intergovernmental structures and federalism in canada1 · intergovernmental structures and...

17
ECPR Conference 2015 – Montreal Intergovernmental Structures and Federalism in Canada 1 Jean-Philippe Gauvin, Université de Montréal Abstract: In its most basic aspect, federalism is primarily practiced through intergovernmental relations (IGR) between officials of federal and subnational units. The literature on Canadian federalism often highlights the importance of yearly meetings for policy coordination and harmonization in the country. However, as with other federations, the ways in which these relations are structured tend to vary a lot between policy sectors. While some sectors hold very informal meetings in a non-consistent manner, others manage relations through IGR councils that have permanent staff and secretariats. Few political scientists have produced systematic studies of the effects of such variation on intergovernmental collaboration. In this paper, we ask whether the varying degrees of IGR institutionalization across policy sectors in Canada affect collaboration between governments. The analysis relies on a systematic comparison of IGR meeting press releases for the years 1997-2014, which combines manual coding and automatic scaling. 1 First draft of the paper. Please do not cite. Author contact: [email protected]

Upload: others

Post on 21-Jun-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

ECPR Conference 2015 – Montreal Intergovernmental Structures and Federalism in Canada1 Jean-Philippe Gauvin, Université de Montréal Abstract:

In its most basic aspect, federalism is primarily practiced through intergovernmental relations (IGR) between officials of federal and subnational units. The literature on Canadian federalism often highlights the importance of yearly meetings for policy coordination and harmonization in the country. However, as with other federations, the ways in which these relations are structured tend to vary a lot between policy sectors. While some sectors hold very informal meetings in a non-consistent manner, others manage relations through IGR councils that have permanent staff and secretariats. Few political scientists have produced systematic studies of the effects of such variation on intergovernmental collaboration. In this paper, we ask whether the varying degrees of IGR institutionalization across policy sectors in Canada affect collaboration between governments. The analysis relies on a systematic comparison of IGR meeting press releases for the years 1997-2014, which combines manual coding and automatic scaling.

1 First draft of the paper. Please do not cite. Author contact: [email protected]

Page 2: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

Intergovernmental Structures and Federalism in Canada

There is a consensus in the literature that in its smallest common denominator, federalism can

be expressed as “shared rule + self-rule” (Elazar 1987), where a federation is a polity that

contain both shared and individual powers between different orders of government. This vision

has enabled many scholars to draw important comparisons between federal countries (Watts

2008; Hueglin and Fenna 2010; Burgess 2006). But while there are similarities as to the federal

characteristics of these countries, it could be argued that there are as many ways to manage

federalism as there are federations. If in some countries, federalism is mainly practiced through

parliament, in others the focus might be on completely different policy arenas.

In order to understand how a federation works, it is necessary ask how exactly federalism is

managed. Through which institutions do intergovernmental relations occur? How do these

institutions work and how do they differ? These questions will let us go beyond the simple

outline of a federation and enables the highlight of processes through which decision-making is

done. This variety of settings needs to be taken into account when studying federal systems.

In this paper, we focus on the case of the Canadian federation. In Canada, federalism is mainly

managed through intergovernmental meetings of line departments, where ministers discuss

key issues and decide on policy priorities (Bakvis et al. 2009; Gauvin et al. 2015). This system is

much different than other systems, mainly due to its British heritage, as well as the relatively

small number of provinces in the country. The paper asks the following question: how does the

way in which federalism is managed influence intergovernmental work? Do differences

between institutions really bear an effect on what is generated by coordination, namely

intergovernmental outputs?

While earlier studies have focused on qualitative analyses of peak meetings (Simmons 2004;

Bolleyer 2006; Papillon and Simeon 2004), we argue that it is important to look quantitatively at

the system as a whole by looking at each meeting independently. Moreover, we need to take all

policy domains into account and look at how their differences influence output. To do this, we

conduct manual and automatic content analyses of intergovernmental communiqués in Canada

for the years 1997-2014.

Page 3: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

The Institutions of Federalism

Federalism is managed through its intergovernmental relations (IGRs). However, these relations

can occur in multiple arenas. As a result, intergovernmental relations are usually defined as

“the various combinations of interdependencies and influences among public officials – elected

and administrative – in all types and levels of governmental units with particular emphasis on

financial, policy and political issues” (Krane and Wright 1998, 1168). As such, the relation

between political and administrative relations is not always clear. Furthermore, many scholars

have observed over the years the development of more interdependencies in federal systems

(Simeon 2001; Lazar et al. 2003; Bolleyer 2006; Agranoff 2011). As these networks grow,

collaborative management becomes harder to grasp in many settings.

In the context of the American literature, scholars often studied intergovernmental relations as

part of public administration. Agranoff (2010) outlines four major epochs of IGR that have key

implications for the administrative body: law and politics, welfare state interdependency,

government partners and collaborative networking. The Law and Politics epoch is associated

with 19th century America, as political powers were still finding their footing. States and the

federal government acted mostly independently in their own jurisdictions. Administrations

were very hierarchical, with municipalities being managed by the States. Some cooperation did

occur, notably in relation to grants for public works (Elazar 1962), but relations were still

limited. Increased collaboration happened during the welfare state era. Creation of major

works, such as railways and the postal system required an impressive amount of coordination

(Carpenter). Tax collection to finance in the Second World War also contributed to the creation

of interdependencies between orders of government through administrative agencies. By the

1960s and 1970s, new partners were involved in programs. Grants, loans and joint ventures in

service delivery became increasingly important. This resulted in a public administration that

went beyond the traditional borders of government agency (Salamon 1995). Finally, recent

years have changed IGRs in the era of collaborative networking. As Agranoff (2011, S71)

explains, “networks of local governments, business associations, and economic development

agents have worked among themselves for four or five decades” and had extensive links with

higher-level governments.

Many lessons can be drawn from this brief overview of American federalism. First, temporal

dynamics are often at play in federal systems. Relations evolve over time and settings can

change to accommodate new actors. Second, there is a relationship between political power

and public administration. In the Canadian literature, however, emphasis has often been put

mainly on the political arena. Unlike many other federal systems, Canada combines both

federalism and a Westminster parliament. This means that it is divided into one federal

government and ten provinces and that each of these governments are elected with a prime

Page 4: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

minister and his cabinet. This importation of the British style of parliament – originally meant

for a unitary system – in a federation has serious implications.

Since parliamentary systems concentrate decisional powers in the hands of executives (Savoie

1999), cabinet members have been particularly important in Canadian federal-provincial

relations. In the absence of formal intergovernmental institutions, federalism has normally

been perceived as practiced by members of the executive branch through what many have

called “executive federalism” (Watts 1989; Smiley 1980; Smiley 1987; Simeon 2006 [1972];

Dupré 1988). This dynamic has evolved parallel to that of the United States. While relations

between governments before the 1960s focused mainly on constitutional issues and law,

economic policies in the Post-war era led to an increasing need for coordination. First ministers

and finance ministers gained important legitimacy in decision-making (Simeon and Robinson

1990).

In this form of federalism, the heritage of a British parliament suggests that federal-provincial

relations are maintained through meetings and conferences rather than through administrative

agencies. These meetings normally occur at regular intervals between various figures, ranging

from public officials to deputy ministers, to ministers and to first ministers themselves. The

relative stability of these relations is quite distinct from what can be seen in other federations,

such as the United States. Indeed, the way in which the parliamentary system interacts with

Canadian federalism requires governments to sit and talk about various issues that are shared

powers with respect to the Constitution, such as environment or agriculture. Another major

difference with other federations is the relatively small number of provinces in Canada. This

facilitates organization of nation-wide meetings and helps officials maintain perpetual

interactions and, eventually, trust ties.

As a result, officials and governments in Canada have developed multiple relations through

meetings in various policy domains. While the literature usually focuses on peak meetings, such

as the First ministers and Premiers meetings, they only represent the tip of the iceberg, as the

core of the intergovernmental work rather is done in line departments (Johns et al. 2007;

Inwood et al. 2011; Agranoff and McGuire 2003; Bakvis et al. 2009; Bakvis and Brown 2010).

But as meetings occur more often, trust ties are developed between public officials. Hence, the

way in which the Canadian IGR system is structured favours a mutual reinforcing mechanism:

meetings of political figures are supported by administrative officials that stay in contact. This

can ultimately lead to transformation of the structures, such as institutionalization of practices

and increase in efficiency. In other words, intergovernmental relations are dominated by civil

servants motivated by problem-solving (Inwood et al. 2011; Montpetit and Foucault 2012). This

echoes Agranoff’s collaborative networking era, where public administration act “not as old-

Page 5: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

fashioned audit-like reviewers or legal compliance checkers, but as working partners […]

engaging in mutual problem solving” (Agranoff 2011 p.S71).

Table 1. Examples of Canadian Ministerial Councils

Atlantic Council of Fisheries and Aquaculture Ministers (ACFAM) Canadian Council of Fisheries and Aquaculture Ministers (CCFAM) Canadian Council of Forest Ministers (CCFM) Canadian Council of Environment Ministers (CCME) Canadian Council of Resource Ministers (CCRM) Canadian council of tourism ministers Canadian Education Statistics Council (CESC) Canadian Endangered Species Conservation Council (CESCC) Canada Parks Council (CPC) Conference of ministers of education of Canada (CMEC) Council of Energy ministers (CEM) Council of Ministers responsible for Transportation and Highway Safety Forum of Labor Market Ministers (FLMM) Interprovincial Sport and Recreation Council (ISRC) Ministerial Conference on Francophone Affairs (MCFA) Northern Development Ministers Forum Wildlife Ministers' Council of Canada (WMCC)

Table 1 gives a list of examples of Canadian ministerial councils. It must be explained that many

policy domains meet through annual conferences without such formality. For example, health is

a major issue in Canada, with departments both at the federal and provincial levels. However,

even if health department officials coordinate regularly, relations are not organized through a

council. This is interesting, as a major multilateral agreement – the 2004 Health Accord – has

been in force for the years 2004-2014. We would expect relations in that domain to be more

formal. Hence, all councils and meeting cycles, or intergovernmental agencies (IGA), are

institutionalized to various degrees. More institutionalized IGAs have structured meetings and

thus meet more regularly. Since this gives opportunities for political figures and public officials

to coordinate and engage in puzzle solving, it could be argued that this could have an impact on

intergovernmental works and outputs.

Hypothesis 1: Institutionalized intergovernmental agencies are more likely to produce multiple

intergovernmental outputs than non-institutionalized IGAs.

Page 6: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

Methods

To test this hypothesis, we measure the size of intergovernmental outputs through different

means. Studies of Canadian federalism usually focus their analyses on visible and tangible

outcomes such as multilateral agreements or specific policy. This is problematic for two

reasons. First, as explained earlier, there is much more to federalism management than

multilateral agreements. Public officials are constantly in contact with their counterparts from

other governments, through face-to-face meetings, phone calls or even emails. It would thus be

impossible to gather a comprehensive account of day-to-day interactions (Inwood et al. 2011;

Johns et al. 2007), as access to these events is limited. Second, multilateral agreements are very

scarce. While smaller agreements have been concluded bilaterally between the federal

government and provinces in a variety of domains, the last major multilateral agreement dates

from the 2004 Health Accord. Since then, the conservative government has preferred bilateral

agreements tailored to each province’s need rather than promote multilateral action and

negotiations (Simmons and Graefe 2013; Gauvin 2012).

Looking at these agreements can enrich comprehension of specific events through qualitative

analysis. However, it does not lead to generalization. We aim to fill a gap in the literature by

suggesting a quantitative analysis of these processes. Gauvin et al. (2015) have shown that

other indicators were possible to visualize intergovernmental relations. In this paper, we rely on

a content analysis of press releases published after ministerial intergovernmental meetings.

Press releases, or communiqués, are usually published at the conclusion of ministerial or

premiers meetings and distributed directly to the press. Deputy minister meetings, however,

never produce press releases, as they are not accountable to the public. These documents can

be considered a very good indicator of what has been decided in a meeting, for various reasons.

Content size can vary, from a few simple paragraphs to many pages, giving some indication of

output size. Topics that were discussed are often mentioned, as well as which priorities should

be put forth in future meetings. For example, a communiqué might mention that ministers met

“with an agreement to continue to work together to address issues of central importance to

achieving equality for women”2. But priorities can also be much more precise, referring to

specific objectives or even sometimes programs.

Other elements can easily be identified, such as the existence of reports or backgrounders, the

tasking of officials, the creation of a new structure or new program, etc. The text itself is often

very neutral or embellished, but the tone is never negative. This makes it harder to identify

negotiation failures or political stalemates. However, the absence of a communiqué can itself

often be considered like a failure, since it was impossible for the ministers present to agree on

2 CICS website: http://www.scics.gc.ca/english/conferences.asp?a=viewdocument&id=1328

Page 7: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

something to tell the press. In the end, we argue that it remains a convincing indicator of

intergovernmental output.

In order to quantify the outputs of the documents, we use two distinct methods. The first was a

manual content analysis of the communiqués for the period 1997-2009. The second was an

automatic computer-based scaling technique called Wordfish.

The manual coding consisted in reading all communiqués and identifying a series of outputs.

Each output was considered a variable and was coded 1 if present or 0 if not present. The

results were then summed up to create an index. The first item is the mention of specific

priorities. The second was the mention of “agree”, such as “ministers agreed to […]”. The third

item refers to the mention of a specific agreement, such as a Canada-wide strategy. The fourth

identifies if a signature was included in the press release. Sometimes, ministers provide a

declaration or an agreement with their signature to show how it is official. Fifth is the mention

of a “spirit of collaboration”. Since most communiqués include such a mention, the lack of such

a phrase can indicate that not much resulted from the meeting. Sixth is the mention of a

document, such as a report or a backgrounder that the ministers agreed to publish. The seventh

item refers to mentions of a new structure, such as the creation of a council in a specific policy

domain that would aim to make meeting in that field more systematic. Finally, the last item

refers to mentions of a new program that has been agreed upon by the ministers.

The index is then created by adding all the values for each press release, giving a score from 0

to 8. The index is then transformed (by subtracting 4) to create an index that goes from -4 to 4.

A meeting that has a higher score had a bigger variety of outputs, while meetings with a lower

score didn’t disclose as much information. Of course, this requires us to assume that less

information means less agreement.

The second technique used to compare the communiqué is the Wordfish scaling model. In the

past ten years, there have been multiple new techniques of automatic scaling used in political

science. Most of them aim to position political parties by scaling a range of political texts, such

as manifestos. This offers a quick and cheap alternative to expert surveys and manual coding,

which can be time-consuming and very expensive. Some of these techniques, like Wordscores

(Laver et al. 2003), require texts to be scored manually and used as reference texts. Other

document positions can then be extracted from the scoring of these reference texts. In the

Canadian context, such techniques have been used to create time-series positioning of political

parties by scoring reference texts with expert manifestos and extracting policy positions from

other years (Gauvin et al. 2013).

Wordfish is a similar technique developed by Slapin and Proksch (2008) that is used to compare

texts between themselves and position them on a single dimension.. Contrary to Wordscores, it

Page 8: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

doesn’t require scored texts to be used as reference, because “it instead assumes an underlying

statistical distribution of word counts” (Slapin and Proksch 2008, 708). To be precise, Wordfish

assumes word frequencies to be distributed randomly as part of a poisson distribution. This

greatly simplifies the estimation, as only one parameter is estimated, λ, which is both its mean

and variance (Slapin and Proksch 2008, 709).

In practical terms, the documents need to be compiled and converted to a word frequency

matrix. This is done by counting the occurrence of each word. In the end, each document

obtains a vector of word frequencies so that the matrix is comprised of rows (documents) and

columns (frequencies). Further text-mining tools can be applied in order to remove stop-words

and stem remaining words. This reduces the amount of error that could be mistakenly scored

by the algorithm. When the matrix is ready, a statistical environment such as R can be used to

estimate a Wordfish model. At this point, the routine compares word usage between the texts

and extracts a latent dimension that explains frequencies. Estimates, along with standard errors

and bootstrapped 95% confidence intervals are also provided. Texts where word occurrences

are similar will get similar scores. However, extremely different documents will obtain very

different scores.

As an example, if three party manifestos are scored, party A might fall in the middle, because it

uses only “neutral” words, such as words that do not distinguish it from the other two parties.

Party B might fall on the left side of the scale, because it is very different than party C which

could fall on the right side of the scale. Party B might have used neutral and social terms, while

party C might have used economic and neutral words thus distinguishing themselves greatly

from each other. When conducting the routine, the user is required to specify two texts (such

as a left-wing manifesto and a right-wing manifesto) to give the results a direction, as it cannot

estimate what is “left” and what is “right”.

As it is currently used in the literature (Lo et al. 2014; Proksch et al. 2011), Wordfish makes

many important assumptions. The first is that word meaning remains the same over time. New

political terms necessarily induce some errors in the results. The second assumption is that

variance in word frequencies lead to an interpretation of political ideology. This latter

assumption might look problematic in the context of our study. Indeed, we are not studying

party positions, but rather intergovernmental press releases. But, as mentioned, Wordfish

scales texts according to their word usage. Hence, the assumption we make here is that

meetings that lead to fewer outputs tend to use similar words in the communiqués.

Consequently, when meetings lead to a wider range of outputs, it will use different words.

Following this logic, Wordfish will score difference in outputs rather than difference in ideology.

Another thing that needs to be said is that all communiqués are not independent. Conference

cycles and council meeting occur in each sector, with mostly the same staff, or at least the

Page 9: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

same administrative culture. In other words, the texts will help us see variation over time for

each policy domain, exactly like Wordfish can help see how party positions evolve over the

years.

Manual coding and Wordfish scoring will provide us with two comparable and interesting

measures of intergovernmental outputs. This leaves us to measure institutionalization of the

different councils and conferences. While different indicators could be used, such as regularity

of meetings, council or forum status, independence of action3, we opt here for a more simple

measure: the presence of a permanent secretariat. Councils that have a physical address, with a

permanent staff are by definition more institutionalized than meetings that occur sparsely with

little or no consistency. Hence, referring to our hypothesis, communiqués published after a

meeting for which there is a permanent secretariat should lead to more outputs than those

that do not have secretariats.

Other controls are included in our models. First, the year will facilitate the comparison over

time. Second, the level at which the meeting occurred will be of importance. We will compare

ministerial and premiers meetings. Finally, the type of meeting will be coded as 1 for Federal-

Provincial-Territorial and 0 for Provincial-Territorial.

Results

When looking more closely at the data, it becomes apparent that communiqué publication

greatly varies between political levels and conference types. Figure 1 provides descriptive

information on the subject. In terms of the type of meetings, it seems interesting that only 56

percent of the 311 provincial-territorial meetings that occurred between 1997 and 2014

produced a press release. However, 75 percent of the 465 FPT meetings at the first ministers

and minister level produced press releases. When looking strictly at the political level, 85

percent of the 74 First ministers and premiers meetings produced communiqués, as opposed to

66 percent of the 706 ministerial meetings.

These figures are interesting, as we would normally expect PT meetings to be more productive

– and thus at least produce a communiqué – because of the absence of the federal government

which could potentially lead to stalemates and conflicts in some policy domains. However, it

seems reasonable that First ministers and Premiers meetings would produce a larger number of

communiqués. These meetings have much more visibility and are often covered by the press.

3 Such measures will be tested in later versions of this paper.

Page 10: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

Figure 1. Percentage Published Communiqués by Type and Level

Data source: CICS

As for the content analyses, a total of 449 communiqués were scored with Wordfish for the

period 1997-2014 and 348 were manually coded for the period 1997-2009. The results are

shown in the figures A1 and A2 of the appendix. These figures present the estimates by policy

topic to account for administrative culture. With only few exceptions, the results are

impressively similar, suggesting that both our manually coded measure and the Wordfish

estimates capture a similar concept. For example, both measures find lower outputs for the

fields of finance, native affairs, housing, local government or even sports and recreation. Also,

in both instances, the higher results occur in agriculture and environment.

In order to test the role of institutionalization – the presence of a permanent secretariat – we

conduct a series of regression analyses, reported in Table 2. Model 1 consists of an OLS

regression on the manually-coded measure, the range of outputs mentioned in the press

release. The model suggests that notwithstanding conference type and political level, meetings

that are part of an institutionalized conference cycle increase output size by .32, on a scale

going from -4 for 4. Likewise, meetings that include the federal have a one-unit increase as

opposed to PT meetings. Model 2 aims to model the heterogeneity of policy domains by doing

a panel regression. More precisely, we use a fixed-effects model, which permits varying

intercept across policy topics. This time, the results are less evident. When looking at the

variation between policy sectors, institutionalization does not seem to bear an effect.

Page 11: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

Table 2. Regression Analysis of Institutionalization

Model 1 (Output

Range - OLS)

Model 2 (Output

Range - FE)

Model 3 (Wordfish -

OLS)

Model 4 (Wordfish -

FE)

Coef. SD. Coef. SD. Coef. SD. Coef. SD.

Institutionalized .32* .16 .04 .22 .61*** .15 .51** .18

Year -.01 .01 -.01 .01 .01 .01 .01 .01

Minister level -1.03 .57 -.78 .62 .17 .27 .63 .54

FPT .98*** .15 .81*** .18 .83*** .13 .30* .13

N. 345 345 449 449

Nb. groups 1 23 1 23

R2 .11 .17 .12 .15

Models 3 and 4 instead use the Wordfish estimates as the dependent variable. These results

are much more consistent with our hypothesis. Indeed, institutionalized meetings have on

average a .61 higher Wordfish score than meetings that lack a permanent secretariat. This

suggests yet again that institutionalization facilitates the production of intergovernmental

outputs. When running a fixed-effects model, the effect remains consistent. As with the

manually-coded measure, conference type is significant in both instances. However, the effect

is less pronounced in the fixed-effects model.

We explore these data even further by simulating predictive margins of the Wordfish estimates

over the presence of a secretariat. Simulations provide substantive interpretation of results by

measuring the predicted outcome at various specific values while at the same time controlling

for other effects. Figure 2 presents these results. As can be seen from the figure, when

controlling for all covariates over time, institutionalized meeting are predicted to have a higher

output size than meetings that lack a secretariat. Every year, this difference is statistically

significant, except for a very small and insignificant overlap for the year 2014. The results also

suggest a slight increase of the effect over time. This could be true if councils became more

institutionalized over time, leading them to have bigger outputs. However, the size of the

confidence interval prevents of accepting this as a convincing effect.

Page 12: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

Figure 2. Varying Effect of Institutionalization over Time

Conclusion

This paper has argued that the study of federal systems needed to take a closer look at how

federalism is managed before drawing conclusions. While American literature has been heavily

influenced by public administration studies, most research in Canadian literature has focused

only on the top-level interactions. This dissonance between the two countries could suggest

entirely different intergovernmental patterns. However, when looking more closely at the

institutions of Canadian federalism, namely the intergovernmental agencies, it seems that the

two countries share more elements than we might have first believed. The regularity of the

meetings over the years is supported by networks of public officials that coordinate and engage

in problem-solving.

The other lesson to draw from this paper is that different IGAs behave differently. Some policy

domains produce more IGR outputs than others. Amongst the most productive lie IGAs that are

more institutionalized than the rest. Hence, structuring federalism management in Canada does

bear some effects. Provinces cannot act in isolation. They need to coordinate, and the way in

which this is done makes the difference.

Page 13: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

References Agranoff, Robert. 2010. "Towards an Emergent Theory of IGR Governance at the Dawn of the

Network Era." In Governance and Intergovernmental Relations in the European Union and the United States, ed. E. Ongaro, A. Massey, M. Holzer and E. Wayenberg. Cheltenham, UK: Edward Elgar.

———. 2011. "Federalist No. 44: What is the Role of Intergovernmental Relations in Federalism?" Public Administration Review Special Issue December:S68-S77.

Agranoff, Robert, and Michael McGuire. 2003. Collaborative Public Management: New Strategies for Local Governments. Washington D.C.: Georgetown University Press.

Bakvis, Herman, Gerald Baier, and Douglas Brown. 2009. Contested Federalism: Certainty and Ambiguity in the Canadian Federation. Don Mills, Ontario: Oxford University Press.

Bakvis, Herman, and Douglas Brown. 2010. "Policy Coordination in Federal Systems: Comparing Intergovernmental Processes and Outcomes in Canada and the United States." Publius 40 (3):484-507.

Bolleyer, Nicole. 2006. "Federal Dynamics in Canada, the United States, and Switzerland: How Subtates' Internal Organization Affects Intergovernmental Relations." Publius 36 (4):471-502.

Burgess, Michael. 2006. Comparative Federalism: Theory and Practice: Routledge.

Dupré, Stefan. 1988. "Reflexions on the workability of executive federalism." In Perspectives on Canadian Federalism, ed. R. Olling and M. W. Westmacott. Scarborough: Prentice-Hall.

Elazar, Daniel. 1962. The American Partnership: Intergovernmental Cooperation in the Nineteenth Century United States. Chicago: University of Chicago Press.

Elazar, Daniel J. 1987. Exploring Federalism. Tuscaloosa, Alabama: Alabama University Press.

Gauvin, Jean-Philippe. 2012. "Les relations intergouvernemental au Canada: gage de succès ou d'échec?" Options Politiques 33 (7):38-41.

Gauvin, Jean-Philippe, Chris Chhim, and Mike Medeiros. 2013. "Did They Mind the Gap? The Distance between the BQ and Quebec Voters Over Time." In Canadian Political Science Association Annual Meeting. Victoria, BC.

Gauvin, Jean-Philippe, Éric Montpetit, and Martial Foucault. 2015. "Intergovernmental Attention and Government Priorities in Canada." In Comparative Agendas Project Conference. Lisbon, Portugal.

Hueglin, Thomas O. , and Alan Fenna. 2010. Comparative Federalism: A Systematic Inquiry. Toronto: University of Toronto Press.

Inwood, Gregory J., Carolyn M. Johns, and Patricia L. O'Reilly. 2011. Intergovernmental Policy Capacity in Canada: Inside the Worlds of Finance, Environment, Trade, and Health. Montreal: McGill-Queen's University Press.

Page 14: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

Johns, Carolyn M., Patricia L. O'Reilly, and Gregory J. Inwood. 2007. "Formal and informal dimensions of intergovernmental administrative relations in Canada." Canadian Public Administration 50 (1):21-41.

Krane, Dale, and Deil S. Wright. 1998. "Intergovernmental Relations." In International Encyclopedia of Public Policy and Administration, vol 2, ed. J. M. Shafritz. Boulder, CO: Westview Press.

Laver, Michael, Kenneth Benoit, and John Garry. 2003. "Extracting policy positions from political texts using words as data." American Political Science Review 97 (2):311-31.

Lazar, Harvey, Hamish Telford, and Ronald Watts. 2003. The impact of global and regional integration on federal systems: A comparative analysis. Kingston: McgIll-Queen's University Press.

Lo, James, Sven-Oliver Proksch, and Jonathan Slapin. 2014. "Ideological Clarity in Multiparty Competition: A New Measure and Test Using Election Manifestos." British Journal of Political Science Published online:1-20.

Montpetit, Éric, and Martial Foucault. 2012. "Canadian Federalism and Change in Policy Attention: A Comparison with the United Kingdom." Canadian Journal of Political Science.

Papillon, Martin, and Richard Simeon. 2004. "The Weakest Link? First Minister's Conferences in Canadian Intergovernmental Relations." In Canada: The State of the Federation, 2002: Reconsidering the institutions of Canadian Federalism, ed. J. P. Meekison, H. Telford and H. Lazar. Kingston: Institute of Intergovernmental Relations.

Proksch, Sven-Oliver, Jonathan Slapin, and Michael F. Thies. 2011. "Party system dynamics in post-war Japan: A quantitative content analysis of electoral pledges." Electoral Studies 30 (1):114-24.

Salamon, Lester M. 1995. Partners in Public Service: Government–Nonprofit Relations in the Modern Welfare State. Baltimore: Johns Hopkins University Press.

Savoie, Donald. 1999. Governing from the Center: The Concentration of Power in Canadian Politics. Toronto: University of Toronto Press.

Simeon, Richard. 2001. "Adaptability and change in federations." International Science Journal 52:145-52.

———. 2006 [1972]. Federal-Provincial Diplomacy: The Making of Recent Policy in Canada . Toronto: University of Toronto Press.

Simeon, Richard, and Ian Robinson. 1990. State, Society, and the Development of Canadian Federalism. Toronto: University of Toronto Press.

Simmons, Julie M. 2004. "Securing the Threads of Cooperation in the Tapestry of Intergovernmental Relations: Does the Institutionalization of Ministerial Conferences Matter?" In Canada: The State of the Federation, 2002: Reconsidering the institutions of

Page 15: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

Canadian Federalism, ed. J. P. Meekison, H. Telford and H. Lazar. Kingston: Institute of Intergovernmental Relations.

Simmons, Julie M., and Peter Graefe. 2013. "Assessing the Collaboration That Was “Collaborative Federalism” 1996-2006." Canadian Political Science Review 7 (1):25-36.

Slapin, Jonathan, and Sven-Oliver Proksch. 2008. "A Scaling Model for Estimating Time-Series Party Positions from Texts." American Journal of Political Science 52 (3):705-22.

Smiley, Donald. 1980. Canada in Questions: Federalism in the Eighties. Toronto: McGraw-Hill Ryerson.

Smiley, Donald V. 1987. The federal condition in Canada. Whitby: McGraw-Hill Ryerson.

Watts, Ronald. 1989. "Executive Federalism: A Comparative Analysis." Institute of Intergovernmental Relations (Queen's University).

———. 2008. Comparing Federal Systems. Kingston: McGill-Queen's University Press.

Page 16: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

APPENDIX Figure A1. Manually Coded Output Range by Policy Topic

Page 17: Intergovernmental Structures and Federalism in Canada1 · Intergovernmental Structures and Federalism in Canada There is a consensus in the literature that in its smallest common

Figure A2. Wordfish Estimates by Policy Topic