static model in time series data

[krzysz1229]

Hi, I'm using eview 6 to do least squares for economic growth and exchange rate regime in US from 1974 to 2004
However, I've got a problem about my model when I tried to regress it, the window said NEAR SINGULAR MATRIX
I've got no one to consult with, I was wondering if anyone could point out my problem and I could have some ideas to solve it.
Many thanks!

The model is
GDP=c+ B1*TT+B2*CVIL1~7+B3*GDP74+B4*GOV+B5*INF+B6*INVGDP+B7*OPEN+B8*POP+B9*POPGR
+B10*SEC+B11FLOAT

the series variables in this model are:
1. GDP - Rate of growth of real per capita GDP
2. TT - Change in terms of trade—exports as a capacity to import
3. CIVIL1~7 - dummy variables for the Index of civil liberties (measured on a 1 to 7 scale, with 1 corresponding to highest degree of freedom), CIVIL1=1 and CIVIL2~7=0 from 1974 to 2004
4. GDP74 - Initial per capita GDP (average over 1970–1973)
5. GOV - Growth of government consumption in last year
6. INF - Annual percentage change in Consumer Price Index
7. INVGDP - Investment to GDP ratio
8. OPEN - Openness, (ratio of [export + import]/2 to GDP)
9. POP - Total population (units)
10. POPGR - Population growth (annual percent)
11. SEC - Total gross enrollment ratio for secondary education
12. Float - dummy variable for exchange rate regimes (THE REGIMES FROM 1974 - 2004 ARE ''ALL'' FLOAT)

When I used the above variables to do regression, the window said NEAR SINGULAR MATRIX.
Even when I excluded c, it's still not working....
And, if I exclude float and civil these two dummies, the model with c is still NEAR SINGULAR MATRIX

what's the problem!!!??? Is it because I have only one country and this model is for panel data??Or, the sample size (31 years) isn't large enough??
It's going to kill me................


The following are other variables I haven't added in the model,

To solve endogeneity of INF on GDP
Instrumental variables are
1. The Index of Central Bank Independence (CBI)
2. Prior Colony (dummy variable) = 1 from 1974 to 2004
3. Latin America (dummy variable) = 0 from 1974 to 2004

To solve endogeneity of FlOAT on GDP
IVs are
1. AREA Land area (sq. km)
2. ISLAND - Dummy variable for countries with no mainland territory.
3. REGEXCH Average de facto exchange rate regime of the region
4. RESBASE Initial Ratio of International Reserves to monetary base (Source: IMF, line 11/line 14)
5. SIZE GDP in dollars over U.S. GDP

[startz]

Looks like you have a constant and a complete set of dummies. That's the dummy variable trap.

[krzysz1229]

Looks like you have a constant and a complete set of dummies. That's the dummy variable trap.

Thanks for your reply!!!
a constant? do you mean the constant is initial GDP74?

[startz]

Little c in your equation looks like an intercept.

[krzysz1229]

i know what the problem i have now, thank you very much indeed.
but, how do i set initial GDP for indicating the presence of conditional convergence on economic growth?
because my data is one single country from 1974 to 2004, the value of initial GDP for this country will be the same through time.
In this way, multicollinearity truns up again in the model. Or should I set t=time this varaible instead?
Thank you!

[startz]

There are probably more elegant methods, but you can just copy the initial value and paste it anywhere you like.