How the U.S. economy resembles a (very) big business.
Campbell, Jeffrey R.
Introduction and summary
This article offers a perspective on analyzing the growth of the
U.S. economy by treating the economy as a very large firm. A
well-functioning economy maximizes households' well-being rather
than firms' profits, so policymakers' objectives and
motivations are not as clear-cut as those of company chief executives.
Still, as in any business, identifying areas of weakness and relative
strength in the economy is inherently valuable in guiding
decision-making.
I present basic tools for measuring different business lines'
contributions to the U.S. economy's growth. Then, I extend the
economy-as-business analogy by using the same tools to measure the
exposure of a large conglomerate to macroeconomic risks. While these
tools are often used to evaluate the strengths and weaknesses of the
economy with the goal of recommending appropriate monetary policy, they
can also he used by profit-maximizing firm managers to better understand
macroeconomic risks to firm performance.
If we consider the U.S. economy to be a large enterprise, how do we
measure its performance? First, what does our company look like? This
very fictional national firm employs all of the workers in the U.S.
economy; owns all machinery, structures, and other productive assets;
and returns its profits to its shareholders (the American public). The
national firm also makes machinery, structures, and materials for its
own account to add to its productive capacity. The national firm has two
customers: the national family (to which every U.S. resident belongs)
and a conglomerate government that encompasses local, state, and federal
governments.
The following two key macroeconomic concepts allow assessment of a
particular sector's contribution to overall economic growth as well
as its sustainability: the fundamental national product accounting
identity and the contributions to growth formula.
These concepts can provide similar insights into a firm's
performance, capturing the contribution of a particular product line (or
group of product lines) to the firm's growth and the likely
sustainability of that contribution.
Applying these concepts to the U.S. economy reveals that
macroeconomic risks arise primarily from sectors such as nonresidential
fixed investment (business investment) that change substantially from
quarter to quarter and also account for a moderately large fraction of
economic activity. Sectors of the economy that are responsible for a
large fraction of national income, such as expenditures on nondurable
goods and services, and whose growth changes relatively little from
quarter to quarter represent small risks to overall economic activity.
Other sectors, such as new home construction, with unstable but
relatively small sales also represent small risks to growth. With these
results in hand, I can assess an individual firm's growth, and the
macroeconomic risks to it, by measuring what fraction of the firm's
sales corresponds to particular sectors of the overall economy.
After developing this methodology, I go on to apply it. The first
application is to a fictional hair salon that has exposure to only one
sector of the U.S. economy, personal consumption expenditures on
services. The second application is to a real conglomerate, General
Electric Company (GE). This application makes use of publicly available
data and is purely for illustrative purposes. A serious evaluation of
any particular company would require much more data than employed here.
In the next section, I present basic concepts from national income
and product accounting, which divides the U.S. economy's production
into different business lines. (Readers already familiar with the
definitions of gross domestic product, or GDP, and its major components
might wish to skip this section.) in the following section, I develop
the contributions to growth formula and use it to understand business
cycle risks to the U.S. economy. Then I develop and analyze
macroeconomic benchmarks for the fictional hair salon and for General
Electric Company.
National income and product accounting
Macroeconomic policy requires quantifying the economic benefits
accruing to the nation's residents over a given interval of time.
In the United States, the U.S. Bureau of Economic Analysis (BEA)
provides one set of such measures with its national income and product
accounts (NIPA) data. These measure the value of market-based
transactions for goods and services produced during the time of
interest. For business cycle analysis, the time of interest is typically
a calendar quarter (January through March, April through June, and so
on). The fundamental national product accounting identity decomposes the
total value of goods and services produced in the nation into distinct
expenditure components. The NIPA data report these for each calendar
quarter.
Figure 1 is taken from a BEA spreadsheet containing the NIPA data.
This is the equivalent of a quarterly sales report for a very large
firm. The top line of data gives the quarterly history of GDP, which is
defined to equal
* The value of all final goods purchased by households and
governments, plus
* The value of all capital machinery and structures purchased by
producers, plus
* The value of goods added to inventories less the value of goods
sold from inventories, plus
* The value of exports minus imports.
With one exception (detailed later), these purchases are all market
transactions. All sales from businesses to households or governments
contribute to GDP, but this definition excludes business-to-business
transactions unless the recipient uses the purchase to augment
productive capital or inventories. This exclusion rule ensures that no
firm's subcontracting decisions (make or buy) have a direct impact
on GDP.
The spreadsheet's remaining lines (in figure 1) report the
expenditure components of GDP. There are four major components, personal
consumption expenditures (line 2), gross private domestic investment
(line 6), net exports of goods and services (line 13), and government
consumption expenditures and gross investment (line 20). The spreadsheet
reports each major component's constituent minor components beneath
it. The definitions of these components and their relationships with
each other can be best understood by examining the national product
accounting identity, which expresses gross domestic product as the sum
of these components. The national product accounting identity comes from
the income statements of the three very hypothetical institutions
mentioned previously--an extended national family to which every U.S.
resident belongs; a national conglomerate firm owned by this family that
is responsible for the production of all goods and services exchanged in
the market; and a conglomerate government that combines federal, state,
and local governments.
The national family
I represent the national family's income for quarter t with
[V.sub.t]. The BEA divides the national family's uses of after tax
income ([V.sub.t] - [T.sub.t]) into personal consumption expenditures
([C.sub.t]) and private savings ([S.sub.t]).
1) [C.sub.t] + [S.sub.t] = [V.sub.t] - [T.sub.t].
Setting C above [V.sub.t] - [T.sub.t] requires the national family
to set [S.sub.t] < 0. That is, consumption in excess of current
inflows requires spending from assets or going into debt.
The BEA further subdivides personal consumption expenditures into
three categories that are somewhat ambiguous but nevertheless useful:
expenditures on services, nondurable goods, and durable goods. That is,
2) [C.sub.t] = [C.sup.S.sub.t] + [C.sup.N.sub.t] + [C.sup.D.sub.t]
Lines 3, 4, and 5 of the spreadsheet in figure 1 report these three
categories. Examples of services are hotel room rentals, movie theater
admissions, and haircuts. Services also include rent paid for the
occupation of residences. The BEA adds to this the implied rent paid by
homeowners to themselves, so a family's choice between
homeownership and renting has no impact on [C.sup.S1.sub.t]. Food, fuel,
and any other goods expected to last less than three years are
nondurable goods; all other goods, except for housing, are durable.
Automobiles and furniture are the most important examples. Durable goods
purchases resemble saving because their ownership enhances the
family's well-being currently and in the future. However, the
BEA's conventions expense them just like purchases that have no
persistent impact on the family.
[FIGURE 1 OMITTED]
All income not spent is, by definition, saved. An individual family
can save by depositing funds in a bank, by purchasing a house (new or
pre-existing), by acquiring stocks and other publicly traded securities,
or by directly lending to another household. Whether these actions
contribute to national saving depends on whether another individual
family's decisions directly offset them. When one family purchases
a pre-existing home from another family, the selling family's asset
reduction offsets the buying family's asset accumulation.
Similarly, one family's mortgage borrowing offsets the lending
family's saving. In both cases, the net contribution to the
national family's saving is zero.
In contrast, a family's purchase of a newly built home does
contribute to the national household's savings because the
transaction's counterparty (the construction firm) is not part of
the national family. Similarly, foreign-financed mortgage borrowing
reduces national saving.
The fact that many of the transactions by which individual
households save are offset by other households' reduced saving must
be kept in mind when considering how the national household can save.
Aside from the purchase of new durable goods, the national family has
four means of saving from its current income to improve its future:
purchasing a new home or improving an existing one, investing in the
conglomerate firm, investing abroad, or purchasing any available
conglomerate government debt. The BEA calls the first vehicle
residential investment. This directly enhances the economy's
productive capacity by augmenting the stock of residential structures.
The effects of investing in the conglomerate firm can be more subtle
because one family purchasing the firm's stock from another family
makes no contribution to national saving. The national family only saves
when purchasing the conglomerate firm's securities at a public
offering.
International investing is intimately entangled with international
trade. Countries trade goods and services to pursue productive
efficiency through comparative advantage and to consume goods and
services that may not be available domestically. Let [X.sub.t], and
[M.sub.t] stand for the values of exports and imports. When [X.sub.t]
exceeds [M.sub.t], we say that the nation runs a trade surplus. In this
case, the national family is saving by extending credit to foreigners in
return for [X.sub.t] - [M.sub.t], the exports that foreigners did not
pay for with an offsetting import. Conversely, if [X.sub.t] falls short
of [M.sub.t], then the country runs a trade deficit. The national family
must cover this either by redeeming previously accumulated IOUs from
foreigners or by issuing new IOUs of its own. Either way, the resulting
reduction in wealth equals [M.sub.t] - [X.sub.t].
The national family's final saving vehicle is the purchase of
bonds issued by the conglomerate government. The government can use the
proceeds of a household's purchase to either repurchase bonds held
by another household (so that the national family's holdings of
government debt remain unchanged), or the government can use the
proceeds to undertake current expenditures. Suppose (counterfactually)
that only American households hold the conglomerate government's
debts and that all such debts are bonds that mature in one quarter. (2)
If [B.sub.t-1] is the face value of bonds purchased by the national
household in the previous quarter, [B.sub.t] is the face value of bonds
purchased in the current quarter, and [R.sub.t] is the interest rate on
these bonds, then the national household's net investment in
government debts equals [B.sub.t]/[R.sub.t] - [B.sub.t-1].
Bringing these four savings vehicles for the national household
together allows us to write national private savings as
3) [S.sub.t] = [I.sup.R.sub.t] + [Q.sub.t] + [X.sub.t] - [M.sub.t]
+ [B.sub.t]/[R.sub.t] - [B.sub.t-1].
Here, [I.sup.R.sub.t] and [Q.sub.t] represent residential
investment and the national family's net equity purchases. An
individual household's total saving could have other contributions
such as the accumulation of other households' debts. However, these
all cancel when adding all households' savings to arrive at the
national family's saving. Equation 3 only contains those
contributions that do not cancel and represent true national saving.
The conglomerate government
Next, consider the conglomerate government, which collects taxes
[T.sub.t] from the national family and combines these with the net
proceeds from the sale of government debt to pay for its current
expenditures. Economists divide these into two categories, transfers and
purchases. A transfer is the granting of funds to an individual with
limited restrictions on their use. The federal government's largest
transfer programs are Social Security, Medicare, and Medicaid. A
government purchase is an exchange of funds for a specifically
contracted good or service. For example, the salaries of government
employees and the purchase of weapons systems are government purchases.
Use [A.sub.t] and [G.sub.t] to represent the value of transfers and
purchases by all governments so that the requirement that
governments' uses of funds equal their sources of funds can be
written as
4) [A.sub.t] + [G.sub.t] + [B.sub.t-1] = [T.sub.t] +
[B.sub.t]/[R.sub.t].
Just as with the national family, transfers from one government to
another (for example, federal grants to states) do not count toward
[T.sub.t]. When the conglomerate government's choices of [A.sub.t],
[G.sub.t], and [T.sub.t] require [B.sub.t] to exceed [B.sub.t-1], we say
that the government is running a deficit.
The national firm
The final institution to consider is the national firm, which
produces all goods and services in the economy. (3) Its assets equal all
productive machinery and structures in the country, along with the
inventories of completed goods and any work in progress, and its sole
liability is its equity. To produce, it employs members of the national
family to operate its productive machinery and maintain its structures.
Any single business seeking to purchase machinery, structures, or
materials faces a make-or-buy decision. By definition there is no other
(domestic) firm to sell anything to the national firm, so it must
fulfill all of its materials and capital needs with its own production.
The national firm's funds come from the sale of goods and
services and from issuing new equity. Let [Y.sub.t] equal the value of
all goods and services produced (both for external customers and the
national firm's own account), and use [K.sup.1.sub.t] to represent
the value of that production in inventory on the last day of quarter t.
Goods in inventory sometimes lose their value. For example, food can
spoil. For this reason, inventories held from quarter t - 1 to quarter t
lose a fraction of their value. We call this the inventory loss rate and
denote it with [l.sub.t]. The national firm's sales combine
receipts from sales of goods produced within the quarter with sales from
goods sold out of inventory, [Y.sub.t] - [K.sup.l.sub.t] +
(1-[l.sub.t])[K.sup.l.sub.t-1].
The national firm's funds from equity issuance are equal to
the national family's net purchases of equity discussed previously,
Q t. Nothing prevents the national firm from repurchasing its shares, in
which case [Q.sub.t] < 0. Assume for simplicity that the national
firm returns all profits to its shareholders through such share
repurchases so that this use of funds is represented as a negative
source. The national firm's other two uses of funds are the
purchase of machinery and structures on its own account,
[I.sup.B.sub.t], and payment of its wage hill, [W.sub.t][N.sub.t]. Here,
[N.sub.t] is the number of hours worked by its employees and W is their
average hourly wage. The national firm's sources and uses of funds
must equal each other, so
5) [Y.sub.t] - [K.sup.l.sub.t] + (1-[l.sub.t])[K.sup.l.sub.t-1] +
[Q.sub.t] = [W.sub.t][N.sub.t] + [I.sup.B.sub.t].
The BEA calls [I.sup.B.sub.t] business fixed investment and calls
[K.sup.l.sub.t] -(1-[l.sub.t])[K.sup.l.sub.t-1] [equivalent to]
[I.sup.S.sub.t] net inventory accumulation. Lines 7 and 12 of the
spreadsheet in figure 1 (p. 31) report these. They both contribute to
the economy's future productive capacity.
The national product accounting identity.
Profits from the national firm that are not reinvested enter the
national family's budget as a negative value for [Q.sub.t].
Similarly, the redemption of foreigners' IOUs enters the national
family's budget as a negative value for [X.sub.t] - [M.sub.t].
Thus, the correct measure of income in equation 1 sums the other two
sources of income for the national family, labor income (which must
equal the national firm's wage bill) and transfers from the
conglomerate government:
6) [V.sub.t] = [W.sub.t][N.sub.t] + [A.sub.t].
The pieces required to assemble the fundamental national product
accounting identity are now in place. To do so, use equations 2, 3, 4,
and 6 to replace [C.sub.t], [S.sub.t], [T.sub.t] and [V.sub.t] in
equation 1. Then, eliminate [W.sub.t] [N.sub.t] from the resulting
equation using the expression for the national firm's profit in
equation 5. Canceling terms that appear on both sides of the equation
and isolating [Y.sub.t] on the right-hand side yields the desired
identity.
7) [C.sup.N.sub.t] + [C.sup.S.sub.t] + [C.sup.D.sub.t] +
[I.sup.R.sub.t] + [I.sup.B.sub.t] + [X.sub.t] - [M.sub.t] +
([K.sup.l.sub.t] - [l.sub.t] [K.sup.l.sub.t-1]) + [G.sub.t] = [Y.sub.t].
The right-hand side equals the value of the national firm's
output, gross domestic product. The left-hand side sums the values of
its distinct uses: the three consumption expenditures, residential
investment, business fixed investment, net exports, and inventory
accumulation. The BEA creates the expenditure side of the NIPA data by
statistically estimating each item in equation 7 separately. The
BEA's estimate of GDP equals this sum.
There are three features of equation 7 worth noting. First, the
national family's receipt of transfers from the government appears
nowhere because transfers appear in both [V.sub.t] and [T.sub.t]. That
is, increasing government transfer programs makes no direct contribution
to national income. (Keynesian macroeconomic theories assert that such
transfers raise income indirectly by reducing savings and encouraging
present consumption. Whether this is indeed the case is the subject of
much ongoing research.) Second, the national family's investments
in government bonds contribute to both [S.sub.t] and [T.sub.t] with
opposite signs, so they also cancel in equation 7. This is because G
represents the conglomerate government's use of national income.
Raising the national family's investments in government bonds
merely allows the conglomerate government to lower current taxes. This
implies that a temporary tax decrease financed with government debt has
no direct impact on national income. (4) Finally, note that the national
family's purchases of the national firm's stock ([Q.sub.t])
contribute nothing to total national saving because the corresponding
increases in the national firm's liabilities cancels them. However,
the national firm's tangible investments ([I.sup.B.sub.t]) remain.
National income shares
A contributor to macroeconomic policy requires familiarity with how
the national family earns and spends its income. To help build this,
table 1 reports the average values for the national product accounting
identity's components relative to GDP decade by decade from the end
of the Korean War (1953:Q4) through the most recent data released by the
BEA (2008:Q1). Personal consumption expenditures represent the majority
of GDP in all decades, and this share has climbed. The share of all
personal consumption expenditures combined equaled about 62 percent
through the 1950s, 1960s, and 1970s. In the 1980s, it climbed slightly
to 64 percent, and this climb continued through the 1990s. For the
current decade, personal consumption expenditures account for about 70
percent of GDP. Nondurable goods' expenditure share fell steadily
from the 1950s through the 1990s (from 30.2 percent to 20.3 percent) and
has since leveled off, while purchases of durable goods have remained
constant at about 8.5 percent. This implies that the well-documented
growth of consumer services has accounted for more than all of the
growth in personal consumption expenditures. Its share of GDP
expenditures has nearly doubled from the 1950s through the current
decade.
The national family directly augments domestically sited productive
capacity with its residential, nonresidential, and inventory
investments. Over the sample period (1953:Q4-2008:Q1), residential
investment has accounted for 4.8 percent of GDP. The post-war boom in
the 1950s drove this to 5.3 percent. This share fell off during the
1960s, rose again in the 1970s to nearly 5 percent, and then fell during
the 1980s and 1990s. For the current decade, it nearly equals its value
for the 1950s. In this sense, residential investment is currently of
unusual importance for the macroeconomy. Nonresidential investment
represented 10.7 percent of GDP over the sample period. This share
climbed from the 1950s through the 1980s (from 9.5 to 12.1 percent) and
then fell to its approximate average level for the 1990s and the current
decade. Unlike most other expenditure shares, it is possible for
investment in inventories to be negative if inventories are sold faster
than they can be replenished. However, the average inventory investment
over the sample period was positive at 0.5 percent of GDP. This share
varied considerably from decade to decade but always remained at or
below 1 percent.
The final two expenditure shares are net exports and government
purchases. It is well known that the United States ran small trade
surpluses in the immediate post-Korean War decades, which then turned to
somewhat larger trade deficits. The current trade deficit (4.8 percent
of GDP) is a distinctive feature of the present macroeconomic situation.
Government purchases represented 21.8 percent of GDP in the 1950s, and
this grew slightly in the 1960s. Thereafter this share fell in two
distinct steps, from the 1960s to the 1970s and from the 1980s to the
1990s. For the current decade, government purchases' share of GDP
equals 18.8 percent.
Some growth accounting
The national product accounting identity paves the way toward an
accounting for the national firm's growth. The BEA spreadsheet in
figure 1 (p. 31) displays the values of product categories in dollars. A
dollar is only worth what it will buy, and inflation has diminished that
purchasing power on and off over the entire post-Korean War period.
Macroeconomic policymakers account for this by examining
inflation-adjusted GDP. Its construction begins with a measure of the
dollar's purchasing power, [P.sub.t]. Dividing [Y.sub.t]) by this
yields real GDP; that is, [y.sub.t] = [Y.sub.t]/[P.sub.t]. The BEA
produces several measures of [P.sub.t]. By construction, these equal one
on average over the year 2000. The corresponding real GDP is measured in
year 2000 dollars. Measuring [P.sub.t] is not straightforward because
there are literally millions of transaction prices. If inflation made
these all move in lock step with each other, then [P.sub.t] could be
constructed using the growth rate of any one of them. In fact prices do
not move together nearly so perfectly. Properly measuring [P.sub.t] in
this case requires accounting for each transaction's share of
expenditures and for the changes in expenditure shares these price
changes induce. A complete description of how the BEA constructs
[P.sub.t] is beyond the scope of this article, but appendix A gives an
overview of the procedure.
Real GDP measures the national firm's growth. For each decade
from the 1950s through the present, figure 2 plots its value relative to
the first observed value in that decade. Clearly, economic growth
dominates these observations. Although the economy contracted twice in
the 1950s, first immediately after the Korean War and again in late 1957
and early 1958, real GDP at the end of the decade was 20 percent higher
than at the close of the war (as shown by the difference between 1.20
and 1.00 in panel A). The 1960s also started with a mild contraction,
but growth equaled a spectacular 50 percent for the decade. The 1970s
and early 1980s saw substantially greater ups and downs in real GDP, but
growth for both decades was about 35 percent. Real GDP contracted at the
beginning of the 1990s, but growth for the decade equaled that for the
1970s and 1980s. The beginning of the current decade also got off to a
slow start, but growth thus far equals 21 percent. This essentially
equals the growth performance in the post-Korean War 1950s.
The quarter-to-quarter fluctuations in the pace of economic growth
make up the business cycle. The national family's long-run
well-being depends most of all on economic growth. Because growth rates
compound, even a slightly higher long-run growth rate can dramatically
improve the national family's welfare. The observed business cycle
fluctuations would be a small price to pay for that, so one might
conclude that the business cycle has little importance: Macroeconomic
policy should focus on maintaining a high level of growth and let the
business cycle take care of itself. Unsurprisingly, this view has little
currency among policymakers because it presumes that economic growth can
be separated from the business cycle. The frequency with which severe
contractions of GDP result in decades of stagnant or negative growth, as
in the United States during the 1930s or Japan during the 1990s,
suggests that this conclusion is overly hopeful, if we suppose instead
that the economy's business cycles sometimes impact its growth
performance, then business cycle policy in those times is growth policy.
A macroeconomic policymaker seeking to understand the business
cycle needs to know how the national firm's various product lines
contribute to it. For this, the contributions to growth formula is
helpful. This writes the growth rate of GDP as contributions from the
growth of the components on the left-hand side of the national product
accounting identity (equation 7). If there were no inflation, we could
derive this formula by dividing both sides of the identity by
[Y.sub.t-1], subtracting one from both sides, and rearranging:
8)[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].
The left-hand side has only GDP growth, while the right-hand side
sums the growth rates of each GDP component multiplied by its share of
expenditures in the previous quarter. Each of these products is the
contribution of an expenditure to GDP growth. The formula shows that the
influence of a given expenditure on overall GDP growth involves both its
own movement and its share of GDP. If two expenditures have the same
percentage reduction, then the one with the smaller expenditure share
will lower GDP less. Of course, abstracting from inflation sets a lot of
macroeconomic reality to the side. Accounting for inflation changes the
contributions to growth formula in two ways: The expenditures'
growth rates are adjusted using expenditure-specific measures of
inflation, as well as slightly modified expenditure shares. The basic
insights of equation 8 remain: An expenditure's contribution to
growth equals the product of its percentage growth rate and its
expenditure share. (Appendix B presents the derivation of the
contributions to growth formula that accounts for inflation.)
Applying the contributions to growth formula to the post-Korean War
NIPA data yields figure 3. Each of its eight panels (one for each
component of GDP) plots the contribution to GDP growth from one
expenditure component. The shaded areas mark recessions as defined by
the National Bureau of Economic Research's (NBER) business cycle
dating committee. These are periods of sustained GDP contraction. The
most recent one began in March 2001 and ended in November 2001.
The first impression from examining figure 3 is that none of the
expenditure components present particularly large risks to GDP growth
individually except for inventory investment. Its contribution to growth
can move wildly from quarter to quarter, but it does not completely
account for fluctuations in GDP growth by itself. The other expenditure
components' contributions move much less over time. The
contributions from personal consumption expenditures on nondurable goods
and on services both evolve smoothly, and government purchases'
contribution appears to be similarly stable. The contribution from
personal consumption expenditures on durable goods moves more (as
expected because the national household can easily delay these
purchases), but it also does not come close to replicating GDP over any
sustained period. The same is true for residential investment, business
fixed investment, and net exports.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
If no particular sector's contributions pose a great risk to
GDP growth individually, then GDP fluctuations must arise from
correlated small risks. That is, even if any given sector can contribute
at most I percentage point to GDP growth, four such sectors all
achieving this maximum rate simultaneously can raise GDP growth by 4
percentage points. This is exactly what happened during the second
quarter of 1978, when annualized GDP growth achieved its maximum
observed value of 15.45 percentage points. In that quarter, business
fixed investment's contribution also hit its maximum value in the
sample, 3.89 percentage points. The contributions from other spending on
the other two capital goods categories (personal consumption
expenditures on durable goods and residential investment) were also
substantially above their averages in that quarter (2.87 percentage
points and 1.08 percentage points versus averages of 0.45 percentage
points and 0.12 percentage points).
Figures 4 and 5 reinforce and refine the point that risks common to
expenditure categories drive most changes of GDP growth. Figure 4 sums
the three capital goods categories' contributions to growth, and
figure 5 sums the contributions of personal consumption expenditures on
nondurable goods and services. Both figures also plot real GDP growth.
Taken together, the three capital goods categories contribute
substantially to changes in GDP growth. For example, in 1980:Q2 their
contribution to growth equaled -10.2 percentage points, while the actual
GDP growth rate equaled -8 percentage points. It is interesting to note
that the contribution of personal consumption expenditures on nondurable
goods and services hit its minimum value (-2.0 percentage points) in the
same quarter; so, there are apparently risks that hit all expenditure
categories. However, the typical movements of the personal consumption
expenditures contribution are much smaller than those coming from
capital goods.
Figures 3, 4, and 5 together show that the most substantial risks
to GDP growth come from common movements in expenditures on capital
goods. Factors that are more specific than that have never substantially
driven the business cycle. For example, the ongoing slump in residential
investment presents only very small risks to GDP growth if the slump is
confined to that sector alone, but it can substantially influence the
overall economy if spillover effects cause other sectors to contract.
Assessing a firm's macroeconomic risks
The analysis in the preceding sections sets the stage for assessing
a given business's exposure to macroeconomic risk. This proceeds in
two steps: Correct the company of interest's recent sales growth
for inflation, and construct a counterfactual benchmark growth rate for
a company with the same distribution of sales across NIPA categories but
with each product's sales growth perfectly synchronized to the
growth of its corresponding NIPA category. The first step determines the
influence of category-specific inflation on the business's growth,
and the second allows us to examine the influence of macroeconomic risks
on similarly positioned businesses in the past.
[FIGURE 5 OMITTED]
Hair salon
Before 1 implement this assessment for a large conglomerate firm, I
begin by considering how to do this for a fictional firm with a single
product: haircuts. Table 2 reports revenues for a fictional hair salon
for four years. Most hair salons receive revenues from both cutting and
styling services and the retail sale of hair care products. To keep this
example simple, suppose that this salon's revenues come entirely
from cutting and styling. These expenditures contribute to personal
consumption expenditures on services (line 5 in the spreadsheet of
figure 1 on p. 31). Table 2's third line reports the NIPA price
index for this category, and its fourth line gives the ratio of the
first to the third lines (multiplied by 100). This equals the
salon's revenues in year 2000 dollars. This simple inflation
adjustment brings the salon's revenue growth down by an average of
3.4 percentage points. Over the three years given, the salon's real
growth equals 2.9 percent. Since the salon's revenues all come from
the sale of services, the real growth rate of personal consumption
expenditures on services is a relevant macroeconomic benchmark. For
these years, this always equaled 2.7 percent. The analysis of panel B of
figure 3 (p. 37) implies that this firm is then exposed to very little
macroeconomic risk.
General Electric Company
Assessing macroeconomic risk for a single-product business requires
little more than looking up the correct price index and real growth rate
from the NIPA data. For a conglomerate firm selling multiple products,
the inflation adjustment and the construction of a macroeconomic
benchmark use the macroeconomic tools discussed previously. I illustrate
this here with an application to publicly available data for one firm,
General Electric Company. Like many other firms, General Electric
reports sales and profits for several operating segments. Each of these
groups together lines of business with similar customers. In its 2006
annual report, General Electric breaks its performance down into six
operating segments.
* GE Infrastructure contains the production and sales of big-ticket
capital items, such as electric generators and jet aircraft engines. It
also encompasses the sale of services (including financial services) to
the capital items' purchasers.
* GE Commercial Finance provides financial services to firms to
finance the purchases of major capital assets.
* GE Money provides financial services to consumers.
* GE Healthcare produces diagnostic equipment and software.
* NBC Universal produces entertainment programming (and thereby
advertising revenue).
* GE Industrial produces white goods (major household appliances
such as refrigerators and stoves), consumer hardware, and some specialty
plastics.
For General Electric, table 3 displays the relevant data. This
table from its 2006 annual report lists each operating segment's
sales (in U.S. dollars) for five years (2002-06). Just as with the
macroeconomic data, I begin with each segment's share of total
revenue, which table 4 lists for 2004. Consistent with GE's
industrial origins, the GE Infrastructure and GE Industrial segments
account for more than half of the company's revenues. Consumer
services (GE Money and NBC Universal) account for just less than 20
percent, and the remaining 25 percent of revenues come from the GE
Commercial Finance and GE Healthcare segments.
The segments' shares allow the calculation of their
contributions to growth. In equation 8, I wrote the rate of GDP growth
(in dollars) as the share-weighted average of its expenditure
components" growth rates. Just so, we can write GE's growth
rate as the share-weighted average of its segments' growth rates.
Table 5 reports this decomposition of growth for 2005 and 2006. This
calculation used segment growth rates adjusted for the effects of those
acquisitions and divestitures mentioned in the 2006 annual report's
managerial discussion of these figures. Overall revenue growth was
similar for GE in those two years at 9.0 percent and 9.6 percent.
Inflation adjustment
The straight decomposition of revenue growth in table 5 ignores the
effects of inflation. If inflation affects the prices of all goods
equally, then taking account of inflation would only lower the measured
growth of total revenues and leave each segment's relative
contribution unchanged. Of course inflation affects some goods more than
others. For example, the prices of capital goods (measured with the
price index for business fixed investment) actually fell in both 2005
and 2006. In this more realistic case, accounting for inflation proceeds
in two steps:
* Assign a GDP expenditure category to each segment; and
* Combine the NIPA price indexes for these categories with the
segments' revenue histories in the contributions to growth formula
in appendix B to calculate each segment's contribution to the
company's real revenue growth.
Choosing which GDP expenditure's price index should be used
for a given segment can be straightforward when a segment's sales
directly contribute to GDP. In the case of General Electric, GE
Healthcare and GE Infrastructure both contribute to business fixed
investment of equipment and software and GE Money's revenues are
part of personal consumption expenditures on services. The sales of
General Electric's other segments do not arithmetically contribute
to GDP because they sell goods and services that other producers use to
create products for sales to final consumers. For example, NBC Universal
sells advertising. Firms use this to enhance the value of the products
they sell to consumers, so it contributes in that way to personal
consumption expenditures. General Electric's other segments sell
products to consumers and producers in multiple NIPA expenditure
categories. Without detailed information on those segments' sales,
the best measure of their prices is the price index for GDP as a whole.
Table 6 summarizes my choices of price indexes for General
Electric's segments.
Table 7 reports the segments' contributions to
inflation-adjusted revenue growth given these price index choices. Just
as with the contributions in table 5, these use acquisition-adjusted
growth rates. Overall inflation was just over 3 percent for both 2005
and 2006, but adjusting GE's revenues for inflation knocks only 2
percentage points off of overall revenue growth (compare the last lines
in tables 5 and 7). This reflects their heavy concentration in capital
investment goods where inflation tends to be lower. To see this, note
that GE Infrastructure and GE Healthcare together do not account for any
of the 2 percentage point adjustment (their values are the same in both
tables 5 and 7).
Macroeconomic benchmarks
The final step in evaluating GE's exposure to macroeconomic
risk is the construction of a relevant benchmark growth series. Here, I
offer two such benchmarks. They both answer the question, How would the
economy have evolved if the mix of products produced were always
identical to GE's mix? Their point of difference is the moment in
time when the product mix is measured. The first benchmark uses the
contributions to growth formula to continuously account for the
company's changing orientation. The second holds the product mix
fixed at the most recent year's values. The first is more useful
for gauging past performance, and the second can be used to ask how the
company would have fared in the past had it been configured as it is in
the present.
The first macroeconomic benchmark arises from the following
calculation: Apply the contributions to growth formula to General
Electric with the NIPA expenditure category listed in table 6 replacing
each segment's real growth rate. This measures how revenues would
have grown if each of GE's segments had tracked its NIPA
counterpart exactly. The result will differ from overall GDP growth
because GE does not produce a representative variety of all goods and
services. Table 8 reports the result of this calculation, and its
comparison with table 7 is instructive. The macroeconomic benchmark grew
4.1 percent in both 2005 and 2006, which is substantially below General
Electric's actual real growth rates of 6.9 percent and 7.6 percent.
The GE Infrastructure segment accounts for nearly 2 percentage points of
this "overperformance" in both years. Three other segments
grew faster than their NIPA counterparts in both years--GE Money, GE
Healthcare, and NBC Universal. Together, they account for 1.5 and 0.9
percentage points of GE's higher growth in 2005 and 2006. There
were only two cases where a GE segment contributed less to growth than
its NIPA counterpart: GE Commercial Finance in 2005 and GE Industrial in
2006.
General Electric differs from the nation's economy in several
respects. It has extensive operations and sales abroad. It specializes
in submarkets within NIPA expenditure categories. For example, medical
equipment, aircraft engines, and generators are only three of the
thousands of capital goods produced in the U.S. Developments in these
specific markets that are out of General Electric's control can
nevertheless disproportionately affect its performance. For these
reasons, the comparison of tables 7 and 8 only starts the conversation
about General Electric's performance relative to its macroeconomic
benchmark. Continuing the conversation requires information on
market-specific developments and operational performance that lies
beyond the scope of this article.
The first benchmark accounts for the business's changing
product mix, using the contributions to growth formula, but this
accounting might actually obscure the information needed for
macroeconomic risk assessment. For example, consider a business that
once produced mostly consumer services but has recently expanded into
residential investment. The first macroeconomic benchmark for it
fluctuated little in recent recessions, but it would be a mistake to
conclude that it has little exposure now to risk. To address this issue,
we can calculate a second macroeconomic benchmark that holds the
operating segments' shares at their most recently observed values.
Figure 6 plots this for General Electric. For comparison's sake,
the figure also plots overall real GDP growth.
Examining the two series during NBER-dated recessions (the shaded
periods) is instructive. A company oriented as GE is now would have been
highly sensitive to macroeconomic risk. The most recent recession (in
2001) was concentrated in precisely the industries GE serves, so its
macroeconomic benchmark growth rate was well below zero for about two
years. During this period, overall real GDP growth was negative for only
two quarters. Although the most recent recession's heavy
concentration in business fixed investment was exceptional, the
comparison of the second macroeconomic benchmark with real GDP growth
during earlier recessions yields the same conclusion. The company's
concentration in industries with customers who can easily delay their
purchases leaves it substantially exposed to macroeconomic risk.
To quantify GE's exposure, we can calculate the standard
deviation of its macroeconomic benchmark growth and compare it with the
standard deviation of GDP growth? For the period plotted, the
macroeconomic benchmark's standard deviation equals 1.22 percent,
while that for GDP growth is 0.91 percent. Thus, the benchmark is
approximately 33 percent more sensitive to business cycle fluctuations
than is the economy as a whole.
Conclusion
This article presented two basic tools for measuring business cycle
fluctuations, the national product accounting identity and the
contributions to growth formula, and applied them to evaluate a
particular conglomerate's exposure to macroeconomic risk. The
application was illustrative only because it omitted factors such as the
scope of the company's international operations and sales that are
important for that company. The macroeconomic benchmarks presented here
can only start a conversation about a business's place in the
larger economy. Finishing it and moving on to action requires more
information and the subjective judgment of those whose wealth is at
stake.
[FIGURE 6 OMITTED]
APPENDIX A. INFLATION ACCOUNTING
Observations of the national income and product accounting
identities' components reveal how the value of goods and services
measured in dollars evolves, but erosion of the dollar's purchasing
power makes these measures insufficient for tracking economic growth.
The BEA fills this gap with measures of how the prices of goods in each
product category change.
If each category had only one good (or service) for sale, then
accounting for inflation would be simple. Let [X.sub.t] represent the
nominal value of some expenditure component, and let [P.sub.t] give the
dollar-denominated price for its single good or service. For example, if
the category in question was nondurable goods and the only good in that
category was bananas, then [P.sub.t] would equal the dollar price of one
pound of bananas and [X.sub.t]/[P.sub.t] would equal the pounds of
bananas purchased in the quarter, also called the quarter's real
expenditure. The real expenditure's growth rate is
A1) [G.sub.t] = [X.sub.t] / [P.sub.t] [P.sub.t-1] / [X.sub.t-1]- 1.
Of course, there are hundreds of thousands of distinct goods or
services that contribute to any given expenditure category.
Nevertheless, the BEA constructs the real growth rate of each
expenditure category by using equation A1. For this, it replaces
[P.sub.t] with an index of prices for goods in the category.
Constructing such an index when there are only two goods in the category
suffices to illustrate the principles involved. Suppose that personal
consumption expenditures on nondurable goods covers only apples and
oranges. Use [P.sup.A.sub.t] and [P.sup.O.sub.t] for their dollar
prices (per pound) and [Q.sup.A.sub.t] and [sup.O.sub.t] for the number
of pounds sold. Two German economists, Etienne Laspeyres and Hermann
Paasche, offered solutions to the problem of combining these data to
measure how the price of fruit changed between the previous and current
quarters. Laspeyres proposed measuring [P.sub.t] / [P.sub.t-1] with a
weighted average growth of individual goods prices. Each weight equals
the expenditure share on that good in the base period.
A2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
The second expression makes clear that Laspeyres equals the cost of
acquiring the previous quarter's purchases at the current
quarter's prices relative to their cost in the original quarter.
Paasche's alternative suggestion is to measure the cost of the
current quarter's purchases relative to the same purchases made at
the previous quarter's prices.
A3) [Paasche.sub.t] = [P.sup.A.sub.t][Q.sup.A.sub.t] +
[P.sup.O.sub.t] [Q.sup.O.sub.t]/[P.sup.A.sub.t-1][Q.sup.A.sub.t] +
[P.sup.O.sub.t-1][Q.sup.O.sub.t].
Both of these suggestions measure the growth of prices, which is
all that equation A1 requires. They both realistically weight prices
based on the national family's actual purchase decisions, but both
of them rely an arbitrary selection of the date at which we measure
their purchases. To avoid this arbitrariness, the BEA measures the rate
of change with the geometric average of the Laspeyres and Paasche
measures. This idea was originally due to Irving Fisher:
A4) [Fisher.sub.t] = [square root of [Laspeyres.sub.t] x
[Paasche.sub.t]].
The expressions for [Laspeyres.sub.t] and [Paasche.sub.t] can both
be easily extended to the case with more than two goods. With these in
hand. the BEA constructs [P.sub.t] by setting its value in the first
quarter of the data to one and setting its values for later quarters
recursively with
[P.sub.t] = [Fisher.sub.t][P.sub.t-1].
To make the series more easily interpretable, the BEA finishes by
multiplying by 100 and dividing by the average value of [P.sub.t] during
a base year (which is currently 2000). The resulting series indicates
how many dollars it requires in any given quarter to buy the same goods
and services that $100 could purchase in 2000. The BEA labels this the
real expenditure series corresponding to [X.sub.t]. In this article. I
use [x.sub.t] [equivalent to] [X.sub.t]/[P.sub.t] to represent this real
series. For example, [c.sup.N.sub.t], [c.sup.s.sub.t], and
[c.sup.D.sub.t] respectively equal real personal consumption
expenditures on nondurable goods, services, and durable goods.
Of course, we can apply the Fisher deflation procedure to create
real values of each national product component. Before working with
these, a note of caution regarding addition is in order. There are two
conceivable ways of calculating the real values of all personal
consumption expenditures. First, one could create a Fisher price index
based on all goods in personal consumption expenditure (call it
[P.sup.C.sub.t]) and then calculate [c.sub.t] = ([C.sup.N.sub.t] +
[C.sup.S.sub.t] + [C.sup.D.sub.t])/[P.sup.C.sub.t]. Second, one could
simply add the components' real values, [c.sup.N.sub.t] +
[c.sup.S.sub.t] + [c.sup.D.sub.t]. Will these answers equal each other?
Yes, if no two goods or services in personal consumption expenditures
have prices that change relative to each other. In the two-good example
used previously, this requires that the number of oranges the national
family must sacrifice in order to acquire one more apple cannot change
with time: [P.sup.A.sub.t-1]/[P.sup.O.sub.t-1] = [P.sup.O.sub.t]. In
this very special case, the Laspeyres, Paasche, and Fisher price indexes
all equal each other; and all prices grow in lock step. Outside of this
very special case. the two calculations of real personal consumption
expenditures will differ. By construction, the first calculation uses
the Fisher price index. The second calculation can only be useful as a
possible shortcut to the first. It is not, so we dispose of it. This
illustrates a general principle: The real analogues of the components in
the national product accounting identity do not sum to teal income
except in the base year.
Calculating the exact Fisher-deflated real analogue to a sum of two
product components requires applying the Fisher procedure to the
original price and quantity observations from both components. Of
course, the BEA does not provide these, so a useful (and accurate)
approximation to this exact calculation is to apply the Fisher procedure
to the components themselves. For example, to calculate approximate real
personal consumption expenditures on nondurable goods and services,
replace [Q.sup.A.sub.t], [Q.sup.O.sub.t], [P.sup.A.sub.t], and
[P.sup.O.sub.t] in equations A2 and A3 with [MATHEMATICAL EXPRESSION NOT
REPRODUCIBLE IN ASCII]. Since the BEA reports real personal consumption
expenditures as well as the real expenditures for each of its three
components, it is possible to assess how well this
"chain-addition" procedure works. The approximate real
personal consumption expenditure differs from its exact counterpart by
at most 3/100 of a percentage point over the post-Korean War sample
period.
APPENDIX B. CONTRIBUTIONS TO GROWTH
By construction, GDP sums the national family's expenditures
in the various categories. To gain a greater understanding of the
sources of growth, one might wish to decompose GDP growth into the
expenditures' distinct contributions. Suppose that the only two
goods produced in the nation are apples and oranges, so that [Y.sub.t] =
[P.sup.A.sub.t] [Q.sup.A.sub.t] + [P.sup.O.sub.t][Q.sup.O.sub.t].
Definition 1
The contribution of expenditures on apples to the growth of GDP
from quarter t - 1 to quarter t equals
100 x ([P.sup.A.sub.t][Q.sup.N.sub.t] -
[P.sup.A.sub.t-1][Q.sup.A.sub.t-1])/[Y.sub.t-1],
the reduction of the GDP growth rate that would have occurred if
the national family had left its expenditures on apples unchanged from
quarter t - 1 to quarter t and income from some source adjusted
simultaneously so that the national product accounting identity
continues to hold good.
The contribution of expenditures on oranges is defined analogously,
and summing them yields 100 x (Y.sub.t] - [Y.sub.t-1])/[Y.sub.t-1], the
growth rate of nominal GDP.
Nominal GDP growth convolves real economic expansion with price
inflation, so calculating the various contributions to it reveals little
about the expansion of the national family's real purchasing power.
For this reason, we might be more interested in the impact on real GDP
growth of holding the quantity of apples fixed. We can call this the
growth impact of real apple expenditures.
Definition 2
The growth impact of apple expenditures from quarter t - 1 to
quarter t equals the actual growth rate of real GDP minus the growth
rate recalculated after replacing [Q.sup.A.sub.t]'s value with
[Q.sup.A.sub.t-1]'s value.
We can make the same calculation for real orange expenditures, but
these two growth impacts will not sum to the growth of real GDP. This
reflects the inconvenient truth noted previously that the sum of real
expenditures does not equal the real expenditure on the sum unless
[P.sup.A.sub.t]/[P.sup.O.sub.t] = [P.sup.A.sub.t-1]/[P.sup.O.sub.t-1].
Hence, the growth impacts of apple and orange expenditures--while
potentially interesting--cannot serve as the basis for an accounting of
real GDP growth.
The single exception to the general proposition that the sum of
real expenditures does not equal the real expenditure on file sum offers
us the possibility of defining contributions m growth by artificially
imposing that all relative prices remain the same. Use
[P.sup.A*.sub.t-1], [P.sup.O*.sub.t-1], [P.sup.A*.sub.t], and
[P.sup.O*.sub.t], to denote these alternative (counterfactual prices)
that satisfy [P.sup.A*.sub.t-1]/[P.sup.O*.sub.t-1] =
[P.sup.A*.sub.t]/[P.sup.O.sub.t]. What values should we assign them?
Consider
B1)[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].
These satisfy the restriction that relative prices do not change,
and either quarter's nominal GDP calculated with them equals its
original value. To show this. start with the definition of nominal GDP
for quarter t - 1 with the alternative prices. Substituting the
alternative prices definitions and manipulating yields
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
Next, replace [P.sup.A.sub.t][Q.sup.A.sub.t-1] +
[P.sup.O.sub.t][Q.sup.O.sub.t-1] with [[pi].sup.L.sub.t][Y.sub.t-1] to
get
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].
The definition of [[pi].sup.F.sub.t] implies that
[[pi].sup.L.sub.t]/[[pi].sup.F.sub.t] = [[square root of
[[pi].sup.L.sub.t]/[[pi].sup.P.sub.t]], so the right-hand side equals
[Y.sub.t-1]. A parallel argument shows that [Y.sup.*.sub.t] = [Y.sub.t],
so changing to the alternative prices leaves nominal GDP growth
unchanged. The Laspeyres. Paasche, and Fisher price indexes constructed
with these prices and the original quantities all equal
[[pi].sup.F.sub.t], so replacing the alternative prices also leaves real
GDP growth unchanged. All of this leads to the following definition.
Definition 3
The contribution of real apple purchases to real GDP growth equals
[P.sup.A*.sub.t-1]([Q.sup.A.sub.t] - [Q.sup.A.sub.t-1])/[Y.sub.t-1]
This added to the analogously defined contribution of real orange
purchases does sum to real GDP growth. This equation defines the
contributions to growth when two components contribute to GDP, but the
same approach also works with a larger number of components: Average a
category's price index in quarter t - 1 with its value in quarter t
divided by [[pi].sup.F.sub.t], divide by 1 + [[square root of
[[pi].sup.L.sub.t]/[[pi].sup.P.sub.t]], and use this for
[P.sup.A*.sub.t-1] in definition 3.
REFERENCE
General Electric Company, 2007, 2006 Annual Report: Invest and
Deliver, Fairfield, CT.
NOTES
(1) This is the only case in which the BEA attempts to measure the
value of a service produced in the home
(2) Of course, foreign governments and institutions hold
substantial US government debt. Within this simple framework, I can
represent these financial investments by supposing that the national
household issues IOUs to foreigners to finance the purchase of
government debt Thus, the convenient restriction that only U.S.
households hold government debt entails no important loss of generality.
(3) For expositional simplicity only, this discussion presumes that
the conglomerate government has no employees and engages in no
production of its own In reality, the BEA adds the values of government
employees' salaries to [G.sub.t].
(4) This assumes that the tax decrease has no effect on the
national household's incentive to work. which is the case when the
lax is collected like a poll tax In the United States. much tax revenue
comes from taxing labor income. Lowering such a tax and financing the
shortfall with debt can temporarily expand income, but even this might
not serve the household's interests.
(5) For any sequence of observations [x.sub.1], [x.sub.2],...,
[x.sub.1], the mean and standard deviation are defined as [bar.x] =
([x.sub.1] + [x.sub.2] + ... + [x.sub.T]) / T and [??] = [square root
of [([x.sub.1] - [bar.x]).sup.2] + [([x.sub.2] - [bar.x]).sup.2]+ ... +
[([x.sub.T] [bar.x]).sup.2]]. If all of the observations are identical,
then the standard deviation equals zero Differences between the
observations raise the standard deviation, so it is a measure of
dispersion.
Jeffrey R. Campbell is a senior economist in the Economic Research
Department at the Federal Reserve Bank of Chicago and a faculty research
fellow in the Economic Fluctuations Program at the National Bureau of
Economic Research. The application using data from General Electric
Company's 2006 annual report is for illustrative purposes only.
In no way does it constitute an endorsement of that company or its
management
TABLE 1 Shares of gross domestic product
Personal consumption expenditures
Nondurable Durable
Date range goods Services goods
(percent)
1953:Q4-1959:Q4 30.2 23.9 8.6
1960:Q1-1969:Q4 27.1 26.3 8.4
1970:Q1-1979:Q4 25.0 28.8 8.6
1980:Q1-1989:Q4 22.7 33.4 8.3
1990:Q1-1999:Q4 20.3 38.5 8.2
2000:Q1-2008:Q1 20.1 41.4 8.4
Fixed investment
Non-
Date range Residential residential Inventory
(percent)
1953:Q4-1959:Q4 5.3 9.5 0.4
1960:Q1-1969:Q4 4.6 9.8 1.0
1970:Q1-1979:Q4 4.9 11.1 0.7
1980:Q1-1989:Q4 4.4 12.1 0.4
1990:Q1-1999:Q4 4.1 10.9 0.5
2000:Q1-2008:Q1 5.2 10.7 0.2
Net Government
Date range exports purchases
(percent)
1953:Q4-1959:Q4 0.3 21.8
1960:Q1-1969:Q4 0.6 22.1
1970:Q1-1979:Q4 -0.2 21.0
1980:Q1-1989:Q4 -1.8 20.6
1990:Q1-1999:Q4 -1.3 18.8
2000:Q1-2008:Q1 -4.8 18.8
Source: Author's calculations based on data from the U.S. Bureau
of Economic Analysis, National Income and Product Accounts of the
United States.
TABLE 2
Revenues and performance of a fictional hair salon
2007 2006 2005 2004
Revenues (dollars) 1,200 1,125 1,050 1,000
Nominal revenue growth (percent) 6.60 7.10 5.00
Price index for services 124.58 120.73 116.73 112.93
Revenues (2000 dollars) 963.24 931.83 899.51 885.50
Real revenue growth (percent) 3.4 3.6 1.6
Growth of macroeconomic
benchmark (percent) 2.7 2.7 2.7
TABLE 3
General Electric Company's operating revenues
Operating segment 2006 2005 2004 2003 2002
(millions of dollars)
GE Infrastructure 47,429 41,803 37,373 36,569 40,119
GE Commercial Finance 23,792 20,646 19,524 16,927 15,688
GE Money 21,759 19,416 15,734 12,845 10,266
GE Healthcare 16,562 15,153 13,456 10,198 8,955
NBC Universal 16,188 14,689 12,886 6,871 7,149
GE Industrial 33,494 32,631 30,722 24,988 26,154
Total segment
revenues 159,224 144,338 129,695 108,398 108,331
Corporate items
and eliminations 4,167 3,618 4,596 5,023 3,636
Consolidated revenues 163,391 147,956 134,291 113,421 111,967
Note: Data are for General Electric Company and its consolidated
affiliates.
Source: General Electric Company. 2006 Annual Report: Invest and
Deliver, p. 53.
TABLE 4
Share of General Electric Company's revenues,
by operating segment, 2004
Operating segment Percentage of revenues
GE Infrastructure 33.7
GE Commercial Finance 15.6
GE Money 11.9
GE Healthcare 9.4
NBC Universal 6.3
GE Industrial 23.1
Source: Author's calculations based on data from General
Electric Company, 2006 Annual Report: Invest and Deliver, p. 53.
TABLE 5
Contributions to General Electric Company's
nominal revenue growth, by operating segment
Operating segment 2006 2005
(percent) (percent)
GE Infrastructure 3.9 3.4
GE Commercial Finance 2.2 0.6
GE Money 1.0 1.4
GE Healthcare 1.0 0.7
NBC Universal 1.0 1.4
GE Industrial 0.5 1.5
Total segment revenues 9.6 9.0
Note: For these calculations, each segment's growth was adjusted
for the effects of acquisitions and divestitures mentioned in the
2006 annual report's managerial discussion.
Source: Author's calculations based on data from General Electric
Company, 2006 Annual Report: Invest and Deliver, p. 53.
TABLE 6
National income and product accounts (NIPA) price indexes
for General Electric Company's operating segments
Operating segment NIPA price index
GE Infrastructure Business fixed investment, equipment
and software
GE Commercial Finance Gross domestic product
GE Money Personal consumption expenditures,
services
GE Healthcare Business fixed investment, equipment
and software
NBC Universal Personal consumption expenditures
GE Industrial Gross domestic product
Note: These are the author's subjective choices of NIPA
price indexes for General Electric's operating segments.
TABLE 7
Contributions to General Electric Company's
inflation-adjusted revenue growth,
by operating segment
Operating segment 2006 2005
(percent) (percent)
GE Infrastructure 3.9 3.4
GE Commercial Finance 1.7 0.1
GE Money 0.5 0.9
GE Healthcare 1.0 0.7
NBC Universal 0.7 1.1
GE Industrial -0.2 0.7
Total segment revenues 7.6 6.9
Note: For these calculations, each segment's growth was adjusted
for the effects of acquisitions and divestitures as described in
the text.
Sources: Author's calculations based on data from the U.S. Bureau
of Economic Analysis, National Income and Product Accounts of the
United States: and General Electric Company, 2006 Annual Report:
Invest and Deliver, p. 53.
TABLE 8
Macroeconomic counterfactual contributions
to General Electric Company's inflation-adjusted
revenue growth, by operating segment
Operating segment 2006 2005
(percent) (percent)
GE Infrastructure 1.7 1.7
GE Commercial Finance 0.4 0.4
GE Money 0.4 0.3
GE Healthcare 0.6 0.6
NBC Universal 0.3 0.3
GE Industrial 0.7 0.7
Total segment revenues 4.1 4.1
Note: For these calculations, each segment's growth was adjusted
for the effects of acquisitions and divestitures as described in
the text.
Sources: Author's calculations based on data from the U.S. Bureau
of Economic Analysis, National Income and Product Accounts of the
United States; and General Electric Company, 2006 Annual Report:
Invest and Deliver, p. 53.
