26

**[Appendix Eq. A1] **

Adjusted R Square .92381

Standard Error 1084.94623

Variable B SE B Beta T Sig T

YR1900 45.033709 7.393652 .078032 6.091 .0000

K84NAT$ 2.267823 .180644 .430355 12.554 .0000

K38NOAG 74.526063 3.189440 .376291 23.367 .0000

K49CAR 800.175208 115.171106 .208561 6.948 .0000

(Constant) -3588.526698 498.833834 -7.194 .0000

N = 617

In addition, we calculate a nation-specific adjustment, in effect a dummy

variable for each nation in the equation.

13

We then use predicted values from

this equation to estimate GDP in each nation-year where *Penn World Table* data

are not available but Banks data are available. These are mostly in the period

1920-1949.

However, these three Banks variables are not available for all nations and years

for which we would like to use them. For those years we use one of several

reduced models. These are (in order of preference):

**[Appendix Eq. A2] **

Adjusted R Square .91793

Standard Error 1126.00151

Variable B SE B Beta T Sig T

YR1900 40.830386 7.647702 .070749 5.339 .0000

K84NAT$ 3.343123 .096683 .634410 34.578 .0000

K38NOAG 72.489619 3.296123 .366009 21.992 .0000

(Constant) -3432.757433 517.186955 -6.637 .0000

and:

**[Appendix Eq. A3] **

Adjusted R Square .91467

Standard Error 1150.01894

Variable B SE B Beta T Sig T

YR1900 45.178285 7.869925 .078679 5.741 .0000

K19GREV 2.587354 .332184 .166798 7.789 .0000

K38NOAG 82.363098 3.271307 .409422 25.177 .0000

K49CAR 1750.614541 74.135355 .458010 23.614 .0000

(Constant) -3562.421958 541.096393 -6.584 .0000

n=646

For a number of nations and years, especially in 1920-1928, data are still

lacking. These we filled in by various ad hoc estimates, mostly by projecting

backwards from later reliable estimates, assuming a 1% growth rate in the

intervening period. Banks data are not available for the World War II period,

generally 1939-1945, so we interpolated from available pre- and post-war GDP

estimates.

*Eastern Europe *

Estimates for Eastern Europe in the Communist period are not generally

available in the *Penn World Table*; however, data are generally available for

1990 or 1991 onwards, after the fall of Communism. Hungarian estimates are

available in the *Penn World Table* for 1970 on (together with estimates for 1961-

13

We also tested for an interaction between national income per capita and time, but found it too small to be

worth worrying about (t = -2.21, p = .03). Electricity consumption per capita (K130ELE3 ) had no

statistically significant effect in this or subsequent models, and so is omitted.