings of the American Accounting Association National Meeting (1978). Mandolini, A. "Discussions on the Municipal Accounting Maze: An Analysis of the Political Incentives." Journal of Accounting Research Supplement, 1977, p. 1. Maschmeyer, R. and R. Van Daniker. "Information Needs of State and Local Officials." Government Finance, November 1979, pp. 15-19. McBride, T. F. "The 'Single Audit' Concept: Issues in Implementation." The Government Accountants Journal, Spring 1980, pp. 29-35. McCormick, R. E. and R. D. Tollison. Politicians, Legislation, and The Economy: An Inquiry Into the Interest-Group Theory of Government. Boston: Martinus Nijhoff Publishing Company, 1981. National Association of State Comptrollers. State Comptrollers-Technical Activities and Functions. Lexington, Kentucky: Council of State Governments, 1986. Newsprint Information Committee. "Newspaper and Newsprint Facts at a Glance 83-84." New York: Newsprint Information Committee, 1984. Olson, M. The Logic of Collective Action. Cambridge: Harvard University Press, 1965. Peltzman, S. "Toward a More General Theory of Regulation." Journal of Law and Economics, August 1976, pp. 211-243.. "The Growth of Government." Journal of Law and Economics, XXXIII, 2, October 1980, pp. 209-280.. "Constituent Interest and Congressional Voting." Journal of Law and Economics, April 1984, pp. 181-210. Posner, R. A. "Theories of Economics Regulation." Bell Journal of Economics, Autumn 1974, pp. 335-358. Ranney, A. The American Elections of 1984. Chapel Hill, N.C.: Duke University Press, 1985. Reynolds, A. L. "Implementation of the Single Audit Approach: A Status Report." The Government Accountants Journal, Summer 1981, pp. 1-7.

-25 -Scammon, R. M-. and McGillivray. America Votes 15. Washington, D.C.: Elections Research Center, 1983. U.S. Bureau of the Census. County and City Data Book [Machine-readable datafile]. Washington: Bureau of the Census, 1977, 1983. U.S. Bureau of the Census. State Government Finances. Washington: Bureau of the Census, 1978a. U.S. Bureau of the Census. Statistical Abstract of the U.S. Washington: Bureau of the Census, 1978b. Zimmerman, J. L. "The Municipal Accounting Maze: An Analysis of Political Incentives." Journal of Accounting Research-Supplement, 1977, pp. 107-0.

POLITICAL ENVIRONMENT NATURE OF POLITICAL COMPETITION EFFECT ON DISCLOSURES Taxed Groups Dominate Subsidized Groups Dominate Neither Taxed nor Subsidized Groups Dominate High Level of Political Competition within Subsidized Groups Low Level of Political Competition within Subsidized Groups Intense Political Competition between and within Taxed and Subsidized Groups GAAP Disclosures No GAAP Disclosures GAAP Disclosures Figure 1. Summary of the Effect Disclosure Practices of States. of the Political Environment on GAAP

TABLE 1 GAAP VS NON-GAAP STATES GAAP STATES 28 NON-GAAP STATES 22) Alaska California Colorado Florida Hawaii Illinois Indiana Kentucky Louisana Maryland Massachusettss* Michigan Minnesota Missouri Montana Nevada New Hampshire New Jersey New Mexico* New York North Carolina Oregon Pennsylvania Rhode Island* Tennessee Utah Virginia Washington Al abama Arizona Arkansas Connecticut Delaware* Georgia* Idaho Iowa Kansas Maine Mississippi* Nebraska North Dakota Ohio Oklahoma* South Carolina South Dakota Texas Vermont West Virginia Wisconsin Wyoming *No response to the State Comtrollers: Technical Activities and Functions Study, 1986

TABLE 2 VARIABLE DESCRIPTIONS GAAP DISCLOSURE VARIABLES: VARIABLE NAME INGRAM NCGA GAAPCODE POLITICAL COMPETITION PROXY VARIABLES' POPG8003 MONY7930 COLLEGE NEWSPR DESCRIPTION Based on disclosure index used in Ingram (1984) Disclosure index based on NCGA pronouncements as of 1984 (a) Dummy variable that indicates ( GAAP compliance Population per Square Mile (b) Per Capita Money Income in 1979 Percent of the Population with Four Years of College Per Capita Pounds of Newspaper Circulated Urbanization Index ) Total amount of grajg5 from the federal government (e) General Obligation Long-term debt ( Salary of the Accounting Administrator Political Competition Ranking Index Political Competitiog)Ranking Index minus GRANTS ranks (p.) URBAN GRANTS G.O.DEBT SALARY CINDEX GTINDEX RANNEY 1976 Interparty Competition Index '~ RAN84 1984 Interparty Competition Index (h' Data Sources: (a National Association of State Comptrollers Technical Activities and Functions, Lexington, Kentucky: Council of State Governments, 1986. U.S. Bureau of the Census, County and City Data Book [Machine - Readable datafile]. Washington: Bureau of the Census, 1983. (c) (Newsprint Information Committee. Newspaper and Newsprints Facts at a Glance, 83-84, 1984. Statistical Abstract of the U.S., 1983. (e)State Government Finances, 1984 and review of annual financial reports. ( Scammon, R. M. and McGillivray. American Votes 15 and 16. Washington, D.C.: Elections Research Center, 1983 and 1985. (a) gJacob, H. and Vines, K. Politics in the American States: A Comparative Analysis. Boston: Little, Brown and Company, 1976. (h)Congressional Quarterly, Inc. Congressional Quarterly Weekly Reports, 1974-1984. )

TABLE 3 DESCRIPTIVE STATISTICS VARIABLE G.. DEIBT GRANTS CINDEX GTINDEX RANNEY RAN84 CPACODE POPG8001 INGRAM NCGA GAAPCODE N 50 50 50 50 50 50 50 50 50 50 50 MEAN 1359854.300 1986600.140 177.000 152.110 0.584 0.542 0.300 4518149.440 6.780 14.230 0.560 STD DEV 1666435.041 2181497.049 67.641 59.207 0.191 0.186 0.463 4715037.753 4.037 10.808 0.501 MEDIAN 650624.500 1350474.000 167.500 141.500 0.549 0.518 0.000 3066433.000 8.500 20.000 1.000 MINIMUM 0.000 318028.000 61.000 55.000 0.320 0.201 0.000 401851.000 0.000 0.000 0.000 MAXIMUM 6902815.000 11092526.000 314.000 272.000 0.993 0.882 1.000 23667902.000 12.000 29.000 1.000

TABLE 4 SPEARMAN AND PEARSON CORRELATION MATRIX BETWEEN GAAP DISCLOSURE VARIABLES AND VARIABLES THAT PROXY FOR THE LEVEL OF POLITICAL COMPETITION G.O.DEBT GRANTS CINDEX GTINDEX INGRAM.5662*** (.4750***).2742** (.3051**).5246*** (.5339***).5314*** (.5482***).0121 (.08813) -.0313 (.0211).1474 (.1503).2816** (.2989**) NCGA.4977*** (.5723***).2500* (.3139**).4458*** (.4138***).4441*** (.4081***).0622 (.1565).0318 (.0723).2036 (.2736*).2276 (.2971**) GAAP CODE.4993*** (.5796***).2764*** (.3127**).5096*** (.5040***).5145*** (.5194***).0889 (.1606).0012 (.0140).1623 (.1623).2568* (.3043**) RANNEY RAN84 CPACODE POPG8001 * Significant at 10 ** Significant at.05 *** Significant at.01 NOTE: Spearman coefficient in parentheses

TABLE 5 CORRELATION MATRIX FOR EXPLANATORY VARIABLES CODEBT CIPACOI)DE GRANTS CINDEX GTINDEX RANNEY RAN84 POPG8001 G. O. DEBT 1.000( -.1016 (1.000) (-.1016). 6459*** (.5768***).5609** (.6287***).4994*** (.5781***).0871 (.2834**).1736 (.3290**).6224*** (.5414***) CPACODE 1.000 (1.000) -.1791 (-.0771).0310 (.0514).0628 (.0741) -.1141 -.1958 -.1884 (-.0696) (-.1890) (-.1134) GIANTS 1.000 ( 1.000).6029*** (.6375***).5016*** (.4773***).0081 (.2800**).0377.9598*** (.2687*) (.9730***) CINDEX 1.000 (1.000).9818*** (.9746***) 1.000 (1.000).0061 (.0955) -.0463 (.0432).0581 (.0631).0028 (.0032).6318*** (.6474***).5172*** (.4974***) GTINDEX RANNEY 1.000 (1.000).8600*** (.8628***) 1.000 (1.000) IRAN84.0754 (.2948**).0840 (.2842**) 1.000 (1.000) POPG8001 * significant ** significant *** significant NOTE: S pea rman at.10 at.05 at.01 coefficient in parenthesis

TABLE 6 OLS REGRESSION RESULTS FOR NCGA DISCLOSURE INDEX REGRESSION COEFFICIENTS PEi'E D [ CT E D SIGN CONSTANT POPULATION G.O. DEBT CPACODE MODE L1 10. 493*** (5.498) (+) -3.07E-7 (-.836) (+) 3.7E-6*** (3.622) (+) MODEL2 2.201 (.569) -5.82E-7 (-1.453) 3.16E-6*** (3.063) 3.617 (1.246) MODEL3 1.393 (.350) -7.2E-7 (-.662) 3.38 (1.07) MODEL4 7.252* (1.546) -2.13E-7 (-.576) 3.7E-6*** (3.592) 4.88** (1.638) MODEL5 9.368** (2.01) -2.15E-7 (-.581) 3.76E-6*** (3.592) 4. 677* (1.548) MODEL6 2.72 (.704) -6.97E-7** (-1.780) 3. 11 E-6*** (3.002) MODEL7 3.371 (.920) -5.6E-7* (-1.513) 3.15E-6*** (3.045) MODEL8 9.792** (2.173) -3.09E-7 (-.831) 3.75E-6*** (3.571) MODEL9 12.290*** (2.84) -3.13E-7 (-.843) 3.84E-6*** (3.62) GRANTS (+) 1.84E-6 (.794) MEASURES C1NDEX GTINDEX RANNE1 Y OF POLITICAL (+) COMPETITION:.0524** (2.001).0589*** (2.282) (+).0751*** (2.637).0598*** (2.245) (-) 2.432 (.343) 1.235 (.172) RAN84 (-) R2.259.356 Adjusted R2.259.299 Root MSE 9.501 9.05 F-Statistic 8.201*** 6.227*** * significant at.10 ** significant at.05 *** significant at.01 NOTE: a. one-tailed t-test b. t-statlstics are in parentheses.228.159 9.911 3.316**.301.239 9.43 4.841*** -1.266 (-.169).300.237 9.44 4.809***.334.291 9.103 7.692***.332.288 9.118 7.617***.259.211 9.601 5.364*** -3.461 (0.464).262.214 9.582 5.448***

TABLE 7 OLS REGRESSION RESULTS FOR THE INGRAM DISCLOSURE INDEX GRGRESSION COEFFICIENTS P KREl)1 CTED SIGN CONSTANT POPULATION G.O.DEBT CPACODE GRANTS MODEL 1 5.256*** (7.216) (+) -1.9E-8 (-.139) (+) 1.18E-6*** (2.983) (+) MODEL2 1.365 (.951) -1.77E-07 (-1.197) 8.8E1-7*** (2.296).9355 (.867) MODEL3 1.133 (.795) 2.18E-8 (.056).8638 (.761) 2.62E-8 (.032) MODEL4 4.814*** (2.663) 1.127 (.079) 1.17E-6*** (2.949) 1.52* (1.321) MODEL5 5.719*** (3.203) 5.26E-9 (.037) 1.21E-6*** (3.031) 1.381 (1.193) MODEL6 1.500 (1.053) 2.08E-7* (-1.441) 8.69E-7*** (2.273) MODEL 7 1.695 (1.263) -1.46E-7 (-1.076) 8.77E-7*** (2.311) MODEL8 5.603*** (3.258) -1.86E-8 (-.131) 1. 18E-6*** (2.96) MODEL9 6.582*** (4.014) 2.34E-8 (-.167) 1.2E-6*** (3.081) (+) MEASURES C1NDIEX GTINDEX RANNEY RAN84 OF POLITICAL (+) COMPETITION:.0268*** (2.752).0284*** (2.993) (+).0344*** (3.38).0299*** (3.064) (-) -.2383 (-.087) -.6106 (-.223) (-) R2.226 Adjusted R2.193 Root MSE 3.626 F-Statistic 6.858*** * significant at.10 ** significant at.05 *** significant at.01.363.306 3.36 6.404***.292.292 6.78 4.631***.256.189 3.63 3.86*** -1.905 (-.664).262.263 3.617 4.008***.352.310 3.354 8.332***.357.315 3.341 8.518***.227.176 3.66 4.496*** -2.554 (-.903).239.190 3.63 4.825*** NOTE: a. one-tailed t-test b. t-statistics are in parenatheses

TABLE 8 LOGISTIC REGRESSION RESULTS FOR GAAP CONFORMANCE REGRESSION COEFFICIENTS PREII) ICTED SIGN CONSTANT POPUI ATION G.O.DEBT CPACODE GRANTS MODEL 1 -.9887** (3.33) (+) -5. OE-8 (.17) (+) 1.48E-6*** (8.69) (+) MODEL2 -3.104*** (6.22) 1.7E-7 (1.28) 1.2E-6*** (5.83).8889 (1.17) MODEL3 -3.687*** (9.85) -8.6E-7** (2.65).4901 (.42) 2.42 (c) MODEL4 -1.392 (1.27) 5.OE-8 (.14) 1.57E-6*** (8.75) 1.248** (2.57) MODEL5 -.0961 (.01) -5.OE-8 (.10) 1.84E-6*** (9.57) 1.909* (1.88) MODEL6 -3.063*** (6.27) 2.OE-7** (1.86) 1.11E-6*** (5.53) MODEL7 -2.925*** (6.56) -1.4E-7 (1.12) 1. 14E-6** (5.85) MODEL8 -.7407 (.47) -5.OE-8 (.14) 1.51E-6*** (8.36) MODEL9.4976 (.22) -4.OE-8 (.09) 1.83E-6*** (9.45) (+) MEASURES CINDEX GTINDEX RANNEY RAN84 OF POLITICAL (+) COMPETITION:.0149** (2.65).0177** (3.96) (+).0080*** (7.42).0178** (4.12) (-) -.1670 (.01) -.4850 (.07) (-) MODEL CHI-SQUARE R (Similar to R2) -2 Log Likelihood CONCORDANT PAIRS SOMER'S DYX * significant at ** significant at *** significant at 12.70***.471 49.36.828.656.10.05.01 19.23***.496 43.71.882.765 14.31***.394 49.97.830.661 14.61***.452 46.57.856.711 -2.950* (1.73) 14.54***.481 44.70.873.747 18.54***.508 44.90.887.774 18. 601***.511 44.70.880.760 12.81***.440 49.29.830.659 -3.3987* (2.43) 13.10***.482 46.65.860.721 NOTE: a. one-tailed t-test b. Coefficient Chi-square statistics are in parentheses. c. Grants is a redundant variable in the model, therefore, no statistics were produced.

TABLE 9 LOW DEBT vs. HIGH DEBT STATES HIGH DEBT 25) LOW DEBT(25) Alaska California Connecticut Florida Georgia Hawaii Illinois Kentucky Louisiana Maryland Massachusetts Michigan Minnesota New Jersey New York North Carolina Ohio Oregon Pennsylvania Rhode Island South Carolina Tennessee Virginia Washington Wisconsin Al abama Arizona* Arkansas* Colorado* Delaware Idaho Indiana* Iowa* Kansas* Maine Mississippi Missouri Montana Nebraska* Nevada New Hampshire New Mexico North Dokota Oklahoma South Dokota* Texas* Utah Vermont West Virginia* Wyoming* NOTE: 1. Median Value of G.O. Debt is $650,624,000 2. *No general obligation debt outstanding per U.S. Census, State Government Finances, 1984

TABLE 10 HIGH COMPETITION vs. LOW COMPETITION STATES BASED ON RAN84 INDEX IGH(29) HIGH LOW(21) LOW Alaska Arizona Colorado Connecticut Delaware Idaho Illinois Indiana Iowa Kansas Maine Michigan Minnesota Missouri Montana Nebraska Nevada New Hampshire New York North Dakota Ohio Oregon Pennsylvania South Dakota Tennessee Utah Vermont Washington Wyoming Alabama Arkansas California Florida Georgia Hawaii Kentucky Louisana Maryland Massachusetts Mississippi New Jersey New Mexico North Carolina Oklahoma Rhode Island South Carolina Texas Virginia West Virginia Wisconsin NOTE: States with RAN84 values greater than 45 are classified as high competition states.

TABLE 11 PROPORTION OF STATES THAT ARE GAAP GIVEN G.O. DEBT OUTSTANDING AND LEVEL OF POLITICAL COMPETITON High Competition Low Competition High Debt Low Debt 9/ 11/ 11 14 7/ 1 18 /7 NOTE: Constructed from data presented in Tables 1, 9, 10.