Section 3: Feature article

View the most recent version.

The role of inventories in recent business cycles

by Philip Cross ¹ 

Introduction

Analysts have long understood and accepted the key role played by inventory movements in the business cycle. Many of the recessions in the decades after World War II were dominated by the “inventory cycle”, as it was called, in which increases and decreases in inventories drove the overall change in output. The pivotal role of inventories was reflected in the inclusion of the ratio of inventories to sales as one component in the composite leading index for Canada.

Statistics Canada has documented elsewhere the long-run decline in the ratio of total inventories to sales, which has fallen in current dollars by more than half from 1.5 in the early 1960s to below 0.7 in recent years. ²  Much of this downward trend is attributable to the better inventory management techniques that accompanied the diffusion of computerized inventory control systems in the 1980s. As a result, inventories played a smaller role in output, which has been one factor in the reduced volatility of GDP in the last three decades.

This paper looks at the role of non-farm business inventories in recent business cycles – to determine whether rising levels of unplanned stock accumulation helped bring on recessions, whether inventory cuts contributed directly to downturns and whether inventory changes helped or hindered recoveries. It does so by looking at inventories from three points of view: changes in their level (up or down); whether the rate of change is accelerating or decelerating; and the ratio of inventory levels to sales. ³  Together, these show how fast firms are responding to cyclical changes in demand. Inventory changes in the 2008-2009 recession are examined in detail, especially their sectoral composition among manufacturers, wholesalers, and retailers.

Inventories and recessions

A quick look at the contribution of inventories to the change in real GDP shows their key role in recessions in the 1970s and 1980s. Inventories by themselves accounted for all of the declines in GDP in the recessions of 1970, 1975 and 1980, and over 80% of the 1981-1982 contraction. This does not necessarily imply that inventories were the determining factor in lower output. Firms cut desired inventory levels as sales slump; the decision by firms to reduce output is made to re-align inventories with the lower level of sales reached during recessions. As noted by researchers at the Federal Reserve Board, “changes in inventory behaviour since the mid-1980s are more likely a consequence of changes in the macroeconomic environment and thus have had a complementary, rather than a leading role, in the decline of aggregate output variability.” ⁴ 

However, inventories have played a less important role in the two most recent recessions. Inventory changes overall had no direct impact on GDP growth during the 1990-1991 recession, while they accounted for slightly under one-third of the drop during the 2008-2009 slump. Research at the Federal Reserve Board found a similar reduction in the part played by inventories in US recessions after 1982. ⁵ 

The ratio of inventories to sales

Inventory increases are not a threat to short-term growth if they are rising at about the same rate as sales: this is normal inventory re-stocking by firms. It is when inventories are misaligned with sales that corrective cuts to output and stocks are warranted. Figure 3.2 shows the ratio of total inventories to sales immediately before, during, and just after the last three recessions starting in 1981, 1990, and 2008. A number of features are notable.

First, the inventory-to-sales ratio was markedly lower in each cycle, with the ratio in 2008 about half its 1981 level. This highlights the long-term downward trend in inventory ratios.

Second, the ratio did not increase appreciably before the recessions began. On the eve of the latest recession, the inventory-to-sales ratio equalled its all-time low of 0.61 in the third quarter of 2008. The ratio edged up from 0.95 to 0.96 in the quarters before the 1990 recession, while it fell from 1.18 to 1.17 in the two quarters before the 1981 downturn (the latter drop would have been larger if the period were extended back to the high of 1.22 reached during the 1980 recession). Output cuts were not triggered by the involuntary accumulation of unwanted stocks, the traditional definition of the onset of an inventory cycle.

Inventories rose slightly relative to sales early in the three latest recessions as sales fell more than inventories. The largest increase in the inventory-to-sales ratio occurred in 2008-2009, when it rose 0.11 points in three quarters. This reflects the extent of the unanticipated collapse of demand in key sectors such as exports, housing, auto sales and business investment as the global financial crisis unfolded. ⁶  By comparison, the inventory-to-sales ratio rose only 0.05 points in the first three quarters of the 1981 recession, which before 2008 was the recession that saw the sharpest slump in demand on record. Because inventory levels were much lower in 2008, the 0.11 point increase was proportionally much higher than the increase recorded in 1982 (17.1% versus 4.6%).

In 1990-1991, the inventory-to-sales ratio did not rise at all, with the exception of a blip in the first quarter of 1991, when the economy was hit by the twin shocks of the Gulf War and the introduction of the GST. This reflects that it was lower sales and not inventories that drove output down during that recession.

It is noteworthy that the inventory-to-sales ratio rose more in 2008-2009 than in the previous two recessions. This suggests firms in Canada overall did not experience unusual difficulties in financing inventories, unlike the disruption elsewhere in the world (especially for trade credit).

The first few quarters of recovery invariably saw rapid declines in the inventory-to-sales ratio, usually to record low levels. In the first four quarters of 1983, the ratio fell 0.16 points, from 1.17 to a then-record low of 1.01; in 1991, 0.08 points, from 0.95 to another record low of 0.87; and so far in the three quarters of recovery starting in 2009 it fell 0.07 points, from 0.72 to 0.65. So while the inventory-to-sales ratio does not rise significantly before or during recessions, firms often use recoveries to push the ratio lower. This very low level of inventories in recoveries helps ensure that growth will be sustained enough to justify ramping up production and hiring.

Outside of outright recessions, the economy has seen three periods of near-recession: in 1986, 1994-1995, and 2001. In all three, real GDP stalled or fell slightly, and the year-over-year growth of GDP fell below 1%. While these three periods were not severe or long enough to technically qualify as recessions, it is worth studying the behaviour of the ratio of inventories to sales during these periods.

In all three slowdowns, the economy-wide ratio of stocks to sales rose by much less than 0.1 points. It edged up by only 0.02 points in 1986 and 2001 and by 0.05 points in 1995. This tight control on inventories helps to explain why in each instance the economy escaped an outright recession. ⁷ 

Inventory levels

Figure 3.3 shows the absolute change in constant dollar inventory levels at annual rates before, during and after the last three recessions. Before each of the last three recessions, the volume of inventories always rose, but the rate of inventory accumulation slowed just before the downturn. In the three quarters before each of the recessions began, inventory accumulation slowed to less than $10 billion at annual rates (a larger increase of over $20 billion occurred in the third quarter of 2007, when the global financial crisis first began to take hold).

In other words, firms began to rein in inventory growth as the downturn in sales neared, rather than accelerate inventory accumulation as they had done in the run-up to recessions in the 1970s. Before the 1970 recession, inventory accumulation doubled to $2.9 billion in 1969. It also doubled to $9.1 billion in 1974, before real GDP fell early in 1975. And inventory accumulation accelerated sharply to $8.6 billion in 1979 before the 1980 downturn.

In each of the last three recessions, the onset of the downturn was reflected immediately in a change in inventory levels held by firms. Despite the drop in sales associated with recession, inventory accumulation stopped or turned negative. ⁸  However, this initial drop in stocks was not as large as sales, and the ratio of inventories to sales often rose slightly, as discussed in the previous section.

Thereafter, firms quickly moved to slash output more than sales, and inventory volumes fell rapidly. The steepest and longest cuts to inventories were made during and after the 1981-1982 recession, when they fell $21 billion over eight consecutive quarters. This was followed by 1990-1993, when inventories fell $22 billion over a record thirteen straight quarters of decline. Inventory cuts were relatively mild and short in the most recent recession, totalling $5.2 billion (or less than their drop during the 2001 slowdown), despite the much higher level of sales than in the 1980s or 1990s.

Inventory levels also usually lag the recovery from recession: that is, they continue to decline even as output begins to recover. This is shown in the dotted lines in Figure 3.3; inventory changes remained negative even after output began to recover in all three cycles, although the rate of decrease slowed markedly after one year (as explained in the appendix, a slower rate of inventory decline contributes to positive GDP growth). Inventories continued to fall in the first three quarters of 1983, long after the recession had ended (in the fourth quarter of 1982). Real GDP stopped falling in the second quarter of 1991, but inventories continued to decline until the fourth quarter of 1993, one reason the recovery in 1992-93 was so sluggish. In 2009, inventories fell in all four quarters of the year, while real GDP turned up in the third and fourth quarters.

Firms apparently use the initial upturn in demand to finish restoring the inventory-to-sales ratio to satisfactory levels. Firms usually begin to raise output once the ratio of inventories to sales have returned to about its pre-recession level. However, the initial boost to output does not match the recovery in sales, and inventories usually fall both in absolute terms and relative to sales early in a recovery. An exception was 2009, when the overall ratio of inventory to sales lagged the recovery as a result of problems specific to the primary sector and manufacturing, which are discussed in the next section.

Sectoral patterns

The small upturn in inventories as the recession began in the fourth quarter of 2008 was limited to manufacturing and mining. Wholesale and retail inventory levels fell steadily right through the recession, a remarkable achievement in view of sales declines of 16% and 4%, respectively, during the winter of 2008-2009 (retail motor vehicle stocks did edge up 1.4% in the fourth quarter of 2008). Wholesalers and retailers began to rebuild inventories in the first quarter of 2010, but manufacturers have continued to cut back. ⁹ 

The sectoral pattern of nominal stock-to-sales ratios in 2008-2009 sheds some light on how this occurred. The sudden slump in demand saw the inventory-to-sales ratio increase in each of wholesaling, manufacturing, and retailing. Both wholesalers and retailers were quickly able to stabilize inventories and then reduce their inventory ratios during the winter of 2009, partly as a result of housing and retail sales rebounding quickly. By early in 2010, the ratio of inventories to sales had reached a record low in retail and wholesale trade.

The inventory cycle was more pronounced for manufacturers in the 2008-2009 recession. Their inventory-to-sales ratio rose a full 0.11 points, versus increases of 0.02 points for retailers and 0.06 points for wholesalers. The ratio for manufacturers peaked at 0.60 in the second quarter of 2009, one quarter after the high-water mark of the trade sectors, and early in 2010 the ratio was still well above its pre-recession level. Almost all of the increase was due to the sharp drop in manufacturing sales, as inventories rose only 2.4% from their low early in 2007.

Inventories held by manufacturers, wholesalers, and retailers account for over 90% of all non-farm business inventories. The remainder are dominated by inventories in the primary sector, mostly mining and utilities (including natural gas). This sector had the most trouble controlling inventories during the 2008-2009 recession. ¹⁰  Inventories rose 14.6% in volume between the third quarter of 2008 and the third quarter of 2009, much of it stockpiles of natural gas. This reflects the severity of the drop in demand and prices during the recession, and the efforts of producers to withhold supply from the market in order to avoid aggravating already weak prices.

This sectoral pattern of changes in the inventory ratio was different from that of the early 1990s (sectoral data are not available before 1989). Then, all three major sectors saw a gradual increase in their respective ratios of inventories to sales, which spiked in the first quarter of 1991 when sales plunged under the combined weight of the Gulf War and the introduction of the GST. Wholesalers were the first to rapidly bring down their ratio of inventories to sales, followed quickly by manufacturers. Retailers struggled to rein in inventories as sales languished throughout 1991, and early in 1992 their stock-to-sales ratio was higher than it had ever been in 1990 or 1991.

Conclusion

Besides the long-term trend to lower inventories, the role of non-farm inventories in the business cycle has changed over time. Inventories no longer rise much more than sales in the run-up to a recession, as new technology and management techniques have allowed firms to better control inventories. Instead, inventories in recent recessions have moved slightly behind cyclical changes in total sales, rising soon after a downturn in output and falling the most early in a recovery. Instead, it is the course of sales that has determined cycles in output in recent decades. More than a decade ago, some researchers suggested “the history of inventory-induced recessions has probably come to an end.” ¹¹  The absence of inventories as a major contributing factor to either the 2001 slowdown or the 2008-2009 recession in Canada supports this conclusion.

Technical appendix

While the role of inventories in the business cycle is well known, inventories are the least understood part of GDP. This is because of the complexity involved in their calculation and their interpretation.

Most components of the National Accounts are calculated in a straightforward manner—taking an easily understood series such as business sales and deflating this nominal series by the appropriate price index to derive the constant dollar value. However, inventories are different. The book value reported by firms requires two steps to derive the change in the current dollar value. The change in inventory values due to price changes is removed to get constant dollar inventories. What remains is then revalued using prices depending on the turnover period of stocks to derive the current dollar value of physical change of inventories.

The interpretation of inventories also requires an extra step, compared with the other components of GDP. For all other components of GDP, the first difference (or the change) in spending shows its impact on overall GDP: if consumer spending rises, unequivocally this will help raise total GDP. However, for inventories, it is the second difference (or the change in the change in inventory levels) that captures its impact on GDP.

Figure 3.7 shows why this is the case. For this exposition, we have created a hypothetical scenario where GDP excluding inventories is the same in periods 1 and 2 at $100 billion. In period 1, inventories rose by $10 billion and so total GDP (GDP excluding inventories plus the value of physical change in inventories) is $110 billion. In period 2, inventories still rise, but only by $5 billion, and therefore GDP falls to $105 billion. Even though inventories rose in both periods, their rate of growth slowed in the second period, lowering overall GDP. Consequently, inventories can reduce GDP even when they do not decline in level, the only component of GDP for which this is true.

It follows that an acceleration in inventory accumulation between two adjacent quarters (or years) will boost GDP. If inventories swing from an increase to a decrease, they will depress growth. If they shift from declines to increases, the change in the change of inventories has to be positive, which serves to raise GDP. When inventories are falling, the “change in the change” rule implies that GDP growth will be lower when the drop in inventories is steepening, while growth will be boosted when the rate of decline slows.

Figure 3.8 shows what happens when inventory cuts slow between periods 1 and 2 from -$10 billion to -$5 billion. With GDP excluding inventories steady at $100 billion in both periods, GDP is $90 billion in period 1 and $95 billion in period 2. So even though inventories are still falling, their lower rate of decline lifts GDP in period 2.

The rate of change of inventories shows only the contribution of inventories in one quarter compared with that in the preceding quarter. When analyzing more than one quarter, one must cumulate the changes in inventories in order to see their impact on GDP. This is different from other series: between the first quarters of 2009 and 2010, personal expenditure rose $45.2 billion (from $885.1 billion to $930.3 billion). This reflects the contribution to nominal GDP growth over that period. However, for inventories, it is the cumulative change over each of these four quarters that matters in measuring their impact on GDP. For example, since reaching a low early in 2009, nominal inventories have moved from an annual rate of decline of $7.5 billion in the second quarter of 2009 to a drop of $0.7 billion in the first quarter of 2010. Some may mistakenly calculate this as an upward swing of $6.8 billion in the contribution of inventories to GDP growth. However, doing so ignores inventory declines in the third and fourth quarters of $4.9 billion and $10.6 billion, respectively. The cumulative rate of change over the four quarters actually was a positive $9.1 billion, which reflects its total impact on nominal GDP growth over the four quarters of recovery.

Next | Previous