ÆäÀÌÁö À̹ÌÁö
PDF
ePub

C. DEMAND FACTORS

Previous empirical studies have recognized that the larger part of the effect of changes in demand makes itself felt in price behavior through the cost channel: the response of input prices to demand yields their contribution to increased output price, their importance depending upon their weight in the total value of the output. Under perfect competition, prices move directly with cost changes, supply and demand are kept equal as markets are cleared, and physical capacity remains fully utilized. Under a strongly oligopolistic, targetreturn or full-cost pricing situation, price movements are again stimulated by cost changes as efforts are made to eliminate deviations of realized from desired profits.

It is primarily in the intermediate case, between perfect competition and strong oligopoly, that measurable demand conditions are seen to affect prices directly in studies of individual industries. Such measures as variations in industrial utilization rates, the ratios of inventories to sales, and changes in orders appear to have significant independent influences on prices.

In the present study, these and other measures were explored. The one measure of demand disequilibrium that was found to be strongly significant throughout the tests was the ratio of real, unfilled orders to physical capacity. If shifts in factor costs, government spending or private tastes result in a market price below equilibrium, unfilled orders will increase and equality of demand and supply will only be achieved through inflation. This effect is felt most strongly in the early phases of an economic espansion before production adjusts to the more vigorous demand. On the other hand, later in the cycle when this ratio falls, there appears to be no significant deflationary effect because of the downward rigidity of prices. Similar asymmetries were found when the change in the ratio of real inventories to capacity was used as the sensor of excess demand in the output market.

Lengthening the response lags with respect to excess demand variables did not noticeably improve the explanatory power of the equations. The only possible difference noted was that, ceteris paribus, equations estimated with longer lags tended to have smaller residuals during the late fifties and early sixties, but generally larger residuals in the late sixties.

The variations in the composition of demands were important during some of the periods, but their impact is primarily felt through the factor prices and unfilled orders. Sectoral demand shifts lead to higher aggregate wage and material costs due to asymmetries and rigidities of the type described by Schultze (1959). Our empirical studies indicated that the largest additional effect on prices in any given quarter was less than one-quarter of one percent.

'As capacity expands, the volume of unfilled orders can expand correspondingly without an inflationary impact. The Federal Reserve Board index of manufacturing capacity was used as the capacity measure. The orders variable was deflated by a fixed weight index of durable and non-durable manufacturing output prices. The weighted index was necessary because durable manufacturing has a disproportionate share of unfilled orders.

Other authors, beginning with George deMenil, further alter such variables by division by a trend connecting cyclical peaks, in an attempt to remove changes in the desired order-sales relation (which they assert to be failing). This correction was also made by us but excluded in the final version because it appeared also to remove information that should be retained.

The potential effects of shifts from a "war" to "peace" economy were tested using various rates of change of the share of defense spending in the budget or GNP. None of these was found to be significant. The effects of the Viet Nam war buildup and the withdrawal are captured by their effects on input prices and on unfilled orders.

D. STABILITY OF THE COEFFICIENTS

Chow tests were performed to test the stability of the sets of coefficients in two formats: first, the standard unit labor cost weights were constrained to those just derived [(0.351, 0.260, 0.195, 0.130, 0.065), (-0.089, -0.044, -0.033, -0.022, -0.011)], and second, new relative weights of current and lagged values were permitted. Also, the pattern of change in the unit labor cost coefficients was analyzed by utilizing regressions over expanding intervals. Table 7 presents the results of these studies. The sets of coefficients are not statistically different in any of the paired periods tested regardless of whether the relative labor costs weights are constrained or not. However, examination of the cost coefficients as the interval is expanded does indicate some movement.

[blocks in formation]

The acceleration of wages occurring in 1967 and 1968 was evidently not matched by an immediate, comparable price inflation. This is indicated by the reduction in the sum of the coefficients in this period

and by the declining relative weight for current increases in standard unit labor cost. This phenomenon as well as the price resurgence in 1970 are possibly best explained by variations in the pressure of foreign competition. Figure 8 below indicates the four-quarter inflation rates of import prices and non-farm, private domestic output prices. Beginning in late 1966, import prices rose very little, implying that further domestic inflation would have significantly reduced the competitive position of American firms. By 1970, the relative inflation rates were reversed as import prices rose sharply. American prices then responded fully to wage inflation.

[blocks in formation]

The basic price equation derived here can be used to analyze the separate factors that explain the increase of prices as seen through the equation. It is important not to impute causality in a static framework to separate factors, because of the simultaneity of prices and wages. Nonetheless, some additional understanding is gained by this form of historical review.

Table 8 presents the historical decomposition for the period as a whole, 1955-1970, and for the major sub-periods. Taking the entire period, prices can be seen to have responded mainly to changes in standard unit costs. Of the average annual rate of increase of prices of 2.36 percent, 1.96 percent is the cost factor. The temporary productivity deviations average out close to zero, since one period's extra gain is the next period's loss. The increases in the ratio of unfilled orders to capacity do add independently to the total price increases because only the pluses are significant; when the ratio returned to normal it did not serve to retard price increases. Finally, the changes in the mix of compensation add another 0.16 percent.

[blocks in formation]

Note: These figures are slightly less than the actual average, due to the omission of the constant term, 0.0026.

Looking at shorter periods, the predominance of the labor cost factor can be seen throughout. But there are some interesting variations in the other factors. The short run productivity swings added substantially to the inflation of 1967-1970, an extra 0.27 percent a year, as the economy switched from an extended period of high growth to four years of growth substantially below the long run trends. During the early phases of the recent upswing, from 1963-66, the productivity factor cut the inflation rate somewhat.

The buildup in the ratio of unfilled orders to capacity had a large effect on prices in the business cycle upswings of 1955-58 and 1963-66. In interpreting these coefficients it should be kept in mind that the impact of these factors on the wage-price system as a whole is considerably more important than on the price equation alone. The extra price increases soon become reflected in increases in wages, which in turn raise prices through higher standard unit labor costs. The impacts of these factors on the entire wage-price system are analyzed in the next chapter.

Chapter V. THE WAGE-PRICE SYSTEM AS A WHOLE:
SIMULATION STUDIES

Because wages and prices are linked together in a simultaneous process, the overall wage-price behavior of the economy can only be ascertained by analyzing the system as a whole, either by explicit mathematical solution of the equations, or by simulation studies. While the equations used here are small in number and of sufficiently simple structure to allow explicit mathematical solution, the simulation method allows retention of the empirically derived lag structure, exploration of the sensitivity of the system to minor variations in specification, and derivation of empirical estimates under alternative economic conditions. This chapter mainly uses the simulation technique. A brief analytical treatment is also presented.

The plan of the chapter is as follows:

First, the model is fleshed out with some supplementary relationships which are necessary to consider it a closed wage-price system.

Second, a historical simulation is run to see if the model can reproduce the actual record. This simulation is used to analyze the two most recent inflations-the episode of the mid-1950's, and the current experience and also to account for the intervening periods of price stability.

Third, the simulation model is used for a series of exercises to assess our present prospects for achievable combinations of unemployment and prices. The model is run forward from mid1971 assuming that there would have been no wage-price freeze, to estimate how long unemployment would have had to remain high to bring the economy back to reasonable price stability.

Fourth, the model is run assuming alternative phase II policies on permissible wage increases. Estimates are also made of the path that wages would pursue given the price result; the difference between the assumed wage increase and the increase that would be produced without constraint for the given price performance is a measure of the additional contribution that policy would have to make to the disinflation process. This exercise is repeated on the price side.

Fifth, the long-run Phillips Curve is traced by simulations which assume that various given unemployment rates persist without change over a sufficient number of years to allow the initial conditions to wear off. Next, a condensed, analytical version of the model is presented.

Sixth, some further sensitivity studies are presented.

A. COMPLETING THE MODEL

Additional links must be provided for the model to complete it. While the model uses the deflator for the non-agricultural, private

« ÀÌÀü°è¼Ó »