This paper analyses the determinants of employment reactions induced by environmental innovations. On the basis of the parameter estimates of the Multinomial Logit and of several Multinomial Probit Models, we show that we have to distinguish between the factors that have an impact on employment increases and employment decreases. The data stem from a telephone survey covering about 1600 firms in five European countries that introduced eco-innovations recently. Environmental product and service innovations increase significantly the probability of creating jobs. Thus, supporting these innovations does not counteract labour market policy. In contrast to this, end-of-pipe eco-innovations increase the risk of destroying jobs, however at a higher significance level. Environmental innovations are skill-biased, they have a significant impact on employment changes if they are substantial and if they are induced by regulations. Firms expecting increasing sales are more prone to increase employment, while firms that want to slash costs by innovation and compete by soft factors decrease employment more frequently.