Technology, Wages and Skill Shortages: Evidence from UK Micro Data

Why have skill shortages continue to persist despite increases in training and the skill levels of the workforce? We argue that technical progress has raised the demand for skilled labour to mach the observed increase in supply. We provide econometric evidence in support of this hypothesis, showing that skill shortages are higher for establishments that use advanced technology in the production process. We also provide econometric evidence that hiring difficulties are inversely related to the relative wage, as theory would suggest. Our results have clear implications for policy. If technological progress continues to be skill biased, policies that address skills deficiencies will only be successful if they produce a continual, rather than a temporary, increase in levels of skills among the workforce. Department of Economics, Queen Mary and Westfield College, University of London, Mile End Road, London E1 4NS Department of Economics and Finance, Brunel University Uxbridge Middlesex UB8 3PH * We thank two anonymous referees for helpful comments. Technology, Wages and Skill Shortages: Evidence from UK Micro Data 1) Introduction Skill shortages seem an endemic feature of the UK economy. The 1997 Skill Needs in Britain Survey (DfEE, 1997), for example, indicates that 18% of employers “felt there was a significant gap between the level of skills their current employees had and those they needed to meet their current business objectives”; this compares with 20% in 1996 and 12% in 1994. Also, 55% of firms in the 1996 survey felt employees were deficient in key computer literacy skills, compared with figures of 64% in 1996 and 53% in 1994. Time series data from the British Chambers of Commerce supports this impression: levels of skill shortages at the 1997 cyclical peak were similar to those at the 1988 peak. These data highlight an important question concerning skill shortages. The amount of training provided by firms and the levels of qualifications held by the workforce have increased in recent years (DfEE, 1997, Robinson and Manacorda, 1997). Why has this increase in the supply of skilled labour not led to a secular reduction in skill shortages? One explanation for this persistence of skill shortages is the speed of technical change. If technological progress increases the demand for skilled labour, so the increase in supply we have observed has been matched by an increase in demand, skill shortages will continue. This hypothesis is consistent with evidence that technical change has increased the relative demand for skilled labour (Haskel and Heden, 1998, Machin, 1996). It is also consistent with data from the Skill Needs in Britain Survey (DfEE, 1997) which indicates 69% of employers in 1997 and 74% in 1996 felt the required level of employee skills is increasing. This paper presents the first formal evidence of a link between skill shortages and technology. We use data from the 1991 Employee Manpower and Skills Practices Survey (EMSPS), a survey of 1600 individual establishments, providing the first comprehensive, nationally representative data source on skill shortages, hiring, turnover, training and other issues. We also use data from the 1990 Workplace Employee Relations Survey (WERS), since every establishment surveyed in EMSPS also appears in WERS. Combining these data sets allows us to investigate the links between technology and skill shortages, controlling for factors such as wages, the composition of employment, training and labour market flexibility. We begin in the next section by discussing our data and presenting descriptive statistics. Overall, 35% of establishments report skill shortages. We also have evidence on other aspects of the skills problem. In our sample, 57% of establishments that recruited workers in the 12 months before the survey reported difficulty in hiring. We also find that 18% of 1 Source: Bank of England Inflation Report, May 1999, chart 3.13. CBI Industrial Trends Survey data show a fall from 43% to 25% when comparing the peaks of 1988 and 1997. However, the CBI data is derived from a survey asking whether shortages of labour will limit output, whereas the BCC survey asks if employers are having difficulty in finding skilled labour. The BCC question seems a more direct measure of shortages. 2 But see Green et al (1995) who dispute that effective training has risen significantly. 3 1990/1 were recession years where real GDP fell. Skill shortages therefore fell, but since output fell in all sectors this should not unduly establishments reported they currently had vacancies that were proving hard-to-fill. These measures are closely related: establishments with skill shortages are significantly more likely to have hard-to-fill vacancies and to have experienced difficulties in hiring workers. Estimates of econometric models to explain skill shortages, hiring difficulties and hard-to-fill vacancies are presented in section 3. The determinants of these variables are broadly similar, which leads us to regard them as different symptoms of the same underlying process. Our most important finding is that “high-tech” establishments (i.e. those where a new product has used “microprocessors or other microelectronic components” or “new materials such as advanced alloys or engineering plastics” in the production process) are more likely to suffer skill shortages. This finding supports the argument that the persistence of skill shortages can be explained by technical change. In addition, establishments that use word processing were more likely to have found hiring difficult and to experience hard-to-fill vacancies. We also present econometric evidence of a systematic negative link between the relative wage and our measures of skill shortages and hiring difficulties. We find little evidence of a link between skill shortages and labour market flexibility and no evidence of a link between skill shortages and training. Our results have clear implications for policy. If technological progress continues to be skill biased (as it has been in the past, Gregg and Manning, 1997), the demand for skilled labour will continue to rise and there will be continual pressure for skill shortages to rise. In order to offset this, the supply of suitably skilled labour must also rise, to keep pace with demand. As a result, policies that are intended to address skills deficiencies will only be successful if they produce a continual, rather than a temporary, increase in levels of skills among the workforce. 2) Data Although widespread interest in the skills issue has lead to a vigorous public policy debate, empirical evidence is surprisingly scarce. Data on skill shortages are available at aggregate or industry level data, largely derived from the CBI Industrial Trends Survey (CBI, 1998), supplemented by other occasional surveys (Hart, 1992). Our data on skill shortages at the establishment level is complementary to these data, since they allow us to investigate variations in skill shortages between workplaces. Our measures of skill shortages, hiring difficulties and hard-to-fill vacancies are constructed as follows: