Having problem correctly estimating a 3sls system of equation with Instrumental variables.

[Mark Baskett]

Hello

I'm currently doing my dissertation in Economics and chose from a topics list at my university to do mine on macroeconomic interdependence. More Specifically, weather FDI stock causes business cycle synchronization. I measured FDI intensity and business cycle syncronisation and other things among sample of 15 countries for a total of 76 country pairs. I chose to use a system of equations similar to Trade, Finance and Specialization by Imbs (2004) https://www.imf.org/external/np/res/...l/pdf/imbs.pdf.

The systems of equations is

(Business Synchronization) _{i j} = (Trade Intensity)_{i j} + (FDI Intensity)_{i j} + (Structural Dissimilarity) + Z_{1i j} + u _{i j}

(FDI Intensity)_{i j} = (Trade Intensity)_{i j} +(Structural Dissimilarity)_{i j} + Z2_{i j} + u _{i j}

(Trade Intensity) _{i j} = (FDI Intensity)_{i j} + (Structural Dissimilarity)_{i j} +Z_{3 i j} +u _ij

(Structual Dissimilarity) _{i j} = (FDI Intensity)_{i j} +(Trade Intensity)_{i j} + Z4 _{I J} +u _{i j}

Where i denotes country I . j denotes country j. Z Denotes exogenous determinants of the dependent variable for that equation

When trying to estimate the model using 3sls with the 3reg command, I noticed I cant pick an instrument for each individual endogenous variable. How can I estimate the system of equations using a different instrument for each endogenous variable?

Thanks

Mark

[Mark Baskett]

Some summary statistics

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
       synch |         76   -.0142061    .0062474  -.0385274  -.0056425
       trade |         76    .0362505    .0403711   .0028487   .2457884
        fdi2 |         76    .0528669    .0545191   .0012282   .2521815
  loggdpprod |         76     9.09946    1.047727   .1387603   9.483682
dissimilar~y |         76     .321165    .4115603   .1414448   3.771808
-------------+---------------------------------------------------------
 logdistance |         76    3.437671    .5263537   2.536558   4.230014
logdisparity |         76    .5937939    .4259253   .0189362   1.783925
        euro |         76    .1052632    .3089314          0          1
      border |         76    .2368421    .4279695          0          1
     customs |         76    .3026316    .4624496          0          1

.

And some correlations

             |    synch    trade     fdi2 logdis~y loggdp~d dissim~y logdis~e
-------------+---------------------------------------------------------------
       synch |   1.0000
       trade |   0.1102   1.0000
        fdi2 |   0.2895   0.6239   1.0000
logdisparity |   0.2026  -0.1453   0.0338   1.0000
  loggdpprod |  -0.0696  -0.0525  -0.4002  -0.0669   1.0000
dissimilar~y |   0.0810   0.0149   0.3665   0.1368  -0.9588   1.0000
 logdistance |  -0.2002  -0.5020  -0.3290   0.1423  -0.1020   0.0533   1.0000

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Many of us won't open files - virus concerns. Se we don't know what you've tried or what Stata gave back.

If you use a system estimator, you just put all the instruments in and let Stata handle it. There is particular reason to tie specific instruments to specific equations. If you really want to, you can also get consistent estimates with 2SLS in ivregress or ivreg2.

Try a new posting with the required information. It is likely that you've made a simple error. However, with only 76 observations, I wouldn't put too much faith in a complex model with an estimator that only has asymptotic properties.