«In the considerable volume of economic literature on public enterprise, one area appears to have been largely overlooked: the impact of public ...»
PUBLIC ENTERPRISE INVESTMENT A N D ECONOMIC STABILITY:
A SIX COUNTRY COMPARISON*
Wayne W. SNYDER,
Center for Research on Economic Development,
University of Michigan (Ann Arbor, Michigan, U.S.A.)
In the considerable volume of economic literature on public enterprise, one area appears to have been largely overlooked: the impact of
public enterprise investment on economic stability. This omission is somewhat surprising, because for several decades economists have demonstrated a growing interest in the relationship of other governmental expenditures and revenues to economic stability. Most of that literature is concerned with the effect of central government budget changes, although in some countries the relationship of state and local government budgets to economic stability has also been studied, and, of course, the automatic or built-in stabilizing effects of social security systems have long been recognized.
The lack of study of the stabilizing effects of public enterprise investment may be due in part to the belief that investment is primarily determined by reasons that have little to do with conjunctural considerations, and that anyway its determination is largely outside the influence of the central government. Both positions are open to question, and in any case public enterprise investment does have a de facto impact, the importance of which is worth studying regardless of whether the government uses or is capable of using it as one means of attempting to achieve economic stability.
Our purpose in this paper is to determine the extent to which public enterprise investment was a sta'bilizing or destabilizing factor during
the eleven-year period from 1955 through 1965 for six countries:
Belgium, France, Italy, Sweden, the United Kingdom, and the United States.l The methods of estimating the impact of that investment are * I am particularly indebted to J.C.R. Dow, who conceived and directed the OECD study o fiscal policy [4J, and to Bent Hansen, with whom I had the pleasure of working f for two years on completing it. I also benefitted from the helpful editorial assistance of Janet Eckstein.
1 A seventh country, Germany, was included in the Hansen study, but no complete data are available concerning its public enterprise investments.
38 WAYNE W. SNYDER the same as those developed for a recent OECD survey by Bent Hansen , which gives the institutional background for budgetary action and analyzes the nature and effects of fiscal policy for seven OECD member countries between 1955 and 1965. This paper supplements the Hansen study by singling out public enterprise investment.
I. Measuring the Effects of Public Enterprise Investment The model developed by Hansen to measure the effects of various types of budget changes draws on the previous contribution of Brown [ l ], Hansen , Lindbeck , and Musgrave . Although Hansen’s model is admittedly simple compared with the large econometric models which have been developed for some countries (partly because he wanted to use a common analysis for each country), it is adequate to measure the relative importance of various kinds of budget changes both within and among countries. Because there were no quarterly data for some countries, his model uses year-to-year changes, and no lags are introduced. The model assumes that private investment, exports, and prices are exogenously determined. Imports are endogenous and for some countries (e.g., Belgium and Sweden) represent the principal leakage of potential budget effects. 2 The combination of the direct and the multiplier effects, or the total
effects, of budget changes are estimated by the following equation:
+ a(1-P) Total effects = [dI (IdPdl 1 --a(l+) where a is the marginal propensity to consume, P is the marginal propensity to import, dI is the annual change in the volume of public enterprise investment, and Idp, is the annual change in the value of public enterprise investment due to price changes. The leakage coefficients a and IA vary, of course, among countries, and consequently the multipliers for the effects due to changes in the volume and price of public enterprise investment differ also. These multipliers allow only for leakages via private savings and imports, and exclude leakages via tax changes and spending induced by the changes in public enterprise investment itself. The effects of tax changes are excluded, because they are included in the impact of general or central government, which is discussed later. If leakages due to tax changes were explicitly included, the effects of public enterprise investment would be diminished for each country but its relative position among the countries would not be affected.
2 For a complete description of the methods used to measure budgetary effects, see Hansen 141, Chapter 1.
PUBLIC ENTERPRISE INVESTMENT AND ECONOMIC STABILITYThe year-to-year variations in the total effects of changes in public enterprise investment, expressed as a percentage of GNP, are given in Table 1. Before we evaluate whether these have been generally stabilizing or not, several comments may be helpful. Several factors account for differences in the typical impact, whether expansionary or dampening, among countries: the relative size of the public enterprise sector and its annual investment, and the relative importance of those two multipliers. For example, France, Sweden, and the United Kingdom all have large public enterprise sectors whose investment normally exceeds 5 per cent of GNP, but the typical impact was nearly 2 per cent of GNP in France and only about 0.5 per cent in Sweden and the United Kingdom. The difference is due partly to the more rapid expansion of the public enterprise sector in France, but it is also due to the smaller multiplier effects in Sweden and the United Kingdom, where import leakages substantially reduce the potential effects of public enterprise investment. The public enterprise sector expanded most rapidly in Italy, where its investment increased from 2 per cent of GNP in 1955 to 4 per cent in 1965. As the sector was relatively small in 1955, the average impact was not as large as in France, but it did amount to about 1 per cent of GNP annually, the second highest among the countries studied. In Belgium the public enterprise sector was smaller than in the countries already mentioned and did not change much during the period covered. In fact, the average impact amounted to only about 0.1 per cent of GNP. The same was true in the United States, where the public enterprise sector was by far the smallest among the countries studied, and where investment amounted to less than 0.5 per cent of GNP. Year-to-year changes were often important, especially when the large multiplier effects are included, so that a typical impact amounted to about 0.5 per cent of GNP, although the numerous times when the effect was dampening more than compensated for the expansionary effects which resulted in the average effect having a slight dampening impact.
Evaluating the Impact of Public Enterprise Investment 11.
No single criterion is clearly the only on,e by which the impact of public enterprise investment on economic stability can be judged. Three of the possible criteria are discussed in this paper. The first evaluation attempts to judge the short-run stabilizing effects by simply determining whether the actual growth rate of GNP would have been more or less stable without the impact of public enterprise investment. Figure 1 illustrates the relevant relationships which are necessary in order to make this assessment: the average GNP-growth rate, the actual rate, and a measure of how GNP might have developed in the absence of the impact of any public enterprise investment (but including the budgetary effects of the order government sector-central and state and
TABLE 1. TOTAL EFFECTS OF ANNUAL CHANGES IN PUBLIC ENTERPRISE INVESTMENT, 195565
local government and the social security system-as well as the combined influence of all other exogenous and endogenous factors). The latter measure is obtained by subtracting the total effects of public enterprise investment from the actual GNP growth rate (see Table 2).
For example, in Belgium during 1955 the total effects amounted to +0.5 per cent of GNP and, ceteris paribus, contributed to pushing the actual growth rate that much higher above the average rate than it otherwise would have b e e n - b y definition a case where the budget impact was destabilizing.
The net stabilizing (or destabilizing) effect over a number of years is simply the accumulated sum of the (gross) stabilizing minus the destabilizing effects; these and other relevant data are given in Table 2.
The net effects were stabilizing in only two countries, France and the United States, and on balance they had a destabilizing impact in the remaining four countries, Belgium, Italy, Sweden, and the United Kingdom. It must be immediately noted, however, that on average the net effects were generally quite small, amounting to no more than 0.1 per cent of GNP annually except in France, where they were more than 1 per cent, and the United Kingdom, where they were about 0 3 per cent.
In general, the rather small net effects resulted from stronger (gross) stabilizing effects being diminished by even larger destabilizing effects in particular years. France again is an exception, as the stabilizing effects were very large and were relatively little diminished by years when the effects were destabilizing. Figure 1 makes clear how little public enterprise investment influenced economic developments in Belgium, Sweden, and the United States, and how substantially they altered the course of events in France and Italy; the destabilizing nature of the effects in the United Kingdom is also evident.
This first criterion for evaluating the net stabilizing effects is most relevant in those countries in which (a) public enterprise investments are strongly influenced by government policy, and (b) this policy is part of the general conjunctural economic policies. Understandably, our six countries vary considerably in these respects. The central government of France maintains the greatest degree of control, which perhaps helps to explain why public enterprise investment was a stronger stabilizing factor there than in any other country. Belgium and the United Kingdom also maintain rather close control, but they allow a greater degree of autonomy in the investment decisions of public enterprises than does France. In Italy, where the public enterprise sector has grown rapidly in importance, the central government has relatively little direct control, except in the case of some public utilities companies, because self-financing provides about 90 per cent of the investment funds. In Sweden, the central government controls rather closely the investment of public enterprises which belong directly to it, but these are few in number and importance; most of the enterprises belong to local government, and the influence over investment decisions is only indirect, through general credit market policies. The same is true of the United States, but the enterprises belonging directly to the central government are so few and small in importance that no separate data about their investments are published. We do know, however, that aside from the stockpiling of strategic materials by the central government, nearly all U.S. public enterprise investment is made by companies responsible to state and local authorities. Because of these substantial differences among the countries studied, it is not at all clear that the basis for appraising the stabilizing effects should be the impact on potential fluctuations in the GNP growth rate, given all other budget effects and exogenous influences.
A second possible comparison gives different results for some countries. We can ask whether in the absence of any budget changes the net effect of changes in public enterprise investment alone has been generally stabilizing or destabilizing. In order to answer this it is necessary to construct a series for GNP growth as it might have developed without any budget impact, by subtracting from actual GNP growth the total effect of all budget changes (public enterprise investment as well as general government). This hypothetical series is called the “pure cycle”, because it attempts to estimate what GNP growth
PUBLIC ENTERPRISE INVESTMENT AND ECONOMIC STABILITYwould have been if there had been no change in any government expenditures or revenues from one year to the next. The pure cycle still incorporates the effects of other government policies (e.g., monetary and direct controls) as well as of autonomous forces (e.g., private investment and exports) and endogenous mechanisms (e.g., leakages through private saving and imports). Hence the pure cycle is not so “pure”, but it nevertheless is a useful analytical tool. Additional graphs illustrating the pure cycle and the relationships similar to those included in Figure 1 are not presented here, because the lines intersect to such a degree that interpreting them becomes more difficult than the help they might provide, but several interesting relationships to the pure cycle can be derived, and these accumulated effects are also included in Table 2.
Potential stabilization can be defined a s the (absolute) difference between the pure cycle and the average GNP growth rate ; potential stabilization for the period 1955-65 is simply the accumulated sum of these annual differences. The effects are defined as stabilizing if they make GNP growth closer to the average than it would have been without them; otherwise they are destabilizing. The accumulated amount of potential stabilization (item 3) is roughly the same for four countries, Belgium, France, Italy, and Sweden; somewhat less for the United Kingdom; and considerably greater for the United States (it is well known that cyclical fluctuations, actual as well as potential, have been substantially larger in the U.S. than in the European countries).