Assessing the effectiveness of early childhood programs in the Netherlands

In this paper I investigate the effectiveness of early childhood programs (ECP) in the Netherlands. By lack of a good experimental setting and instrumental variables, an alternative method, developed by Altonji, Elder & Taber (2005), is used. Because negative selection is likely to occur in ECP enrollment, I assume equal selection between observables and unobservables. ECP effects are assessed on cognitive skills and pupil characteristics in grade 2 and 5 in three steps. The main conclusion is that ECP-children have lower school results than non-ECP children. However, a large selection is shown in the ECP enrollment. It is important that previous studies are evaluated well, as most of them did not incorporate the selection effect well, but concluded that ECP is ineffective. Moreover, nothing can be said about the marginal development of the children. Strong conclusions about ECP effectiveness can therefore not be made.