CKM replication

Dear professor Pfeifer

I currently studying Business Cycle Accounting methods using Professor Pfeifer’s code. I have a question about the data construction process for MLE estimation. In the original code of CKM, they constructed output(y), investment(x), labor(l), and government spending(g) following the below codes.


Plotting ‘Data_CKM_corrected’, I still observe trends of y, x, i, and g. Did the ‘Data_CKM_corrected’ was constructed the same as the original code of CKM or another method? If different, could you elaborate on the method of data construction given per capita y, x, l, g.

I am trying to understand observational equations using ‘A Guide to Specifying Observation Equations for the Estimation of DSGE Models’ but I cannot figure out what method should I use. I will really appreciate your help.

best regards

I follow exactly their approach. The problem is that once the growth rate is fixed to 1.6 percent per year as described in the paper, this trend does not completely remove the trend in the data once the problem with the durables is corrected. For an actual implementation (as opposed to a replication), you would linearly detrend the data.

Thanks for your kind explanation professor!