Since 1973 European Economies have faced a period of slow growth aggravated by a new recession in 1980 which has dramatically increased unemployment. It has also particularly affected investment whose volume declined in absolute terms between 1979 and 1983.

Whilst in the period of rapid growth, exports and investment were the driving forces of expansion, however in the mid 70’s growth was sustained rather by public and private consumption. But since then public deficits have constrained governments to adopt austerity policies reducing their expenditures and slowing down as much as possible transfers to households.

Simultaneously, the weakening of firms’ financial situation has led governments to endeavour to render the distribution of value added more favorable to profits, which has brought pressure on real wages. This, together with limitation on the volume of transfer payments, has eroded consumer’s purchasing power.

These policies have been quite effective in terms of inflation rates and external deficits. However, the medium term prospects for unemployment remain around 10%, if not more, of the active population.

Daniel Weiserbs has produced a very interesting model of investment behaviour in the six major countries of the EEC. As is usual in econometric studies he has had to steer a careful course between what is feasible and what is desirable. In particular for this study the data constraints are severe, since the data are annual and only for a 21 year period. In view of such constraints it is very important to have a clear economic interpretation of the coefficients to be estimated and in this paper we are supplied in most cases with a convincing story.

The model is however derived in a fairly ad hoc manner and this can sometimes have the disadvantage of hiding the full implications of the assumptions made. For example we know from work by Lucas (1967), Gould (1968) and others that if we assume that an industry or sector acts, as a whole, like a representative profit-maximizing firm with costs of adjusting its capital stock, then the use of a partial adjustment mechanism as specified in equation (6) implies that the representative firm has static expectations – that is future growth, as seen from a point in time, is expected to remain constant for all time into the future. Now since the expectation of growth one period ahead is in fact specified to be equal to growth this period, the model is implicitly assuming that, at each period, firms expect growth to continue at its present rate for ever – which must have been a depressing thought in 1981.

Usual approaches to the investment decision of the firm consider either that the firm is always constrained on its output sales or that it never faces any constraint at all. The paper is an attempt to enlarge this traditional framework in two directions. First by combining the two previous standpoints and by considering that the firm is sometimes constrained on its output (effective demand) and sometimes not constrained at all (notional demand); next by including other kinds of constraints as well, like a credit or a labor constraint. The outcome is a distinction between five possible «regimes», corresponding each to a different explanation of observed investment.

Irrespective of Malinvaud’s theoretical considerations about profitability (see Malinvaud (1980 and 1982)), this paper is, together with a companion article of Gérard and Vanden Berghe (1984), the second paper I am aware of dealing with investment in disequilibrium situations.

The first paper is one of Nishimizu, Quandt and Rosen of Princeton University, entitled; «The Demand and Supply for Investment Goods: Does the Market Clear?», (1980) (not yet published), in which they apply the popular Maddala-Nelson Maximum – Likelihood Procedure (see Maddala and Nelson (1974)) to estimate the model parameters for a postwar quarterly model on Japanese manufacturing data. They specified and estimated a model of the market for investment goods alone, that explicitly incorporated a supply side and that allowed for the possibility that this market might not clear instantaneously. The most striking results were that the long run Japanese supply curve of investment goods is virtually horizontal, and that the Japanese investment market appears to be characterized by equilibrium.

And now comes this second paper, but also allowing for possible rationing on the output, labour and credit markets.

During the last twenty years there has been an increasing tendency to derive firms’ decision rules from explicit inter-temporal optimizing models. Such quests for choice-theoretical foundations has been regarded both as logically necessary and empirically fruitful in specifying the equations describing input demand, production, pricing and inventory accumulation. In the more recent past the research in this area has followed different directions. A group of studies has relied on the introduction of convex adjustment costs to justify the forward-looking nature of input demand decisions and has analyzed the implications of the rational expectations assumption in this context. The papers by Sargent (1978), and Nickell (1984) on employment, by Abel (1980) on investment, and by Meese (1980) on both are representative of this approach. The papers by Pindyck and Rotemberg (1983a), (1983b) proceed from the same assumption but differ from the rest because the first order conditions of the maximization problems are estimated directly without obtaining a closed form solution for the firm’s decision rules. Another set of studies has instead developed the implications of the dual approach for factor demand, usually adopting the assumption of a cost-minimizing firm facing a putty-putty technology. The contributions by Jorgenson and Fraumeni (1980), and Berndt and Wood (1979), (1975), to mention just a few, belong to this category. A third group of studies has concentrated on prices and output decisions, disregarding the input demand aspect of the problem. Rotemberg (1982), (1983) and Maccini (1977), (1978) are good examples of this line of research. Finally, in Maccini (1981) and (1984) there is an attempt at analyzing investment in fixed capital and inventories together with prices and output decisions in the presence of convex costs in adjusting the capital stock.

This paper is interesting and ambitious since it tries to provide both on the theoretical and empirical levels an integrated account of firm’s behaviour under the assumptions of an imperfectly competitive output market and of a putty-clay technology.

In my opinion subject to caution, if the author succeeds quite well at the theoretical level, the empirical results are more questionable.

One of the difficulties found in building an integrated model of firms’ behaviour is to take into account the different time horizons concerned by the firms’ enterprises.

The intertemporal characters of these decisions result from several factors:

That article is part of an investigation on investment behaviour in Belgium both at aggregate and sectoral level. More precisely it aims to test the hypothesis that investors become more sensitive to expectations regarding future sales opportunities when the economic conditions are going down. Consequently the relative weight of that determinant of investment with respect to the so called notional determinant, i.e. here the real cost of capital, increases when business conditions deteriorate.

Indeed the application of standard investment functions, like the Jorgenson one – see a.o. Hall and Jorgenson (1971) – or the one due to Eisner and Nadiri (1968), to Belgian data suggests that the estimated coefficients not only vary from one sector to another but also strongly depend on the period on which they are estimated. However some economic logic seems to arise from that instability.

Therefore, in order to obtain more accurate predictions regarding the effects of a change in the cost of capital, e.g. a change in the interest rate or in tax parameters, or of a change in expected demand, a methodology is needed which captures that variability of the coefficients of the investment functions.

The authors derive some interesting conclusions about the relative importance of the cost of capital and output during periods of economic crisis and relative prosperity. In essence as a result of output constraints during periods of economic recession the level of output is seen to dominate the demand for capital goods with real cost of capital playing a minor role. These conclusion result from a theoretical development of investment behaviour that allows for the introduction of rationing when a sector is operating under an output constraint. The following points arise from a detailed study of the paper.

Initial experiments with standard investment functions of the Jorgenson & Eisner-Nadiri type give the familiar result that they are unstable. If attention is turned to the results presented in Tables 2 and 3 it can be clearly seen that elasticities and mean lags are very sensitive to small changes in the data set. This can be interpreted as indicating the severe problems of mispecification that are present but may also result from severe multicollinearity between the explanatory variables or the presence of an outlier in the data set. If the values of σ and μ are used to calculate v, the degree of homogeneity of the production function, it is worrying to find that it takes unrealistic values with a large number of estimates exceeding 2 and several exceeding 20. In addition, the interpretation of the mean lag should be treated with caution. The mean lag is meaningless if, as for some sectors appear likely from the information given, the value of βj alternate in sign. The use of the median lag may however overcome these problems.

In this paper we have a threefold objective. Firstly, to identify the principal macroeconomic determinants of U.K. private sector investment. Secondly, to examine the effects of private investment on aggregate supply. Finally, we attempt to integrate the effects of public sector investment and pricing on private sector output.

Investigators have modelled private sector or manufacturing investment behaviour in a variety of ways. The neoclassical tradition, associated with Jorgenson (1963), explains investment expenditures in terms of the lagged capital stock, expected output and a distributed lag on the cost of capital. The most recent effort along these lines on U.K. data is the paper by Bean (1981). Although he found significant negative effects of the change in interest rates and the real cost of capital on the change in investment he failed to obtain satisfactory estimates of long-run cost of capital effects on the level of investment. Thus investment was determined solely in relation to output in the steady state. Alternatives to the neoclassical formulation have been proposed in the form of the multiplier-accelerator model and Tobin’s q model, in which changes to the capital stock incur adjustment costs. The former, which have been widely used in macroeconomic models, have generally failed to predict turning points in investment. Oulton (1981) and Dinenis (1984) have had moderate success with q models for the U.K. The major difficulty encountered is that of rationalising the existence of a lag structure for q, the ratio of market value to the current replacement asset cost.

The conclusions of this paper are interesting and controversial. Its conclusions that the promotion of investment can only come about by means of reducing real wages, and, the public sector borrowing requirement means that Keynesian policies being recommended for the reduction of unemployment are inconsistent, in that increases in these two variables will lead according to the results of this model to a reduced level of investment, and hence a lower rate of growth of output, and a lower rate of job creation. These results being true in the short run as well as in the long run. In my view it has not been conclusively shown that these results are true for the British Economy, and I intend in the rest of my comments to explain my reservation about the model and its results.

As these results are derived solely from a view of the data obtained from their model, a first approach is to examine their model critically. At first the model seems unexceptable. In fact in many ways it embodies a degree of structure unusual in such a study. However, the structure is not complete, one could not expect it to be, and it is here that some problems may arise.

The investment equation is determined as one from a set of factor demand equations derived from profit maximisation subject to a KLEM technology. Output is somewhat curiously defined but for the moment let us pass over that and concentrate on the problem max

Current unemployment is at unacceptably high levels for both the segment of the population concerned and policy-makers in most industrialised countries. The causes of the rise in world-wide unemployment are usually attributed to the severe supply shocks in the 1970s (the sharp increase in oil and other raw material prices), the associated redistribution of income from countries with high propensities to spend to countries with low propensities to spend, inadequate policy responses and rigidities in industrial economies responsible for slow adjustment to a different environment(‘)• Rigidities seem to be more severe for the countries of the EEC than for Japan and the US and unemployment consequently has risen more sharply in the EEC. The main rigidity can be seen in the high growth of real wage costs. Instead of attenuating the supply shocks through moderate wage increases the effects of these shocks have been accentuated in the EEC by the wage cost explosion of the 1970s. An approximate empirical measure for these effects is provided by the sharp increase of the real wage cost gap in EEC countries which, however, is an underestimate since firms have responded by reducing employment to improve productivity.

The title of the paper presupposes that there is a choice only between investment or employment subsidies in job creation. Perhaps this was not the intention of the author but by the bottom of p.2. it is clear that «first best» remedies for unemployment have been dismissed as undesirable or impractical. Before moving on to discuss the body of the paper I think that an important issue has been brushed aside here. Why not tackle the wage rigidities that are held responsible for the high unemployment directly? Why not dismantle the existing structure of subsidies on capital and taxes on labour as a first step? Perhaps then the nature of the unemployment problem will be more clearly seen. Determination to remove factor price distortions which arise either because of tax-subsidy arrangements or restrictive labour practices can overcome the objection that this strategy will take too long.

It is plain that even if one accepts the raison d’être of the paper there is also a lack of information with which to apply a Marginal Employment Subsidy (MES). The natural rate of unemployment is unobservable and hence only approximately known. Also there is the problem of structural unemployment caused by a mis-match of skills of the unemployed with vacancies created by an MES. It is difficult to see a 55 year old male steel-worker being re-employed as an assistant in a sweet shop, but something of this sort is implied by the analysis. Also, there are the risks associated with an MES. The first is the risk of creating an artificial situation of excess labour demand for certain skills or in certain industries causing an increase in real wages which would be readily mimicked by other sectors of the economy. Second, there is the risk that in creating a further distortion to factor prices on the pretext of a short-term remedy, the subsidy would, like its predecessors, become a permanent feature and prevent desirable labour market adjustments taking place in future years. Third, there is the risk of giving expanding firms a cost advantage in their industry which, as their market share grows, becomes self-perpetuating.