Introduction

The relationships between business and politics are a classic theme for social scientists. They usually document the prerequisites for and the magnitude and effects of corporate ascendency on elected officials, who vote on the rules that regulate economic activities, as well as on the authorities that implement those rules.Footnote 1 In management, most research is devoted to the effects on business performance of political connections (Goldman et al. 2009, 2013; Houston et al. 2014). These connections usually refer to campaign contributions, external lobbying expenses or the executives’ prior positions.

Research conducted in Canada has shown that political connections positively influence the financial performance of firms (Dicko and Khemakhem 2015), although little is known about how political connections help firms that operate in Canada gain access to external resources, such as credits or contracts. The current study examines to what extent political connections help Canadian firms obtain government contracts.

The federal, provincial and municipal governments are rich, and they contribute much to the domestic economy (general government spending accounts for more than 40% of the GDP). In sectors like construction or professional, scientific and technical services, public entities are the biggest clients. Moreover, public contracts are public information in Canada, which makes it possible to construct datasets and run statistical tests.

Business and Government Interdependence

It is a truth universally acknowledged that organisations depend on their environment (Pfeffer and Salancik 1978). Drawing from the resource-dependence theory and informed by the sociological literature on social capital (Lenoir 2016) , we consider board members’ multiple affiliations (past and present) as channels through which firms tend to access the resources they need and to influence the competitive and regulatory environment in which they operate.

From this perspective, political connections are crucial. Firms always depend on decisions and policies, which are made by elected officials, bureaucrats and judges. These public actions affect the costs of the factors of production, as well as the characteristics and prices of the goods and services the companies buy or sell, in addition to their uses of accumulated profits.

Such a dependence does not imply that politics trumps business. Politics is neither strictly autonomous from business influence nor absolutely subordinated to corporate power. Public authorities enact the rules of capital accumulation that institutionalise private enrichment. On the other hand, public authorities depend on corporations for resources, especially for financial means (but also jobs), which are essential to the perpetuation of the form of government on which the positions of elected and public officials depend (Benquet et al. 2019).

Consequently, business and government are interdependent, but the relationship between corporate executives and public officials may not necessarily be symmetrical. They vary according to the institutional settings that define different varieties of capitalism (Amable 2004) and according to the rules and values that determine legitimate, acceptable, improper or illegal quid pro quo between public officials and representatives of private interests.Footnote 2

The morphology of national ‘elites’, i.e. the social characteristics and trajectories of those who occupy positions of power, matters too. For instance, in France, there is a powerful minority of high civil servants who opt for even more lucrative careers in corporations (Denord et al. 2018). Yet, the full revolving-door phenomenon tends to be less common than in the US, where individuals can switch back and forth between the public and private sectors. A contrario, in Norway, it seems that the judiciary and its people remain quite distinct from other domains of activity (politics, business, administration, etc.) that constitute the Norwegian field of power (Hjellbrekke et al. 2007).

The Parliament of Canada Act contains no provision about switching from politics to business or vice versa. It simply stipulates that a person cannot accumulate business activities and political functions. In the province of Quebec, a person can cross over from business to politics without restriction. However, the situation differs when it comes to moving from politics to business. Under the Quebec Code of Ethics and Conduct of the Members of the National Assembly, no cabinet minister may hold the position of director or officer of a legal person, partnership or association. Moreover, cabinet ministers, who are all elected officials, may not within 2 years after leaving office accept any appointment to a board of directors or as a member of any body, agency, enterprise or other entity that is not a State entity. However, both at the federal and the provincial levels, a member of Parliament, member of the Provincial Parliament, member of the National Assembly or member of the Legislative Assembly may be a shareholder of a company. As well, in these jurisdictions, switching from politics to business and vice versa remains relatively easy.

Research Hypotheses

Scant research exists on the relationship between political connections and winning government contracts, and most studies focus on the US case. Agrawal and Knoeber (2001) found that firms that had a higher proportion of sales to the government also had more politically connected directors on their boards. The study was based on a sample of 264 manufacturing firms on the Forbes 800 list in 1987. Wang (2014) concluded that a fifth of the total revenue of politically connected firms was generated by Department of Defence contracts, whereas such contracts accounted for less than 10% of the total revenue of non-connected firms. This result was based on a sample of 112 public companies out of the top 500 recipients of US Department of Defence contracts in 2008. For the European Union, Laurens (2015, p. 127–166) documents the same positive relation between lobbying expenditures and European subsidies.

In the Canadian context, Dicko and Khemakhem (2015) showed that politically connected listed companies tend to perform better (measured by return on assets and return on equity) than non-politically connected firms. However, despite frequent scandals that indicate corruption and collusion in sectors like construction, there is no systematic study of the effect of firms’ political connections on public-procurement contracts.Footnote 3 We put forward the following main research hypothesis:

  • 1: There is a positive relationship between political connections and winning government contracts in the Canadian context.

This hypothesis leads to two sub-hypotheses:

  • 1a: Politically connected firms win more government contracts than non-politically connected firms.

  • 1b: Politically connected firms win higher-value government contracts than non-politically connected firms.

Sample, Data and Model of Analysis

Our study examined the largest listed companies that constitute the main Canadian stock-index (S&P/TSX). After eliminating financial and insurance companies, mainly because of the particular structure of their financial statements, our final list contained 220 firms. The list of these companies was downloaded from the Compustat commercial database (accessible through the Wharton Research Data Service) for the period from 2010 to 2014 inclusively. Financial data were retrieved from the same source, while data on firms’ political connections were hand-collected from the commercial BoardEx database. Information about government contracts was derived from the Government of Canada’s Public Works and Government Services website.Footnote 4

We used the following model to test our hypotheses:

$$ \mathrm{GC}=\mathrm{a}+\mathrm{b}1\mathrm{PC}+\mathrm{b}2\mathrm{FS}+\mathrm{b}3\mathrm{FP}+\mathrm{b}4\mathrm{Debt}+\mathrm{b}5\mathrm{I}+\mathrm{b}6\mathrm{U}.\mathrm{S}.+\varepsilon $$
(9.1)

Where:

GC = government contracts

a = constant

PC = political connections

FS = firm size

FP = financial performance

Debt = indebtedness

I = industry

US = US listing

ε = error term

 

In this model, all our explanatory variables could be endogenous, i.e. determined by other explanatory variables or with the error term. For example, studies, including ours (Dicko and Khemakhem 2015), have shown that political connections have an impact on firms’ financial performance. To deal with this endogeneity problem, we decided to perform a two-stage least squares estimation. Our second model is the following:

$$ \mathrm{FP}=\mathrm{a}+\mathrm{b}1\mathrm{PC}+\mathrm{b}2\mathrm{GC}+\mathrm{b}3\mathrm{FS}+\mathrm{b}4\mathrm{Debt}+\mathrm{b}5\mathrm{I}+\mathrm{b}6\mathrm{U}.\mathrm{S}.+\varepsilon $$
(9.2)

Study Variables

The model contains one dependent variable (government contracts) and one independent variable (political connections). We have also included several control variables that could potentially influence the firms’ revenues. In fact, a government contract secured by a firm constitutes part of its revenues. Financial literature recognises certain variables, such as firm size, financial capacities (equity and debt) and past financial performance, as having an influence on a firm’s revenues.

Dependent Variable: Government Contracts

In this study, government contracts consist of purchases of goods and services made by the federal government. A dummy variable (winning government contracts) is used to measure whether the firm received a government contract between 2009 and 2014 (‘1’ if the firm had a government contract and ‘0’ otherwise). Another numeric variable (value of contracts) is used to measure the value of the government contracts obtained.

Independent Variable: Political Connections

In this study, a firm is politically connected when at least one of the members of the board of directors, the executive director or the majority shareholder, is currently serving or has served in politics as a minister, cabinet member or Member of Parliament, in federal or provincial governments. The dummy variable, political connections, is assigned the value ‘1’ if the firm is politically connected and ‘0’ otherwise.

Control Variables

  1. 1.

    Financial performance. Past studies have shown that three measures of financial performance could be influenced by political connections: return on assets (ROA), return on equity (ROE) and market-to-book value (MTB). In this study, these three variables are used to represent financial performance. ROA is the result of earnings before interest and tax (EBIT) divided by total assets. ROE is EBIT divided by shareholders’ equity.

  2. 2.

    Firm size. To measure the firm size, we compute the natural logarithm of its total revenue, to remove the magnitude statistical effect.

  3. 3.

    Industry. This factor is measured by a dummy variable based on the North American Industry Classification System and having the following values:

1. Agriculture, forestry, fishing and hunting;

2 Mining, quarrying, and oil and gas extraction

3 Utilities

4. Construction

5 Manufacturing

6 Wholesale trade

7 Retail trade

8 Transportation and warehousing

9 Information and cultural industries

10 Finance and insurance

11 Real estate and rental and leasing

12 Professional, scientific and technical services

13 Management of companies and enterprises

14 Administration and support, waste management and remediation services

15 Educational services;

16 Health care and social assistance

17 Arts, entertainment and recreation

18 Accommodation and food services

19 Other services.

 

This industry classification is all but obvious.Footnote 5 For instance, companies that specialise in construction and engineering, such as the Canadian giant SNC-Lavalin, are classified in the category ‘professional, scientific and technical services’ rather than ‘construction’.

  1. 4.

    Indebtedness. Capital structure is an extremely important factor that can influence corporate strategies and operations. In this study, this variable is measured by the ratio of long-term debt to shareholders’ equity.

  2. 5.

    US listing. Owing to Canada’s close business ties with the United States, many Canadian companies are also listed on US financial markets. Since certain regulations in the US (e.g. on corporate governance) are more stringent than in Canada, Canadian companies behave differently when listed on US financial markets (Khemakhem and Dicko 2013; Dicko and Khemakhem 2015). We therefore decided to include this variable in our study, using a dummy variable with the value ‘1’ if the company was listed on US markets and ‘0’ otherwise.

Analyses

Descriptive Statistics and ANOVA Results

In the following, we describe our results referring to the tables at the end of the texts presenting the outcomes of the ANOVA statistical procedure, which consists of testing whether the differences between the means of variables in two datasets are significant.

Table 9.1 presents information about all the S&P/TSX firms analysed in this study, including financial and insurance companies, showing that this industry accounts for 11.6%. After eliminating them, the final list of firms examined is mainly comprised of extractive and manufacturing companies.

Table 9.2 presents descriptive statistics. They indicate that the mean of firm size tends to be higher for politically connected firms than for non-connected firms. We note the same pattern for economic and financial performance indicators (ROA and ROE) and indebtedness. The mean value of contracts of politically connected firms is also a lot higher than the mean value of contracts of non-connected firms. In addition, politically connected firms are listed on the US market more often than non-connected firms (96 vs 79). Conversely, the mean of market-to-book value of non-politically connected firms is higher than that of politically connected firms (2.183. 1.978). This could be explained by the fact that the book values of non-connected firms are smaller; which confirms that politically connected firms tend to be bigger.

Table 9.3 illustrates that 48.8% of S&P/TSX companies are politically connected. Only 11.3% of all companies received government contracts. Of those companies that received government contracts, 7% are politically connected and 4.3% are not. While 14.4% of the politically connected firms received government contracts, only 8.8% of the non-politically connected ones did. Thus, the number of politically connected firms that won government contracts is greater than the number of non-politically connected firms that won such contracts (77 vs 47 over the five-year of study. This equal to 15 vs 9 in terms of number of companies).

Table 9.4 shows a very significant difference between politically connected and non-connected firms in terms of firm size (p value = 0.000), US listing (p value = 0.012), industry (p value = 0.000) and winning government contracts (p value = 0.002), as well as a slightly significant difference in the value of these contracts (p value = 0.066). However, no significant difference was noted between politically connected and non-connected firms in terms of performances, market-to-book value or indebtedness.

All these results tend to confirm our sub-hypotheses (H1a and H1b), as well as the significance of size in most of the quantitative characterisations of capitalist firms.

Correlation Analyses and Endogeneity Test Results

In our model (Eq. 9.1), all explanatory variables could be correlated between themselves or correlated with the error term. To determine the extent of collinearity and endogeneity problems, we decided to perform tests using correlation analyses. These analyses identify the variables that are correlated between themselves, particularly independent variables. A correlation coefficient of 50% or more is generally considered a critical level that could lead to collinearity problems.

Table 9.5 shows a positive and very significant correlation between political connections and winning government contracts (0.095**). In addition, winning government contracts is very significantly correlated to firm size (0.270**), ROA (0.128**), ROE (0.107**), market-to-book value (0.083**), indebtedness (0.080**) and industry (0.230**).

Of the independent and control variables, political connections are positively and very significantly correlated to firm size (0.278**) and to industry (0.211**). Industry is significantly correlated to firm size (0.127**) and ROA (0.138**). Firm size is significantly correlated to ROA (0.219**), to ROE (0.099**), to indebtedness (0.062*) and to the value of contracts (0.107**). ROA is also significantly correlated to ROE (0.137**), while industry is significantly correlated to the value of contracts (0.113**). However, since they are below 50%, these correlation coefficients are not critical enough to create collinearity problems.

Results of the Two-Stage Least Squares (2SLS) Estimation

As previously indicated, we resorted to a two-stage least squares estimation. Two sets of analyses were performed: one with the dummy variable, winning government contracts, and the other with the value of contracts, each as the dependent variable.

Table 9.6 shows the results of the 2SLS analyses with winning government contracts as the dependent variable. It indicates that political connections are positively and significantly associated with winning government contracts (p value = 0.034). This confirms Hypothesis 1. A similar result is obtained for firm size (p value = 0.001) and ROE (p value = 0.000). Indebtedness (p value = 0.005) and ROA (p value = 0.003) are negatively and very significantly associated with winning government contracts, whereas no significant association is noted in terms of US financial market listing (p value = 0.945), industry (p value = 0.848), or market-to-book value (p value = 0.229). The model is very significant (model p value = 0.000), with an adjusted R2 of 27%.

Table 9.7 presents the results of the 2SLS analyses with the value of contracts as the dependent variable. It shows that political connections do not significantly influence the value of contracts awarded to companies (p value = 0.189). Of the control variables, firm size (p value = 0.038) and ROE (p value = 0.000) positively and significantly influence the value of contracts awarded, while this influence is negative for ROA (p value = 0.000). Market-to-book value (p value = 0.791), US listing (p value = 0.354) and industry (p value = 0.393) have no significant influence on the value of contracts. The model is very significant (model p value = 0.000), with an adjusted R2 of 33.6%.

Discussion and Conclusion

The statistical results and estimations confirm that there is a positive relationship between political connections and winning government contracts among Canadian listed firms. Political connections positively and significantly influence the receipt of government contracts.

According to Hypothesis 1a, politically connected firms win more government contracts than non-politically connected firms. Hypothesis 1b posited that politically connected firms win higher-value government contracts than non-politically connected firms. The two-stage least squares estimation indicates that political connections do not significantly influence the value of contracts awarded. Political connections tend to have more influence on the securing of government contracts rather than on the value of contracts.

Our study is the first to demonstrate a direct relationship between firms’ political connections and government contracts in the Canadian context. Methodologically, it is based on the statistical treatment of data that comes from public sources of information. From listed companies’ financial documents and public-contract registries, we elaborated crude indicators (political connections as well as the number and value of public contracts) to assess the capacity of economic actors (large, publicly listed and politically connected firms) to influence public decisions (public contracts). The fact that almost one out of two large, publicly listed, companies in Canada have government officials on their board, among their executives or as main shareholders is a reminder that doing big business is always political. The results of this study provide empirical support to the structural theorisation of economic power. The results also confirm that political connections are not only about lobbying on the rules of the game, but also on securing resources, like public contract procurements. Some former officials tend to act like business providers. The resources they mobilise and the conditions for their efficient brokerage remain unclear and deserve further investigation. Doing so pursues the goal of the ‘sociology of elites’ (Khan 2012): making explicit under what conditions one type of resource can be converted into other forms of capital (Desan 2013).