The purpose of this chapter is to set out the definitions, parameters, data sources and methodologies of this comparative investigation. It will begin by building on the discussion of the previous chapter by specifying the operational definition for
intergovernmental agreements that will serve as the foundation for the rest of this study. This will be followed by a brief explanation of the qualitative approach to the analysis and comparison of agreements. Next, the sources of data for the seven factors which are hypothesized to have an effect on agreement formation will be described. Finally, the chapter will conclude by explaining how the data were obtained for the federations that were included in this study as well as the difficulties presented by other potential cases that could not be incorporated into this analysis.
An Operational Definition for Intergovernmental Agreements
The literature review of the preceding chapter indicated that intergovernmental agreements are understudied in many national contexts and virtually neglected in a comparative setting. This discussion also built on the foundation of existing scholarship to ground this investigation in an institutional paradigm. While defining the study of intergovernmental agreements as an investigation of intergovernmental institutions is a necessary step in crafting hypotheses and establishing the theoretical basis of this work, it provides limited guidance in clarifying exactly the type of data that must be gathered. As intergovernmental agreements cover a broad range of structures from informal contact between two government officials, such as a telephone conversation, all the way to national constitutional amendments, a specific definition is a necessary step for any
investigation. Unfortunately, there are few models to draw upon in order to craft an operational definition for intergovernmental agreements that can be used as a starting point for this research. As such, this comparative inquiry will propose its own definition and parameters for the study of agreements, rather than attempt to rely on any existing notions.
This study will define an intergovernmental agreement as a single policy project that is formal, written and national in scope. By being a single policy project, this restricts an agreement to a single topic, but not necessarily a single document. This is best illustrated by an example: over the last decade, the Canadian federal government has signed a series of bilateral immigration agreements with each of the ten provincial
administrations. While there are ten separate documents, they are all part of a single federal effort to coordinate immigration responsibilities with the provinces. Under this
definition, such a project would be counted as a single agreement, rather than ten.86 This
stipulation also excludes renewals of the same agreement. The requirement that agreements must be formal and written simply means that they must be publically
accessible and have an agreed-upon text.87 This is a methodological necessity rather than
a dismissal of informal instances of intergovernmental agreements. It is possible to identify and study public, written agreements and compare them among a number of federations; it would be difficult, if not impossible to effectively assemble anything
86 This approach also avoids "agreement inflation" for federations that have a larger number of states or
provinces or those which use a "hub and spoke" model to national agreements, with the federal government signing similar - if not quite identical - agreements with each subnational government rather than
participating with all units in a single document.
87 Note: this definition of agreements makes no distinction between agreements based on their legal status.
Agreements that can be enforced via a judicial process are as valid as those which cannot be, provided they meet all of the criteria.
approaching a complete database of informal connections and accords in a single country,
let alone several.88
The reason for studying national intergovernmental agreements - as opposed to all formal accords, regardless of the number of participants - is two-fold. First, as described in the theory discussion, national agreements represent changes made to an entire
federation and thus, are interesting elements of a federal system's evolution. This is true whether agreements are formed solely among the constituent units or when the national government is involved (state-state agreements and federal-state agreements). Two - and more germane to this methodology chapter - studying all intergovernmental agreements, even bilateral ones, across multiple federations would be exceedingly difficult. For those federations for which there is not a pre-existing and accessible database of all agreements (i.e. almost all of them) a federation's complete listing of agreements could only be constructed by obtaining complete records from every subnational government within a country. The likelihood that every province, state, canton or länder in a federation
possesses a complete and publically accessible record of all agreements is highly unlikely
for even one country, let alone seven.89 For methodological as much as theoretical
reasons, the focus of this comparative inquiry is on national agreements only.
Because of these considerations, only agreements which include more than 90% of a federation's subnational governments will be counted in this study. This threshold was selected to ensure that all the agreements included in this analysis had an
overwhelming majority and fully-national participation rate without discounting large
88 As an example of the difficulty, consider Zimmerman's description of informal agreements in the United
States, where "thousands" of these have been created. Zimmerman, Interstate Cooperation, 163-165.
89
This assumption was only confirmed by this research. In only one case, the United Kingdom, was there anything approaching a complete record of agreements for each of the constituent units.
agreements because they did not achieve strict unanimity. For example, an agreement in the United States that had been signed by forty-nine states could certainly be described as "national", but if the threshold of this study was unanimity, it would be disqualified based on the actions of one government in fifty.
The final parameter that must be established is the time frame for this comparative inquiry. The general time period that will be studied begins in 1945, at the end of the Second World War, and concludes in 2008. This is the "general" time period because only a handful of federal systems have existed since 1945. Even a federation as
paradigmatic as Germany has not functioned as a federal system for the entire period, to say nothing of newer examples such as South Africa or Bosnia-Herzegovina. Because of the potential differences in time periods between the federations in this comparison, a yearly average (by country) will feature in this analysis in addition to the consideration of the complete records of national agreement formation. This average of agreements created per year will serve as the principle means of comparing the number of agreements found in each federation and ranking them relative to the other cases in this study.
Using these definitions and parameters, records of national intergovernmental agreements were found for the following seven federations: Australia, Canada, Germany, South Africa, Switzerland, the United Kingdom and the United States of America. The final section of this chapter will discuss the selection of these cases, the sources of the data, and the other cases for which sufficient information could not be found.
Studying Intergovernmental Agreements: Quantitative versus Qualitative Measurements
The principal goal of this investigation is to increase our understanding of intergovernmental agreements and attempt to determine how a federation's institutional characteristics affect the frequency of their creation. This will be accomplished by analyzing a federation's record of agreement formation - the number of agreements, potential patterns in the frequency of their creation, and, where possible, the policy areas they occupy - and comparing this against the seven institutional features hypothesized to affect agreement formation. This approach should not be mistaken for a form of
quantitative methodology that would rely on statistical methods to determine correlations. With only seven federations, there are simply not enough data to examine the statistical
relationship between the number of agreements and various institutional features.90 Even
if the data set was large enough, a purely quantitative approach would not be desirable because features such as the constitutional division of powers are better addressed
through a qualitative methodology which can adequately explore and compare the various nuances of federal constitutional arrangements.
Although this is not a statistical analysis, the central measurement of
intergovernmental agreements is numerical as opposed to a more qualitative approach. The measurement of the frequency of agreement formation was chosen instead of another measurement, such as the importance of an agreement (in monetary or constitutional terms) or the longevity of certain agreements for two reasons. First, the data required to list an agreement's title, the date it was agreed upon, and its signatories are the most basic, and thus the most accessible, information that could be sought regarding agreements. The intention behind this decision was that this approach would allow data to be gathered for
90 Seven federations still represents more than a quarter of the approximately twenty-five federal countries
that currently exist (plus or minus three, depending on the definitions used). Yet, even if 80% of these federations could be included, the data would still be too limited for proper statistical analysis.
a much larger number of federations than might be possible using a method which
required greater detail.91 Second, focusing on the number of agreements removes the
need to establish some kind of qualitative measurement that might privilege certain types of agreements over others. For example, rather than examine intergovernmental
institutionalization by analyzing the number of agreements, one might instead examine the proportion of funds which are dependent upon agreements or perhaps the number of laws, federal or state, which arise from an agreement. Even assuming that such
information could be found, it might produce a perspective on institutionalization that is more applicable to some federations than to others. If one federation's welfare programs are governed by a series of agreements, but a second federation has the same for its criminal justice system, is the first federation necessarily more institutionalized if money is used as a unit of measurement? Counting the number of institutions, while by no
means perfect, avoids this difficulty.92 Each agreement made is a decision point -
political actors in a country can choose to institutionalize or not. The quantitative data provide a benchmark for comparison among countries. They also permit systematic study of the relationship between agreement formation and the measurable (or estimable)
position of the federation on the independent variables. Thus, the frequency of
91 Given the difficulty in finding even basic information regarding agreements in several countries, this
proved to be a prescient concern. See the last section of this chapter for more details regarding the search for agreements.
92 While there is not enough information to conclusively demonstrate this, there seems to be some
correlation between the frequency of agreement formation and their importance. Looking at the federations' records of agreements (found in the appendices) and the types of agreements formed suggests that those countries which form the most agreements, also seem to be forming agreements that are "larger" in terms of money spent and legal/constitutional consequences. At least among these cases, there is not a good example - with the possible exception of the nascent federal system in the UK - of a federation with a small number of agreements that exclusively address highly significant (defined in legal or monetary terms) policy matters.
agreements serves as the best and least ambiguous method of evaluating
intergovernmental agreements and the institutional environment of a particular federation.
Testing the Hypotheses: Sources of Data for the Seven Variables
Chapter Two provided an explanation of seven factors that may have an effect on a federation's propensity to form national intergovernmental agreements. While this discussion provided reasons why these variables would affect agreement formation, it left the question of how they will be operationalized largely unanswered. In the best case scenario, consistent and unambiguous measurements would exist across all seven federations which would aid in the analysis and comparison of the cases. Although this may be possible in the consideration of certain variables such as the number of
subnational governments, it is not sufficient, possible or even desirable for all of them. For instance, it would be difficult to fully summarize and explain the nuances of
decentralization in a federation only through the use of a single numerical variable. Thus, a comprehensive approach relying on multiple sources, instead of simply one "silver bullet", provides a broader foundation for considering these factors.
For all seven of the variables, secondary sources provide a large proportion of the information used in the analysis. While intergovernmental agreements may be
understudied, there have been a number of works on other features of federalism such as the constitutional division of powers, second chambers and intergovernmental relations that provide useful insight into these variables. This investigation will utilize both comparative works on federalism as well as country-specific studies. Where possible, quantitative measurements - of factors such as government spending - have been used to provide an additional metric that can apply similarly in all cases. Each variable will be
investigated separately at first, before considering some of the relationships between these features in Chapter Eleven (Comparative Analysis).
1. The Degree of Constitutional Overlap
In addition to the other secondary sources, the analysis of the effect of
constitutional overlap also utilized one comparative measurement. In Comparing Federal
Systems, Ronald Watts includes an appendix comparing "the distribution of powers and
functions in selected federations".93 This chart lists more than fifty different common
policy areas and it notes which order of government, federal or state (subnational), has
jurisdiction, as well as whether jurisdiction is shared.94 While this is not a perfect
comparison - several powers, such as nuclear energy, must be inferred from other enumerated powers - it allows for standardized evaluations of several federal constitutions. With respect to overlap, this comparison allows for a universal measurement of how many policy areas (in both raw totals and percentages) in each federation involve the activities of more than one order of government. Again, this information requires further explanation and context from other sources, but provides at least one means of simplifying and generalizing a topic that is normally difficult to quantify and compare.
Aside from an over-reliance upon this measurement, the one notable difficulty is the limited number of federations included in Watts' original appendix: while five of the seven federations studied here are included, South Africa and the United Kingdom are not. An attempt was made to replicate Watts' exercise using the division of powers found
93 Watts, Comparing Federal Systems, 3rd ed., 193-198. 94
Watts' comparison distinguishes two types of shared jurisdiction: "concurrent" and "separate, but overlapping" responsibilities of both orders of government.
in those two countries in order to provide the data for a full comparison among all seven
federations.95 While this may have created some slight inaccuracies (for example, Watts'
original work seems to have relied more upon inferring powers from broad authority), the results are both unambiguous (both South Africa and the United Kingdom have a clear predisposition to either overlap or centralize) and consistent with the other five cases in terms of the total policy areas enumerated. In this way, it was possible to incorporate the two missing cases into this comparison, allowing it to serve as a common measurement of constitutional overlap.
2. The Degree of Centralization in the Constitutional Division of Powers
As with overlap, the centralization variable relies on Watts' comparison of federal constitutions and their divisions of powers. The process is effectively the same, with the sole difference that while overlap is concerned with shared jurisdiction, evaluating the degree of centralization puts the focus on areas of exclusive jurisdiction. While it is not a definitive measurement, a federation with a large number of policy areas held exclusively by the national government, and fewer by the subnational governments, is likely trending towards centralization. Again, this is measured by both considering the raw totals of the jurisdictions as well as the percentages of the total policy areas. This observation is even more salient when studying intergovernmental agreements: the more policy areas in which the federal government can act upon unilaterally, the fewer areas which may require some form of intergovernmental coordination, including agreements.
95 As the United Kingdom has no single, written constitution, the division of powers found in the
Devolution Acts of each national administration were used. Please see the chapter on the UK for more information.
3. The Size and Status of the Federal Spending Power
As one of the two variables with an explicit financial element, the size and status of the federal spending power lends itself to comparative quantification. Specifically, the size of the spending power can be measured by determining how reliant subnational governments are upon federal transfers. As the volume of these transfer payments increase, the opportunities for intergovernmental collaboration will do so as well. The primary measurement of this was the data gathered from the International Monetary
Fund's (IMF) World Financial Yearbook. This annual publication lists a category
measuring the percentage of subnational revenue that is received from federal grants. This source provided data for six of the seven cases - as the United Kingdom is not officially a federation, information regarding transfers between London and the devolved administrations was not included. Fortunately, the data on the UK`s transfers were
available within the budgets of the governments of Northern Ireland, Scotland and Wales. Although this may not be measured by exactly the same criteria that the IMF used, it provides some basis for comparison. Additional sources, notably Ronald Watt's extensive
discussion of federal transfers, were used to bolster and confirm this data.96
While the size of the spending power was easily quantifiable and comparable, the status of this power - its usage, legal standing and role in intergovernmental relations - is equally important and requires a more nuanced approach. Secondary sources and federal constitutions provided the information necessary to understand the workings of the spending power in each case.
4. The Size and Scope of the Welfare State
Much like the size of the federal spending power, any evaluation of the welfare state requires some kind of comparable quantification. Although no attempt is made to try to determine a precise relationship between each additional dollar spent on welfare and the likelihood of more agreements - such an effort would be quixotic at best - the relative levels of spending do provide a clear and comparable measurement, if not a