Printer Friendly

Economic stability in the 1990s: the implications of improved inventory control.

THE WAY THAT movements in inventory investment dominate business cycles is well known. In the six most recent recessions, for example, changes in inventory investment accounted for an average of 98 percent of the peak-to-trough changes in gross domestic product. At other stages of the business cycle, too, GDP and inventory investment often move together.

Many business analysts have suggested that improvements in inventory control, such as just-in-time and production-to-order, have reduced the influence of inventory investment on business cycles. We investigated this issue in our recent empirical study.(1) We tested the hypothesis of whether inventory control has improved and we found that it has. We concluded, however, that improved inventory control has increased the volatility of inventory investment and, therefore, the sensitivity of the business cycle to swings in inventory investment.

In the following paragraphs, we summarize our inquiry into questions about improved inventory control. We then review the implications of our findings for inventory cycles, and we suggest what improved inventory control implies for economic stability in the years ahead.



To test the hypothesis of improved inventory control, we use a standard partial stock-adjustment model of inventory investment.(2) In this model, inventory investment in any period, or the amount by which inventories at the end of a period exceed those at the beginning, is the sum of planned and unanticipated inventory investment.

The "partial stock adjustment" description of the model refers to the planned component of inventory investment. Users of this model assume that in any period, businesses close some but not all of the gap between the actual inventories they carry into the period and the inventories they would like to have at the end of that period. More specifically, businesses plan their inventory investment in any period to be a proportion of this gap. That proportion is of special interest, because it determines how fast inventories will be adjusted when there is a discrepancy between actual and desired stocks. This proportion, or the "speed of adjustment" parameter, is one of two inventory control parameters that we use to represent inventory investment behavior.

Desired stocks of inventories are not observable, of course. In order to use the partial stock adjustment model to test for improvements in inventory control, therefore, we make some simplifying assumptions. We assume that, at the margin, desired stocks change in proportion to expected sales, and that expected sales in the current period equal actual sales in the past period. This proportion, the desired marginal ratio of inventories to sales, is the second inventory control parameter whose behavior over time is of interest.

We estimate values of the two inventory control parameters in periods before and after 1980 to determine whether inventory control has in fact improved. We use constant dollar data for manufacturers, for retailers, and for wholesalers. Further, we look for changes in manufacturers' control of their inventories of finished goods and in their control of their inventories of materials and work in process. We disaggregate total business inventories to this extent because inventory behavior changes in different ways for different reasons in different sectors.

Our estimates of the two inventory control parameters for periods before and after 1980 are summarized in Table 2. Improved inventory control shows up more clearly in manufacturing than in trade. In improving their controls of inventories of finished goods and other stocks, manufacturers increased the speed with which they adjusted inventories and decreased their desired marginal ratio of inventories to sales. It is clear from this evidence that manufacturers now control their inventories much more tightly than they did before 1980. In wholesale trade, both parameters suggest improved inventory control, although the changes are not statistically significant. In retail trade, the results are even less conclusive; the speed of adjustment rose, a sign of improved inventory control, but the desired marginal inventory-sales ratios also rose, suggesting a deterioration of inventory control. We feel that these adverse findings for the inventory-sales ratio in retail trade may reflect structural changes in this sector, including the increasing market share enjoyed by the large discount and retail warehouse chains.
Table 2
Values of Inventory Control Parameters
in Earlier and Later Periods
 Desired Marginal
 Speed of Adjustment Inventory/Sales
 Earlier Later Earlier Later
 Period(*) Period(**) Period(*) Period(**)
 Materials and
 Work in Process 0.12 0.48 1.77 0.52
 Finished Goods 0.09 0.37 0.35 0.08
Retail Trade 0.28 0.47 1.62 1.84
Wholesale Trade 0.13 0.20 1.44 1.19
* 1967:2 through 1980:4 for manufacturing; 1967:2 through
1979:2 for trade
** 1981:1 through 1991:2 for manufacturing: 1979:3 through
1991:2 for trade


We turn now to the implications of improved inventory control for inventory cycles. To draw some conclusions about these implications, we simulated inventory investment under alternative assumptions about the behavior of sales and the values of the inventory control parameters.

In the first set of simulations, we assume sales grow at a constant rate, drop in one quarter and then resume growth at the former rate. These simulations indicate that the improved inventory control implied by the new parameters as compared with the old ones leads to a larger inventory swing with this one-time negative shock to sales. With the new control parameters, inventory liquidation is initially larger in response to the dip in sales and is completed in fewer quarters.

For a second set of simulations, we assume sales behave the way they have in past recessions, on average, i.e., sales fall a total of about 5 percent over a four-quarter period and subsequently rise by more than 3 percent over the next two quarters. As was the case under the prior simulations, these simulations show again that, with the new parameters, inventories are liquidated faster while sales are falling, but recover faster once sales begin to recover.

Finally, we simulate inventory investment over the most recent recession experience using actual sales data. Sales over the period are characterized by a sharp dip in 1990:4 and 1991:1 totaling over 5 percent, followed by an increase of less than 2 percent over the next three quarters, 1991:2 through 1991:4. Once again, the simulation obtained using the new parameters produces a larger inventory cycle than the simulation using the old parameters.

The accompanying chart illustrates the fit of our results to the actual data. The inventory investment series that is simulated with the new parameters lines up reasonably well with the actual investment data. We therefore conclude that, in spite of our simplifying assumptions concerning inventory control behavior, our estimates and simulations accurately capture inventory fluctuations in the economy.

In summary, all of the simulations strongly suggest that improved inventory control has increased the volatility of inventory investment since 1980. That is clearly the case, as is shown on Table 3, which documents the increase in the variances of inventory investment. Thus, contrary to popular belief -- or, at least, to popular belief before the recent recession -- inventory investment is more, not less, volatile today. Leaner inventories are not a sufficient condition for less variability in inventory investment because increases in speeds of adjustments can more than offset decreases in inventory-sales ratios. When the two key parameters both show improved inventory control, they have opposite effects on inventory volatility. The net effect on variability is an empirical question answered by Table 3, which shows increases in inventory investment variances in all four manufacturing and trade categories.
Table 3
The Variance of Inventory Investment Before and After 1980
 Before 1980 After 1980
 Materials and Work in Process 3.76 4.85
 Finished Goods 0.83 1.81
Retail Trade 2.52 8.15
Wholesale Trade 1.69 2.98


We now tie up the package by drawing some implications from improved inventory control for economic stability in the rest of this decade. This stretch is not a long one because of the continued and even greater importance of inventory investment to business cycles.

There is much recent evidence that improved inventory control can lead to greater instability in production. Not long ago, for example, the assembly of Saturn automobiles came to a halt almost immediately when workers struck a key Saturn supplier. The strike stopped shipments of parts that usually arrived just-in-time. But with just-in-time, of course, the risk of not-at-all is always there. When the buffer-stock role of materials inventories is eliminated, production becomes more susceptible to interruptions in supplies.

Given the evidence of improved inventory control, we conclude that goods production is likely to grow at a less stable rate in the years ahead than it would have in the absence of these improvements. Thus, if sales behavior is uneven, our findings suggest that gross domestic product will grow even more unevenly despite lower inventory-sales ratios, because inventories will be more rapidly adjusted to the changes in sales.

In addition and importantly, changes in inventory investment are likely to be associated with sharper initial declines in GDP. This conclusion follows from our findings that the speeds of inventory adjustments have increased. The other side of this coin, of course, is the sharper recoveries in GDP that can be expected to accompany initial surges in sales. Again, faster speeds of inventory adjustment suggest that inventories will follow sales even more closely than they have in the past.

Our findings are, of course, not the last word. We have more research to do ourselves, and we are well aware of how alternative models of inventory investment and alternative assumptions can produce different results. We know also about other inventory simulation studies that reach different conclusions. For example, in a recent paper, Lovell(3) contrasts the dynamic behavior of two kinds of economies: one, a just-in-time (JIT) economy in which firms produce from perfectly timed arrivals of materials; the other, a traditional economy in which firms produce from stocks of materials. His simulations produce greater stability in the JIT economy. In an earlier paper, however, Lovell(4) found that a JIT economy was extremely unstable and subject to economic gridlock -- a condition in which all production ceases for lack of one or more materials.

While we find exercises like those of Lovell instructive, our conclusion remains consistent with the evidence that improved inventory control has, so far at least, added more volatility to the business cycle.

Dan M. Bechter is a Vice President and Stephen Stanley is an Assistant Economist in the Research Department, Federal Reserve Bank of Richmond, VA. This paper is adapted from one presented at the 34th Annual Meeting of the National Association of Business Economists, September 13-16, 1992, Dallas, TX.


1 Dan M. Bechter and Stephen Stanley, "Evidence of Improved Inventory Control," Federal Reserve Bank of Richmond Economic Review, V. 78:1 (Jan/Feb 1992): 3-12.

2 This model was first applied to inventory behavior by Michael C. Lovell, "Manufacturers' Inventories, Sales Expectations, and the Accelerator Principle," Econometrica 29 (July 1961): 293-314.

3 Michael C. Lovell, "Simulating a 100% Just-in-Time economy." International Journal of Production Economics, 26 (1992) 71-78.

4 Michael C. Lovell, "Input-output simulations of inventory fluctuations," paper presented at 5th International Symposium on Inventories, Budapest, Hungary (August 1988).
COPYRIGHT 1993 The National Association for Business Economists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Bechter, Dan M.; Stanley, Stephen
Publication:Business Economics
Date:Jan 1, 1993
Previous Article:Privatization in the former Soviet Union - one year later.
Next Article:Consensus forecasts in planning.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters