| 
View
 

05 Instrumental Variables

Page history last edited by editor 13 years, 1 month ago

 


 

Introduction


Instrumental variable methods allow consistent estimation when the explanatory variables (covariates) are correlated with the error terms of a regression relationship. Such correlation may occur when the dependent variable causes at least one of the covariates ("reverse" causation), when there are relevant explanatory variables which are omitted from the model, or when the covariates are subject to measurement error. In this situation, ordinary linear regression generally produces biased and inconsistent estimates. However, if an instrument is available, consistent estimates may still be obtained. An instrument is a variable that does not itself belong in the explanatory equation and is correlated with the endogenous explanatory variables, conditional on the other covariates.1

 

In linear models, there are two main requirements for using an IV:

 

  • The instrument must be correlated with the endogenous explanatory variables, conditional on the other covariates.
  • The instrument cannot be correlated with the error term in the explanatory equation, that is, the instrument cannot suffer from the same problem as the original predicting variable.1

 

 

Methodology


STATA: ivreg

 

  •  Simple IV estimate2
  • Syntax: Need to specify D (instrumented) and Z (the instrument)

 

. ivreg Y (D=Z)

 

Instrumental variables (2SLS) regression

 

      Source |       SS       df       MS              Number of obs =      20

-------------+------------------------------           F(  1,    18) =    7.64

       Model |         357     1         357           Prob > F      =  0.0128

    Residual |       136.2    18  7.56666667           R-squared     =  0.7238

-------------+------------------------------           Adj R-squared =  0.7085

       Total |       493.2    19  25.9578947           Root MSE      =  2.7508

 

------------------------------------------------------------------------------

           Y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

           D |        8.5    3.07544     2.76   0.013      2.03874    14.96126

       _cons |       69.7   1.945079    35.83   0.000     65.61354    73.78646

------------------------------------------------------------------------------

Instrumented:  D

Instruments:   Z

 

 

 

Example Applications


 

The Impact of Public Expenditure on Agriculture Outcomes in Nepal

IFPRI | Dillon, Sharma, Zhang | PDF

 

Problem:

The study attempts to determine the impact of 'Public Expenditure' (on agriculture extension & subsidies, irrigation, roads) on 'Per capita spending' to assess the contribution of public expenditure on agriculture.

 

1) Theoretize model with Y, D, Z

Y (Outcome) = Per capita spending

D (Treatment) = Level of Public expenditure

Z (Instrument) = Conflict Killings (highly correlated with public expenditure, but not with the error term)

 

2) Regress D and Z to see if reg is statistically significant

The instrument was first regressed with D and was found to be statistically significant.

 

3) If SS, then ivreg Z on Y, along with covariates

The ivreg regression was performed (note D is divided into 3 sub-parts):

lnY = f(AGexp, IRexp, RDexp, landarea, landelevation, districtpop, rainfall, region, belt)

 

This process showed these results:

  • Rural roads are one of the most productive expenditures with marginal benefits that range from 5.43%-30.25% on consumption per capita
  • Estimates of the impact of irrigation expenditures are also very high: 3.92-9.61% on consumption per capita
  • BUT, agriculture spending is much lower (1.36-2.98%)

 

  • HENCE:
    • returns on agricultural extension & subsidies are low
    • returns on roads and irrigation are much more
    • ... roads & irrigation should be the areas of target
    • ... need to find out why returns on agric extension & subsidies are so low

 

 

 

 

 

 

 

Reference


(1) Wikipedia: Instrumental Variable

(2) Course materials from Experimental and quasi-experimental methods in program evaluation 2012. Alan Yang. Columbia - SIPA.

 

  1. Instrumental Variables Estimate: STATA
    1. First, estimate the percentage of units who are D=1 when Z=1
      1. This is just the coefficient of the treatment (Z)

. regress D Z

 

      Source |       SS       df       MS              Number of obs =      20

-------------+------------------------------           F(  1,    18) =    3.60

       Model |          .8     1          .8           Prob > F      =  0.0739

    Residual |           4    18  .222222222           R-squared     =  0.1667

-------------+------------------------------           Adj R-squared =  0.1204

       Total |         4.8    19  .252631579           Root MSE      =   .4714

 

------------------------------------------------------------------------------

           D |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

           Z |         .4   .2108185     1.90   0.074    -.0429133    .8429133

       _cons |         .4   .1490712     2.68   0.015      .086813     .713187

------------------------------------------------------------------------------

  1. Next, compute the ITT estimate, which is the regression of Y as a function of Z

regress Y Z

 

      Source |       SS       df       MS              Number of obs =      20

-------------+------------------------------           F(  1,    18) =    2.39

       Model |        57.8     1        57.8           Prob > F      =  0.1396

    Residual |       435.4    18  24.1888889           R-squared     =  0.1172

-------------+------------------------------           Adj R-squared =  0.0681

       Total |       493.2    19  25.9578947           Root MSE      =  4.9182

 

------------------------------------------------------------------------------

           Y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

           Z |        3.4   2.199495     1.55   0.140    -1.220967    8.020967

       _cons |       73.1   1.555278    47.00   0.000     69.83248    76.36752

------------------------------------------------------------------------------

  1. We then divide ITT (3.4) by the percentage of “Compliers” (.40)
    1. 3.4/.40 = 8.5
  2. Another way to estimate the IV estimate:
    1. Regress D on Z (1st stage)
      1. Save predicted values (Dhats)
    2. Regress Y on Dhat (2nd stage)
      1. The coefficient on Dhat is the IV estimate
    3. The standard error on that coefficient would be inappropriate however because it doesn’t account for the fact that Dhat is a prediction
      1. Stata’s ivreg command gives the correct s.e. (see below)

. regress D Z

 

      Source |       SS       df       MS              Number of obs =      20

-------------+------------------------------           F(  1,    18) =    3.60

       Model |          .8     1          .8           Prob > F      =  0.0739

    Residual |           4    18  .222222222           R-squared     =  0.1667

-------------+------------------------------           Adj R-squared =  0.1204

       Total |         4.8    19  .252631579           Root MSE      =   .4714

 

------------------------------------------------------------------------------

           D |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

           Z |         .4   .2108185     1.90   0.074    -.0429133    .8429133

       _cons |         .4   .1490712     2.68   0.015      .086813     .713187

------------------------------------------------------------------------------

 

. predict pre_1, xb

 

. regress Y pre_1

 

      Source |       SS       df       MS              Number of obs =      20

-------------+------------------------------           F(  1,    18) =    2.39

       Model |        57.8     1        57.8           Prob > F      =  0.1396

    Residual |       435.4    18  24.1888889           R-squared     =  0.1172

-------------+------------------------------           Adj R-squared =  0.0681

       Total |       493.2    19  25.9578947           Root MSE      =  4.9182

 

------------------------------------------------------------------------------

           Y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

       pre_1 |        8.5   5.498737     1.55   0.140    -3.052418    20.05242

       _cons |       69.7   3.477707    20.04   0.000     62.39361    77.00639

------------------------------------------------------------------------------

  1. STATA: ivreg
    1. Simple IV estimate
    2. Syntax: Need to specify D (instrumented) and Z (the instrument)

. ivreg Y (D=Z)

 

Instrumental variables (2SLS) regression

 

      Source |       SS       df       MS              Number of obs =      20

-------------+------------------------------           F(  1,    18) =    7.64

       Model |         357     1         357           Prob > F      =  0.0128

    Residual |       136.2    18  7.56666667           R-squared     =  0.7238

-------------+------------------------------           Adj R-squared =  0.7085

       Total |       493.2    19  25.9578947           Root MSE      =  2.7508

 

------------------------------------------------------------------------------

           Y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

           D |        8.5    3.07544     2.76   0.013      2.03874    14.96126

       _cons |       69.7   1.945079    35.83   0.000     65.61354    73.78646

------------------------------------------------------------------------------

Instrumented:  D

Instruments:   Z

 

 

Comments (0)

You don't have permission to comment on this page.