Smoothing Out the Business Cycle – Federal Reserve Bank of Chicago

The following publication has been lightly reedited for spelling, grammar, and style to provide better searchability and an improved reading experience. No substantive changes impacting the data, analysis, or conclusions have been made. A PDF of the originally published version is available here.

As cliches go, “The economy has changed significantly in recent years” ranks near the top in economics. The
irony is that economic models—the formal ones that academics use or the rules of thumb that traders and
businessmen employ—are almost always based on the premise that nothing has changed.

The problem is that history is all we have to build empirical models. There are no data on the future. So,
examining how the economy has changed is the best—and only—way to try to discover how the economy’s behavior may
change in the future. Put another way, understanding how the economy has changed can provide insights into how
conventional wisdom will probably be wrong.

The good news

The U.S. labor market, for example, has changed dramatically in the last 35 years. As figure 1 shows, the
percentage of the labor force that is employed in durable-goods manufacturing has declined significantly. In
1955 employment in durable-goods industries was 9.5 million workers, 19% of the total workforce. Today, the
number is 11.5 million, but this is only 11% of the total workforce.

Figure 1. In jobs, manufacturing is down, services up…

Figure 1 is a line graph showing the percent of employment share held by the service, manufacturing, durable manufacturing, and finance industries from 1955 to 1989. Finance has increased its employment share very slightly over this period, from about 5% in 1955 to about 6% in 1989. Manufacturing and durable manufacturing have fallen over this period, from about 34% and 19%, respectively, to about 18% and 11%. The service industry has grown, from an employment share of about 12% in 1955 to about 25% in 1989.

Why is this important? As figure 1 also shows, durable-goods manufacturing is where the big swings in employment
usually happen during recessions. Other industries do not experience such changes.

The industries that have picked up the slack left by the decline in durable-goods manufacturing
employment—services and financial services—are not as sensitive to aggregate business fluctuations, that is,
they are less cyclical. In fact, services seem to display a profound disinterest in the rest of the economy,
growing steadily recession or not. These sectors combined now account for 33.7 million workers, nearly 31% of
the workforce, up from 8.5 million workers and 17% of the workforce in 1955.

It is easy to underestimate the importance of these changes. With a smaller share of the workforce employed in
cyclical industries, recessions should have less impact on employment. This means that personal income and
unemployment will be less affected. This, in turn, reduces the impact of a recession on personal consumption.

Clearly, someone who is still employed will spend more than someone who is not, but the effects go much further
than that. Spending can fall because people are worried about losing their jobs, but people in less cyclical
industries will worry less. This means that across large numbers of people, spending levels will be maintained
at higher levels than in previous periods of economic uncertainty. This is probably part of the explanation of
why the 1987 October stock market crash had so little effect on overall spending. The scare factor had less bite
than most analysts thought.

This relative increase in job security, occurring over the last 35 years, may also help explain why saving rates
have fallen. If you are less worried about a rainy day, you will save less for it. This is reinforced by the
fact that service workers tend to suffer fewer bouts of extended unemployment than durable-goods workers.
Service organizations do not usually generate the extended bouts of unemployment that the closing of a steel
mill or an automobile plant can engender.

These changes have been reinforced by a number of other changes in the U.S. labor market. The increase in the
number of two-income families is a good example. Two-earner families have a more secure income stream. If one
earner is unemployed, the family still has one income left plus unemployment insurance. As a result, faced with
the same level of overall economic uncertainty, they would save less and maintain spending at higher levels than
a single-income family. All these changes mean that spending is likely to be significantly less volatile in
response to a given recessionary impulse than in the past. And, if this is true, then recessions will be
smaller. Such smaller recessions will then feed back into and reinforce the effects described above.

There have also been some changes in the way businesses behave that will have some beneficial effects on the
business cycle—two, in particular. First is the increasing use of part-time and contract workers to meet swing
needs. These workers are better prepared for disruption in employment and as a result are better able to adjust
to temporary unemployment. Business, on the other hand, feels less need to keep these workers employed when
demand falls. This makes it easier to keep inventories in line—a second benefit.

Indeed, it is in the handling of inventories that business has made the biggest changes, in terms of how future
business cycles may look.
Computers have not always produced the expected gains in productivity; however, almost no one disputes the
tremendous strides that they have made possible in inventory control.

“Just in time” has become a kind of business mantra. Inventories as a percent of sales have been remarkably
stable during this business cycle, especially if autos are excluded. This closer management of inventories has
some very positive consequences for the business cycle. If inventories are lean, then downturns, to the extent
they still occur, are likely to be much smaller and shorter.

Historically, much of the dislocation caused by cyclical fluctuations in the economy has been generated by sales
shortfalls without accompanying slowdowns in production. Large inventories of both finished and raw materials
built up quickly. This, in turn, caused orders to fall dramatically, all at once. As a result, widespread
disruptions in the production process followed. Today, production responds much more quickly and smoothly to
falling demand. Less inventory builds up. What does, can be worked off more quickly. As a result, order and
secondary production impacts are smaller.

The maybe good news

However, before the business cycle is given the last rites, a number of other factors need to be looked at.
Personal consumption makes up 66% of GNP, but it has always been the most stable segment. A lot of the business
cycle “action” takes place in business investment. Here, things have also changed, but the impact of these
changes is less clear. From the standpoint of GNP, there has been some change in the structure of the economy
away from cyclical sectors, but it has been modest relative to the changes in shares of employment.

Figure 2 shows the relative share in nominal GNP of various sectors of the U.S. economy using the same scale as
figure 1. Far less change is evident here than in figure 1.

Figure 2. But the change is less severe, measured by output

Figure 2 is a line graph showing the percent of GNP share held by the service, trade, finance, durables, and nondurables industries from 1955 to 1987. Durables and nondurables have decreased over this period, from GNP shares of just under 15% and 20%, respectively, to about 12% and just under 10%. The service industry’s share of GNP has remained relatively stable, at just under 20%. Trade and finance have both increased to about 20% in 1987, from respective starting points in 1955 of about 10% and about 14%.

The basic difference between the employment view and the production view is that the manufacturing sector, in
general, and durable manufacturing, in particular, have experienced solid improvements in labor productivity,
improvement not experienced by other sectors.

The importance of this is difficult to evaluate. If something causes the cyclical industries to reduce output,
it will have approximately the same primary effects on aggregate investment and other business
expenditures that such events have had in the past. But the employment and inventory data seem to suggest that
the secondary effects will be smaller. Thus, the sum impact on the aggregate economy will hinge on the
relative size of the primary and secondary effects. On net, smaller and shorter fluctuations seem likely.

More disturbing is that the distribution of investment expenditures has not changed at all across industries.
Table 1 shows the breakdown on investment by industry and by type of goods purchased. The “Who buys” columns
show almost no change over this period, indicating that investment may not be any less cyclical today than in
the past, except to the extent that the changes discussed above create more stable demand during downturns.

Table 1. Who buys what

Who buys
(Plant and
equipment expenditures, percent of
total by industry)

1970
1988

Manufacturing

Durable goods
20
18

Nondurable goods
20
20

Nonmanufacturing

Mining
2
3

Transportation
8
5

Public utilities
16
11

Commercial
33
43

What
(Private purchases of
producer durable equipment,
percent of total purchases)

1970
1988

Equipment type

Informational
21
32

Industrial
30
23

Transportation
24
21

Other
24
22

Thus, while consumer-based industries may be more insulated from aggregate business fluctuations than in the
past, investment-goods industries may be just as vulnerable to recessions as in the past. Perhaps even more so,
because the dampening elsewhere will make recessions look smaller overall, without necessarily diminishing their
impact on the investment-goods industries.

Another interesting fact is contained in the table. Examination of the “What” columns indicate that while who is
doing the purchasing has not changed, a greater proportion of investment money is now being spent on computers
and telecommunications equipment. The full implications of this are not clear. Much of the past data on the
computer and telecommunications industries are from periods when these items were being purchased as part of
first-time modernization programs. Today, a much larger percentage of purchases are for upgrades of existing
equipment. Upgrades and replacements are more easily postponed and thus subject to greater cyclical pressures.
Thus, it may be that the high-tech sectors will displace autos and steel as the cyclical industries of the
1990s.

The bad news

Perhaps the greatest changes in the last decade have taken place in the financial markets. The cyclical
implications will not be known for decades. Yet one change stands out—the tremendous increase in corporate debt.
This is a new and very big question mark for the economy as a whole. The financial markets are now more exposed
to the risk of corporate default on obligations. Equally important, businesses stand more exposed to the
vagaries of the financial markets.

In light of recent research done at this Bank and elsewhere reaffirming the importance of internal sources of
funds for investment, this increased exposure of industry to the financial markets means that while many events
may have less impact on the aggregate economy now than in the past, events in the financial markets may be far
more important.

Figure 3 shows interest expense as a percent of internal funds available for investment since 1955. In that year
interest-rate expense equaled 14% of internal funds. Thus, a doubling of interest-rate expense would absorb 14%
of internal funds, leaving the firm still able to pursue 86% of the investment activity funded internally.
Today, interest-rate expense equals 82% of internal funds available for investment. If funding costs doubled, an
82% cut in internally funded investment would be needed.

Figure 3. Corporate debt burden zooms

Figure 3 is a line graph showing interest expense as a percent of internal investment funds from 1955 to 1987. This percentage has increased dramatically over the period, from 14% up to 82%.

Conclusions

Changes in the structure of the economy indicate that consumer-led sectors will probably be less sensitive to
business fluctuations than in the past, while investment-goods industries may face little change. This
combination may create a situation whereby some measures, such as unemployment or consumption, the economy
appears reasonably robust, while investment sectors may feel a recession. In fact, some have suggested that this
is exactly the situation the U.S. economy is in today. Further, increased corporate exposure to financial
markets make disruptions in the corporate debt market far more important than in the past.

MMI-Midwest Manufacturing Index: Current expansion

The Midwest Manufacturing Index is a line graph comparing manufacturing activity in the Midwest to activity in the U.S. Midwestern manufacturing activity flattened in October/November, while activity in the nation overall increased.

Manufacturing activity in the Midwest flattened in October and November, after rallying from a mid-year low.
Off a slight 0.1% from October, the November MMI was 1.5% above its most recent low point in July.
Transportation equipment (autos) and primary metals (steel), both about 2% below their mid-year levels,
continued to be a source of weakness, as auto producers struggled to reduce inventories. Most other
durable-goods industries were up in November.

While the Midwest followed a pattern similar to the nation, manufacturing activity nationwide was up a full
percentage point in November. The relative weakness of the Midwest reflects in part the importance of auto
production.

Opinions expressed in this article are those of the author(s) and do not necessarily
reflect
the views of the Federal Reserve Bank of Chicago or the Federal Reserve System.