Ideology versus Economics during the Federalist Period
by Cyril Morong
ABSTRACT: This paper investigates the economic and ideological factors influencing Congressional Roll Call Voting during the Federalist Period in the United States, 1789-1801, by applying a discrete choice analysis to that period's roll call votes. This research is relevant because it helps explain the causes of change in the institutional structure of our economy, a necessary step for developing a theory of institutional change throughout history. The major institutional question addressed in the Federalist Period was the balance of power between the states and the federal government. This research combines elements of the work done by Kalt and Zupan on voting in the U. S. Senate in the 1970's and the work done by McGuire and Ohsfeldt on the framing and ratification of the constitution. The roll call votes for each bill are analyzed with a logit model with a congressman's vote. There are three types of explanatory variables: personal ideology, constituents' economic interests and constituents' ideology. This approach uses the Principal-Agent model. Ideology is defined in the Higgsian sense of a belief system that individuals use as a guide in their personal economic actions, their public policy preferences, and group identification. The results show that both economic interests and ideology were both important factors. Ideology was more important on legislation that lacked clear cut economic implications while economic variables were more important on bills that did. Ideology must be used by economists to explain public policy. The founding fathers formed two political parties, Federalist and Republican, based on economic and ideological differences. Both factors are also important in determing why the institutional structure changes.
Introduction
The public choice literature has now produced a long list of studies examining the relative impact of ideology and economic interests on the voting behavior of legislators (several of these are discussed below). Perhaps the most important of these is Kalt and Zupan (1984). The question that the literature addresses is "does ideology matter?" Some have concluded that it does not while others have concluded that it might and that further study is needed to find out how and why. Although this seems to be the burning question, none have been concerned with the problem of defining ideology. It is assumed that there is one definition that is appropriate and that all statistical results can be interpreted through this definition. In fact, most studies use the definition of ideology from the Kalt and Zupan (K-Z) work mentioned above. No alternative definitions are mentioned and there is no discussion on the relative merits of competing definitions. This will uses an alternate definition of ideology to interpret the statistical results from the analysis of congressional voting that took place during the Federalist Period. It is the one developed by Higgs (1987). It offers a broader view of human motivations than that typically used by economists and consequently has implications for how regression results might be interpreted and the question of "does ideology matter?"
This paper does five things. The first is to summarize the definitions of ideology that have been used in the literature. The second is to outline the Higgsian definition, its differences from and similarities to the other definitions, and to identify its implications for the interpretation of regression results. The third is to formulate several hypotheses based on this definition. The fourth is to present evidence related to those hypotheses. Finally, that evidence is discussed in the context of the literature and the assumptions that economists make about human behavior. One conclusion is that ideology plays a larger role when defined differently and more broadly. A second is that regression results must be carefully interpreted because the broader definition makes it unclear if a variable measures ideology, economic interests, or some combination of both. Email me for the the theoretical model and how ideology works in voting models.
The Federalist Period
During the Federalist Period, the balance of government power between the federal government and the states, perhaps the key issue of the Constitutional Convention, had to be sorted out. Although the drafting and ratification of the Constitution represents a significant change in that balance, much legislation was passed in the period of 1789-1801 that reflected issues similar to those dealt with at the Philadelphia convention. Some of the important legislation of this period also revolved around the financial plan of Alexander Hamilton.[1] Charles Beard considered this plan to be the logical payoff to the basic principles in the Constitution.[2] The period 1789-1801 is particularly relevant since it represents the era of power for the Federalists who were the strongest supporters of the Constitution. Therefore, any legislation passed during this time is important because it shows not only how the Constitution was being interpreted, but also how it was implemented. The interpretation and implementation of the Constitution certainly show the degree of power that was actually given to the national government relative to the state governments.
In light of the different interpretations of ideology seen in the literature, the question of the relevance of applying these models to such a far removed historical period must be addressed. In one sense, the work of McGuire and Ohsfeldt shows that the public choice voting models help explain the motives of the founding fathers. But more importantly, the founding fathers themselves did recognize, at least implicitly, that the principal agent relationship could cause problems for an "ideal" form of government. Two quotes from the Federalist papers illustrate this. The first, from No. 62 reads "It is a misfortune incident to republican government, though in a less degree than to other governments, that those who administer it may forget their obligations to their constituents and prove unfaithful to their important trust."[3] This clearly indicates that they knew of shirking and expected it to happen. The second, from No. 46 reads "A local spirit will infallibly prevail...in the members of congress ...Measures will too often be decided according to their probable effect, not on the national prosperity and happiness, but on prejudices, interests, and pursuits of the governments and people of the individual states."[4] The latter quote clearly indicates that they also expected that legislators would place a high value on the needs and interests of their constituents. They knew that both were important and therefore it is completely legitimate to attempt to model both motives in any analysis of congressional voting for this time period.
Choosing an ideology rating
Personal ideology of legislators is usually measured as a percentage of "correct" votes on a group of issues that may or may not be associated with each other. A vote is correct if it coincides with the viewpoint of the organization constructing the rating. For example, if the Americans for Democratic Action or ADA (a political watchdog group) finds a particular congressmen to have voted "liberal" 75% on of the votes in their sample, it gives him a score of 75. No such ideological rating currently exists for congressmen of the Federalist Period. Since the major issue of the period was the size and strength of the federal government, an ideological rating based on roll call votes for military questions is appropriate and indicative of a congressman viewpoint on the issue of of a strong versus a weak federal government. There are several justifications for using this measure. One is that two other major issues, the settling of western lands and foreign policy (mainly remaining neutral in European affairs) were directly affected by the size of the military and therefore the size and strength of the federal government. A second is that some historians of the period feel that Hamilton wanted a strong military to enforce his Federalist policies against states that may have resisted them.[5] Stourzh clearly shows Hamilton's belief in a strong peacetime military to keep the peace and preserve order.[6] Third, the Senate passed a bill giving the president authority to augment the army to execute the laws of the Union on February 16, 1799.[7] And last, there were two instances in which armed revolts against taxation were put down by the military. The first, which occurred just before the movement for the Constitution began (and helped to hasten it) is known as Shay's Rebellion. The second occurred in 1794 and is known as the Whiskey Rebellion because the protesters were opposed to paying the Whiskey Tax. The ideological rating then is a congressman's percentage of votes cast for a larger and stronger military or national defense. [8]
Since this study is using the Higgsian definition of ideology, its application to the Federalist Period must be explained. The important addition to the definition of ideology by Higgs is the fact that it helps to facilitate group membership. It will be assumed here that in the in the Federalist Period the the group membership that was facilitated by ideology was membership in the two political parties, the Federalist and Republican. One of the hypotheses to be tested is that ideology should become more important in explaining how congressmen voted when parties became important than before they were important. Email me for an explanation of the empirical model. The model used was a logit model since legislators could vote yes or no.
The Relevant Data
The data and variables used in the statistical analysis were picked for their general relevance to the larger issue of the Federalist Period, for their specific relevance to the bills analyzed and its availability. Data availability is a problem in many studies and probably even more of a problem in historical studies such as this one. This section discusses why these variables and data were chosen. Appendix B contains the details of each variable, the data used to compute it, how the computation was done and why that variable was relevant and its meaning. Here tables will be presented that contain brief definitions of these variables and their relation to the question of a strong national government. The first lists the variables used in the first stage of the two stage process that was discussed in the earlier section on the general model. The second lists the variables used in the second stage. Each variable is designated after its description as being expected to be negatively (A) or positively (P) related to the question of a strong national government. The specific reasons for this are listed in Appendix B. There is no reason to consider AGE and MF pro-national or anti-national. The variable WEALTH was not designated because it is not clear how it would be affected by a strong national government.
Table 1
Stage One Variable Definitions
______
Variable Definition
AGE The average age of male voters in each state.
MF The male-female ratio in each state.
POPDENS The population density in each state and district in 1790. P
URBAN The percent of each state's and district's population that lived in urban areas in
1790. P
ENG The percent of each state's and district's population that were of English ancestry.
DIST The distance form the center of a congressional district to the nearest navigable coastline. A
MILITIA The quota of Militia assigned to each state, per capita. P
WEALTH The per capita wealth of each state.
TAX A dummy variable that indicates if being a taxpayer was required by the state to
vote. A
PROP A dummy variable that indicates if being a property owner was required by the state to vote. A
______
Table 2
Stage Two Variable Definitions
______
Variable Definition
CARR The average annual revenue collected from the carriage tax in each state. A
CUST The average annual revenue collected from customs in each state. P
DIRTAX The per capita tax assessed in each state due to the Direct Tax of 1798. A
DIST The distance form the center of a congressional district to the nearest navigable coastline. A
EXP The average annual per capita value of exports from each state. P
FUND The per capita amount of money received by each state due to the Funding and Assumption Act. P
SLAV The number of slaves per 100 white persons in each state and district in 1790. A
SUG The average annual per capita gross revenue collected from refined sugar in each state. A
TON The average annual tonnage per capita in each state employed in the coasting
trade. P
VELL The average annual per capita gross revenue collected from stamped vellum, parchment and paper in each state. A
WEALTH The per capita wealth of each state.
WHIS The average annual per capita gross revenue collected from domestic distilled spirits and stills.
RATING The percent of pro-military votes cast by each congressman. P
MILRES The ideology residual for each congressman. P
FIT This the constituents' part of RATING. P
______
Estimation
The bills for which congressional voting was analyzed are listed in Appendix A. Each is described there and is designated as being pro-national or anti-national depending on whether it would increase or decrease the power of the national government.
The first stage in the statistical analysis is to estimate coefficients in an ordinary least squares regression on the ideology variable RATING (this allows fitted values to be calculated which can then be subtracted from the original RATING to get the residual and the constituents' part of RATING). The variables that went into this regression are seen in Table 1 (a more comprehensive explanation of the variables is in Appendix B). These variables were chosen in part because of their relevance to the Federalist Period but also because they are similar to the type of variables used in the literature. They represent a demographic profile of the constituency. The equation estimated in the first stage is