Political discourse in recent years has upheld skill shortages and worker mismatches as serious problems facing the labour market. This paper reviews the literature on skill and skill mismatch and provides some new evidence. We measure mismatch by comparing education to occupational skill requirements in two ways: vertical mismatch, in terms of years of schooling, and horizontal mismatch, in terms of field of study. Using the Survey of Labour Income Dynamics (SLID), and Labour Force Survey (LFS) data, we find that vertical skill mismatches are dependent largely upon the classification of management positions. Census data show that horizontal mismatch depends on field of study and position in the lifecycle career path. We further show that aggregate labour productivity depends particularly upon the extent of mismatch among university graduates.

Skill mismatches are commonly defined in two ways. Vertical mismatch} examines the discrepancy in the level of education between workers and their occupation by measuring years of education attained by the worker and ``required" by the occupation. Horizontal mismatch concerns itself with specific skills that are endowed from particular degree programs. Using the SLID and LFS data, we implement regression analysis to provide evidence on mismatch in Canada. Classifying the occupation distribution according to O*NET, we document trends in the occupational and worker skills over time. Our results suggest that the definition of management occupations may play a large role general skills shortages.

We also document the extent of horizontal mismatch using data from the 2006 Canadian Census. We construct a measure of mismatch between the skills associated with a field of study and the skills required in an occupation. Using regression analysis from SLID and Census data, we find that mismatches in specific skill are concentrated among workers in particular disciplines, including humanities and physical sciences, and appear to occur across both sexes and at all levels of education.

The implication of a skill mismatch is foregone productivity - either because the worker is endowed with unused skill, or because the firm is not producing at full capacity. We measure these implications empirically using regression analysis on aggregate panel data at the industry-year level. Our results suggest that vertical mismatch among university workers has a negative effect on productivity, and that mismatch among community college workers may be mislabelled if many of these workers are in management positions where they may have developed considerable job experience.