1

I am building a log-log least squares dummy variable model to estimate product demand elasticities based on a kMeans clustering mechanism (5 total clusters). The model also has other independent variables but aren't relevant to the question. The clusters are based on groups of similar prices. When I have worked with dummy variables previously, it was important to never include all n dummies, otherwise the model wasn't of full rank. In this case, I need coefficients for all cluster price terms (Ln.Price.Cluster1...n). Is there a way to get the full n coefficients using dummy variables?

Michael Westerman
  • 13
  • 6
  • 1
    It's easier to help you if you include a simple [reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) with sample input and desired output that can be used to test and verify possible solutions. Perhaps you just want an intercept free model? Normally you add "-1" to your formula to do that. But without knowing anything about the code you are actually running, it's hard to say for sure. – MrFlick Apr 22 '19 at 17:40
  • Thanks - it is an intercept free model. Are there concerns using an intercept free model to estimate a fixed effects model? – Michael Westerman Apr 22 '19 at 17:46
  • No. It just changes your interpretation of the parameter estimates when you have categorical variables. – MrFlick Apr 22 '19 at 17:52

0 Answers0