Liechtenstein Politische Schriften
form, does not confirm the notion of an incorrect specification (p > 0.05 for all F-Tests with different numbers of fitted terms). Furthermore, a plot of the recursive residuals shows proof that the parameters of the re- gression equation are rather stable. Note that stability tests have been performed for model specification (7).58 A similar argument is also valid for the dummy variables. Other functional forms (like multiplicative relationships between certain dum- mies and other independent variables) are imaginable and may be possi- ble, but there is no clear theoretical prediction behind these alternative specification possibilities. One would surely run the risk of something approaching data mining when testing for a lot of such alternative speci- fications. We therefore apply the linear regression model and, given the evidence of stability tests, are rather sure that it is a reasonable specifica- tion, although we of course do not want to give the impression that im- provements to our model by adding and/or removing variables or trying different functional forms are 
impossible. 3.2.5 Per capita income and government size The influence of per capita income on government size in Table A.6 is especially striking. In contrast to Alesina and Wacziarg (1998) per capi- ta income enters positively and is significant on the 1% level in model specifications (2), (3) and (4). Take column (2) for instance. A doubling of per capita income leads to a 0.92% increase of government consump- tion according to the regression59; in contrast to the period 1985–89, in which according to Alesina and Wacziarg a doubling of per capita in - come would have been associated with a 0.66 (marginally significant) decrease of government consumption. A univariate regression between per capita income and government consumption yields an adjusted R2of 0.093 according to our data and is significant on the 1% 
Empirical evidence 58Note further that the Durbin-Watson statistics of all model specifications are satisfactory. 59One has to exercise caution in implying strict causality. There is a good theoretical basis for supposing that per capita income influences government consumption, but one has to bear in mind that government consumption also influences national in - come through national accounts.


Sehr geehrte Benutzer,

aufgrund der aktuellen Entwicklungen in der Webtechnologie, die im Goobi viewer verwendet wird, unterstützt die Software den von Ihnen verwendeten Browser nicht mehr.

Bitte benutzen Sie einen der folgenden Browser, um diese Seite korrekt darstellen zu können.

Vielen Dank für Ihr Verständnis.