Abstract
New Zealand has a unique accident insurance system that pays the direct costs of all injuries and compensates workers up to 80% of their earnings for any time that they are unable to work. To estimate the effect of injuries on labor market outcomes, the authors use Statistics New Zealand's Linked Employer-Employee Database (LEED) covering the period April 1999 to March 2004 using a difference-in-differences matching approach. They use two alternative matching criteria to construct control groups of non-injured workers whose pre-injury characteristics are similar to those of injured workers: the first uses only characteristics of the workers whereas the second exploits the nature of the LEED and matches workers in the same firm. Findings indicate that injuries resulting in more than three months of earnings compensation have negative effects on future labor market outcomes that do not decline with time post-injury. The results are broadly similar using both matching criteria, providing more validity to the authors' findings.