PurposeThe phenomenon of overeducation and the magnitude and persistence of the imbalance impact are analysed for the Spanish labour market from 2006 to 2013.Design/methodology/approachThe authors present random-effects probit estimations comparing individuals and their short-term and long-term labour mismatches.FindingsThe results support the existence of long-term persistence (status in the previous year) and short-term persistence (status at the beginning of the observed period) in overeducation. Precariousness in the labour market, measured by temporality or by the strong destruction of employment, could force individuals to choose a job below their qualification. Additionally, the phenomenon of overeducation is shown to have increased in the period 2010–2013 in relation to the period 2006–2009 independently of the region considered, though those regions with higher unemployment rates display greater imbalances.Research limitations/implicationsAlthough the results come from two different samples, it is possible to conclude that overeducation is a phenomenon that tends to perpetuate over time in Spain.Practical implicationsOne of the issues of greatest interest that is crucial to assess the relevance of the spreading of overeducation is whether overeducation can be considered as a temporal mismatch, in which case the seriousness of the problem would not be so important, or, on the contrary, as a persistent one, in which case, governments should take it into account in their education reform programmes.Originality/valueOvereducation persistence has been studied in countries such as the United States, Canada, Switzerland or Germany; however, in Spain, there are hardly any studies. Spanish labour market has certain specificities that make the analyses relevant: the high unemployment rates and high elasticity of employment with respect to the economic cycles. Under these circumstances, workers could opt for more stable positions that require a lower qualification than the one they have. This option could be even more convenient during crisis. Additionally, the article includes a disaggregated analysis by Spanish regions. The differences in the unemployment rates within and between regions are significant (some of them had at the beginning of the crisis an unemployment rate close to 7%, while in others it exceeded 12%) which allows the authors to study the phenomenon in different contexts.