Who Spends More: Left or Right?
Economic Note showing the absence of correlation between the governing party's ideology and the evolution of public spending as a share of GDP
Public policy debates are often coloured by ideological preconceptions. For example, we expect political parties on "the left" to have a tendency to increase public spending when they are in power, and parties on "the right" to have a tendency to reduce it. This perception clearly stems from official statements that emphasize different goals. But what is the reality? Do these different ideological leanings systematically translate into increases in the size of the state under left-wing governments and reductions in the size of the state under right-wing governments?
Technical Annex for the Economic Note
Media release :: Do governments of the left spend more? Not necessarily, concludes an MEI publication
|Interview (in French) with Vincent Geloso (Radio-Canada, March 26, 2013)||Interview (in French) with Michel Kelly-Gagnon (ARGENT business news network, March 26, 2013)|
Economic Note prepared by Michel Kelly-Gagnon, president and CEO of the MEI, and Vincent Geloso, a Ph. D. candidate in economic history at the London School of Economics and an economist at the MEI.
Public policy debates are often coloured by ideological preconceptions. For example, we expect political parties on “the left” to have a tendency to increase public spending when they are in power, and parties on “the right” to have a tendency to reduce it. This perception clearly stems from official statements that emphasize different goals.
But what is the reality? Do these different ideological leanings systematically translate into increases in the size of the state under left-wing governments and reductions in the size of the state under right-wing governments? We have observed the evolution of the size of government for three states over the last 40 years to see if this is really what happens: the Canadian federal government, the Quebec government, and the American federal government.
Left and right are quite simplistic terms that can be interpreted in various ways. They are nonetheless widely used and do describe a certain political reality.
As is common in political science, we define them as follows, considering for our purposes solely the economic dimension: a left-wing party will be more likely, in its official statements, to propose additional government intervention in the economy and measures that will lead to increases in public spending, while a right-wing party will be less likely to propose such measures and will rely more on market mechanisms to solve economic problems.
In each of the three cases under examination, only two parties held power during the period studied. We ignore all other parties that were present on the political scene but never took power. The problem is therefore not to determine whether a party is “truly on the left,” centre-left, etc., but rather to determine which of the two parties in question is more to the left and which is more to the right according to the general criteria explained in the preceding paragraph.
The distinction between left-wing and right-wing parties is clear enough in the case of Canada and in the case of the United States. The Conservative Party of Canada (formerly the Progressive Conservative Party) is almost unanimously considered to be the party that is more to the right, while the Liberal Party of Canada is considered to be the party that is more to the left (once again, among the two parties having held power during the period under study). In the United States, the Republican Party is the right-wing party, while the Democratic Party is the left-wing party.
The case of Quebec is less clear, however. The two main parties are first and foremost coalitions of federalists or separatists, attracting people of all ideological stripes. Nonetheless, the Parti Québécois (PQ) has always defined itself as a social democratic party and maintains closer ties with unions, while the Liberal Party (QLP) has a reputation of being more in favour of free markets and closer to the business world. We based ourselves on this generally accepted perception to describe the PQ as the left-wing party and the QLP as the right-wing party.
The variable used to describe the growth or reduction in the size of government is the ratio of public spending to gross domestic product (GDP). It is important to note that this criterion is relative and not absolute.
In a growing economy, a government has growing resources at its disposal even if it maintains a stable ratio of expenditures to GDP. It can therefore finance new programs or expand those that already exist without increasing the relative size of the government. Even if the ratio of spending to GDP decreases, this therefore does not necessarily mean – in fact very rarely means – that the government is spending and intervening less in absolute terms, but rather that its size is not growing as fast as the economy is.
Like any variable, the ratio of spending to GDP does not always provide a perfectly accurate picture of reality. For example, a period of recession can cause the ratio to increase (by reducing the size of the economy) without implying any particular budgetary decisions on behalf of the government. It is nonetheless one of the most relevant criteria, and one of the most frequently used.
Despite the limits of such an analysis, if the popular perception of political dynamics is generally correct, we should at least expect to see a minimally coherent trend emerge over the course of the four decades examined.
Results for the three governments
In reality, there is no systematic relation, for any of the three governments, between the left-wing or right-wing ideologies of the parties in power and the evolution of public spending as a share of GDP. In all three cases, it is actually left-wing governments that most reduced the relative size of government, and in one of the three cases (the United States), it is a right-wing government that most increased it.
In Canada, the periods that stand out in terms of the greatest increase and the greatest reduction in the relative size of government were both periods of Liberal government (see Figure 1 on iedm.org).
While Canadian government spending accounted for 16.9% of GDP when Pierre Elliott Trudeau came to power in 1968, it had climbed to 23.6% when his successor John Turner lost the election of 1984. This represents a 40% increase.
The ratio fell somewhat, by 5.4%, under the Conservative governments of Brian Mulroney and Kim Campbell. It was then reduced considerably, by 32.5%, under the governments of Jean Chrétien and Paul Martin, dropping from 22.4% to 15.1% of GDP. Finally, the size of the Canadian government since the arrival of the Harper government experienced a significant upsurge during the economic crisis, followed by a decline that has brought it back to practically the same level it was at six years ago.
In the case of Quebec as well (see Figure 2 on iedm.org), it is governments led by the party that is more to the left that were responsible both for the strongest growth and for the greatest reduction in the relative size of the state. The governments of René Lévesque and Pierre-Marc Johnson grew the state by 16.4% from 1976 to 1985, whereas those of Jacques Parizeau, Lucien Bouchard and Bernard Landry shrank it by 19% from 1994 to 2003.
The first Bourassa government increased the size of the state by 15.6%, almost as much as his PQ successor did. As for the second Bourassa government and the Charest government, the size of the state remained relatively stable under their governance.
The American president who most grew the relative size of the government was a Republican president, George W. Bush, who increased it by 39% (see Figure 3 on iedm.org). The one who shrank it the most was a Democratic president, Bill Clinton, who decreased it by 14.3%. The American case therefore shows the exact opposite of what could have been expected.
The current president, Barack Obama, has slightly reduced the size of government, by the same proportion as Ronald Reagan. However, this must be weighed against the fact that he was starting from a record high for the last 40 years and that this level remains much higher than it was under all of the presidents who preceded George W. Bush during this period.
An empirical study carried out in the early 1990s across 15 industrialized countries and covering a period of 28 years comes to similar conclusions. According to the authors, although majority governments of the left have a tendency to spend a little more than those of the right after several years in power, the role of parties is minimal and “a change in the composition of government is not systematically followed by a shift in public spending.”
How can we explain the fact that parties do not necessarily enact policies that correspond to their programs once they get into power?
One of the theories most widely used to explain this phenomenon is that of the median voter. In an ideal model, if all voters are placed on a left-right axis according to their preferences, each of the two parties will garner the support of half of the voters. Those whose opinions are firmly planted on the right or on the left can be counted on to stay put, while the median voter, the one situated right in the centre of the axis whose opinions are neither to the left nor to the right, has no clear loyalty.
In order to win the extra votes that will allow them to obtain a majority and to keep it, both parties will therefore need to court voters who are a little further along the axis on the other side of this median voter. A right-wing party will need to convince centre-left voters to support it, and vice versa for a left-wing party. In the end, the positions of the two parties will converge toward the centre.
We can also imagine more circumstantial reasons. The prevailing climate of opinion (for or against government intervention) during a given period can push governments, whatever their ideological preferences, to adopt policies that will please the majority. Economic circumstances also impose some constraints. A period of economic difficulties can push a government to spend more. Conversely, a new government grappling with a high level of public debt will have little room to manoeuvre when it comes to increased spending.
In sum, we cannot presume that a party of the left that gets into power will necessarily increase the relative size of government or that a part of the right will certainly reduce it, nor can we assume that a current or past government has made decisions in accordance with these expectations. This observation allows us to draw two lessons regarding public policy debates.
First of all, in order to properly understand economic reality, we must go beyond ideological clichés and the official statements of political actors and focus instead on facts and statistics. Otherwise, we risk falling prey to false beliefs (like the perception, widely shared not so long ago, that the government of George W. Bush had significantly reduced the size of the American state, whereas the reality is the exact opposite).
Finally, this shows that it is possible to have autonomous debates on public policies unconnected to partisan positions, since it is not necessarily the parties that seem most likely to support a policy that might ultimately put it in place.
The practical and theoretical implications of these observations are important. They deserve serious thought on the part of whoever wants to contribute to the study or the development of public policies, and also on the part of media representatives who cover these themes.
 André Blais, Donald Blake and Stéphane Dion, “Do Parties Make a Difference? Parties and the Size of Government in Liberal Democracies,” American Journal of Political Science, Vol. 37 (February 1993), No. 1, pp. 42-43.
 The separation of powers between the Presidency and Congress, and within Congress between the Senate and the House of Representatives, obviously makes the situation more complex. A much deeper analysis could be carried out by considering different cases in which one or the other of the two parties control these institutions. Here, we take into account only the party of the president.
 See the table in the annex on the MEI’s website for the complete data for each government.
 André Blais, Donald Blake and Stéphane Dion, op. cit., note 1, p. 57.
 Gérard Maarek, “La tyrannie de l’« Électeur médian »,” Sociétal, 2nd quarter 2005, No. 48, pp. 20-25; Roger D. Congleton, “The Median Voter Model,” in C. K. Rowley and F. Schneider (eds.), The Encyclopedia of Public Choice, Kluwer Academic Press, 2003.
 See also Geneviève Tellier, “Public Expenditures in Canadian Provinces: An Empirical Study of Politico-Economic Interactions,” Public Choice, 2006, No. 126, p. 380; Thomas R. Cusack, “Partisan Politics and Public Finance: Changes in Public Spending in the Industrialized Democracies, 1955-1989,” Public Choice, 1997, No. 91, p. 391.
Sources for figures: See annex on the MEI’s website.
* Conservative Joe Clark was Prime Minister for just nine months, from June 1979 to March 1980, between two Liberal governments led by Pierre Elliott Trudeau. Given the short duration of his mandate, we decided to include him in this 16-year period dominated by Trudeau governments instead of creating a distinct period.