Publications Empirical Studies

 5.Discoordination  Hypothesis

1
Both variants of investment cycle theory, i.e. the reinvestment and the investment lag theories have one common feature: They assume that fluctuations of output are a direct outcome of fluctuations of inputs, primarily of capital input. This may be seen from Klacek's and Kupka's attempts to put Goldmann's model into a mathematical form. Their models contain a simple single-input production function, Y = F(K), so that the fluctuations of output must come from the fluctuations of capital stock which in turn could result only from fluctuations of investment (they disregard scrappage). This approach is consistent with the common view that the supply oriented Soviet-type economies always work with full utilization of capacity. This view seems to be supported by the nonexistence of unemployment and by the fact that production plans are almost always taut.

Nevertheless, such an interpretation is misleading. It can be very simply demonstrated that the primary inputs - labor and capital - fluctuate during the cycle much less than does output, which indicates that the degree of utilization of factors of production, or rather the efficiency with which factors are utilized, cannot be constant over the cycle. This can easily be tested by estimating total factor productivity from a production function, for example of a Cobb-Douglas type with disembodied technical change containing quadratic and cubic terms.

Y = AKbL1-b exp(r1t + ½ r2t2 + 1/3 r3t3 + e)

2   The fitted values of output Y* can then be interpreted as "normal capacity output", that is, the output which would be produced with fully utilized factors and with normal trend in factor productivity. It is further apparent that as long as the actual output is larger or smaller than the fitted value, the actual factor productivity deviates from its trend value. The residual  u = logY - logY* or better its antilogarithm  eu = Y/Y* measures the ratio of actual output to "normal capacity output". If our hypothesis that the fluctuations of outputs cannot be fully explained from fluctuations of inputs is true, then the residual u and its antilogarithm should follow a cyclical pattern.

Our attempt to estimate simultaneously all the parameters of the production function was not successful for the known problem of East European time series - multicollinearity. Because we are not interested here in precise values of the capital elasticity b, we have simplified the task by choosing our previous estimate of b for Czechoslovakia (b = .35) and then estimating just the parameters r1, r2, r3.Of course, we kept r3 and r2, in the equation only when they turned out to be statistically significant.

The estimates of trends in total factor productivity from the production function were run for the total economy, industry, construction, and agriculture and, alternatively, with gross or net products as left hand variables. These are the estimates mentioned already at the beginning of this paper. The estimated parameters as well as fitted values Y*, residuals, and exponentiated residuals are shown in the Appendix. Exponentiated residuals are also shown in Diagram 1. They very clearly demonstrate considerable fluctuations in total factor productivity during the cycle which are much larger than the fluctuations of capital stock and labor

How can we explain cyclical fluctuations in total factor productivity? We suggest that the plausible explanation is to be found in the periodic changes in the degree and quality of coordination of the economy. The economy goes through periods when everything is all right, plans are well balanced and can be easily overfulfilled, no shortages in the material supplies occur etc. These are the periods when the ratio of actual output Y to the capacity level Y* is increasing. Then, for some reason, the economy begins to discoordinate. Plans cannot be fulfilled or even be well balanced, scarcities and bottlenecks begin to grow, and the ratio of actual to capacity output begins to decline. That this had actually been happening is not difficult to prove; the question is, why it has been happening. In this respect we are inclined to believe, together with Kalecki, Goldmann, and others, that the reasons may be found in the overreaction or wrong reaction of planners. Let us see whether the elements mentioned so far can be put together with some additional elements so as to generate growth cycles.

First, we have to find out why planners persistently repeat their errors by tending to overstrain the economy. The main reason, in our view, is that planners do not know what the "equilibrium" growth path of the economy is. They can approach it only by trial and error. This would be a feasible method, if it were not combined with a peculiar feature of planners, namely to think generally in terms of percentage rates of growth. This attitude, as can easily be shown, results in a tendency to overstrain the economy.

Suppose that the equilibrium growth path of a particular economy is determined by the capacity output as defined by the fitted values of the production function. In the Figure below, EGP represents the equilibrium growth path thus defined.  If actual output in period t is A and in period t + 1 is B, then planners have the impression that the economy is doing well. A relatively high rate of growth was easily realized and an ambitious planner might think about increasing the planned rate of growth.  However, let us assume that planners are less ambitious and are satisfied with the moderate goal of keeping last year's pace. They, therefoe, impose a plan on the economy that is based on the same rate of growth as last year's performance. This might again be a feasible target for the economy such as indicated by point C in the Figure above. However, if the planners now continue to remain as "moderate" as in the two preceding years, that is, if they again impose the same rate of growth on the economy, they will obviously he pushing the economy beyond its capacity limit. Thus, overambitious plans, in the sense of increasing the target rates of growth, are not needed in order to explain the overstrain of the economy. It is sufficient if planners follow the "moderate" goal of repeating what has proved successful in the previous year. Of course, if planners would react to a successful year by putting target rates of growth higher, the argument developed so far would hold afortiori.

5  In fact, it is possible to show that planners in Czechoslovakia behaved just in such a way. They adjusted the planned rates of growth almost always with a certain lag after a change in the actual rates of growth, and sometimes they clearly overreacted. In the early 1950's, the actual rates of growth of national income were 10 percent and the planned rates of growth only 7.6 percent. As a result, planners revised the plan targets of the Five Year Plan and increased the planned rates of growth to 13.6 percent. However, this rate was never achieved. After two years of declining actual rates of growth, Czechoslovak planners cut considerably the plan for 1955, i.e. for the year in which the actual rate of growth was already up.

The information contained in published planned growth rates does not fully reveal how the central authority reacted to good or bad performance of the economy. Since the central economic decision makers' ambition mainly means that they attempt to attain high rates of growth of output, we should expect to find clear evidence that this ambitiousness manifests itself in the primary source of growth, i.e. in the accumulation of capital stock. The time series of investment should, therefore, mirror the planners' reaction to good and bad years, respectively. We can test this hypothesis by estimating a special form of investment function. Let us assume that there are two, rather than one, investment functions describing planners' investment decisions. The first one corresponds to the decision which would be made if the economy were on its equilibrium path. The other one corresponds to the deviation from the equilibrium path.

6  Let us consider the simplest version of these functions:

I* = ao + a1Y*                 I** = bo + b1(Y - Y*)

For total investment, it must then be true that

I = I* + I** = ao + bo + a1Y* + b1(Y - Y*)

I = go + g1 Y* + g2 Y

7   where   go = ao + bo ;      g1 = a1 - b1 ;       g2 = b1

Y* = capacity output calculated as a fitted value of the production function

I* = the corresponding "normal" investment

I** = investment induced by (Y - Y*).

For total investment and output (measured by national income), we obtained the following estimates:

Table 4: Investment Function

 Type Estimated parameters (t-statistics) Auto- regres. Coeff. R2 D.W. F go g1 g2 OLSQ -14.167 -.055 .417 .994 1.12 2051.8 (-12.281) (-1.225) (9.420) CORC -14.322 -.078 .441 .452 .995 1.54 2189.0 (-6.815) (-1.266) (7.336) (2.535)

8
The results are statistically quite good, except for a low t-statistic of the coefficient  g1. However, the very large F-statistic shows that the joint explanatory power of the right-hand variables is highly significant, which indicates that the low t-statistic of g1 is probably a result of multicollinearity.

From the estimates of g1 and g2 we can derive:

 OLSQ CORC a1=.362 b1=.417 a1=.364 b1=.441 (62.996) (9.420) (38.275) (7.336)

The economic interpretation is clear: Planners do respond quite differently to the equilibrium situation and to the disequilibrium situation. Their "marginal propensity to invest" in the equilibrium situation is I/Y* = a1 = .36   Their "marginal propensity to invest" out of deviations of output from the equilibrium path is I/Y** = b1 = .42 to .44

9  This clearly demonstrates the overreaction of planners: If the economy is above the trend they invest overproportionally more than in the equilibrium situation; if the economy is in the downswing they invest overproportionally less.

Formally, this type of behavior resembles what is known as the accelerator in the macrotheory of market economies. We have, however, to keep in mind that the reaction of planners is not demand determined, but depends on available output, i.e. supply.

The overreaction of planners in both directions confirms what Goldmann and others have observed. From what we have said about the role of the gestation period, we can add here a new feature in the explanation of the cycle. In Goldmann's theory, its length of 8 -10 years resulted in one peak directly creating the next. We claimed that this is not in accordance with our empirical findings which indicate a considerably shorter investment lag of roughly one year. If this finding is correct, then planners get a much quicker response to their investment decisions than was assumed so far. In fact, they seem to get their response within the same upswing or downswing. This positive response encourages planners to remain on their way already long after they have overshot the balanced growth path of the economy, and vice versa.

 OK Economics was designed and it is maintained by Oldrich Kyn. To send me a message, please use one of the following addresses: okyn@bu.edu --- okyn@verizon.net This website contains the following sections: General  Economics: Economic Systems: Money and Banking: Past students: http://econc10.bu.edu/okyn/OKpers/okyn_pub_frame.htm Czech Republic Kyn’s Publications American education