The Size Of State and Economic Growth

The Size Of The State and Economic Growth

The corollary of the Smithian prescriptions must be that, whereas it will be generally accepted that some state intervention is a necessary condition for long-term growth, an increase in the proportion of resources absorbed by government will exercise both a negative effect on the rate of economic growth and growing misallocation of resources. The latter proposition has been extensively investigated by public choice liberals employing the theory of rent-seeking. The former has received less attention though it has received more prominence with the marriage of hypotheses with empirical testing.

That well known proposition known as Baumol's Law postulates that the opportunities for employing new technologies in service sectors (notably in government) are limited whereas wage rates in these sectors are not likely to fall out of line with those in manufacturing sectors, if the normal forces of competition operate in the labour market. It follows that, even if the proportion of real resources absorbed by government is constant, government expenditure will rise as a proportion of Gross National Expenditure (GNE). It follows further that if the proportion of real resources rises, a progressive rise in the ratio of G to GNE will be associated with a lower rate of economic growth than would otherwise be the case.

The proposition that purely technological factors account for the differences in economic performance of the public and private sector seems suspect. I would not deny that welfare and environmental services, commonly produced within the public sector, contain a large element of personal service so that the physical presence of labour input is a necessary constituent of output. However, I refuse to believe that, in the age of 'info-tech' and computers, the technical limitations on factor substitution are of paramount importance in the public sector. A more plausible argument for the existence of labour-intensive public services lies in the barriers that can be erected to prevent or to control the speed of introduction of new technologies. In the private sector, given competition, product and process innovation are necessary for firms to survive. The pressures to introduce new technologies are much less in circumstances where, as with government undertakings, there is a monopoly or near-monopoly of the output of a particular service and the product is not sold on the market. The barriers to improving productivity in the public sector therefore reflect the behaviour of rent-seekers, particularly bureaucrats, and do not have their origin in technological constraints.,'

My criticism of Baumol's thesis has been elevated into a more general hypothesis that echoes Adam Smith's complaint of 'want of parsimony' in public spending. Buchanan (1980) argues that the incidence of rent-seeking and the associated diversion of resources is a function of the increased size of government resulting in a decline in economic growth. The effect is cumulative, for the growth in the relative size of the public sector itself reduces incentives in the private sector through higher and higher taxes. In other words, the more an economy moves away from the Smithian ideal of limited government input to promote free trade, the lower its rate of growth; if this were true, then the economic prescriptions of Classical liberalism should command wide appeal.

Table 1 Growth rates of real GDP per capita in relation to the share of government expenditures in GDP, 1960-1980

This hypothesis should be testable and, indeed, there is an attempt by Scully (1989) to do so. Here I only report his results in summary form (Table 1). He uses World Bank data for 115 market economy countries which record economic growth, as measured by the growth rate of real GDP over the period 1960-1980, alongside the growth in the share of government in GDP. The relation between government share and economic growth was estimated using linear OLS and the error term is assumed to be normally distributed, it being claimed that tests for heteroscedasticity validate this assumption. In equation (1), the 'simpliste' version of the hypothesis is presented, in which only the government share variables appear as regressors. Both the size of government share (GOVT 60) and the change in that share (CHGGOVT) are of the right sign, which accords with earlier studies, However, as Scully points out, account has to be taken of the change in the capital/labour ratio over the relevant period. Public investment has a positive effect on the growth rate, though less productive than private investment. To neglect this factor may be to overstate the negative effect of the government share on growth. In equation (2) the compound rate of growth in the capital/labour ratio (KLGWTH) is introduced as a regressor but GOVT60 and CHGGOVT still have the correct sign and are statistically significant in a one-tail test at better than the 1 per cent level.

I have only reproduced the results which are sufficient to further discussion of the economics, and certainly not the econometrics, of such an approach. Here are the points for discussion that occur to me:

(i) As Scully recognizes, his approach really requires that the economy-wide growth rate should be a weighted average of the growth rates of each sector, with separate production functions estimated for both private sectors and the public sector. The increased size of the government sector and its effect on the growth rate could then be directly measured. Scully rejects this approach, simply because data are not available. There is a much more important problem than this. Ignoring any attribution of the relation between growth in public investment and output in the private sector, how is one to measure the real output of public sector services which are not priced? Conventional measures based on input changes 'adjusted' to take account of productivity changes are notoriously arbitrary.

(ii) Government expenditure as a percentage of GDP cannot embrace all the influences of government actions on the private economy. As is often argued, government by statute and regulation can exercise a major effect on private incentives and the allocation of resources. At the very least, any complete empirical study should take account of non-budgetary forms of government intervention. It cannot be assumed that growth in regulation is complementary to the growth in overall government expenditure. It could be; but account would then have to be taken of the influence of the composition as well as the amount of government expenditure on the complexity of the regulatory system. Matters are further complicated by the propensity of Western governments embarking on privatization programmes to subject privatized concerns which were previously public monopolies to strict regulatory control. The size of the public sector relative to the rest of the economy will decline, cet. par., but will its influence be diminished? (iii) One should be careful not to conclude that a Western-type economy which embarked on a vigorous programme of cutting government services in order to improve growth prospects would necessarily achieve its aim. As Maddison (1991, pp. 79-80) has emphasized, extensive state action to equalize the primary distribution of income has been a prerequisite for widespread acceptance of capitalist property relations and the operation of market forces, making them more legitimate by removing most of the grievances which motivate proponents of a socialist alternative. This is not to deny that potential efficiency gains would not accrue from reductions in social transfers and subsidies with disincentive effects on individuals and firms respectively, but the perceived distribution effect could engender strong opposition to major cutbacks. While it is arguable whether Maddison is right in claiming that the efficiency losses from the growth in public spending are likely to be insignificant, one can certainly agree with him that they are modest in comparison with those experienced in command economies where the level of state intervention has been very much higher.