Year-fixed effects and Country-specific effects

[Sophie Starck]

Hi,

I am currently writing my thesis and am studying the relationship between social infrastructure and economic growth. I have a panel dataset (countries) and want to outreg the regressions. I have divided the time period into three different intervals, but I do not know how to include year-fixed effects (I am studying countries at different points in time) and country-specific effects.

This is the code that I have:
* Column 1, length 1-14 (without capital accumulation growth)

qui xtreg g_gdp length GB1 g_prim1 g_death patent1 polity2 length_g_prim1 length_g_death length_patent1 length_polity2 g_gov_con g_trade g_pop initial_gdp if inrange(length, 1, 14) i.year, fe i.country, fe
qui outreg using Test2.doc, nocons se bdec (3) starlevels(10 5 1) sigsymbols(*,**,***) starloc(1) summstat(r2 \ rmse \ N) summtitle(R-squared \ Root mean square error \ Observations) summdec(2 2 0) title(Table 2. \ The Effect from Social Infrastructure on Economic Growth \ Dependent Variable: GDP Growth) ctitle("", "" \ "Explanatory Variables" \ "" "(1) Year 1-14") blankrows nolegend varlabels replace

Although it does not work when I add the i.year, fe i.country, fe.

Does anyone know how I solve this?

[Andrew Musau]

xtreg, fe

already includes the country effects provided that your xtset command was

xtset country year

So you just need to include the year effects using

i.year

so

qui xtreg... g_pop initial_gdp i.year if inrange(length, 1, 14), fe cluster(country)

(Also, if you have 30+ countries, cluster your standard errors at the country level).

[Sophie Starck]

Thank you!