TIME TO IMPLEMENT AND AGGREGATE FLUCTUATIONS

In this paper, we slightly modify the canonical RBC model by adding an implementation lag in the technology, to reproduce the fact that cons umption is a leading indicator of output. Indeed, we assume that agent s have today some information on the realization of the technology sho ck that will occur tomorrow. Today, inventions are known, and give inf ormation on the innovations which will be implemented tomorrow. The mo del is estimated and tested on US data by a simulated method of moment s, using some suitable reconstructed data that are compatible with tha t model. The results are in favor of this implementation lag hypothesi s: the model is not rejected by US post-war data.