During the last decade an increasing amount of studies have investigatedthe relationship between air pollution and human health effects. In thisstudy we investigate how these effects in turn induce reduced labourproductivity in terms of sick-leaves, which is an important factor inassessment of air pollution costs in urban areas. For this purpose weemploy a logit model along with data on sick-leaves from a large office inOslo and different air pollutants. Our results indicate that sick-leaves aresignificantly associated with particulate matter (PM₁₀), while theassociations with SO₂ and NO₂, are more ambiguous. We also tryto estimate the induced social costs in terms of lost labour productivity andincreased governmental expenditures, although these estimates are moreuncertain.