Many cities in California have passed Inclusionary Zoning policies (also known as below-

market housing mandates) as an attempt to make housing more affordable. These policies require

developers to sell some units below the market price on a percentage of the new homes built. For

example, in a development of 10 new homes each with market value $850,000, the developer

may have to sell 5 of the units at $180,000. Kang examine the effects of such policies on house

prices and number of housing units available using 1990 (before policy impact) and 2000 (after

policy impact) census data on California cities. Use means.dat for the following questions.

(i) Using only the data for 2000, compare the sample means of LNPRICE and LNUNITS for

cities with an Inclusionary Zoning policy, IZLAW=1, to those without policy, IZLAW=0.

Based on these estimates, what is the percentage difference in prices and number of units

for cities with and without the law? [Hint: use the simple rule that 100[ln(

ln( )] is

?

1

)

?

?

0

the approximate percentage difference between and .] Does the law achieve its

?

0

?

1

purpose?

(ii) Use LNPRICE and LNUNITS in difference-in-difference regressions, with explanatory

variables

D

, the indicator variable for year 2000; IZLAW, and the interaction of

D

and

IZLAW. Is the estimate of the treatment effect statistically significant, and of the

anticipated sign?

(iii)To the regression in (b) add the control variable LMEDHHINC. Interpret the estimate of

the new variable, including its sign and significance. How does the addition affect the

estimates of the treatment effect?

(iv)To the regression in (c) add the variables EDUCTTAIN, PROPOVERTY, and LPOP.

Interpret the estimates of these new variables, including their signs and significance. How

do these additions affect the estimates of the treatment effect?

(v) discuss the essential results in parts (a) through (d). Include in

\fdifference-in-difference regressions The REG Procedure Model: did Dependent Variable: Inprice Number of Observations Read 622 Number of Observations Used 622 Analysis of Variance Sum of Mean Source DF Squares Square F Value Pr >> F Model 3 23.23168 7.74389 26.31 <.0001 error corrected total root mse r-square dependent mean adj r-sq coeff var parameter estimates standard variable df estimate t value pr> |t) Intercept 12 06457 0.03254 370.79 *.0001 d 0.22054 0.04602 4.79