A Leading Indicator for Employment using Big Data

This paper uses job ad information from Peru's three main employment search websites to estimate an employment leading indicator. The information from more than 25 thousand job ads per day, posted by more than 15 thousand firms, allows us to classify ads by industry and predict the evolution of employment in the commerce, manufacturing, and services sectors. The results show that this indicator has better properties for predicting the level of employment relative to alternative models based on the mean quadratic error criterion.