This paper identifies Okun´s law in Colombia between 1984 and 2016 using a Vector Error Correction Model (VECM) as there is evidence of a long-term relationship between the unemployment rate and the GDP. Results suggest that after a one percent increase in GDP, the unemployment rate is reduced by 0.45 percentage points in the long run (after sixteen quarters). In addition we inspect for nonlinearities using a threshold cointegration model (TVECM). Results suggest the existence of two regimes a low and a high one. The high regime starts at   the late nineties and is associated with a more flexible labor market. Under this regime, a 1% increase in GDP, reduces the UR 0.6 percentage points after eighteen quarters. By contrast, under the low regime the response is 0.2 percentage points after eight quarters.