Publications

Learning a Living: First Results of the Adult Literacy and Life Skills Survey

Chapter 3
Skills and Valued Economic and Social Outcomes

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

Skip to text

Summary
3.1 Overview and highlights
3.2 Skills and valued economic and social outcomes
3.3 Using skills to predict economic and social outcomes
3.4 The earnings and employment advantage
3.5 Skills and participation in community activities
Endnotes
References
Annex

Text begins

Summary

This chapter examines the relationships between adult skills and valued economic and social outcomes. Researchers have often explored the link between educational attainment and labour market outcomes, but fewer studies have looked at the role skills play in determining labour market experiences. The evidence presented in this chapter goes along with a growing number of studies which have found that literacy and numeracy skills can have a positive influence on earnings and access to full-time employment, even when taking into account the effects of education and experience. Skills may also be important in predicting a variety of social outcomes. Results on participation in community groups and voluntary activities are presented with particular attention given to the relative effects of skill on these important social outcomes.

3.1 Overview and highlights

This chapter examines the relationships between adult skills and valued economic and social outcomes. It consists of three sections. The first describes the importance of considering labour market outcomes and their relationships with educational attainment and skill proficiencies. This section also examines the distributions of skills and education levels across all countries and domains. The second section examines the earnings premiums associated with having high levels of skill and education, and compares these premiums across countries. This section also explores the relationships between skills and earnings levels, complementing recent research studies that have uncovered a high education, low earnings phenomenon. In addition, the relationship between skills and the likelihood of being employed full-time is examined. Finally, results on participation in community groups and voluntary activities are presented with particular attention given to the relative effects of skill on these important social outcomes.

The main findings presented in this chapter are:

On average, individuals with medium to high prose or document literacy proficiencies (levels 3 and 4/5) in Bermuda, Canada, Hungary, the Netherlands, New Zealand, Norway and the United States, earn significantly more than individuals with lower skills.
With respect to numeracy, highly skilled respondents in Bermuda, Canada, Hungary, Italy, the Netherlands, New Zealand, Norway and the United States also earn significantly more on average than those with low numeracy skills.
Educational attainment also has a significant and positive effect on earnings in all countries. The earnings premium experienced by highly educated workers is largest for those who access at least some post-secondary education.
For most of the countries, skills significantly decrease one's chances of earning less than half the median earnings, even when controlling for experience, gender, community size, immigrant status, parents' education and level of education.
For most of the countries, individuals with high levels of skill have a better chance of securing stable, full-time employment. Adults in Canada, Hungary, New Zealand, Norway and the United States with higher prose skills are on average 1.2 to 1.5 times more likely to have secured full-time employment over the year prior to the survey, when compared to their counterparts with low skill levels.
In all countries, higher skilled individuals are significantly more likely than those with low skills to engage in community groups or organisations, even when controlling for level of education, age, community size, gender, children present in the home, income, and parents' education.
Overall, skills have strong and consistent effects across all skill domains and in all countries on the likelihood of individuals engaging in unpaid voluntary activities.

3.2 Skills and valued economic and social outcomes

3.3 Using skills to predict economic and social outcomes

As 'knowledge-based' economies emerge worldwide, workers are increasingly expected to acquire high levels of education and skill to enjoy labour market success. Those lacking the knowledge and competencies to satisfy the increasingly technical demands of the new economy may find themselves in part-time, less lucrative employment or trapped in longer bouts of unemployment. Researchers have often explored the link between educational attainment and labour market outcomes, but fewer studies have looked at the role skills play in determining labour market experiences, largely because of the scarcity of surveys collecting data on direct measures of skill.

Formal education plays an important role in the development of skills and, in most countries, has been shown to be the strongest predictor of literacy proficiency among a large number of antecedent variables (Desjardins, 2004; Organisation for Economic Cooperation and Development (OECD) and Statistics Canada, 2000; OECD and Statistics Canada, 2005; Raudenbush and Kasim, 1998). Yet it should be recognised that although education and skills are interdependent, this relationship is not perfect. Attaining a high level of education does not necessarily guarantee having a high level of skill, nor does it ensure a given skill proficiency for the duration of one's life career. Unlike educational attainment, which can only improve as one gets older, an individual's skill set can deteriorate over the life course.

Before examining how skills are related to valued outcome variables, it is important to take a first look at how skills and education are distributed within each country. Figure 3.1 shows the percentage of the populations aged 16 to 65 years with moderate to high skill levels and the percentage of populations with moderate to high education for each country in the Adult Literacy and Life Skills Survey (ALL) (2003 and 2008). The education variable is based on respondents' highest level of formal education and is derived from the International Standard Classification of Education (ISCED) 1997 categories. Respondents are classified into three categories – having less than upper secondary education, having completed upper secondary education, and having at least some post-secondary education – while the skill variable is dichotomised into those with moderate to high skill (Levels 3 and 4/5) and those with low skill (Levels 1 and 2).^{Note 1} Attaining higher levels of formal education as well as achieving higher levels of skill have been formally and informally deemed as benchmarks for securing a good job and having the abilities to successfully carry out the tasks required of 'knowledge workers' (Statistics Canada and OECD, 2005).

Figure 3.1 Distribution of the highly skilled and highly educated

Overall, in terms of the proportions of moderate to highly skilled respondents in the adult population aged 16 to 65 years, the Netherlands and Norway consistently rank higher than most other countries across all skill domains. Canada consistently ranks third on prose and document literacy but drops to fourth place on numeracy and problem solving. Bermuda ranks well on the prose literacy scale but drops relative to the other countries on the other three domains. New Zealand consistently scores in a middle position on all skill domains. Switzerland has relatively elevated proportions of moderate to highly skilled respondents on the numeracy and problem solving scales but does less well in terms of prose and document literacy. Italy, Hungary and the United States typically show lower proportions of highly skilled individuals on all domains. Hungary is an exception in that its highly skilled respondents perform about average on the numeracy scale.

Figure 3.1 also shows how the distribution of formal education varies from country to country. Some have higher proportions of secondary graduates than others. The percentage of respondents aged 16 to 65 years with at least some post-secondary education in the ALL ranges from only nine per cent in Italy to just over 57 per cent in Bermuda. In addition to Bermuda, Canada (46%), New Zealand (44%), Norway (38%) and the United States (35%) also have relatively high proportions of highly educated respondents. A middle group consisting of Hungary, the Netherlands and Switzerland shows moderate levels of highly educated individuals, ranging from 23 to 30 per cent.

3.4 The earnings and employment advantage

Educational attainment and experience in the labour force are typically used in social science research as indirect measures of human capital in explaining earnings differences between population groups. However, a growing number of studies have found that literacy and numeracy skills also have a strong, positive influence on earnings, over and above the effects of education and experience (Finnie and Meng, 2006; Green and Riddell, 2001; Murnane et al., 1995; OECD and Statistics Canada, 2005; Osberg, 2000). In fact, skills may more closely approximate productivity differences by providing a more direct measure of one's knowledge and competencies (Stern and Tuijnman, 1994). Some economists and sociologists have noted that the effect of education on earnings is less tangible and direct than often assumed — claiming that educational credentials send signals to prospective employers about the potential productivity or competence of a job applicant (e.g., Arrow, 1973; Spence, 1974) or messages about socio-economic status and cultural capital (e.g., Bourdieu and Passeron, 1977; Collins, 1979) rather than provide a direct indication of cognitive abilities per se.

This section examines how much of an earnings advantage higher skilled (Levels 3 and 4/5) individuals experience over those with low skills (Levels 1 and 2), and whether the relative earnings advantage of high skilled individuals changes across countries and/or varies across skill domains. Consistent with previous research studies using The International Adult Literacy Survey (IALS) data and first round ALL countries (Green and Riddell, 2001; OECD and Statistics Canada, 2005; Osberg, 2000), the evidence indicates that literacy skills also explain a part of the differences in earnings, even when controlling for level of education and years of experience.

In terms of the prose and document literacy domains, Figures 3.2.1 and Figure 3.2.2 show that individuals with at least Level 3 skills in Bermuda, Canada, Hungary, the Netherlands, New Zealand, Norway and the United States, on average, earn significantly more than individuals with low skills. This relationship holds even when controlling for experience, gender, community size, employment status, immigrant status, parents' education, and level of education – other well-established factors that have been shown in previous studies to influence earnings. Only two countries, Italy and Switzerland, do not show statistically significant effects on earnings of having high levels of literacy skills.^{Note 2}

With respect to numeracy, highly skilled respondents in Bermuda, Canada, Hungary, the Netherlands, New Zealand, Norway and the United States earn significantly more on average than those with low numeracy skills (see Figure 3.2.3). In Italy, unlike the prose and document literacy domains, there is evidence of an earnings premium for having high numeracy skills.

Figure 3.2.4 presents the results of a regression analysis of an earnings model that specifies a measure of problem solving skill. Much like the results for the other domains, significant and positive effects on annual earnings for respondents with high levels of skill are found in most countries.

Figure 3.2.1 and 3.2.2 Earnings premiums for holding medium to high levels of education and skill

Figure 3.2.3 and 3.2.4 Earnings premiums for holding medium to high levels of education and skill

As expected, the results shown in Figures 3.2.1 to Figures 3.2.4 indicate that educational attainment has a significant and positive effect on earnings in all countries.^{Note 3} The earnings premium experienced by highly educated workers is largest for those who obtained some post-secondary education. All other variables held constant in the model, respondents who attained post-secondary education earned on average nearly 23 per cent more in Bermuda, to just over 67 per cent more in Switzerland than those who did not graduate from upper secondary school. This wide range likely reflects the relative supply of educated workers in each country and may point to differing cultural values and perceived usefulness of educational credentials and/or varied labour market demands across countries.

In many countries, skills also account for part of the differences in earnings, even when controlling for level of education and other important factors. Contrary to the wide range of premiums associated with education level, earnings premiums related to skills proficiencies are smaller and vary much less across countries. On the prose scale, for example, Figure 3.2.1 shows that workers with higher skills on average earn about 10 or 11 per cent more in Hungary, Norway and the Netherlands, about 17 per cent more in the United States, about 20 per cent more in Canada and New Zealand, and nearly 24 per cent more in Bermuda. For document skills (see Figure 3.2.2), the premiums are slightly higher in Bermuda, Hungary and the United States, and remain about the same in the other countries.

In terms of numeracy, Figure 3.2.3 shows that workers with Level 3 or 4/5 skills in Italy earn about nine per cent more on average. The earnings premium for highly skilled workers in New Zealand and the United States increases to just over 20 per cent, while Canada's earnings premium is about 18 per cent. For the other countries the findings are similar to those noted above for the prose and document literacy domains.

Finally, a similar fact emerges when examining the earnings premiums associated with medium to high problem solving skills (see Figure 3.2.4). However, in New Zealand and Norway, the earnings premium associated with high problem solving skills is greater relative to the other skills domains. For Bermuda, the opposite is true, as the earnings premium decreases slightly to just over 20 per cent.

Although it is typically the case that obtaining higher levels of education leads to higher earnings, some well educated individuals earn significantly less than their country's median earnings. The ALL survey provides an opportunity to explore the relative distribution of functional skills across high and low earnings categories.

Figure 3.3.1 and 3.3.2 Distribution of the population earning half the median earnings or less by skill

Figure 3.3.3 and 3.3.4 Distribution of the population earning half the median earnings or less by skill

Figure 3.3.5 Distribution of the population earning half the median earnings or less by skill

Figures 3.3.1 to Figures 3.3.4 and Tables 3.3.1 to Tables 3.3.4 show the per cent of the labour force populations by skill and earnings level.^{Note 4} There is some evidence to suggest that having higher skills decreases one's chances of earning substantially less than the median country earnings. Figure 3.3.1 compares the percentage of low and medium to highly skilled workers who earn half the median earnings or less on the prose scale. Overall, in most countries, between 20 and 25 per cent of the highly skilled population earns half the median earnings or less. The exceptions are Bermuda at 14 per cent, Italy at 10 per cent and Hungary at seven per cent. The variation between the two skill groups ranges from no significant difference in Norway to about an eight per cent difference in New Zealand, suggesting that skill level may have a marginal effect on entering this low earnings category. These patterns also hold for the other three skill domains, although most countries show higher percentages of those with advanced problem solving skills having low earnings.

Figure 3.3.5 shows the differences by education level across the same earnings categories. Three broad groups emerge from the findings. New Zealand and Norway show the highest percentage of highly educated respondents with low earnings, with 18 and 17 per cent respectively. Bermuda, Canada, the Netherlands and the United States have smaller relative proportions (about 11 to 15 per cent), while Hungary, Italy and Switzerland have the smallest percentage of highly educated, low earners. Having a high level of education appears to have a strong marginal effect on earnings at or less than half the median earnings in all countries.

Figure 3.4.1 and 3.4.2 Likelihood of medium to high skilled adults earning more than half the median earnings

Figure 3.4.3 and 3.4.4 Likelihood of medium to high skilled adults earning more than half the median earnings

While having a high level of education consistently reduces the likelihood of earning half the median income or less, there is weak evidence to suggest that having high skills produces similar advantages. Tables 3.4.1 to Tables 3.4.4 and Figures 3.4.1 to Figures 3.4.4 reveal that for most countries, skills significantly decrease one's chances of earning less than half the median earnings, even when holding variation associated with variables such as experience, gender, community size, immigrant status, parents' education and level of education constant in the model.^{Note 5} For example, on the prose literacy scale, workers with higher skills are on average about 1.3 to 1.8 times less likely to earn less than half the median earnings.

On the problem solving scale, the differences between groups are slightly larger for most countries, and in Bermuda, workers with Level 2, 3 or 4 skills are more than twice as likely to earn more than half the median. In other words, highly skilled workers in Bermuda are about 50 per cent as likely to earn less than half the median. There are some exceptions, however. Most notably, in Italy, Switzerland and the United States, no significant advantages for highly skilled problem solvers emerge across all domains.

Figure 3.5.1 and 3.5.2 Likelihood of medium to high skilled adults being employed full-time for the previous year

Figure 3.5.3 and 3.5.4 Likelihood of medium to high skilled adults being employed full-time for the previous year

The final charts in this section explore the relationship between skill and full-time employment status. Figures 3.5.1 to Figures 3.5.4 present the likelihood of highly skilled workers being employed on a full-time basis over the 52 weeks preceding the interview. For the prose, document and numeracy skill domains the findings generally are the same. For most of the countries in ALL 2003 and 2008, population groups having high levels of skill have higher chances of securing stable, full-time employment. For example, Figure 3.5.1 shows that even after controlling for age, gender, children present in the home, and education level, adults in Canada, Hungary, New Zealand, Norway and the United States with higher prose skills are on average about 1.2 to 1.5 times more likely to have secured full-time employment over the year prior to the survey, compared to their counterparts with Levels 1 or 2 skills. Although the general patterns hold, the differences between low and moderate to high numeracy skill groups in countries like Hungary, Italy, the Netherlands and New Zealand are greater than for the other three domains, while in Canada, the likelihood of securing full-time employment changes relatively little across the prose, document and numeracy domains. For problem solving, the group differences are slightly smaller than for numeracy skills for most countries (as indicated by the smaller odds ratios). However, both Bermuda and Switzerland show no statistically significant differences across skill groups on each of the scales.

3.5 Skills and participation in community activities

Social capital theorists have long argued that engaging in community activities outside the workplace is important for the quality of life in democratic societies. A high level of social capital manifests itself in greater social trust, social cohesion, norms of reciprocity, higher civic and political participation, more organisational involvement, and volunteerism (Putnam, 2000). Although many factors contribute to varying levels of civic and social engagement, educational attainment has been consistently shown to be the most important determinant (Huang et al., 2009; OECD, 2009; Putnam, 2000; Schellenberg, 2004; Schuller and Desjardins, 2007). For example, in all 32 countries participating in the 1991 World Values Survey, Schofer and Fourcade-Gourinchas (2001) found education as well as employment status to be particularly strong indicators of voluntary association membership. Strong education effects are perhaps not surprising, since schools play a formative role in the establishment of social networks, beliefs, attitudes and social norms (Coleman, 1988).

Skills may also be important in predicting a variety of social outcomes. A study using the The IALS 1994 data alluded to this possibility, as highly skilled individuals in most countries were found to be more likely to participate in voluntary community activities (OECD and HRDC, 1997). This section examines the relationship between the skill domains and two measures in the ALL survey that serve as indicators of social capital: participation in community groups and organisations, and participation in unpaid voluntary activities.

Figure 3.6 Distribution of the population engaged in community groups or organizations

Before examining the impact of skills on social outcomes, it is important to first examine how the distribution of social capital, measured by the two variables mentioned above, varies across the ALL countries. Figure 3.6 shows the percentage of the population aged 16 to 65 years, by country, engaging in community groups or organisations in the 12 months preceding the interview.^{Note 6} Overall, the results for Bermuda, New Zealand, Norway, Switzerland and the United States show between 65 and 70 per cent of their populations engaged in community groups or organisations in the previous 12 months. In Canada and the Netherlands, around 55 to 58 per cent participated, while in Hungary and Italy significantly smaller proportions of the population engaged in these types of activities.

Figures 3.7.1 to Figures 3.7.4 and Table 3.7 show the unadjusted and adjusted odds ratios obtained from logistic regression models predicting the probability of engaging in community groups or organisations for medium to highly skilled adults. In all countries, skills show strong marginal effects. Moreover, even when controlling for level of education, age, community size, gender, children present in the home, income, and parents' education, higher skilled individuals are significantly more likely than those with low skills to engage in community groups or organisations. This strong, positive relationship holds for all countries on the prose, document and problem solving scales, and for all countries except for Italy on the numeracy scale.

Figure 3.7.1 and 3.7.2 Likelihood of medium to high skilled adults engaging in community groups or organizations in the previous 12 months

Figure 3.7.3 and 3.7.4 Likelihood of medium to high skilled adults engaging in community groups or organizations in the previous 12 months

Figure 3.7.1 displays the odds ratios for high and low skill groups on the prose scale. It is quite evident that the magnitudes of the differences across skill level vary from country to country. While the effects are statistically significant (p<0.10) for all countries, the adjusted odds ratios range from as low as 1.3 for Hungary to just over 1.7 for Switzerland. In other words, individuals with Level 3 or 4/5 prose literacy skills are nearly 1.7 times more likely than those with Level 1 and 2 skills to engage in community groups or organisations, even when holding a number of other factors constant. Although the strength of the effects varies across countries (as indicated by the different odds ratios), the findings provide some evidence that skills play an important role in predicting participation in civic, non-political activities.

Figure 3.8 Distribution of the population engaged in unpaid volunteer activities

Figure 3.8 shows the country distributions on the indicator measuring participation in unpaid voluntary activities.^{Note 7} Between 50 and 60 per cent of respondents in Bermuda, Canada, Switzerland, New Zealand, Norway and the United States participated in such activities, while the percentages of those who participated in the Netherlands (23 per cent), Italy (21 per cent) and Hungary (14 per cent) were significantly lower than in the comparison countries.

Figure 3.9.1 and 3.9.2 Likelihood of medium to high skilled adults engaging in unpaid volunteer activities in the previous 12 months

Figure 3.9.3 and 3.9.4 Likelihood of medium to high skilled adults engaging in unpaid volunteer activities in the previous 12 months

Figures 3.9.1 to Figures 3.9.4 and Table 3.9 present the unadjusted and adjusted odds ratios obtained in logistic regression models predicting the effects of age, gender, level of education, community size, and number of children present in the home, income, parents' education and respondents' assessed skills on the probability of engaging in unpaid volunteer activities. Overall, skills have strong and consistent marginal effects across all skill domains and in all countries (as indicated by the unadjusted odds ratios). Moreover, the results indicate that, on average, medium to highly skilled individuals are significantly (p<0.10) more likely to engage in unpaid voluntary activities than those with low skills. When controlling for a number of factors, the effects of skill on voluntary participation remain strong and highly statistically significant for nearly all countries, even though the strength of the relationships varies by country and across skill domains. In Canada, for example, adults with high problem solving skills are more than twice as likely to engage in voluntary activities compared to those with low skills.

The next chapter will present the results of detailed analyses of data obtained through the numeracy skill assessment that was fielded in ALL for the first time. It will also describe the conceptual and measurement frameworks that underpin this particular assessment.

Endnotes

For problem solving, the low skill category includes respondents at Level 1, whereas the moderate to high skill category includes those at Levels 2, 3 and 4.
The R-squared values for the linear regression models range from 0.17 to 0.56, indicating that these models capture the antecedents of earnings more accurately in some countries than in others.
The percentages displayed in Figures 3.2.1 to Figures 3.2.4 are based on the fitted values of annual earnings for respondents with high skill compared to those with high education. Each of the percentages for these respondent groups is obtained by substituting the sample means and proportions for all except for the variable of interest in the estimated regression equations displayed in Figures 3.2.1 to 3.2.4.
The analyses were restricted to respondents who were in the labour force at the time of the survey.
Additional data analyses (not shown) reveal that when a control for occupation is included in the models, the skill effects weaken somewhat for most countries. This suggests that skills may be more closely tied to occupational outcomes, and may indirectly influence the likelihood of having low earnings.
This indicator is derived from a series of measures collected in the ALL survey that asked respondents to provide information on whether they participated in a political organisation, sports or recreation organisation (e.g. Baseball League, Tennis Club, etc.), cultural, education or hobby group (e.g. Theatre Group, Book Club, Bridge Club, etc.), a neighbourhood, civic or community association or a school group (e.g. Parent / Teachers Association, your neighbourhood community association), group associated with a community of worship (e.g. a youth group associated with a church), or any other group or organisation in the 12 months prior to the survey.
This measure is derived from a series of questions in the ALL that asked respondents to provide information on whether or not they participated in the following activities as an unpaid volunteer through a group or organisation: fundraising; serving as an unpaid member of a board; coaching, teaching or counseling; collecting food or other goods for charity; and any other activities such as organising / supervising events, office work or providing information on behalf of an organisation.

References

Bourdieu, P. and J.-C. Passeron (1977), Reproduction in Education, Society and Culture (Second edition). Sage, London.

Coleman, J. (1988), "Social Capital in the Creation of Human Capital", American Journal of Sociology, Vol. 94(Supplement), pp. S95-S120.

Collins, R. (1979), The Credential Society: An Historical Sociology of Education and Stratification, Academic Press, New York.

Green, D.A. and W.C. Riddell (2001), Literacy, Numeracy and Labour Market Outcomes in Canada, Catalogue No. 89-552-MPE, No. 8, Statistics Canada and Human Resources Development Canada, Ottawa.

Huang, J., H. Maassen van den Brink and W. Groot (2009), "A Meta-analysis of the Effect of Education on Social Capital", Economics of Education Review, Vol. 28, pp. 454-464.

Murnane, R., J.B. Willett and F. Levy (1995), "The Growing Importance of Cognitive Skills, Wages, and the Changing U.S. Labor Market", Review of Economics and Statistics, Vol. 77, No. 2, pp. 251-266.

OECD and Human Resources Development Canada (1997), Literacy Skills for the Knowledge Society: Further Results from the International Adult Literacy Survey, OECD Publishing, Paris and Hull.

OECD and Statistics Canada. (2005), Learning a Living: First Results of the Adult Literacy and Life Skills Survey, OECD Publishing, Paris and Ottawa.

OECD (2009), Education at a Glance: OECD Indicators, OECD Publishing.

Osberg, L. (2000), Schooling, Literacy and Individual Earnings, Catalogue No. 89-552-MIE, No. 7, Statistics Canada and Human Resources Development Canada, Ottawa.

Putnam, R. (2000), Bowling Alone – The Collapse and Revival of American Community, Simon and Schuster, New York.

Raudenbush, S.W. and R.M. Kasim (1998), "Cognitive Skill and Economic Inequality: Findings from the National Adult Literacy Survey", Harvard Educational Review, Vol. 68, No. 1, pp. 33-79.

Ross, F. and R. Meng (2006), The Importance of Functional Literacy: Reading and Math Skills and Labour Market Outcomes of High School Drop-outs, Catalogue No. 11F0019MIE, No. 275, Statistics Canada, Ottawa.

Schellenberg, G. (2004), 2003 General Social Survey on Social Engagement, Cycle 17: An Overview of Findings, Catalogue No. 89-598-XIE, Statistics Canada, Ottawa.

Schofer, E. and M. Fourcade-Gourinchas (2001), "The Structural Contexts of Civic Engagement: Voluntary Association Membership in Comparative Perspective", American Sociological Review, Vol. 66, No. 6, pp. 806-828.

Schuller, T. and R. Desjardins (2007), Understanding the Social Outcomes of Learning, OECD Publishing, Paris.

Stern, D. and A. Tuijnman (1994), Adult Basic Skills in OECD Countries: Policy Issues and a Research Agenda, OECD and NCAL International Paper IP94-01, OECD Publishing, Paris.

Contributors

David Zarifa, Statistics Canada

Simone Greenberg, Statistics Canada