Re: st: policy simulation

[Shikha Sinha <shikha.sinha414@gmail.com>]

Dear all,

I need help in predicting with more than 100 dummies. For example, I
want to perform the following steps to capture the policy effect of
everyone having access to ORT (ORT=1).

(a) y= b0+b1*x1+b2*x2+b3*ORT +b4*d1 +b5*d2.....+b100*d1 (there are 100
dummies for d)- as mentioned by Dan Waldo in the previous post
(b) predict y_hat
(c) gen policy effect= y_hat-(b1*x1 + b2*x2 + b3  +b4*d1 +b5*d2.....+b100*d1

Is there any shorter way to generate (c) without writing out 100 coeff of d?

Thanks,
Shikha


On Wed, Jul 25, 2012 at 7:12 AM, Dan Waldo <dan_waldo@yahoo.com> wrote:
> Maarten's observation from yesterday is quite correct.
> But if you mis-spoke and you really want to look at the effect on your dependent variable of everybody having ORT=1, a clunky way to do this (assuming that X1 and X2 are uncorrelated with ORT) would be to use the predict command to generate the estimated y (say, y_hat) and then
>
> . gen policy_effect =y_hat-(b1*x1 + b2*x2 + b3) [you'd pull the coefficients from the e(b) vector ]
>
>
>
> On Tue, Jul 24, 2012 at 2:46 AM, Shikha Sinha wrote:
>> I want to conduct a counterfactual policy simulation in state. I
>> estimate the following model:
>>
>> Y = b1X1 + b2X2 +b3ORT +e
>>
>> ORT is the main independent variable and is binary. The estimated b3
>> coefficient is 0.06. Now I want to answer the counterfactual question:
>> what would have been b3 if everyone got access to ORT? How do I get
>> this?
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/