Consumption and the recession of 1990-1991

Digitized

by

the

Internet

Archive

in

2011

with

funding

from

Boston

Library

Consortium

Member

Libraries

working

paper

department

of

economics

massachusetts

institute

of

technology

50 memorial

drive

Cambridge,

mass.

02139

Consumption and

the

recession of 1990-1991

Olivier Blanchard

MIT and

NBER

No.

93-5

Jan.

1993

J.T.

UBRAR

AUG

1

2

1993'

Consumption

and

the recession of

1990-1991.

Olivier

Blanchard,

MIT

and

NBER

*

January

15,

1993

Abstract

The

paper looksatthe recession of1990-91. Itreaches twoconclusions:

Notall recessions have thesamedynamics. Decreases in output which

come

from

other sourcesthanconsumptionshocks tendtobeshortlived,andfollowed by sharp

recoveries. Decreases in output which

come

from consumption shocks, decreases

in consumption given income, tend to be longer lasting, and followed byweaker

recoveries.

In contrastto its predecessors, the recession of 1990-91 was causedprimarilybya

consumption shock. Thus, conditionalon the natureofthose shocks, the slow

re-coverycomesasnosurprise. Asoftheendof1992,theeconomywasnearly exactly

onthe trackgivenby adynamicforecastbasedoninformation uptothefirstquarter

Recession 2

Duringthelasttwoquarters of1990andthefirstquarter of 1991, the

US

economy

experienced negativegrowth.

These

three quartershave

now

been designatedas

the 1990-1991 recession by the

NBER.

But both in the year preceding

and

the year following, growth was

anemic

as well.

Growth

since the beginning of 1989

hasaveraged

0.7%

atan annual rate, afarcry from its post-1973

mean

of 2.2%.

Incontrast toitspredecessors, this recessiondoes not have an obvious proximate cause.

And

precisely because of that, explanations abound, ranging, to cite only the leading candidates, from a longexpansiondying ofold age, to

consumer

de-pression, a

mix

ofdebt overhang

and

therealizationoflowergrowthprospects, to

thecreditcrunch, thecombinationofimprudentbehavior byfinancial institutions and the tightening ofregulation, to the

end

of the

Cold

War

and thedecrease in

defense spending, tostructural adjustmentsrequiredby globalcompetition*.

The

purpose ofthis paper is to look at the data, putting just

enough economic

structure

on

the econometricstopinpoint ifnot the deep, atleast the proximate

causes of the recession.

The

story

which

emerges is relatively clear. By far, the

main

proximate cause of the recessionwas a "consumption shock" a decreasein

consumption

in relation to its normal determinants. Because the effectsofsuch shocks are longlasting, this also explains why, incontrast topreviousrecoveries,

the recoverywas aslow and

weak

one.

The

issueleft

unanswered

iswhether this

"consumption shock" was

due

to animal spirits/taste shocks, orsimply to

antici-pations of the slowgrowth to come. Circumstancial evidencepoints to a role of

Recession 3

I.

Looking

at the

Components

of

GNP.

A. Estimating a

VAR.

A

simplefirstpassis toestimate the jointbehaviorof the

components

of

GNP, and

look at the residuals. I thus

decompose

real

GNP

(Y), as the

sum

of

consump-tion

on

non

durables

and

services (C),

consumption

expenditures

on

durables

(CD), residential investment (IR),

non

residential (INR) investment,

govern-ment

spending (G), inventory investment (1NV)

and

net exports (NX),

lb

achievestationarity, I log-difference C,

CD,

IR,

INR

and

G,

and

divide

INV

and

NX

by trend output, obtained byfitting abroken exponential trendto

GNP,

with

a breakin 1973.

I then run a

VAR

for the seven transformed variables, for the period 1959:1 to

1992:3, with 3 lags ofeach variable, a constant and a post-73

dummy

in each

equation. Issuessuchastreatment oftrend,or the incorporation of cointegrating

relations, while important if

we

were to look at impulse responses, are oflittle

import

when

the focusis, as here,

on

the residuals toeachequation.

Given

the 7 estimated equations,an auxiliary equationis

needed

tocharacterize thebehaviorof

GNP

and

itsresidual.

Given

thenon-linear transformations of the individualseriesusedtoinducestationarity, there aretwo waysofdoingthis.

The

first is to construct

GNP

residuals for each period as a weighted average of the

residuals ineachequation, multiplied by the time varying ratiosofeach

compo-nentto

GNP.

The

otheristo regress log-differenced

GNP

on

thesetofright

hand

side variables of the

VAR,

and

use the residuals from this equation. This second approachyields resultsvery similartothe first,

and

issimpler to implement. This

Recession 4

B. Identifying the Shocks.

Denote

the residuals to each equation by lower case letters, for

example

c for

the

consumption

equation, yfor the-auxiliary-outputequation.

These

residuals are simply forecast errors.

They

are, not surprisingly, generally positively corre-latedacrossequations. Thisreflectstheir joint

dependence on

common

underly-ing shocks, as well as their direct

dependence on each

other.

To

theextent that

consumption

respondsto

income

within the quarterforexample, partof the con-sumption residual, c,reflectsalsotheshocksto othercomponents.

We

can

make

some

progress

and

get closer to the underlying shocks by

making

two identifyingassumptions.

The

firstassumption isthat, within thequarter, the

components

of

GNP

depend

on

each other only through

GNR

Thus, I assume

that, forexample, cdependsonly

on

y-theresidualto theoutputequation-, not

separately

on

ir,inr

and

soon.

The

second assumptionisthatg,the

government

spendingresidual, isexogenous.

Under

thosetwoassumptions, gcanbe usedasan instrument toestimate theeffectsof y

on

eachof the othersixresiduals.

Denote

the residuals from these instrumentalvariable regressions by e, thus using e_c for

consumption

forexample,

and

referto

them

asthe "shocks".

These

new

residuals are stillcross-correlated, but lessso thantheoriginalones.

The

two identifying assumptions are crude,

and

forecast errors in

government

spending arenot a very powerful instrument. Nethertheless, the estimates of the

contemporaneous

effectsofy

on

its

components

make

sense. Using sample

mean

values togo fromestimatedelasticities to derivatives, the estimates imply that a

one

dollar increasein

GNP

increases

consumption

of

non

durables

and

services

by 12 cents, durable

consumption

by 11 cents, residential investment by-1 cent,

in-Table 1.

The

source ofshocks

Accumulated

normalizedshocks

Quarter

_Uv/°v)

E(Cc/^c)

EO.rMr)

1989: 1 0.00 0.00 0.00 1990:2 2.32 -2.50 0.25 1990:3 1.38 -2.29 -1.15 1990:4 0.79 -4.07 -1.61 1991:1 -0.39 -5.42 -2.87 1991:2 0.18 -5.35 -2.17 1992:3 0.46 -6.12 -0.15

The

estimation of the shocks, the e's, isdescribed in thetext.

The

accumulated

seriesfollow

random

walks with zero

mean

and

unitstandard deviation.

Thus

for

example, starting at 0.00 in 1989:1, the distribution of the

sum

in 1992:3 has

Recession 5

vestmentby -15 cents.

The

next step is then to look at the sequence ofshocks, the €_x's, and look for

unusually large individual shocks, or for sequences ofshocks of the

same

sign.

A

convenient

way

to

do

so is to look at the accumulated shocks to the various

components, each normalized byits estimatedstandard deviation, starting from

some

date

which

Ichoose tobethefirstquarter ofslowgrowth, 1989:2.

The

main

results are presentedin table 1.

The

first

column

gives the

sum

ofnormalized

GNP

residuals, y/cr_y.It

shows

that,

despite the fact that output growth was below its

mean

in 1989

and

early 1990,

residualstotheoutputequationwere, duringthat period,smallbutpositive. Thus,

the

VAR

explains the

slowdown

ofoutput pre-1990:3 through the internal

dy-namics of the

economy

rather than through adverse shocks. It

would

be worth

following thisfinding through; I have not

done

ityet.

The

second

and

third

columns

give the

sum

ofaccumulated shocks to

consump-tion of

non

durables

and

services,

and

to residential investment respectively.

These

arethe onlytwo

components

of

GNP

which

show

largenegativeshocks

(al-though, interestingly,

even

forthose, no single residualislarger inabsolute value

than 2 standard deviations).

A

negative shock to residential investment

domi-nates thefirst quarter of the recession, 1990:3.

Two

negativeshocks to

consump-tiondominate the next twoquarters.

While

the largest shocks take placeduring the recession, thewhole period since 1989:2 is

dominated

by negative shocksto

consumption.

The sum

is still large in 1992:3.

The

rest of the paper focuses

on

Recession 6

II.

Looking

at

Output

and

Consumption

A.

From

7 to 2 Variables.

7-variable

VAR's

are unwieldy, their dynamics hard to understand

and

harder to

describe. Fortunately, examination of the

dynamic

structure of the

VAR

above

suggests a short cut. Tests of the significance ofeach variable in each equation

(which,becauseofspaceconstraints,arenot reported)

show

thestrong predictive

power

of

non

durables and services

consumption

for nearly all

components

of

GNP; and

the

weak

predictive

power

of most other components. This suggests

that

we may

not lose too

much

by focusing

on

a bivariate system in output

and

consumption. Thisis

what

I do in thissection.

I thus estimatea bivariatesysteminthelogarithm of

consumption

of

non

durables

and

services,

and

thelogarithm of

GNR

There

is

weak

evidenceof cointegration,

that the ratio of

consumption

to

income

is stationary. I thus specify the

VAR

in

terms offirstdifferences, allowing forthree lags ofeachfirst-differenced variable,

the lagged value ofthe log

consumption

to

income

ratio, a constant

and

a post-1973

dummy

2.

B.

From

Residuals to

Shocks

Estimation of the system yields residuals for

consumption

and

income, c

and

_y

respectively. 1 follow the

same

approach asbefore to get to

consumption

shocks,

using the unexpected

component

of

government

spending as an instrument ina regressionofc

on

y.Thisyieldsanelasticityof.12 of

consumption

toincome, thus

an increase of

consumption

of7cents within the quarter fora dollar increasein income.

The

"consumptionshock", ec, isdefined astheresidualofthisregression.

1 i

o

i

o

i

O

i

o

o

03

O

•vl IS3 en CD

L

i

O

b

o

p

CO

8

S

m

3

3

I

1

o

1

o

o

o

o

o

o

o

-fc.

o

N3

o

o

o

N3

o

o

en

o

CD U3

O

<n 1 l I l -i i r Tl

o

m

o

>

—1

>

O

—1

c

>

I

-* t 1 i | i | / | I t :; t... . ':/-y

~M

0-0*: /

W+M

::

:M

}\

1 \ 1 \ 1 \ \ \ • \ 1 V 1 \ f \ I \

\

\

\

\

\

\

\

\

\

\ J \ \ v I \ 1 \ \ \ I

\\

\

8

I

I

I

I

>i

cz

70

N>

1

o

1

o

o

o

o

o

o

o

o

o

o

o

o

-b.

o

en

o

CO

V

\ A ':

V

• \ 4 / t

o

70

m

O

>

in

>

o

—i

c

>

O

8

i <>>

m

Recession 7

A

regressionof y in turn

on

ec yields an elasticityof.51 (thus an increase of96

cents in

income

for an increase in

consumption

ofa dollar, a multiplierslightly

below one within thequarter), and an

"income

shock" as the residual, e

y.

Sum-marizing, the

mapping

from residuals toshocks is-givenby:

c

=

A2y

+

e c

y

=

.51ec

+

e_y

An

examination of thesequenceofrealizedshocksgives a picture similarto that

in table 1.

The

first quarter of the recession, 1990:3 is

dominated

by an

income

shock,

which

we

know

from earlier can be traced mostly to

non

residential

in-vestment.

The

next two quarters are

dominated

by

consumption

shocks. This

lastfindingisthekeytounderstanding

why

therecoveryhas

been

soweak.

To

see why, I proceed in twosteps, first looking at the

dynamic

effects ofboth

income

and

consumption

shocks,

and

then looking at

dynamic

forecasts ofoutput as of 1991:1.

C.

Impulse

Responses,

and

the

Weak

Recovery.

The

effects ofone-standard-deviation negative

consumption and income

shocks are

drawn

inFigure 1.

The

figure hastwo

main

features.

The

first is thatshocks to

income

haveonlytransitory effects

on consumption and

income; their effectis

largely

gone

within two years3.

The

second is that shocks to

consumption

have

long lasting,

hump

shaped, effects

on

output,

and

to alesserextent

on

consump-tion4.

Recession 8

for output based

on

the bivariate

VAR

using information up

and

including the

last quarter of the recession, 1991:1 (Thus, both lines are the

same

up to

and

including 1991:1).

What

isstriking isnot only

how

weak

the recovery has been

since 1991, butalsothat a

weak

recoveryisalso

what would

have

been

predicted by the bivariate

model

as of the

end

of 1991:1. Indeed the actual recovery has

beenslightly fasterthanthe

dynamic

forecast... (Note that the

dynamic

forecasts

are independentofouridentificationof theshocks.).

For comparison purposes, Figure 3 performs the

same

exercise for the previous

recession, giving output, actual and forecast as of the last quarter ofthat

reces-sion, 1982:4.

Note

that, in that case, the

same

bivariate

model

largely predicts

the strong recovery

which

followed 5.

Given

our decomposition of the shocks, and the characterization ofimpulse

re-sponses, the key to Figure 2

and

Figure 3 is easily given.

More

so than in

pre-vious recessions, thedecrease in output in 1990-91 was due to adverse shifts in consumption.

Those

shifts have long,

hump

shaped effects

on

output,

and

their

dynamic

effects explain

why

the recovery has been slow.

There

is an important

lesson: recessions are not necessarily followedby fastrecoveries; this

depends

on

thenature ofthe shocks

which

triggered the recession.

HI. Foresight, or

Animal

Spirits ?

I have so far established that the recession was associated with large negative

"consumptionshocks", thatsuch shocks have long lasting effects

on

output,

and

that thisexplains

why

therecovery hasbeenso slow.Thisraises

however

thenext

Recession 9

from ?

And,

inparticular,

where

did those of 90-91

come

from ?

One

interpretation, call it "foresight", isthat

consumption

shocks aresimply the

reflectionof anticipationsbyconsumersof othershocksand theireffect

on

future income.

The

interpretationofimpulse responsesinFigure 1isthenthat

consump-tion shocks are simply followed by thechanges in

income

which

triggered

them

in the first place;

consumption

shocks are a mirror, not a cause.

Another

inter-pretation

however

isthat

consumption

shocksreflect inpart

movements

in con-sumptionnotduetochangesinexpectations offutureincome. Reasons

may

range

from increasing prudence, to changesin intertemporal preferences, to-stepping outside ofthe usual maximizing

model-

sudden realizations ofpast

overborrow-ing, panic, and so on.

Under

that interpretation, call it "animalspirits", impulse responses in Figure 1

show

how

shifts in

consumption

lead through a combina-tion of

dynamic

multiplier

and

accelerator effects to a

hump

shapedresponse of output.

Can

one tell these two interpretations apart ?

To some

extent, one can.

Under

theanimalspiritsinterpretation,

and

under the plausibleassumption thatanimal

spiritshavelittle or

no

long runeffect

on

output, the impulse response ofoutput

to a

consumption

shock should eventually return to zero. This isclearlyviolated

inFigure 1, suggesting that

consumption

shocks must reflect in part foresightof

shocks with

permanent

effects.

The

foresight interpretation

on

the other

hand

imposes constraints

on

the relation of the

consumption

response to the

subse-quent response ofincome, and ofthe shapeof the

consumption

impulse response

itself. Following the route of putting

more

theoreticalstructure

on

thedata totry

to disentangle foresight and animal spirits

would

take

me

too far. I shall instead

Recession 1

nature of the negative

consumption

shocksof 1990-91:

I see theoverall evidence, from the timing of

monthly

declines in the index of

consumer

confidence (from the

Michigan

Surveyofconsumers), in the indexof

leadingindicators,

and

incommercialforecastsof output,aspointingto

more

than

consumer

foresight atwork. First, incontrast to earlierrecessions, the decline in confidence was largely prior to

-and

much

stronger than

would

have

been

pre-dictedby-either thedecline of leadingindicators or commercialforecasts of the

recession.Second, perhapsnotcoincidentally,thefirstlargedeclineinconfidence

inAugust 1990,wasassociated with

an

importantbutlargely

non economic

event, theinvasion ofKuwaitbyIrak. Third, afterhavingdropped,

consumer

confidence remainedverylow inthe following twoyears,

much

lower than

would

have been

predicted

on

thebasisofhistorical relationswith aggregatevariables.

Each

piece ofevidencecan, with

some

effort, be reconciled with theforesight interpretation;

together however,I find

them

suggestive ofa role foradropinconfidence

coming

Recession I 1

References

Cochrane, John, "Permanent and Transitory

Components

of

GNP

and

Stock Prices",

mimeo,

UniversityofChicago,July 1992.

Sinai,Allen,"What's

Wrong

with the

economy

?", Challenge,

forthcom-ing, 1992

Walsh, Carl,

"What

caused the 1990-1991 recession ?",

_mimeo,

Federal

Recession 1

2

Footnotes

*

_Department

_{ofEconomics, MIT, 50}

Memorial

Drive,

Cambridge

MA

02139.

I thank the

NSF

for financial support, Roger Brinner, Ricardo Caballero,

Julio Rotemberg, AllenSinai

and Lawrence

Summers

for usefuldiscussions

and

comments.

1

Two

usefuldiscussionsofpotential causesaregiven byAllen Sinaiand Carl

Walsh

.

2

Such

systemshave

been

estimated by

many

others.

John Cochrane

in par-ticularfocusesalso

on

thedifferent

dynamic

effectsof

consumption and

in-come

shocks. Hisapproach toidentificationis

however somewhat

different

fromthat used in this paper.

3

The

fact that the long run effect ofa negative shock is positive is mildly

embarassing, but thislongrun effectisnot significantlydifferent from zero.

4 Thispoint isalsoemphasized byCochrane.

5 Results for the post -recession episodes of 1971

and

1975 are qualitatively similar to those for 1982, with fast predicted and actual growth following

Date

Due

291994

MIT LIBRARIES DUPl