MICHAEL
C. LOVELL
WesleyanUniversity
The
Trends
Profit
Picture:
and
Cycles
HAS THEREBEEN a squeeze on profits during the past decade? And if
profits have been under pressure, what has been the primary source of the
squeeze? These are important questions, not only for stockholders, but
also for the economy as a whole. If George Perry was correct in stating
that profits are an important determinant of money wages in manufacturing, a squeeze on profits might serve to dampen wage demands, which
would mitigate inflationary pressure. If profits, by stimulating investment
spending, trickle down into greater productivity, a squeeze on profits may
lead to less output per man-hour, which would intensify inflation. And if
William Baumol is correct in arguing that the profits realized by the firm
are more than a residual, a squeeze on profits has unfortunate implications for price movements and employment. According to Baumol's proposition that firms maximize sales volume subject to a profit constraint, a
squeeze on profits is likely to induce a rise in prices along with a simultaneous contraction in output and employment.' Recently, complaints
about a profit squeeze, coupled with a concern that investment levels are
inadequate and the decline in productivity growth, have brought intensified pressure for a reduction in the corporate tax rate and a more generous
investment tax credit.
1. William J. Baumol, Economic Theory and OperationsAnalysis, 4th ed. (Prentice-Hall, 1977), chap. 15. Herbert A. Simon's satisficing theory suggests that a
profit squeeze eats into organizational slack and encourages a deeper look at the
corporation'saccustomed practices;it may lead to a tighteningof pricing policy and
a search for ways of increasing the efficiency of operations. See Herbert Simon,
"Theories of Decision-Making in Economics and Behavioral Science," American
Economic Review, vol. 49 (June 1959), pp. 253-83.
0007-2303/78/0003-0769$00.25/0
) Brookings Institution
770
BrookingsPaperson EconomicActivity,3:1978
ArthurM. Okun and GeorgeL. Perrypoint out in their 1970 paper,
"NotesandNumberson the ProfitsSqueeze,"that,beforethe mid-1960s,
the shareof corporateprofitsin grossnationalproductcouldbe explained
quitepreciselyby fluctuationsin the percentageshortfallbetweenrealized
and potential GNP.2 In this report I decomposethe movementsof a
variety of alternativeprofit measuresinto trend and cyclical forces. It
turnsout thatwhethera profitsqueezehas actuallyoccurred,its intensity
andthe extentof profitrecoveryin recentyearsdependslargelyuponhow
thismultifacetedconceptis definedandmeasured.
ProfitMeasures
Profitscan be measuredin relationto equity,the capitalstock,or value
added. They can be measureddirectlyfrom tax accountingrecordsor
with correctionsfor the diversityof conventionalaccountingprocedures
and with adjustmentsfor some distortingeffectsof inflation.Profitscan
be narrowlydefinedas the returnon equity capitalor more generically
measuredto includethe interestcosts on borrowedcapital.
Table 1 lists fourteenalternativeconceptsfor measuringprofits.The
eight profitmeasureslisted acrossthe firsttwo rows are for the shareof
profitsin the income generatedby nonfinancialcorporatebusiness.The
first four of these measures(Pa, Pla, PATa, and PIATa) are adjusted
for variationsin accountingprocedures;all inventoriesare measuredto
approximatelast-in-first-out(LIFO) accounting,and the capital consumptionallowanceis based on replacementratherthan historicalcost
and on the estimatedeconomiclives of assetsratherthanthe servicelives
allowed for tax purposes.Two of the adjustedprofitsharesinclude interest payments;two measuresare net of the corporateprofitstax. The
PIATa profit measure-profits plus interest after tax-corresponds to
the "genuineprofit"conceptusedby WilliamNordhausin his study,"The
FallingShareof Profits."3
The measureslisted in the second row (Pc, Plc, PATc, and PlA Tc)
are based on the conventionalaccountingfiguresused by businessfirms,
2. BPEA, 3:1970, pp. 466-72.
Falling Share of Profits,"BPEA, 1:1974, pp. 1693. William D. Nordhaus, "bThe
208.
771
MichaelC. Lovell
Table 1. AlternativeConceptsfor MeasuringProfits
Concept
Description
Profits
Profits
Profits
plus
Interest
after
Profits corporate after
plus
profits corporate
interest
tax
profitstax
Shareofprofits(nonfinancial
corporations)
With inventoryvaluationand capital
consumptionadjustments
Using conventionalaccounting
procedures
Pa
Pla
PATa
PIATa
Pc
PIc
PATc
PIATc
...
ROEATm
...
RPIa
...
RPIATa
...
RPIc
...
RPIATc
Rate of returnon equity(manufacturing
firms)
Using data from FTC-SECquarterly
financialreportsaand conventional
ROEm
accountingprocedures
Rate of returnrelativeto capitalstock
(nonfinancialcorporations)
With inventoryvaluationand capital
consumptionadjustments
Using conventionalaccounting
procedures
Sources: See appendix for sources and for a more detailed description of the profit concepts.
a. Federal Trade Commission-Securities and Exchange Commission, Quarterly Financial Report for
ManufacturingCorporations;currently, Federal Trade Commission, QuarterlyFinancial Reportfor Manufacturing, Mining and Trade Corporations.
at least for tax purposes.PIATc correspondsto the "nominalshare"concept of the Nordhausstudy.
The two profitconceptsin the thirdrow, ROEm and ROEATm, are
rates of returnon equity as reportedby U.S. manufacturingfirmsto the
and Exchange Commission
joint Federal Trade Commission-Securities
publication (since 1971, Federal Trade Commissiononly), Quarterly
FinancialReport. These are both nominalbook measuresreflectingthe
diversityof accountingpractices.The ROEATm concept was used by
Perry in his study of the determinants of money wages.4
4. George L. Perry, Unemployment, Money Wage Rates, and Inflation (MIT
Press, 1966). A similar measure is used by Laurence S. Seidman in '4Tax-BasedIncomes Policies," BPEA, 2:1978, pp. 301-48.
772
BrookingsPaperson Economic Activity,3:1978
The four profitmeasuresin the bottomtwo rowsof the table areratios
of profits plus interest for nonfinancialcorporationsrelative to Musgrave'sestimatesof the value of equipmentstock plus structures.5The
RPJa measurecorrespondsto thatusedby FeldsteinandSummersin their
study,"Isthe Rate of ProfitFalling?"6
Whilenone of the profitmeasuresis withoutmerit,some may be more
informativethan others.Duringtimesof inflation,conventionalaccounting estimatesare subjectto a varietyof distortions;some of these distortions may understateprofits,while othersmay exaggeratethem. First-infirst-out (FIFO) inventoryaccountingand historicalcost depreciation
overstateprofits.Accelerateddepreciationandthe neglectof the inflationinduceddecline in the real value of the firm'sfinancialobligationsconConceivably,thesetwo offsettingtypesof error
tributeto understatement.
It is especiallydifficultto characterizethe
out.7
may more or less cancel
ROEmandROEATmconceptsbecausethe profitsaremeasuredrelative
to conventionalaccountingestimatesof corporateequity, and because
both numeratorand denominatorare affectedby inflation,but not in a
directlyoffsettingway.
The adjustedprofitseries,identifiedwith an a in table 1, incorporate
two nationalincome accountingrefinementsof the conventionallymeasuredbook figures.The capitalconsumptionadjustment(CCA) makes
two correctionsthat go in opposite directions.It raises profitsby eliminatingaccelerateddepreciationand,of greaterimportancein recentyears,
lowersthemby convertingfromhistoricalcost to replacementcost depreciation. The inventory valuation adjustment(IVA) corrects for the
vagariesof inventoryaccountingproceduresby calculatingprofitsuniformlyas thoughall firmsused LIFO accounting.This latter correction
5. John C. Musgrave, "Fixed Nonresidential Business and ResidentialCapital in
the United States, 1925-75," Survey of CurrentBusiness, vol. 56 (April 1976), p.
49. His data are updated in the August 1976 and August 1977 (vol. 57) issues of the
Survey of CurrentBusiness,pp. 64 and 57, respectively.
6. Martin Feldstein and Lawrence Summers, "Is the Rate of Profit Falling?"
BPEA, 1:1977, pp. 211-27. In contrast to the Feldstein-Summersmeasure, the denominator does not include the value of nonfinancialcorporate landholdingsor the
stock of inventories. Consequently,the computed measure overestimatesthe rate of
profit.The value of land may be approximatelyproportionalto the value of included
assets. The value of inventories may have a different cyclical movement than the
value of included assets.
7. See George M. von Furstenbergand Burton G. Malkiel, "FinancialAnalysis
in an Inflationary Environment," Journal of Finance, vol. 32 (May 1977), pp.
575-88.
MichaelC. LoveHl
773
has been quitelargeduringinflationaryepisodes: in 1974 the adjustment
reducedthe conventionallymeasuredbook profitfiguresby 30 percent.
While the inventoryand capital consumptionadjustmentsmade by
the national income accountantslead to more conservativeprofit estimates, the adjustedprofitfigureserr in the directionof understatement.8
Inflationcauses profit figuresto understateand interestfiguresto exaggerateactualreturnsbecausethe figuresdo not net out the decliningreal
valueof financialobligations.Wheninflationis generallyanticipated,borrowershave to pay highernominalinterestrates to compensatelenders
for the real capitallosses they will sufferthroughthe predictederosionof
the purchasingpower of periodic interest and amortizationpayments.
The nationalincomeaccountsexaggeratethe returnto the lenderbecause
the higherinterestcost is countedas income,while the inflation-induced
real capital losses are neglected;profit attributedto owners is understated because the decline in the real value of the firm'sliabilities is
neglected,while the added interestcost is recognized.These considerations mean that measuresof gross profits includinginterest payments
(those in the second and fourthcolumnsof table 1) are probablymore
reliablethanmeasuresof profitsnet of interest.
Whilethe IVA maylead to an improvedmeasureof profitsby approximatinguniformLIFO, it poses certainconceptualproblems.A firmborrowing funds to finance speculativeinventoryholdings may suffer an
apparentdiminutionof LIFO accountingprofitseven if the anticipated
price rise eventuallymaterializes.9An alternativeto LIFO accounting,
the "constant-dollarFIFO" procedure recommendedby Shoven and
Bulow,10counts the capital gain on inventoryholdings as profitsto the
8. John B. Shoven and Jeremy I. Bulow, "InflationAccounting and Nonfinancial
Corporate Profits: Financial Assets and Liabilities," BPEA, 1:1976, pp. 15-57.
Also see Sidney Davidson and Roman L. Weil, "InflationAccounting: Implications
of the FASB Proposal," in Henry J. Aaron, ed., Inflation and the Income Tax
(Brookings Institution, 1976), pp. 81-114.
9. Consider a company that correctly anticipates rising prices of raw materials.
The firm increases its inventory of purchased materials from a customary threemonth supply to a four-month supply. During the accounting periods over which
the company maintains its extended inventories, LIFO profits are reduced because
the added carrying costs, including interest, are treated as current expense, while
the capital gain is not counted until the speculative position is liquidated. The national income accounts never capture the capital gain because they do not catch the
base change when stocks are liquidated.
10. John B. Shoven and Jeremy I. Bulow, "InflationAccounting and Nonfinancial CorporateProfits: Physical Assets,"BPEA, 3.1975, pp. 589-97.
PlcPc PIaPa
PATc PATa
ROEm
See
Sources:RPIATa
ROEATm
Variable'
table
a.
RPIc RPla
Table
2.
1 See
for
appendix.
Correlations
Pa
0.831
0.913
0.539
0.903
0.823
0.822
0.374
0.214
0.825
0.975
1.000
definitions
of
and
0.871
0.291
0.382
0.838
0.965
0.656
0.873
0.836
0.848
1.000 Pla
Summary
variables.
0.530
0.860
0.148
0.069
0.834
0.635
0.479
0.556
1.000
0.923
0.921
0.971
0.358
0.777
0.502
0.250
1.000
0.261
0.960
0.758
0.532
0.459
0.934
1.000
PATa
Statistics
for
Pc
Plc
Alternative
Simple
Profit
0.743
0.408
0.879
0.383
0.483
1.000
PATc
Measures,
correlations
0.243
0.476
0.329
0.807
1.000
ROEATm First
0.031
0.563
0.411
1.000
0.862
0.754
1.000
ROEm
Quarter,
RPIa
1949-Fourth
0.407
1.000
RPIATa
Quarter,
1.000
RPIc
1976
17.31
18.82
9.43
22.95
12.30
10.00
8.43
20.76
18.66
19.39
17.26Mean
Summary
3.101
3.795
2.197
3.861
3.486
1.644
2.072
2.345
2.954
2.744
3.772
deviation
statistics
Standard
MichaelC. Lovell
775
extent that they result from prices of the specificgoods held by a firm
rising more rapidlythan the generalinflationrate, as measuredby the
domesticspendingdeflator.A largedifferenceexistsbetweenthe estimate
of Shoven and Bulow and the officialLIFO estimatesof profitfor some
years, such as 1974, when the price of goods held in inventoryincreased
much more rapidlythan the generalprice level. For 1974, insteadof the
official IVA writedownof book profitsof nonfinancialcorporationsby
$35.1 billion, the constant-dollarFiFO writedownis only $16.2 billion;
thus Shovenand Bulow arguethat the nationalincome accountestimate
of 1974 nonfinancialcorporateprofitsis understatedby $18.9 billion, or
17 percentof before-taxprofits.Nonetheless,constant-dollarFIFO adjusted profitsare not an appropriateguide for a firm contemplatingthe
augmentationof stocksfor speculativereasonsbecausethe relevantdecision comparesthe rateof pricechangeanticipatedfor the itemsto be held
with the marginalcarryingcost of inventories,includingfinancing.
As these considerationsindicate,theremay be no single "ideal"profit
measure.Ratherthan single out a preferredmeasurefor exclusivescrutiny, this report looks at the common and distinguishingfeaturesof a
varietyof profitmeasures.
HistoricalOverview
Over the years the fourteen time series on profits have been only
moderatelysynchronized.Table 2 reportsthe correlationsamongeleven
of the profit measures.Generallythe degree of correlationis not spectacularlyhigh,whichsuggeststhat the apparentimportanceof profitvariables in empiricalstudiesof wage and price behaviormightbe sensitive
to the particularprofitmeasureemployed."The two measuresof manufacturingprofit,ROEm and ROEATm,althoughclosely correlatedwith
each other, are not tightly related to the measuresthat refer to nonfinancialcorporationsand arenot measuredrelativeto equity.Amongthe
nonfinancialcorporatemeasures,those in the samerows of table 1-that
is, those based on similar accountingprocedures-are rather highly
correlated.
11. My preliminaryinvestigation tentatively suggests that the strongest influence
on money wages may be exerted by the share of profits plus interest net of the
corporate profits tax (PIATc), rather than by ROEATm, the FTC-SEC profit rate
customarily employed in empirical studies of wage behavior.
20
a.
30
-
20
-
1945
See
Sources:
Figure
Percent
1.
table
I See
lTree
for
appendixc.
definitions
1950
of
'PAT
0Pic
the
Measures
of
the
profit
Profit
measures.
1955
Share
of
PATc~
~
~
~
~
Income,
Using
1960
Conventional
1965
Accounting
Procedures,
1970
4.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4
1946-77&
4.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1975
PIC
~~~~~~~~~~~~~~~~~~AT
MichaelC. Lovell
777
Three measuresof the profitshare of income are plotted in figure 1.
Whetherthe profitshareis measuredgross or net of interestor beforeor
afterthe corporateprofitstax, the pictureis one of declini g profitshares
from the end of World War II throughthe 1960s, followed by an upward trend towardpartialrecoveryin the currentdecade. The more inclusivemeasureincorporatinginterestpaymentsandprofitsshowsa much
moremoderatedecliningtrendover the yearsbecauseof the growingimportanceof interestpaymentson borrowedcapital.'2This growinginterest sharereflectschangesin capitalizationratios as well as risinginterest
costs. The portion of the increase associatedwith rising interest costs
is, in general,accompaniedby a decline in the real value of outstanding
debt.
Profit movementslook quite differentwhen the tax accountingmeasures are adjustedfor inflationaccordingto the proceduresused in the
GNP accounts.The Pa profitsharein figure2 involvesthe uniformevaluation of inventorieson a LIFO basis and the consistentmeasureof the
capital consumptionallowance.While the Pa and the conventionalPc
seriesusuallymove quiteclosely together,there are sizablediscrepancies
duringthe inflationof the post-WorldWar II and KoreanWar periods
and the 1970s. Inflationaryprofiteuphoriais considerablydampenedin
the Pa series.The conventionalaccountingseries,Pc, exaggeratedprofits
relativeto the figuresadjustedfor inflationin the GNP accounts.In contrastto ROEATm,both Pa and Pc show a substantialsqueezeon profits
duringthe Vietnamperiodof the 1960s, followedby a partialrecoveryin
the 1970s.
Trendsand Cycles
A convenientpartitioningof the effectson alternativeprofitmeasures
of shiftingtrend and cyclical movementsis providedby the regressions
reportedin table3. The basicformof the regressionsis:
(1) log (profitmeasure)= ki + k2TREN47
+ k3TREN65 + k4TREN70 + k5GAP+ e,
wheree is an errorterm.
12. The lower standard deviations reported in table 2 for the more inclusive
profit measures including interest payments are primarily the consequence of this
reduction in trend.
30
-
a.
See 1945
Sources:
table
I See
PC
for
,
0
appendix.
~~~~~
\,Ip
10
definitions
1950
of
\
the
i
Plgure
Percent
2.
Profit
Share
of
20
A
Income,
profit
Using
/
,I
1955
Alternative
.1
Accounting
1960
_
and
Profit
Concepts,
1965
-~~~~~~~~~~~~~'.
1946-77&
t
/~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1970
measure-
1975
#41~~~#
.-~~~~~~~~~~~~~~~~~~~~~~~~~~
111
MichaelC.Lovell
779
The effect of cyclical forces and stagnationare capturedby the GAP
variable,which is the percentageshortfallbetween actual and potential
The
GNP as estimatedin the 1978 EconomicReportof the President."3
the
coefficientson the GAP variableare alwaysnegative,reflecting general tendencyfor profits,howevermeasured,to deteriorateduringrecession. Profit shares, returnon equity, and profits relativeto the capital
stock all deterioratein periodsof economicslack. The magnitudeof the
GAP regressioncoefficientsindicatesthe relativesensitivityof the various
profitmeasuresto the businesscycle. Those measuresthat are gross of
interestpaymentsare less cyclically sensitivethan the net profit series.
And the after-taxprofit measuresare generallyless cyclically sensitive
thanthe before-taxprofitmeasures.The tendencyof the corporateprofits
tax to smooth the profit stream may reflect a tendency for legislated
changesin corporatetaxesto be appropriatelycountercyclicalratherthan
anybuilt-inflexibilityin the tax structureitself.
The other terms in the regressionall relate to trend. The TREN 47
variable allows for steady compound interest growth or decay since
WorldWar II at an annualrate estimatedby regressioncoefficientk2.'4
The last two trend variablesallow for a possible change in the growth
trendduringthe Vietnamperiodin the 1960s (TREN 65) and againin
the decadeof the 1970s (TREN 70). Coefficientsk3and k4measurethe
extentof the change.To illustrate,in the firstregressionof table 3, there
is a mild downwardtrendup to 1965 of k2 = - 1.2 percentper year.Beginningin 1965, the annualtrendsteepensto k2+ k, = -7.7 percent.In
the 1970s the profitslide is reversed,with the trendrising at an annual
k2 + k3 + k4 = 1.7 percent. The trends for the three periods for each
of the profitmeasuresare recordedat the bottomof the table.
13. Economic Report of the President, January 1978, p. 84. A quarterlypotential GNP series was derived by interpolating the annual figures log linearly. The
Durbin-Watsonstatistics obtained by running regression 1 are extremely small, indicating substantial autocorrelation in profits beyond what can be explained by
cyclical forces. The p coefficient measures the serial correlation of these residuals.
The regressions reported in table 3 are performed on data subject to a CochraneOrcutt transformationto correct for the autocorrelated errors. Because 1 is a descriptive rather than a structural equation, the t-statistics reported directly below
the regression coefficientsmust be interpretedwith caution.
14. The TREN 47 variable is 0.25 in the first quarterof 1947, 0.5 in the second
quarter, 0.75 in the third, and so on. That is, for the qth quarter of the yth year
TREN 47 = q/4 + y - 47. Because the log of the profit measure is the dependent
variable, coefficientk2 is the annual rate of growth compoundedcontinuously.
K~~~~wCh
Ch
q
%O
Wt I
I
oo
o}
T"
,T
el
N
0
en
en
r-
C?
c
?
0
r-
eq 0
m
8
N
4
4
e on 'T
r-
O O
~~~~o
e o
"-
Iuot'TT
q
tn
'T
4-
_
2
0
)
%O0- ,
)
Ne
d~
0
o I I I I I _III
en 0
r-
8 0 00
c
it
00
O
0 O0
O
en
en
0
" ot*
0
0 F
%O %
XO o0
en
t00 _
t-
C
o-
I
_I
Wo#
0
oq
t-
en
m
ON 0 0
0Ien
I
I
4O-4
0
-
oo?o^-X@
o
Wo
^
v)
I~ I
00
o
~~~~~"t
%O
en
O~~~~~a
_ I I I II_en
O:
N
I_
II
d Q n OO-4O
00e 0
s
r
d
qN
>I
I
o-2FooB
O OOO
m
en
ON
00
No
en~~~~~~~~
en
o
^
m
Ot
~
m
?
^oo
N
t
t
t
X
^
N
N
?N
o
WI en
ON
C
~~~~~~~~~~~~~~~~~~~~~O
^ m
^
F
O
*- o
.e00
^
(aUE
4
d- O m oo l#~n
^ d- N m oo m N N^ q
^ o oo o o o F o NOoo^
o_X
en '
..
en
0
00
..
N
...
7t
o
'*- o
-
CI*
0
as
-
o~
o
4.
C
t,
0
o0
V_o_4O_otOOoOOo,
o
N I I I I _III
oooO
a.
e 0
o
4)j a
>
~~~~~~~~~~~~~00
^
00^O000ox?
en 0
C4~~~~~~~~~~~~
~o
as
C
ON
O;OzOt?OOm
e
sXk+
$
tn
'T
'T
0 W.b
r
4: *:g+ +
00
c:
j
,
:
?
~oe
4
~~~~~~~~c
t;
gE
k4
782
BrookingsPaperson EconomicActivity,3:1978
The regressionspresentedin table 3 help in describingshiftingprofit
trendsnet of the cyclicalforcescapturedby GAP. Severalpointsdeserve
attention.First, most profit measuresshow a long-rundecliningtrend
over the decades(k2 < 0), but thereare threeexceptionsto this observation. The exceptions-PA Ta, PIATa, and RPIATa-are all after-tax
profitmeasures,and all arenet of the IVA and CCAcorrections.Because
interestpaymentsexpandedsecularly,the grosssharemeasures,including
interest as well as profits, tend to deteriorateless over the long run.
Second, the Vietnam period steepened the declining trend in profits
(kI2+ kI3< 0). Two exceptionsto this depressingsqueezeare the FTCSEC manufacturingseries,ROEm andROEATm,both of whichare unadjustedfor the distortionsof inflation.This meansthat if the profitmeasureROEATmbelongsin the wage equation,then the profitsqueezedid
not help to mitigateinflationarywage pressures.Third,the 1970s generally have not been characterizedby a furtherdeteriorationof the profit
picture.Only one of the fourteenmeasures,PATa, has been subjectto
continuingdeterioration.All profitmeasuresusingconventionalaccounting show particularlyvigorousgrowthon a cyclicallycorrectedbasis.
Productivityand Inflation
A thoroughunderstandingof profitmovementsrequiresthe investigation of pricing strategy,wage determinants,material and energy cost
movements,and the forces underlyingthe changingrate of productivity
growth.Such an investigationis beyondthe scope of this report.But the
proximatecausesof profitvariationsmay be obtainedby augmentingthe
profit-trendregressions.The ad hoc regressionsin table 4 add the annual
ratesof productivitychangeand inflation.15
Changesin labor productivitywould have no effect on profitsif cost
savingswere immediatelypassed throughto consumersin lower prices
or absorbedby largerwage increases.The regressionsin table 4 reveal
that this is not what happens;rather,increasesin labor productivityimprove profits,with the exceptionsof the two measuresof manufacturing
returnon equity, ROEm and ROEATm. The regressionsalso suggest,
15. The nonfarm implicit price deflator and average labor productivityare used
to measure inflation and productivity because these series are available quarterly
for the entire period since WorldWar II.
MichaelC. Lovell
783
againwith two exceptions,thatrisingpricesaugmentprofits.'8If priceincreaseswere no more than an immediatepass-throughof concurrentincreasesin labor and other costs, the rate of changein priceswould have
a negligibleeffecton profits.The positivecoefficientsof the price-change
variableestablishesthat this is not what happens;in fact, priceincreases
contributeto higherprofits.It is not surprisingto observethatchangesin
the rate of productivitygrowthand inflationinfluencethe profitpicture.
Whatis most intriguingaboutthe regressionsin table4 is thatso manyof
the trendvariablesretain a majorrole in explainingmovementsin most
of the profit measures, even when the changing rate of productivity
growth,inflation,andthe GAP areincludedin the regressions.If the trend
movementswere generatedby these variables,addingthemto the regressions wouldhave causedthe trendexplanatoryvariablesto dropout. But
the trendcoefficients,althoughmodulated,remainstrongin table 4.17 It
must be concludedthat althoughthe slowdownin productivitygrowth,
inflation,and the GAP have all had an impacton profits,these variables
do not sufficeto explainthe dramaticdropin profitsduringthe latterhalf
of the 1960s and the recoverytrendof the 1970s.
Table 5 illustratesthe relativeimportancefor profitabilityof the business cycle, the slowdownin productivitygrowth,and inflationfor one of
the profitseries.The firstcolumnis the actualPla series,the profitsplus
interestshare net of the IVA and CCA. The remainingseries are synthetic. The first syntheticseries shows how the Pla profit share would
have moved in the absenceof the businesscycle. The GAP adjustment
smoothsthe recession-inducedtroughsin the profitshare.But while part
of the movementin the profitshareis cyclicallydetermined,the cycle is
not the entire story of the changingprofit share. The next column adjustsfor the effect of the slowdownin productivitygrowthas well as the
16. The coefficients on price changes are larger in the regressions with conventionally measured profits than in those with the IVA and CCA adjustedseries; this
is presumably because the adjusted measures give a conservative picture of profits
relative to the others in times of inflation.
17. Daniel M. Holland and Stewart C. Myers, in examining "Trends in Corporate Profitabilityand Capital Costs," a study to be published by the Committee
for Economic Development, estimate trendsin the rate of returnon the capital stock,
which they calculate as the ratio of profits plus interest net of the IVA and CCA to
plant and equipment plus inventories. They obtain a t-statistic on trend of -0.26
for after-tax profits, which resembles the RPIATa profit measure; for their beforetax measure, which resembles RPla, their t-statisticis -1.26. They do not allow for
bends in their trendline.
p
Period
Degrees
of
A2
Log
p
Summary
Log GAP
log
ALP
log
ALP(-4)
Durbin-Watson
statistic
p(-4)
freedom
and
Table
TRENTRENTREN
4.
Variable
Intercept
65
70
47
Independent
summary
Effect
of
variable
statistic
Shifting
108
1949:1 0.801.420.898
(1.32) (6.10) (-6.67)(2.49) (-2.96)(-1.87)(47.85)
3.050
0.6737 2.4210 -0.02570.0628-0.0558-0.0093
log
Pa
Trend,
-1977:3
Cyclical
log
1080.831.54
1949:1
0.784
(3.03) (6.85) (-6.59)(2.08) (-2.15)(-1.24) (-29.11)
-1.549
1.1660 2.0370 -0.01950.0434-0.0339-0.0052
Pla
-1977:3
Movements,
log
1080.681.50
1949:1
0.517
(-1.70) (3.88) (-1.01)(0.22) (-1.19)(0.36) (15.21)
2.151
-2.58604.7090 -0.01060.0124-0.04840.0038
-1977:3
Productivity,
PATaProfit
and
log
108
1949:1 0.801.760.476
(-0.73) (4.22) (-0.64)(0.06) (-0.51)(0.97) (-21.47)
-2.377
-0.64392.8980 -0.00430.0028-0.01670.0084
measures
Inflation
on
PIATa
-1977:3
(dependent
Alternative
108 1.77
1949:1 0.80 0.908
(3.44) (4.67) (-7.05)(3.66) (-2.53)(-3.78) (53.95)
3.193
1.6380 1.7280 -0.02530.0857-0.0443-0.0175
log
Pc
variables)"
Profit
-1977:3
log
108
1949:1 0.801.760.748
(4.46) (5.15) (-6.26)(2.91) (-1.53)(-3.26) (-27.97)
-1.417
1.8270 1.6380 -0.01930.0583-0.0230-0.0129
Plc
-1977:3
log
1080.881.82
0.67
1949:1
-1977:3
(2.15) (4.11) (-3.99)(2.21) (-1.13)(-2.12)(-17.70)
-2.025
1.34501.9730 -0.01980.0945-0.0377-0.0196
PATc
Measuress
p
a.
.R2
Log
log
Logp
Seeappendix Period
Degrees
Sources:
Summary
ALP
for
of
notes
log
to
Durbin-Watson
ALP(-4)
Derived
sources
statistic
table
of
freedom
p(-4)
3. from
the
text
basic
108
data.
equation
1949:1 0.881.780.654
1,
-1977:3
GAP
TRENTRENTREN
Variable
Intercept
65
70
47
log
(-2.52)(1.21) (-0.08)(-1.47)(-21.31)
(2.92) (4.56)
-2.028
1.51901.8270 -0.01050.0432-0.0022-0.0113
PIATc
expanded
to
log
108
0.841.36
0.899
include 1949:1
(2.63) (-4.34) (-5.45)(0.71) (0.52) (-2.67)(49.10)
1.1920-1.5180 -0.01920.01900.0106-0.01463.360
-1977:3
ROEm
average
log
labor
108
1949:1 0.950.860.821
(2.22) (-4.00) (-4.02)(1.14) (0.32) (-1.97)(23.94)
0.8477-1.1760 -0.01260.04040.0092-0.01812.785
-1977:3
ROEATm
productivity,
ALP,
and
log
109
1948:1 0.871.520.898
the
(-9.20)(1.23) (-1.59)(-1.16)(38.31)
(3.93) (6.38)
2.997
1.70602.3090 -0.03460.0404-0.0379-0.0072
RPla
-1976:4
implicit
log
price
109
1948:1 0.861.780.603
deflator,
-1976:4
p,
(-2.47)(-0.26)(-0.23)(0.13) (14.88)
(0.29) (4.57)
0.24933.3140 -0.0185-0.0163-0.01060.00162.223
RPIATa
both
log
for
109 1.68
1948:1 0.84 0.891
the
(-8.22)(2.04) (-1.12)(-2.87)(40.71)
(5.52) (4.50)
3.165
2.78301.9060 -0.03530.0682-0.0265-0.0175
RPIc
-1976:4
nonfann
(business
sector.
1948
1090.891.65
0.561
:1
-1976:4
See
log
(-4.99)(1.10) (-0.10)1.91) (-16.84)
(4.31) (4.28)
-2.018
2.60202.1430 -0.02620.0536-0.0036-0.0184
RPIATc
786
BrookingsPaperson EconomicActivity,3:1978
Table 5. PIa Shareof Profits,Actualand withCycle, Productivity,andInflation
Adjustments,1949-77a
Percent
Synthetic
Adjustedfor
businesscycle
Adjustedfor
Adjustedfor
cycle and
cycle,producproductivity tivity,and inflation
Year
Actual
1949
0.22
0.26
0.27
0.26
1950
1951
1952
1953
1954
0.24
0.24
0.21
0.19
0.19
0.26
0.23
0.20
0.19
0.20
0.26
0.24
0.21
0.19
0.21
0.26
0.22
0.20
0.19
0.21
1955
1956
1957
1958
1959
0.22
0.20
0.19
0.17
0.20
0.22
0.20
0.20
0.19
0.21
0.23
0.21
0.21
0.20
0.22
0.22
0.20
0.20
0.20
0.21
1960
1961
1962
1963
1964
0.18
0.18
0.20
0.21
0.22
0.20
0.20
0.21
0.22
0.22
0.21
0.21
0.22
0.23
0.23
0.21
0.21
0.22
0.22
0.22
1965
1966
1967
1968
1969
0.23
0.22
0.21
0.20
0.18
0.23
0.21
0.20
0.20
0.18
0.23
0.22
0.21
0.20
0.19
0.23
0.21
0.20
0.19
0.17
1970
1971
1972
1973
1974
0.15
0.16
0.17
0.16
0.14
0.16
0.17
0.17
0.16
0.15
0.17
0.18
0.18
0.17
0.16
0.16
0.17
0.17
0.16
0.14
1975
1976
1977
0.16
0.17
0.17
0.19
0.20
0.19
0.20
0.20
0.19
0.17
0.19
0.18
Average
0.193
0.202
0.210
0.199
Standarddeviation
0.028
0.026
0.026
0.029
Sources: Actual, see appendix; synthetic, see text note 18. Annual averages were computed from
quarterly data.
a. See table 1 for definitdonof Pla.
MichaelC. Lovell
787
cycle; the last series adjustsfor productivity,cycle, and the estimated
effect of inflation."8Inspectionof these last two columns suggeststhat
changesin labor productivityand prices have only a minorinfluenceon
the historicalmovementsof the profitshare.The markeddeclinein profits
duringthe latterhalf of the 1960s and the partialrecoveryin the 1970s
remainseven afteradjustmentfor the effectsof the GAP, the slowdown
in productivitygrowth,and inflation.
Summary
The variousprofitmeasuressurveyedin this report displaydiversity
in movementover the years;they are not highlycorrelated.Certaincommon themesnonethelessstandout. First,by all measures,the profitability
of nonfinancialcorporationsdeclinedgraduallyover most of the period
since WorldWarII. Second,by almosteverymeasure,profitswere decisively squeezedduringthe last half of the 1960s. Thus the VietnamWar
period pushed down profitability.Third, profitabilityhas been on a recoverypath in the 1970s, althoughthe degreeof recoupmentvariesconsiderablyamongthe variousmeasures.
The admittedlyad hoc regressionsdescribingthe movementsof the
variousprofitmeasuresleave a majorpuzzle for futureinvestigation.It
is not surprisingto find that profits are sensitiveto cyclical forces and
thus subjectto deteriorationin slack times. As expected,a slowdownin
productivitygrowthis bad for profits.And thereis no surprisein the finding that rising prices in the corporatesector help profits.The puzzle is
that the cycle, fluctuationsin productivitygrowth,and price movements
do not explainmorefully the trendmovementsin the variousprofitmeasures.The puzzlemay arisebecausethe bunchingof investmentexpenditureshas implicationsfor subsequentprofitmovementsthat are not fully
capturedby the GAP and productivityvariables.The trends,including
the squeezein the late 1960s, may be associatedwith variablesthat can
be identifiedin subsequentwork but that are not includedin the regres18. The syntheticprofitsharesare calculatedfrom the regressioncoefficientsfor the
Pla series in table 4. The cyclicallyadjustedshare is
CAS = exp (log PIa + 0.01656GAP);
and
the cyclically
productivityadjustedshareis
=
CPAS exp flog CAS- 3.484[logALP - log ALP(-4) - 0.0251;
and the cyclically, productivity,and inflation adjustedshare is
CPIAS = exp flog CPAS - 0.9438 logp - log p(-4)]j.
BrookingsPaperson Economic Activity,3:1978
788
sions in table 4. My suspicionis that a betterexplanationof these profit
trendswill requirea more detailed structurallook at the way in which
pricesrespondto changesin labor and energycosts.
APPENDIX
Data Sources
ALLDATA
arefromthe NBER (NationalBureauof EconomicResearch)
data bank, except where otherwisenoted. NBER symbols are used in
explainingthe constructionof the profitseries. The profitseries are for
the nonfinancialcorporatesector, except for ROEm and ROEATm,
which cover only manufacturing.
Profit
measure
Pa
Pla
PATa
PIATa
Construction
Corporateprofits with IVA and CCA/domestic income of nonfinancial
corporations(GJJVA/GJY)
Corporateprofitsplus interest/domesticincome(GJJVA+ GJINT)/GJY
Corporateprofitsafter tax/domestic income (GJJVA - GJTAX)/GJY
Profits and interest after tax/domestic income (GJJVA + GJINT
-
GJTAX)/GJY
Profitshare,businessmens'convention(GJPBT/GJYBCwhereGJYBC=
GJPBT + GJINT + GJCOMPis conventionallymeasuredincome)
PIc
Profitplus interestshare(GJPBT + GJINT)/GJYBC
Profitsafter tax, businessmen'sconvention (GJPBT - GJTAX)/GJYBC
PATc
PIATc
Profits and interestafter tax (GJPBT + GJINT - GJTAX)/GJYBC
Before-tax return on equity, approximatedby calculating ROEBT
ROEm
ROEAT* GJPBT/(GJPBT- GJTAX)
ROEATm After-tax return on equity, FTC,19 smoothedin accordancewith Perry's
procedure
RPIa
Profits with IVA and CCA/capital stock (GJJVA + GJINT)/IPXQ,
where IPXQ is Musgrave'sestimate of equipment plus structuresof
nonfinancial corporations (Survey of Current Business, April and
August 1976, and August 1977)
RPIATa
Profitsplus interestaftertax/capitalstock (GJJVA+ GJINT - GJTAX)/
IPXQ
RPIc
Profits, businessmens'convention/capital stock (GJPBT + GJINT)/
IPXQ
RPIATc
Profits, businessmens'convention, plus interest after tax/capital stock
(GJPBT + GJINT - GJTAX)/IPXQ
19. After-tax return on equity is from Federal Trade Commission-Securitiesand
Exchange Commission, Quarterly Financial Report for Manufacturing Corporations, and (since 1971) FTC, QuarterlyFinancial Report for Manufacturing,Mining
and TradeCorporations,variousissues.
Pc
8
MichaelC. Lovell
789
Discussion
THEPANEL discussedsome of Lovell's alternativeprofitmeasures.John
Shovensaid thatprofitmeasuresthatfail to includeinterestpaymentsare
biased. And the bias is most serious in those measuresthat purportto
adjustfor inflationbecausethey adjustonly the assetside of the accounts.
Inventoriesanddepreciableassetsarecorrectedfor the effectsof inflation
while capitalgains on the firm'sliabilitiesare neglected.Shovenbelieved
this mightexplainwhy the seriesthatincludeinteresthave smoothertime
trends.Lovell concurred.ArthurOkun found the returnon equity the
least defensiblemeasureof profitabilitybecausethe denominatorhad no
clear meaningin a time of changingprices. LaurenceSeidmannoted,
however,that the after-taxrate of returnon equityshowedno evidence
of a squeezeduringthe past decadewhen wages acceleratedand argued
thatit, ratherthanthe adjustedindicatorsof profitabilitydevisedby economists,mighthave been relevantto wagedetermination.AlthoughLovell
found that argumentplausible,he reportedthat the returnon equityhad
not performedas well in wage equations as conventionallymeasured
profitsplus interest.Albert Rees said there was little evidencethat any
measureof profitabilitybelongedin a wage equation.He believedprofits
proxiedfor other cyclical variableswhen they were used in econometric
wageequations.
Shoven discussed Lovell's complaintthat constant-dollarFIFO accountingdid not measurethe gainson speculativeinventoryholdings.He
arguedthat no single systemof accountingcould serve all measurement
purposes. He reasoned that cost accountingshould give a backwardlookingsummaryof whathad happenedto the firmratherthan a view of
prospectiveprofitability.Constant-dollarFIFO leads approximatelyto a
measureof the real gain fromholdinginventoryin the past.
Okun questionedLovell's interpretationof the price-changevariable
in his profitsequationas measuringthe effectof inflation.It was not surprisingthat a risingdeflatorfor the corporatesectorimprovedprofitability. But whethergeneralprice inflationdid so would dependon whereit
originated.Inflationthat camefromsupply-sidedisturbancesto materials
costs mightnarrowprofitmarginsinstead.Lovell repliedthatdistinguishing the sourceof the inflationwitha profitequationwouldbe difficultand
that addingmaterialscosts to an equationwith prices and productivity
virtuallyreducedit to an identity.