Changes

'''20/07/2017''' - Cleaned hubs data. Performed preliminary analysis for hubs in stata. See the section on analysis in page [[HubsAnalysis 2017]]. '''21/07/2017''' - review of propensity score matching. '''24/07/2017''' - Review of propensity score matching and the Hochberg and Fehder (2015) paper. '''25/07/2017''' - Review of instrumental variables in panel data methods. '''26/07/2017''' - in order to eliminate any endogeneity in the categorical variable "hub", a potential instrument is real estate taxes. If the real estate taxes are higher, there might be a higher demand for Hubs. Data on mortgage based real estate taxes was downloaded from the census data, using the American Communities Survey. A fixed effects IV regression shows that taxes don't perform as satisfactory IVs in this model.  '''31/07/2017''' - Estimated hazard rate using '''stcox''' in stata. Relevant do-file is '''hubs_load_data'''. It is located in Z>Hubs>2017>hubs_data. '''1/08/2017''' - Explored stata routines for performing matching based on manually supplied scores. '''teffects''' only matches on automatically estimated probabilities. '''psmatch2''' has more scope of taking a manually estimated variable to match on. '''2/07/2017''' - Estimated matched sample using Cox proportional hazard ratios. Relevant do-file is '''hubs_load_data'''. It is located in Z>Hubs>2017>hubs_data.