«January, 2009 Abstract. Theoretical and empirical work seems to support the idea that tight labour markets driven by resource booms induce young ...»
Long term consequences of natural resource booms for human capital accumulation
Herb Emery, Ana Ferrer 1 and David Green 2
Abstract. Theoretical and empirical work seems to support the idea that tight labour markets
driven by resource booms induce young individuals to leave school for employment as the
opportunity cost of schooling increases. Resource booms may have an adverse long term effect on
the productivity of the labor force if workers under-invest in human capital rather than merely altering the timing of schooling. We explore the validity of this thesis using the Albertan 1973oil-boom, which created a period of rapid growth both in wages and employment in Alberta relative to the rest of Canada, and analyze the effect of this oil-boom on the long-term human capital investments and productivity for birth cohorts that were of normal schooling ages before, during and after the oil boom. Our findings suggest that resource booms may change the timing of schooling but do not seem to affect greatly the total accumulation of human capital.
Department of Economics, University of Calgary Department of Economics, University of British Columbia Introduction Does natural resource wealth reduce future income levels by crowding out human capital formation? Standard models of human capital acquisition predict that a decline in the relative skill premium will induce individuals to leave school since the opportunity cost of schooling rises. This effect may be even more pronounced in the case of resource booms. Because resource industries traditionally employ low skilled workers, high wages arising from resource booms may crowd out human capital formation by pulling young individuals out of school. While evidence to date shows that resource booms reduce school enrolment, this may or may not imply long run problems.
Whether it does depends on whether the short run reduction in enrolment reflects permanently lower levels of school attainment; a mere interruption to schooling that does not change ultimate educational attainment; or access to a source of finance that could ultimately lead to higher levels of schooling. We explore the long term effects on human capital formation of natural resource booms using the Alberta 1973-1981 oil-boom. The 1973 to 1981 OPEC oil crises generated high oil prices in Alberta which created a period of rapid growth both in wages and employment in Alberta relative to the rest of Canada. 3 A series of papers investigate the immediate impact of resource booms and changes in labour market conditions more generally on education choices in developed economies. Black, Mckinnish and Sanders (2005b) examine the impact of the 1973 OPEC oil crisis on coal prices in two mining states in the US. The oil embargo created a prolonged boom in the coal industry that increased the opportunity cost of education among low skill youth in these states. Their analysis of the effect of changes in skill premia on high school enrolment during the boom and subsequent crash suggests that persistent shocks to skilled wage differentials substantially reduce high school enrolment. More generally, several studies have looked at the effect of economic conditions in general and unemployment rates in particular on high school drop out rates. Although earlier studies by Duncan (1965) and Rumberger (1983) found contradictory evidence in this regard, a general consensus has emerged that favourable economic conditions reduce high school enrolment (Neumark and Wascher 1995, Rees and Mocan 1997, Beaudry, Lemiuex and Parent 2000), and high school completions (Goldin and Katz 1998).
The idea that resource booms affect human capital investment in Alberta seems supported by casual observation In Alberta, males are opting to “drop-out” of high school to work in the oil patch during the recent boom. Calgary Herald April 23, 2006, and Globe and Mail, July 29, 2008.
Whether sharp changes in short run economic conditions have longer term impacts on educational outcomes, however, remains an open question. 4 On first consideration, it appears that large, sudden changes in the price of natural resources will have long term impacts on education by raising the opportunity cost of schooling. This, in fact, is one of the claimed channels for “resource curses”. 5 That short and long run outcomes might diverge, though, gains credibility when one considers the literature that shows that transitory resource booms do not appear to have permanent effects on labour market outcomes. Carrington (1996) investigates the adjustment of the Alaskan labour market between 1974 and 1977 when the Trans-Alaska Pipeline was under construction. He finds that flexible wages and elastic labour supplies implied that this particular short run demand shock had no long run impacts. Coe and Emery (2004), using wage data for 13 Canadian cities that spans the oil price shocks of the 1970s and 1980s, find no evidence that regional labour demand shocks result in permanent changes in relative real wages across provincial labour markets in Canada for building trades. As Lemieux and Card (2000) suggest, given that people can go back to school later in life, these temporary shocks may not have an effect on ultimate educational choices. Conversely, in a world in which students face financial constraints on educational investment, a temporary resource boom could allow some individuals to finance more education than would otherwise have been possible for them. In that case, we could observe short run reductions in enrolment as the individuals work during the resource boom to accumulate savings, but longer term increases in educational outcomes relative to the counterfactual case with no resource boom.
In this paper, we use the same resource boom episode as Carrington (1996) and Black, Mckinnish and Sanders (2005b) (the 1970s oil crises) to explore the impact of a resource boom on long run educational outcomes. Our focus is on education outcomes in the province of Alberta, which has the large majority of Canada’s oil reserves. Our analysis differs from that in Black et al.
(2005a,b) primarily because we focus on impacts on ultimate educational attainment as measured by levels of school completion rather than immediate impacts on enrolment. Further, the possibility of returning to school after the boom also changes the focus of our analysis from high school related outcomes along (i.e., enrolment and graduation rates) to higher levels of educational Other types of short run shocks (e.g. famines or war) are likely to have long lasting effects (Meng and Qian, 2009).
Gylfason et al. (1999) and Gylfason (2001) and argue that low growth and income levels in resource abundant economies could be due to low human capital accumulation. They observe that public expenditures on school and school enrolment rates are inversely related to natural resource abundance.
attainment. Once we consider the possibility that earnings in a resource boom could be used to finance post-secondary education, the impact of the boom on high school graduation rates becomes complex. Hence, it is of interest to consider not only people who are at the margin of graduating or not graduating from high school but also people who may have decided not to continue on to postsecondary education at the time of the boom. Indeed, our main results show that more males dropped out of high school during the boom but that the proportion attending post-secondary education also fell, implying that the proportion listing their highest level of education as “high school graduate” at the time did not change. Only by considering higher levels of educational attainment can we make sense of the apparent lack of response in high school graduation to the boom.
We use a variety of data to assess the human capital accumulation of the cohorts of Albertans most affected by the oil boom. The 2003 International Adult Literacy Survey (IALS) allows us to look at the schooling attainment and literacy achievement of these cohorts compared to the rest of Canada. The IALS offers an in-depth look into skills accumulation because of the inclusion of a direct measure of cognitive skills through literacy tests. Moreover, the IALS includes a question on where the individual attended high school, which is very useful for our investigation. We also use Census data to construct synthetic cohorts which we follow over time to track the evolution of their schooling achievement over the years and assess the long term consequences of the oil boom.
Overall, our results indicate that resource booms may change the timing of human capital accumulation, but they do not have negative consequences on ultimate levels of schooling. If anything, it appears that resource busts are the problem for resource abundant economies as we find that following the collapse of oil prices, human capital formation in Alberta fell behind that of the rest of Canada.
The following section provides background information on the Alberta oil boom. Section 3 describes the data we use for analysis. Section 4 discusses our results and section 5 concludes.
2. The Alberta oil boom During the 1970s, world oil prices increased as a result of what have been called the first and second OPEC oil crises (Figure 1). In 2002 purchasing power terms, the price of crude oil increased from $16 per barrel in 1972 to $99 per barrel in 1980. Prices started to fall after 1981 reaching $75 per barrel in 1982, and $60 per barrel in 1985. World oil prices collapsed to $30 per barrel in 1986 when OPEC’s pricing agreement unwound. In Canada, the decline in oil prices was accentuated by the federal government’s National Energy Program (NEP), introduced in 1982. The NEP was an attempt to shield the Canadian manufacturing sector from the effects of higher oil prices, effectively sharing the resource rents from Alberta oil with the rest of the country. To do this, under the NEP the Canadian price of oil was mandated to be half of the world price (Emery 2006). That this policy was not implemented until after world oil prices were declining reflected lags in policy making.
It has been well documented that the Alberta economy’s boom in the 1970s and early 1980s, and subsequent bust, resulted in dramatic changes in its labour market and incomes relative to the rest of Canada and the other western Canadian provinces. 6 One reason the oil boom was particularly influential for Alberta was the level of investment that followed the rising oil price. As Figure 2 shows, at the height of the oil boom, on a per capita basis, investment expenditures in Alberta were more than double those in Ontario and the neighbouring province of Saskatchewan.
With falling oil prices after 1980, the announcement of the NEP and a sharp recession, investment in Alberta plummeted back to the per capital levels of the other provinces.
The boom translated into increasing employment opportunities in Alberta relative to the rest of Canada, particularly for males (see Figure 3a). Employment rates in Alberta for males aged 16 and over were over 80% during the late 1970s until 1981 - 6% points higher than Ontario, which had the next highest employment rate. By 1983 Alberta’s employment rate was at similar levels to Ontario’s. Employment rates for females were higher than in the rest of Canada as well, but they remained higher after the boom, suggesting that this is part of a long term pattern rather than an effect of the boom (Figure 3b). Figure 4 shows that personal incomes in Alberta increased relative to the Canadian average and in comparison to the other prairie provinces of Manitoba and See Emery (2006) and Emery and Kneebone (2008). Mansell and Percy (1990) showed that the boom and bust conditions in Alberta were more pronounced than in the rest of Canada and in comparison to oil producing states in the US. Alberta accounts for nearly 80% of Canada’s oil production and even today, Alberta remains remarkably dependent on energy exports compared to the other provinces. In 2009, exports from Alberta were 40% of the provincial GDP, two thirds from mining and oil and gas extraction. Over 80% of these exports go to the US (The Canada West Foundation (2010, chapters 10 and 11) http://cwf.ca/CustomContentRetrieve.aspx?ID=1207055).
Saskatchewan. Reflecting the collapse of oil prices and investment spending in Alberta after 1980, Alberta’s income advantage relative to the other provinces was gone by the mid-1980s. 7,8 Our interest is in the impact of this boom on post secondary education trends. In Canada, students complete high school after 12 grades of school which is for the most part publicly funded. 9 Since 1954, all provinces set the minimum age for leaving school at 15 or 16 (Oreopoulos 2006). After the basic 12 years of education, provinces have similar post-secondary education systems comprised of colleges that provide technical/vocational education, and universities. In Figure 5, we present post-secondary enrolment rates by province. Two points are readily apparent from this figure. One, Alberta and the other western Canadian provinces (British Columbia, Saskatchewan and Manitoba), which are all resource abundant, have persistently lower postsecondary enrolment rates compared to Ontario. Second, during the second OPEC oil price shock, 1978-1982, post-secondary enrolment in Alberta was lower than in the other western Canadian provinces. With the end of the boom after 1982, the level of enrolment in post-secondary education returned to a level comparable to the other western provinces.
The negative relationship between post-secondary enrolment in Alberta and oil prices is made apparent in figure 6, which shows oil prices on the left hand axis and the difference in post secondary enrolment between Alberta and the rest of Canada on the right hand axis. In rough terms, the peaks of the oil price series correspond to the valleys in the enrolment gap, suggesting that, similar to the case of coal prices in the US, Alberta’s oil boom had an impact on enrolment rates. 10
3. Data Description Weekly wages were also higher in Alberta. Figure 4 in Emery, Ferrer and Green (2011) shows weekly wages in Medicine Hat (Alberta) and two other similar sized towns in Saskatchewan (Prince Albert) and Ontario (Pembroke).
Medicine Hat weekly wages spike during the oil boom relative to the other two cities. This is particularly striking in the comparison with Prince Albert since Saskatchewan is the adjacent province to Alberta and, apart from the oil boom the two provinces share many similarities.