Vous êtes sur la nouvelle plateforme d’Érudit. Bonne visite! Retour à l’ancien site


Social Investment by Union-Based Pension Funds and Labour-Sponsored Investment Funds in CanadaL’investissement social des fonds de pension gérés par les syndicats et des fonds d’investissement des travailleurs au CanadaLa inversion social de los Fondos sindicales de pensión y de los Fondos de inversión patrocinados por los trabajadores en Canada

  • Jack Quarter,
  • Isla Carmichael,
  • Jorge Sousa et
  • Susan Elgie

…plus d’informations

  • Jack Quarter
    Ontario Institute for Studies in Education, University of Toronto, Toronto, Ontario.

  • Isla Carmichael
    Ontario Institute for Studies in Education, University of Toronto, Toronto, Ontario.

  • Jorge Sousa
    Ontario Institute for Studies in Education, University of Toronto, Toronto, Ontario.

  • Susan Elgie
    Ontario Institute for Studies in Education, University of Toronto, Toronto, Ontario.

The research for this paper was funded by a three-year grant from the Social Sciences and Humanities Research Council of Canada.

Corps de l’article

The first objective of this study is to measure the extent of social investment by labour-based pension funds and labour-sponsored investment funds in Canada. In undertaking this research, it became apparent that there was no existing index to measure the extent of social investment by a fund. Creating an index was a necessary step before measuring the extent of social investment and undertaking the second objective[1]—to understand the factors related to it. The existing research does not indicate why some funds adopt a social investment strategy and others opt for conventional investment. Are such differences simply leadership-driven or do other factors such as union culture and member education influence the outcome? This study is a first step in attempting to understand the characteristics of funds that engage in social investment.

Although there is a growing body of research on social investment, this research is predominantly from the U.S. and does not relate social investment to the involvement of unions. This current study, by comparison, is focused on funds in Canada that have union sponsorship—that is, union-based pension funds and labour-sponsored investment funds. The latter are a different type of fund than a pension fund, but in the context of this study have the common feature of labour involvement. The growth of union interest in the investment of pension funds and their participation in labour-sponsored investment funds reflect a change of attitude on the part of labour (Quarter 1995), which is an important component of the context for this study.

The following sections discuss this change of attitude among unions and key issues in the social investment movement. In particular, we highlight unions’ ambivalence with respect to involvement in social investment and internal divisions over labour-sponsored investment funds. These divisions have led to some funds with genuine union sponsorship and others for which labour associations have served as a front, and therefore are referred to as “rent-a-union funds.” There follows a discussion of the various manifestations of social investment and the increasing number of examples of union involvement. We also review the rate of return issue and the related question of fiduciary responsibility. These concerns appear to place greater constraints on pension funds than on labour-sponsored investment funds. As such, it is expected that labour-sponsored investment funds are more likely to engage in social investment than pension funds. Similarly, the expectation for social investment is greatest for genuine labour-sponsored investment funds rather than rent-a-union funds. We then present the methodology and key results before discussing the central issues in light of our results.


By the late 1970s, unions began to show an increased interest in how pension funds were being invested. The context for that interest was rising unemployment, stagnant wages, restrictions of the rights of public sector workers to strike, and the internationalization of finance. Within the Canadian Labour Congress, earlier discussions culminated in a resolution at the 1986 convention that “endorse[d] the goal of organized Canadian workers achieving greater control and direction of the investment of pension funds” (cited in Baldwin et al. 1991: 10). This resolution was reinforced by a similar one adopted at the 1990 convention. To assume greater control, unions have taken two types of initiatives. One is the direct sponsorship of pension funds, involving about 14% of the membership of pension plans in Canada, largely in the building trades and in industries such as textiles (O’Grady 1993). The other is joint trusteeship, typically in the public sector, the building trades, forestry, transportation and some retail industries. The Ontario Federation of Labour argued for co-determination of pension funds in a 1988 brief to the Rowan Commission (Task Force on the Investment of Public Sector Pension Funds 1987). In Ontario, OPSEU (Ontario Public Service Employees Union, the large public sector union) has achieved joint trusteeship of its major pension plans (Carmichael 1996, 1998). CUPE (Canadian Union of Public Employees), which represents the employees in 30 of the top 100 plans in Canada, is pushing for that objective (having succeeded with other unions in co-trusteeship of the large Hospitals of Ontario Pension Plan). Trusteed pension funds in Canada had $522 billion of assets midway through 1999 (Statistics Canada 1999). Investments by pension plans represent about 25 per cent of the equity in Canadian businesses, including the largest corporations, and that figure continues to increase (Deaton 1989). The Caisse de dépôt et placement du Québec and the Ontario Teachers Pension Plan, both with in excess of $50 billion of assets, are among the largest pools of capital in Canada.

In addition to pensions, organized labour has also become involved in the provision of venture capital through labour-sponsored investment funds. Unlike pension funds that are invested predominantly in government bonds or blue chip equities, labour-sponsored investment funds represent risk capital that is designed to meet gaps in markets for small- and medium-sized firms in particular provinces, as defined by the fund, and possibly in particular sectors of the market, if the fund is specialized (Quarter 1995). For that reason, labour-sponsored investment funds are also referred to as venture capital funds. In general, these funds make long-term investments that bear a greater risk than the equities purchased by pension funds. However, labour-sponsored investment funds, like pension funds, are required by law to diversify their investments and to minimize the risks. To encourage participation in labour-sponsored investment funds, participants receive tax credits (federal and provincial) of 30 per cent of their investment.

From 1983, when the Quebec Federation of Labour started the Solidarity Fund (Fonds de solidarité des travailleurs du Québec), an increasing number of unions have set up labour-sponsored investment funds. A 1995 study found there were 17 funds with more than $1.8 billion of assets, $582 million of investments, and 362,350 investors (Canadian Labour and Business Centre 1995, formerly the Canadian Labour Market and Productivity Centre). Most provinces have one labour-sponsored investment fund organized by the central labour federation. However, in Ontario, where the Ontario Federation of Labour split over whether or not to support such a fund, many labour-sponsored investment funds have evolved. Moreover, some of these are sponsored by organizations that have a questionable status as a union (for example, the Canadian Football League Players Association). As noted above, these latter funds have been labelled as rent-a-union funds to denote the fact that the so-called labour associations that have organized them serve as a front for a conventional investment firm that wants to take advantage of the tax assistance available to labour-sponsored investment funds. In the subsequent data analysis, we will differentiate between genuine labour-sponsored investment funds and rent-a-union funds.

It is estimated that labour-sponsored investment funds are providing nearly one-half of the venture capital in Canada. As with the investment of pension funds, organized labour’s interest in such funds is based upon dissatisfaction with how capital is being invested. A report prepared for the Canadian Labour Congress in 1991 states: “A strong common current that has animated both the direct involvement of trade unionists in investing… is a strong sense that capital investment markets left to their own devices and operating under the direction of traditional managers have not served the interests of working people adequately” (cited in Baldwin et al. 1991: 11).

The social investment movement

At the same time as the labour movement was increasing its interest in how capital was being used, there was a parallel movement regarding the socially responsible investment of capital (Ellmen 1989; Quarter 1992; Reder 1995). Although this phenomenon was not specifically tied to the labour movement and had a broad base of support that included religious organizations, social investment groups and some businesses (for example, The Body Shop and Ben and Jerry’s [Quarter 2000]), the social investment movement has interacted with strategies of the labour movement.

Although there are differing manifestations of social investment, all involve the inclusion of social standards in investment decisions (Bruyn 1987; Carmichael 2000; Ellmen 1989; Kinder, Lydenberg and Domini 1998). In other words, investment decisions are not simply based on the rate of return (the typical standard), but also social criteria (for example, impact on the community) that may interact with the rate of return. The problem with this definition is it also allows for the inclusion of right-wing criteria such anti-gay screens used by some U.S. funds. Therefore, the current study utilizes the additional criterion suggested by some researchers (Bruyn 1987; Carmichael 2000; Lowry 1991; Zadek, Pruzan and Evans 1997) that social investment should challenge conventional corporate behaviour. Social investment is also referred to as ethical investment and these two terms are used interchangeably in this article.

There are at least three distinct forms of social investment: asset screening, asset targeting and asset managing.[2] Each suggests a method for handling the assets of a fund. We shall discuss each in turn. Asset screening involves the application of social screens, either negative or positive, to investments. Negative screens or sanctions occur where the funds prohibit particular investments. South Africa prior to the move to majority rule was one of the earliest examples; tobacco and armament companies are more current targets. The prohibition of investments in South Africa was utilized by a number of unionized pension funds and other labour investment vehicles prior to majority rule. Some labour-sponsored investment funds (e.g., the Working Opportunity Fund sponsored by the British Columbia Federation of Labour, the Crocus Fund of the Manitoba Federation of Labour and the Solidarity Fund) have applied social screens to the firms that they evaluate for investment.

Where asset screening is positive, investment is directed to a fund with a positive social goal: for example, to encourage the quality of the environment (Desjardins Environment Fund) or to pursue more general ethical objectives (for example, the Summa Fund). One difficulty with this approach is that it is based on normative criteria within a particular industry and the overall standard within an industry might not be very positive. In general, unions have not been the sponsors of investment funds with positive objectives. However, there is a union-screened fund (MFS Standard Trust) in Washington that is channelling its investments to firms with a positive labour record.[3] Moreover, CalPERS (the California Public Employees’ Retirement System), with assets of about (U.S.) $170 billion, has recently instituted a more comprehensive screen based on the Global Sullivan Principles (Sullivan 1999). These principles emphasize a broad range of environmental and social justice criteria.

A second form of social investment is asset targeting or economically targeted investment (Carmichael 2000; Jackson 1996). In this strategy, a fund targets one or two per cent of its assets for specific social goals (for example, affordable housing for low-income earners). In Canada, construction unions in British Columbia have engaged in economically targeted investment by establishing development companies (for example, Concert Properties) to which they channel a small portion of their investments. The construction unions, which benefited from the booming West Coast real estate market through the first half of the 1990s, employ their own members through this strategy, thereby increasing the pay-in to the pension plan. In the U.S., the AFL-CIO has set up a housing trust for a similar purpose.

Asset management or shareholder action, the third form of social investment, involves both individuals and funds that are concerned about issues typically related to the governance of companies in which the fund invests. Activist shareholders raise these issues for discussion and propose strategies for change. In Canada, the impetus for shareholder activism has come largely from religious organizations (Hutchinson 1996). However, individuals also engage in such actions; for example, lawyer Yves Michaud’s campaign to force the banks to address issues of executive compensation as well as issues related to governance. In the U.S., some of the large public sector pension plans (for example, CalPERS) have established a reputation for using this practice (Smith 1996). Labour-sponsored investment funds often insist on participating in a company’s governance as a condition for investment.

Even though there are examples of labour involvement in social investment strategies, unions have also been reluctant to engage in such practices. There are at least two major reasons for this reluctance: first, a concern that applying social criteria to investment can adversely affect the rate of returns; second, the tendency on the part of labour trustees to defer to management. Each of these issues shall be discussed in turn.

Rate of Return

This concern is most pronounced with respect to pension plans because the so-called “prudent man” rule suggests that the trustees are required to seek the best possible rate of return for the beneficiaries. Canadian legal opinion on this matter has been heavily influenced by the 1984 decision of the Court of the Queen’s Bench of England, the widely cited Cowan v. Scargill case, in which union trustees for the coal miners’ fund insisted that there not be investments in energy industries in direct competition with coal. Justice Megarry, writing for the court, ruled against the union trustees, stating: “When the purpose of the trust is to provide financial benefits for the beneficiaries, as is usually the case, the best interests are normally their financial interests… the trustees must not refrain from making the investments by reasons of the views they hold” (Cowan v. Scargill, [1984] 2 All E.R. 750).

As a result of the Megarry ruling, which has been cited in Canada as well, trustees of pension funds have been very cautious about making investments that do not maximize the return to beneficiaries. In summarizing Canadian legal opinion on the issue, Waitzer (1990: 10–11), the former chair of the Ontario Securities Commission, issues a warning that probably reflects the norm for social investment of pensions in Canada: “If ethical choices do not lower investment returns, the practical (and legal) reality is that trustees are unlikely to face judicial interdiction, regardless of their motivation. If investment returns are lowered, trustees are in trouble.”

Unions in Canada have struggled with this issue. The policy statement passed at the 1992 convention of the Canadian Labour Congress hedges a bit on the rate-of-return issue when it states that: “Unions which achieve greater control of pension fund investment should seek to broaden the range of criteria involved in investment decisions, consistent with securing an adequate [emphasis added] rate of return.” (Canadian Labour Congress 1993: 9).

Andrew Jackson, a former senior economist at the Canadian Labour Congress, takes the point further and suggests that a fund “invest for a positive rate of return but that [it] does not have to compete with best rate of return” (Jackson 1993: 2). CUPE refers to a “good rate of return” (Beggs 1993: 3), whereas the OPSEU Pension Trust, a jointly trusteed pension plan for Ontario government employees, stipulates a ’reasonable’ rate of return (OPSEU Pension Trust 1996). In legal circles, there are also some who argue that a broader range of benefits for the participants of a plan than the rate of return should be considered in determining appropriate investments (Ravikoff and Curzan 1980).

In addition to legal concerns, there are also practical and political consi-derations. Evidence that pension plans are yielding a lower rate of return than RRSPs, for example, would provide justification for employers already eager to rid themselves of the responsibilities associated with pensions plans and also create dissatisfaction among plan members. With respect to labour investment funds, there is not the same legal prohibi-tion, but the practical concerns mentioned above do apply. If the funds are not yielding a competitive rate of return to investors, they will eventually lose their appeal. This type of criticism has already been directed at labour-sponsored investment funds, both from labour critics (Gindin 1989; Stanford 1999) and from academic critics (Suret 1993).

An implicit assumption in the argumentation regarding social investment strategies is that they are likely to reduce the rate of return. Yet there does not appear to be evidence to support this point of view. In Canada the only evidence about social investment and rate of return appears to be anecdotal. In the U.S., there is some systematic research related to shareholder activism (one type of social investment strategy). A comprehensive review of the U.S. literature on pension fund activism and firm performance suggests that there is no substantial effect (Wahal 1996). That review indicates that the firms targeted by CalPERS experienced a small increase in stock values, whereas non-CalPERS targets did not change significantly. Wahal suggests, therefore, that there might be an effect associated with that particular fund rather than funds in general. There is some evidence that public funds that are subject to political interference may be forced to make investments that do not yield the best rate of return (Romano 1993).

In Canada, anecdotal evidence exists that ethical investment funds are above-average performers for mutual funds. For example, the Social Investment Organization has reported that the Ethical Growth Fund, with a screened portfolio, has performed “as well or better than” non-screened mutual funds, with an average annual compounded rate of return over 10 years of 12.5 per cent (Social Investment Organization 1998: 3). Over the same period, the Ethical Growth Fund outstrips the TSE 300 by 1.1 per cent. However, given the interest in the issue, there is very little systematic research.

Managerial Control

Even though unions have been increasingly assertive in assuming the trusteeship and sponsorship of investment capital, their role in management has been limited. Few unions take a direct role in the management of their pensions. One of the exceptions is the British Columbia Carpentry Workers’ Pension Plan, which has been managed directly by elected members of the union since its inception (Carmichael 2000; Quarter 1995). In cases where union representatives are either sole or joint trustees, normally they do not assert themselves. Deaton (1989) argues that trustees of pension funds, including union trustees, often defer to management’s advice and fail to assume the level of independence in decision-making that they have the right to exercise. The reasons for this are not entirely clear. It could be a lack of confidence in their abilities, particularly where a fund involves large amounts of money. Other possible explanations include: lack of training; acceptance of the view that introducing social criteria is likely to reduce the return on investment; and lack of interest in the importance of social criteria in selecting investments. These issues merit further investigation. Even labour-sponsored funds not constrained by the “prudent man” principle have tended to downplay a social investment strategy. In their promotion literature, they emphasize rate of return; a few, however, also promote their social objectives.

The current study

This study is exploratory in that it deals with issues for which there is not yet a body of research in Canada. Although there is a public perception that pension funds do not typically engage in social investment, there is no empirical research that formally assesses this phenomenon. General evidence of the social-investment phenomena and of the increasing union interest in social investment was discussed above. The attitude change among unions is indicated in such practices as the decision to seek joint trusteeship of pension funds, the decision by some unions on the West Coast to participate in economically targeted investment, and some union involvement in shareholder action and screens. However, there is also opposition and ambivalence within the labour movement. One of the objectives of this study then is to determine the extent of labour’s involvement in social investment of pension funds and labour-sponsored investment funds in Canada. We therefore conducted a broad survey of pension funds and labour-sponsored investment funds in Canada, the first of its kind to determine the extent of social investment and union involvement. We first established an index to measure social investment and then analyzed the factors related to it.

As noted above, given the lack of constraint other than the rate of return, one might expect labour-sponsored investment funds to engage more readily in social investment than pension funds. However, at this point, there is no formal evidence of this. Similarly, one might expect that genuine labour-sponsored investment funds are more likely than rent-a-union funds to engage in social investment.


A national sample of pension funds with assets of at least $50 million was drawn using the Canadian Pension Fund Investment Directory (Toronto: Maclean Hunter), which lists 504 pension funds of that size. All of the funds were contacted and 189 (38%) agreed to participate. Of these, 48% were from Ontario, 26% from Quebec, 17% from the four Western provinces and 9% from the four Atlantic provinces.

Similarly, all 20 of the labour-sponsored investment funds were contacted and 10 (50%) agreed to participate.[4] Six (60%) of these were from Ontario, two from the West, one from Quebec and one from the Atlantic. Since 70% of the funds are located in Ontario, the sample was regionally representative.[5]

Parallel surveys[6] were designed for the pension funds and the labour-sponsored investment funds to assess their social investment practices, background characteristics (for example, size, number of members/investors, age), barriers to social investment, attitude to social investment, and rate of return. Obtaining responses to the rate-of-return items was problematic; we, therefore, decided not to use these data. The survey was conducted by phone with either the fund’s manager or a knowledgeable substitute in a senior role.

As shown in Table 1, there are several notable differences between the pension funds and labour-sponsored investment funds. The pension funds in the study had the following characteristics: a median age of 30 years; median assets of $331.5 million; a median of 3,800 members (including retirees); 66% were defined benefit plans and another 25% combined defined benefits with defined contributions; and 71% of the members were unionized.

The labour-sponsored investment funds were newer, with a median age of 4 years, and also smaller, with median assets of $103.5 million. The plans had a median number of 16,250 investors, with average holdings of $4,100 and a median income of $50,000. Sixty-one percent of the investors were males and 71% (the median) of the funds’ investors were unionized.

The Social Investment Index

The following eight items were used to gauge the degree of social investment. Does your fund have formal screens of ethical criteria for its investments? Has your fund withdrawn investments or boycotted particular investments (for example, tobacco companies) for social reasons? Has your fund invested in particular corporations because in addition to their business performance they also represent social values that you would like to support? Has your fund invested in other funds that emphasize social criteria (for example, ethical or environmental mutual funds)? Does your fund submit shareholder proposals? Has your fund invested in community economic development? Has your fund invested in regional economic development? Has your fund invested in affordable housing?

Table 1

Background Characteristics of the Pension Funds and Labour-Sponsored Investment Funds

Background Characteristics of the Pension Funds and Labour-Sponsored Investment Funds

Note: Dashes in the cells indicate that this information was not collected.

-> Voir la liste des tableaux

These items were part of a pool of 12 that were included in the survey because they had face validity. To determine whether these items could be viewed as a scale, and whether any of the items should be dropped because of an inconsistent relationship with the other items, a Cronbach’s alpha for internal consistency was computed. The alpha was .53, signifying a pattern of relatively low but positive inter-correlations among the items. Items with a negative correlation with the other items were discarded.

The range in the degree of endorsement was from a high of 11% for the item on provincial/regional economic development to a low of 3% for the item, investing in other funds that emphasize social criteria. In other words, the responses reflected a consistently low degree of social investment. Because of the low variability in the responses, the alpha level was not surprising. The eight items measuring social investment were treated as a scale; hence referred to as the Social Investment Index.


The Degree of Social Investment among Pension Funds

Of the 189 pension funds in the sample, 130 (69%) responded negatively to all of the items on the Social Investment Index; 35 (19%) responded positively to one item; 11 (6%) were positive on 2 items; 9 (5%) were positive on three items; 3 (2%) on four items; one (.005%) to 5 items; and none to more than 5. If the level of agreement is scaled from 0 on 5, the mean is only 0.53 and the standard deviation is 0.98. In other words, 87% of the sample either have no social investment or are in agreement with only one of the items on the Social Investment Index.

Factors Affecting Social Investment Strategies

To test the relationship between the Social Investment Index and the set of explanatory variables, we used primarily non-parametric measures of association because of the skewed distribution of the index and some of the other variables as well. The correlations reported below are Spearman correlations (rs). The initial set of analyses involved pension funds only.

An ANOVA was conducted in order to determine whether there was a difference between the level of social investment, as measured by the Social Investment Index, and the region of Canada where the fund is located. The ANOVA for the Social Investment Index by region was significant (F(3, 185) = 2.96, p < .05). Tamhane post-hoc tests indicate that funds located in the Western provinces have a lower mean score (M = 0.16, SD = 0.57) than Quebec (M = 0.72, SD = 1.14) or Ontario (M = 0.62, SD = 1.01) and are not different from the Atlantic provinces (M = 0.25, SD = 0.98). Thus, sampled funds in central Canada have higher means than those in other parts of Canada.

There is some evidence that larger funds were more likely to engage in social investment than smaller ones. Size of fund assets (ranging from $50 million to $48.9 billion) is positively correlated with the Social Investment Index (rs = 0.225, N = 170, p < .005), and fund membership (ranging from 6 to 640,000) also tends in the same direction (rs = 0.12, N = 189, p < .01).

Several other characteristics of the pension plans are unrelated to the Social Investment Index. These are: age of the fund (ranging from 1 to 111); type of plan (defined benefit; defined contribution; and a combination of the two); and the contributory status of the fund (whether or not the majority of members are required to contribute). However, the percentage of the fund’s members that are unionized was approaching significance (rs = .164, N = 121, p < .07).

Several characteristics of the pension fund’s boards were related to the Social Investment Index. Funds with union representatives on the board are more likely to engage in social investment than those without (rs = .157, N=189, p < .05). In addition, the presence and composition of an investment committee was related to the Social Investment Index (F(4, 192) = 3.11, p < .05). Even though the Tahame post-hoc tests could not locate significant differences between the cells, funds without an investment committee had a mean of only 0.35 (SD = .74), whereas funds with both union and employee representatives had a mean of 1.42 (SD = 1.9), suggesting that a union and employee presence on the investment committee was associated with greater social investment.

Several items related to having a supportive environment for social investment were significantly related to the Social Investment Index. These were:

  • the leadership taken by sponsoring organizations in encouraging a social investment strategy (M(Yes) = 1.1, M(No) = 0.48), t(177) = 2.12, p < .05);

  • the attitude to social investment among the trustee/directors; senior management and the sponsoring organizations (rs = 0.24, N = 178, p < .001); and,

  • training of the fund’s trustees/directors (rs = 0.19, N = 189, p < .01).

Interestingly, 78% of the funds engaged in some form of training, predominantly in fiduciary responsibility, capital markets and pension law. However, only 29% were trained in labour issues and 13% in social investment. When the variable training in social investment is treated separately, it is unrelated to the Social Investment Index. This is possibly because there are so few funds that train their trustees in social investment.

The next set of analyses assessed particular barriers to social investment. Several variables—concern about a reduced rate of return, inadequate government incentives, investment too high a risk and illiquidity of investment—were not statistically significant or had very low endorsement levels. However, conflict with fiduciary responsibility was negatively related to the Social Investment Index (rs = –.16, N = 189, p < .05), indicating that those funds that had a greater concern about fiduciary responsibility had lower social investment scores.

In summary, in spite of the highly skewed response pattern on the social investment items, signifying a minimal amount of social investment by the pension funds in this study, a number of background characteristics were related to the Social Investment Index. These were: the region where the fund is located; the fund size; whether there were union representatives on the board of trustees/directors; representation on the investment committee by both union and employee representatives; encouragement from sponsoring organizations; the attitude to social investment among the trustees/directors; management and sponsoring organizations; the training of trustees (though not specific training in social investment); and concerns about fiduciary responsibility.

Pension Funds and Labour-Sponsored Investment Funds

The next step was to compare pension funds with the ten labour-sponsored investment funds in the survey. The results bear out the ex-pectation that more social investment would occur among labour-sponsored investment funds. Whereas 69% of pension funds engage in no social investment, and the mean on the Social Investment Index was only 0.53 (SD = 0.98), the mean for the labour-sponsored investment funds is 3.0 (SD = 2.54), indicating that on average they responded positively to three of the social investment categories. When a t-test is applied to the social investment scores, the results are highly significant (t(197) = 6.93, p < .005).

The results indicate that the difference between pension funds and labour-sponsored investment funds on the Social Investment Index is due largely to the labour-sponsored investment funds that are genuine. That group had a mean of 4.6 on the Social Investment Index as opposed to 1.4 for the rent-a-union funds (SDs = 2.6, and 1.14)—a difference that proved significant (t(5) = 2.51, p < .05). In other words, the scores on the Social Investment Index can be arranged on a continuum: pension funds (0.53); rent-a-union labour-sponsored investment funds (1.40); and genuine labour-sponsored investment funds (4.60).

Pension funds and labour-sponsored investment funds also were compared on the variables included in the survey. The significant variables were:

  • pension funds were older than the labour-sponsored investment funds (t(53) = 14.86, p < .005);

  • labour-sponsored investment funds had a higher score on the attitude towards social investment (t(8) = 4.76, p < .001), and also had a higher proportion of union representatives on their board (t(11) = 7.46, p < .005);

  • various forms of training were more readily available for pension fund trustees and directors than for those serving a similar role in labour-sponsored investment funds (t(197) = 1.98, p < .05), possibly reflecting the greater availability of resources in well-endowed organizations like pension funds; and,

  • pension funds generally expressed greater concern than labour-sponsored investment funds on the barrier variables—conflict with fiduciary responsibility (t(188) = 15.01, p < .005), reduced rate of return (t(188) = 11.07, p < .005), investment too high a risk (t(188) = 8.03, p < .005), inadequate government incentives (t(188) = 8.22, p < .005), and investment being too illiquid/long-term (t(188) = 2.26, p < 05).

Given the small size of the sample of labour-sponsored investment funds and the two categories of these funds, the factors affecting social investment in these funds was not dealt with separately. However, these funds were included with the pension funds in an ordinal regression procedure to discover which variables would be independent predictors of the Social Investment Index. The ordinal procedure was selected over an ordinary least squares regression because of the nature and distribution of the dependent variable. Only variables found significant in preceding analyses were entered in the model, which was significant (c2 (9, N = 165) = 44.4, p < .001) as shown in Table 2. The item parameter estimates in the table are an indication of the strength and direction of the effect of each variable on the Social Investment Index (with effects of other variables partialled out) where the Wald statistic and associated probability indicate significance.

The following variables maintained statistical significance and are therefore independent predictors: the type of fund (pension or labour-sponsored, labour-sponsored being higher than the defined benefit, defined contribution, and a combination of the two); attitude to social investment on the part of the organization’s leaders (that is, the more positive their attitude, the greater the social investment); the size of a fund’s assets (the larger the assets, the greater the social investment) and representation by both independent and unionized employees on the investment committee (that is, this type of committee predicted higher scores than all other membership patterns except those with employee representation without union representation, from which it was not different). Other variables that were significant in the bivariate analyses with social investment (encouragement from sponsoring organizations, union representatives on the board of trustees/directors, concerns about fiduciary responsibility, too illiquid investment, and the training of trustees) were not statistically significant in the ordinal regression procedure.

One other procedure was applied to the data to determine whether the funds that had high social investment scores had any distinct characteristics in relation to the overall data set. To make the group meaningful, only funds with a score of 4 or higher on the Social Investment Index were included. This group was limited to only 7 funds: four pension and three genuine labour-sponsored investment. When profiled against the overall data set, there did not appear to be a distinct pattern to the characteristics of these 7 funds. The only striking difference was the 7 funds that were high on social investment had a much higher score on the Attitude towards Social Investment item, indicating that the leadership and sponsors of those funds were more positive in their orientation. We shall return to this point in the discussion.

Table 2

Summary of Ordinal Regression Predicting Social Investment Index

Summary of Ordinal Regression Predicting Social Investment Index

Note: Dashes in the cells represent reference categories used in the ordinal regression analysis.

IC = Investment Committee


p < .05


p < .01


p < .001

-> Voir la liste des tableaux


As was expected, the data indicate that there is very little social investment among pension funds in Canada. Labour-sponsored investment funds are more likely to engage in social investment, particularly if they have genuine sponsorship by unions. While this may be due in part to the mandate associated with these funds, there was also variability among labour-sponsored investment funds. Given that genuine labour-sponsored investment funds are more likely to engage in social investment than the rent-a-union-funds, it appears that it is not simply the mandate of a fund but the attitude of the fund’s sponsors (that is, unions) that is a critical factor in social investment.

There is some other evidence that union involvement is facilitative of social investment. For example, in the case of pension funds, there is a positive correlation between union representation on the board and the Social Investment Index and a similar positive relationship for the investment committee. However, we cannot jump to the conclusion that union involvement per se is critical to social investment. If that were the case, we would have expected more social investment in general among these funds, given that all had a union membership. Moreover, we would have also expected a significant correlation between the percentage of a pension fund’s members that are unionized and the degree of social investment. The result was approaching significance, but not significant, suggesting that even though there is some evidence that union involvement is facilitative of social investment, other factors are of importance. This weak relationship may be attributable in part to the relatively small variance both among the Social Investment Index (69% had no social investment) and the percentage unionized (since all of the funds in the study had some unionized members).

Nevertheless, it appears that it is not unions per se but having a supportive framework that determines whether or not an organization engages in social investment. For example, the question “Did any of the sponsoring organizations take the lead in encouraging a social investment strategy?” correlated positively with the Social Investment Index. Similarly, the attitude to social investment among the trustees/directors, senior management and sponsoring organizations correlated positively with social investment, as was training of the fund’s trustees/directors, not specifically for social investment, but training more generally. It is probably not coincidental that among pension funds engaged in at least some form of social investment, they are disproportionately located in Quebec, which seems to have a more supportive environment for this type of work. Both the major fund for public pensions, the Caisse de dépôt et placement du Québec, and the major labour-sponsored investment fund, Fonds de solidarité des travailleurs du Québec, have a mandate that ties them to provincial economic development and to the development of local communities within the province.

The mandate of the fund might be viewed as part of the supportive framework. Labour-sponsored investment funds have a mandate that on the surface appears more supportive of social investment than pension funds. Among pension funds, one of the factors that correlated with the Social Investment Index was conflict with fiduciary responsibility—a concern for pension fund managers. When compared to labour-sponsored investment funds, pension funds had significantly higher scores on conflict with fiduciary responsibility, concern about reduced rate of return, investment that is too high a risk, and inadequate government incentives. These findings seem to support the view that the mandate for labour-sponsored investment funds is more supportive of social investment than for pension funds.

This study does not address in detail the dynamics of the seven funds that have high scores on the Social Investment Index. As noted above, with the exception of the strikingly high score on the attitude to social investment by the leadership of these organizations, their profile was similar to the other funds. An in-depth analysis of such funds could usefully be the subject of further research, particularly case study research. However, a related study by Carmichael (2000) of pension funds in British Columbia engaged in economically targeted investment through Concert Properties identifies three critical characteristics: leadership (that is, someone in a key organizational role in the union); support and expertise; and education, particularly the education of trustees.

The factors proposed by Carmichael are not dissimilar to those identified in this study. For social investment to be undertaken, a supportive framework is required, both intra-organizational starting at the top and in society at large. The evidence from this study suggests that some organizations achieve internal support, as reflected in the high score on the attitude to social investment item, but support does not exist in society at large. Except for Quebec, where economic development is linked to nationalist aspirations, the predominant culture is to focus exclusively on the rate of return. Fiduciary responsibility provides a justification for this focus, even thought there is sufficient ambiguity in the legal rulings to leave the door open for social investment (Campbell and Josephson 1983; Lane 1990; Ravikoff and Curzan 1980; Scott 1987; Waitzer 1990).

Nevertheless, the macro culture in which funds operate affects the likelihood that a fund’s leadership will encourage social investment, in any form. In that regard, a parallel can be drawn to workplace democracy. Although there are some companies in Canada that embrace democratic practices, their adoption becomes more probable when the overall culture, including the legislative framework, is supportive. In Western European countries such as Germany, for example, employee representation on the board of directors is the norm (Adams 1986; Knudsen 1994). Corporations operating within that framework (even corporations that would be resistive in North America) behave accordingly.

In the field of social investment, an example of where the overall culture supported social investment activity was the disinvestment in South Africa under apartheid. In those circumstances, it was normal for investment funds, including pension funds, to behave accordingly. Therefore, even though factors such as a supportive environment within an individual organization can make a difference to the likelihood of social investment, a broad change is made more probable by political leadership that provides a supportive legislative framework. In Canada, particularly outside of Quebec, that support does not exist. Therefore, only a small number of organizations with atypical leadership engage in social investment.

Not only is there a lack of support for social investment strategies among business leaders and government, the same can be said of organized labour. As indicated in the first part of this article, unions are becoming more interested in how the pension monies of which their members are beneficiaries are being invested and some unions are pushing for joint trusteeship and involvement in investment strategies. There is a growing recognition that these vast pools of capital are an important part of corporate investment in Canada and that there may be the potential to shape corporate investment strategies and related policies. Yet there is ambivalence within the labour movement about becoming involved in issues related to investment and outright opposition on the part of some unions. Traditionally, unions have represented the interests of labour against capital, and social investment does involve unions in an aspect of capitalism. This argument is made by representatives of the Canadian Auto Workers, who have been among those that have vigorously opposed union involvement in investment (Gindin 1989; Stanford 1999), either through labour-sponsored investment funds or through pension funds. The implicit assumption is that unions can stand outside of the system and confine themselves to labour issues only. However, unions do negotiate pensions for their members and these pensions have to be invested in order to pay the benefits that are negotiated. Irrespective of unions’ involvement in the process, a pool of capital is being invested and re-invested on behalf of unionized workers who are dependent upon the performance of these funds.

If the pension plan is defined benefit, as many union plans are, technically speaking at least, the only concern of the union is the ability of the plan to generate sufficient revenues to pay the benefits. However, if the investments yield an insufficient return to meet that goal or if investments generate a huge surplus, unions become involved. That is not the same thing as influencing the investment policies of those plans, but it is not far removed. Social investment takes unions one step further and involves them in formulating actual investment policies. Many unions are reluctant to take that step, particularly when it involves recommending policies as to what types of criteria should be used for corporate investments—the so-called screens that are used for ethical and social investment.

Unions have been more willing to become involved in shareholder action strategies, where corporate behaviour is called into account either by representatives on the board or, more typically, by critics at annual general meetings. This type of action might be more in line with the oppositional role of unions in relation to capital.

Some unions, particularly in the construction industry, have been willing to use a small portion of a pension fund’s assets for economically targeted investment. In Canada, the classic example is Concert Properties on the West Coast, involving 26 unions primarily in construction. This strategy has proven practical in generating both jobs for union members and also in yielding a good rate of return for the pension plan (Carmichael 2000).

As long as labour is ambivalent on whether or not to become involved in social investment, there is unlikely to be a major change towards social investment policies for pension funds. For labour-sponsored investment funds, there is more support for social investment strategies, but also much ambivalence on the part of various unions. In essence, labour is experimenting in a new domain and has yet to decide on a definitive course of action.


One of the difficulties of conducting research on social investment in Canada is the lack of information in the public domain. In the U.S., by comparison, there are stronger legal requirements for pension funds to make public their financial performance and investments. Nevertheless, this study creates a foundation upon which other researchers can build. The Social Investment Index can serve as the basis for subsequent research and the factors that this study has related to social investment can be researched further. One of the limitations of this study is the lack of insight into the dynamics of the funds that are engaged in significant social investment. These funds, as noted, are outliers in the sample, but are nevertheless deserving of more in-depth case studies—something to which we will turn in the next phase of our research.

Parties annexes