Laserfiche WebLink
With a large enough dataset, we could simply compute empirical frequencies for all con- <br />ditional choice probabilities. However, since there are many states, not all OCPs in our data <br />are measured precisely. We therefore use kernel smoothing on the empirical frequencies to <br />improve the prediction error. We smooth over distance, neighborhood tenure, and age. We <br />use a Gaussian kernel. Distance is measured between the midpoints of zipeodes. Neighbor- <br />hood tenure equals the number of years the renter has lived in that zipcode. Young renters <br />are those under the age of 40, while mature/old renters are those 40 and older. We use k -fold <br />cross validation to set the optimal bandwidths with k=5. <br />5.3.2 Identifying the Parameters of the Model <br />We set R = .85.14 We estimate the various parameters of the model by estimating equa- <br />tion (9) and (11) for appropriately chosen values of (01_1,Bt_1) and (x, x'). Intuitively, by <br />examining the differential behavior of individuals in certain states of the world and follow- <br />ing certain types of deviations, we can isolate the impact of the different parameters of the <br />model. We begin by constructing a regression equation for -ym, ar, and A2. These are the <br />(mature) rent utility parameter and the parameters of the transfer function. Normally, we <br />would be confronted with a significant endogeneity problem in estimating these parameters <br />since market rents Rt (j, 0) in neighborhood j are likely correlated with the amenity value <br />w t unobservable to the econometrician. <br />We overcome this essential endogeneity problem by exploiting the quasi -experimental na- <br />lure of the 1994 San Francisco rent control ballot measure. This law change quasi -randomly <br />assigned renters within a given neighborhood j to rent control status. As mentioned, we <br />focus exclusively on this population for our regressions. <br />Now let Ot_r = (j,T, 1, M) and Fl t_1 = (j,T, 0, M) for some j E J. We furthermore set <br />X = x' = S and let j* be any element of ,7. In words, we consider two mature households who <br />both lived in neighborhood j in 1994 and have not moved as of year t. The two households <br />14This choice is consistent with the evidence provided in De Groote and Verbovcn (2016), who estimate a <br />household discount factor of .87. <br />28 <br />