What is a Recession and Who Decided When It Started?

^Pr^ep^ar^{ed f}^o^r^M^e^m^b^ers^an^d^Com^mⁱ^t^t^ees^of^C^ong^r^e^ss

The National Bureau of Economic Research (NBER) recently announced that the economy had
reached a cyclical peak and that a recession had begun in December 2007. A recession is one of
several discrete phases in the overall business cycle. The term may often be used loosely to
describe an economy that is slowing down or characterized by weakness in at least one major
sector like the housing market. When used by economists, “recession” means a significant decline
in overall economic activity that lasts more than a few months. NBER business cycle dating
committee is the generally recognized arbiter of the dates of the beginnings and ends of
recessions. With all statistics it takes some time to compile the data, which means they are only
available after the events they describe. Moreover, because it takes time to discern changes in
trends given the usual month-to-month volatility in economic indicators, and because the data are
subject to revision, it takes some time before the dating committee can agree that a recession
began at a certain date. It can be a year or more after the fact that the dating committee announces
the date of the beginning of a recession. Just as it took some time for the business cycle dating
committee to determine when the recession began, it is likely to be some time after the fact that
they determine when the recession has come to an end.

Introduc tion ..................................................................................................................................... 1
What is a Recession?.......................................................................................................................1
Who Decides When the U.S. Economy is in a Recession?.............................................................2
How Do We Know When a Recession Starts?................................................................................2
Rhetorical and Analytical Significance...........................................................................................4
Table 1. Dates of Recent Business Cycle Turning Points and the Dates They Were
Announced by the NBER.............................................................................................................3
Author Contact Information............................................................................................................4

On December 1, 2008, the National Bureau of Economic Research confirmed what many had
suspected for some time – that the economy was in a recession. They announced that the
economy had reached a cyclical peak, and that a recession had begun in December 2007. The
term “recession” may often be used loosely to describe an economy that is slowing down or
characterized by weakness in at least one major sector like the housing market. When economists
use the term, however, they try to do so consistently. Recessions typically have common
characteristics and so economists try to identify the beginning and ending dates of recessions in
order to further their overall understanding of the economy.

A recession is one of several discrete phases in the overall business cycle. The beginning of a
recession is known as a business cycle “peak,” and the end of a recession is referred to as a
business cycle “trough.” In 1946, Arthur Burns and Wesley Mitchell published a study of
business cycles and offered a definition intended as a guide for further study:
^B^u^sⁱⁿ^e^s^s^c^y^cles^{are a t}^y^{pe of}^f^l^u^c^tu^a^tion^f^oun^{d in}^th^{e agg}^r^eg^a^t^{e econ}^o^mⁱ^{c activ}^it^y^ofⁿ^a^tⁱ^oⁿ^s
^t^h^at^org^aⁿⁱ^{ze t}^h^ei^r^w^o^rk^m^aⁱⁿ^l^yⁱⁿ^b^u^sⁱⁿ^e^s^s^eⁿ^t^e^rpri^s^e^s^:^a^c^y^cl^{e cons}ⁱ^s^t^s^of^e^x^pansⁱ^ons
^o^ccu^r^rⁱⁿ^g^{at ab}^o^u^{t th}^{e sa}^m^e^ti^m^eⁱⁿ^m^aⁿ^y^ecoⁿ^o^mⁱ^{c activ}ⁱ^tie^{s, f}^o^llo^w^e^d^b^y^si^m^ilar^l^y^g^eⁿ^e^r^a^l
^reces^sⁱ^ons^{, con}^t^ractions^{, a}ⁿ^d^revⁱ^v^a^l^s^w^h^ic^h^m^e^rg^eⁱⁿ^{to t}^h^e^ex^pans^ion^p^h^as^{e o}^f^t^h^eⁿ^e^xt¹
^c^y^cle.
This definition requires both expansions and recessions to be apparent in many economic
activities at about the same time, which would seem to exclude an economy exhibiting weakness
in a single market.
More recently, economists at the National Bureau of Economic Research (NBER), issued a memo
with a slightly more precise definition:
^A^reces^sⁱ^on^is^{a s}ⁱ^gn^ifican^{t dec}^lin^{e in}^ecoⁿ^o^mⁱ^{c acti}^v^it^y^s^p^{read acros}^s^th^{e econ}^o^m^y^{, las}^ting
^m^o^re^th^aⁿ^a^f^e^w^m^oⁿ^t^hs^,ⁿ^o^r^m^all^y^vⁱ^sⁱ^{ble i}ⁿ^real^GDP^{, real i}ⁿ^com^e^,^e^m^plo^y^m^eⁿ^t^,ⁱⁿ^d^u^s^t^rial
^produ^ction^,^an^d^w^h^oles^ale^-^ret^a^{il s}^a^les^.^A^reces^sⁱ^on^be^gⁱⁿ^s^j^u^s^t^af^{ter t}^h^{e econ}^o^my^reach^es^a
^peak^of^activit^y^aⁿ^{d en}^ds^as^t^h^{e econ}^o^m^y^reac^h^e^s^its^tro^u^gh^.^B^e^t^w^eeⁿ^trou^gh^an^{d peak}^{, t}^h^e
^econ^o^m^yⁱ^sⁱⁿ^aⁿ^ex^paⁿ^s^ion^.^E^x^pan^s^ion^is^t^h^{e n}^o^r^m^a^l^s^t^{ate of}^th^{e econ}^o^m^y^;^m^o^s^t^reces^sⁱ^on^s²
^{are brief}^an^{d th}^e^y^h^a^v^e^been^{rare in}^recen^{t decades}^.
This is the generally accepted view among economists of what constitutes an economic recession.
There is also a commonly cited “rule of thumb” that is referred to in the press. That rule is that a
recession is two consecutive quarterly declines in real gross domestic product (GDP). But this
rule does not always apply. For example, there was a recession beginning in March 2001 and
¹
^A^r^thur^{F. B}^u^rⁿ^s^a^{nd W}^e^s^l^ey^C^.^Mitc^he^ll,^Me^a^s^ur^ing^B^u^sⁱ^ne^s^s^C^y^c^l^e^s^{, N}^a^tiona^{l B}^u^r^e^a^u^of^Ec^on^omⁱ^c^R^e^s^e^a^r^c^h^{, 194}^6,
^{p. 3}^.
²^Busine^{ss Cy}^c^l^e^Da^ting^Co^m^m^itte^e^,^Me^m^o^{, Na}^tiona^{l Bure}^a^u^of^Ec^o^nom^ic^Re^se^a^r^c^h^{, Ja}^{n. 7}^,²⁰⁰⁸^,³^p^p^{. A}^v^a^ila^ble^on
^the^inte^rⁿ^e^t^a^t^h^ttp:^//w^w^w^.nbe^r^.^or^g^/^cy^c^l^e^s^/ja^n08bc^dc^_m^em^o.htm^l^.

ending in November 2001 that was not characterized by two successive quarterly declines in real
GDP.
In any case, an important distinction is that a recession is a period of declining output and not just
a period of slower economic growth. It is possible for GDP growth to be positive yet so slow that
the unemployment rate rises. This is sometimes referred to as a “growth recession.”

Among economists, the NBER is the generally accepted arbiter of business cycle turning points.³
The NBER is a private nonprofit and nonpartisan organization that was founded in 1920. In the
beginning its focus was on the macroeconomy, business cycles, and long-term growth, but now it
seeks to promote research on a wide variety of topics. For many years, the NBER itself
determined the dates of swings in the business cycle. In 1978, however, a separate business cycle
dating committee was formed. The members of the committee are appointed by the president of
the NBER, and they are now responsible for determining the dates of the beginnings and ends of ⁴
recessions. The current members of this committee are
• Robert Hall, Chair – Director of NBER’s Program of Research on Economic
Fluctuations and Growth, and Professor at Stanford University;
• Martin Feldstein – President Emeritus of NBER, and Professor at Harvard
University;
• Jeffrey Frankel – Director of NBER’s Program in International Finance and
Macroeconomics, and Professor at Harvard University;
• Robert J. Gordon – NBER Research Associate, and Professor at Northwestern
University;
• James Poterba – NBER President, and Professor at MIT;
• David Romer – Professor at the University of California, Berkeley;
• Victor Zarnowitz – Senior Fellow at the Conference Board, and Professor
Emeritus at the University of Chicago.

One important indicator of economic conditions is growth in real gross domestic product (GDP).
GDP statistics are compiled each quarter, and thus there is a time lag between the first month that
is reflected in the data and the release of the data. For example, the first release of data for the
first calendar quarter of a given year does not occur until late April. The data from that release is
³
^T^h^e^{NBER w}^e^bsite^{is a}^t^http://w^ww^.nbe^r.org^.
⁴^A^w^o^r^k^ing^pa^pe^r^pu^blis^he^{d by}^th^e^B^u^r^e^a^u^of^Ec^on^om^ic^Aⁿ^a^l^y^s^is^f^oun^{d t}^h^a^t^“^[^t^]^he^N^B^ER^da^ti^ng^c^o^m^m^itte^e^’^s
^m^e^thod^olog^y^a^ppe^a^r^s^{to be}^v^e^r^y^r^obus^t^.^”^Se^e^B^r^uc^e^T^.^G^rⁱ^m^m^,^“A^l^t^e^r^na^tiv^e^Me^a^s^ur^e^s^of^U^.^{S. Ec}^on^o^m^ic^A^c^tiv^it^yⁱⁿ
^B^u^sⁱ^ne^s^s^C^y^c^l^es^a^{nd B}^u^sⁱ^ne^s^s^C^y^c^l^e^D^a^ting^,^”^B^u^r^e^a^u^of^Ec^on^om^ic^Aⁿ^a^l^y^sⁱ^s^W^o^r^k^ing^P^a^pe^r²⁰⁰⁵^-⁰⁵^,^A^u^g^.²⁰⁰^5.

subject to revision in each of the next two months and may be revised later on as well. It is not
inconceivable that a first release of data that showed a decline in real GDP would later be revised
to show an increase. Even so, those using the rule of thumb that two successive quarterly declines
in real GDP constitutes a recession would have to wait for the release of the second quarter data
in August to establish that a recession began at the start of the year.
Because of the time lag associated with the release of GDP data, and because business cycle
turning points are associated with months rather than quarters, the dating committee relies on a
number of monthly economic indicators. Among the more important monthly indicators the
committee looks at are personal income, employment, and industrial production. Even in the case
of monthly indicators, it may require several months of data to establish a change in trends.
When there is a recession, not all of the economic indicators may show a change in trend at the
same time. Historically, some indicators, such as housing starts and the stock market, tend to slow
or decline in advance of a recession, and some, like the unemployment rate tend to react to
changing conditions with a lag.
With all statistics it takes some time to compile the data, which means that the statistics are only
available after the events they describe. Moreover, because it takes time to discern changes in
trends given the usual month-to-month volatility in economic indicators, and because the data are
subject to revision, it takes some time before the dating committee can agree that a recession (or
an expansion) began at a certain date. Table 1 shows, for recent business cycle peaks and troughs,
the date of the turning point and the date when the committee issued a release identifying the date
of the turning point.
Table 1. Dates of Recent Business Cycle Turning Points and the Dates They Were
Announced by the NBER
^{Date Turning Point was}^{Months After Tu}^rning
^{Turning P}^oint^{Date of Turnin}^{g Point}^annou^{nced by}^NBER^Point
^peak^Janua^{ry 1}⁹⁸⁰^June^{3, 1}⁹⁸⁰⁵
^trough^{July 1}⁹⁸⁰^{July 8, 1}⁹⁸¹¹²
^peak^{July 1}⁹⁸¹^Janua^{ry 6,}¹⁹⁸²⁶
^trough^November¹⁹⁸²^{July 8, 1}⁹⁸³⁸
^peak^{July 1}⁹⁹⁰^{April 25, 1}⁹⁹¹⁹
^trough^March¹⁹⁹¹^December^{22, 1}⁹⁹²²¹
^peak^March²⁰⁰¹^{November 26, 20}⁰¹⁸
^trough^November²⁰⁰¹^{July 1}^{7, 20}⁰³²⁰
^peak^{December 20}⁰⁷^{December 1,}²⁰⁰⁸¹²
^Source:^{National Burea}^{u of E}^{conomic Research.}
The longest delay between the beginning of a new phase of the business cycle and its
announcement was when a recession was found to have ended in March 1991 but was not
announced until 21 months had passed. The shortest delay was five months after the expansion
ended in January 1980. Of the eight examples shown here, four were not announced until at least
a year had elapsed.

While all recessions have unique characteristics, they also have many common aspects. Thus they
are the object of economic analysis both individually and collectively. Because they are
undesirable, economists study them in the hope that they can advise policymakers how to avoid
them. To do so, it is important to agree on a chronology, and it may not be an inconvenience to
economists that it takes time to establish one.
Policymakers, on the other hand, are more concerned with the present and the immediate future.
If they hope to avert or mitigate the consequences of recession, they cannot wait for an “official”
declaration. By then the recession may be history.
Although there can be a significant delay between the onset of a recession and the dating
committee determination, there is often little doubt that the economy is, or has been, in recession ⁵
well before the announcement. For policy measures to have mitigating effects, they must be
timely. Policymakers may not have the luxury of holding themselves to as strict a definition of
recession as economic analysts.
^B^r^ian^W^.^Cash^ell
^Specialis^{t i}ⁿ^Macroecon^o^mⁱ^c^P^o^licy
^bcas^h^e^l^l^@^c^rs^.l^oc.g^ov^{, 7}^-⁷⁸¹⁶

⁵
^Ju^{st as it t}^o^o^k^so^m^e^tim^e^f^o^{r th}^{e b}^u^sⁱⁿ^{ess cy}^{cle d}^a^tin^g^co^mm^itte^{e to}^d^e^termⁱⁿ^{e w}^h^en^th^{e recessio}ⁿ^b^e^g^aⁿ^,^{it is lik}^ely^to
^b^e^so^m^e^tⁱ^m^e^a^f^t^e^{r t}^h^{e f}^a^ct^t^h^at^t^h^e^y^d^e^t^e^r^mⁱⁿ^{e w}^h^en^t^h^{e recessi}^oⁿ^h^a^{s co}^m^e^t^o^an^en^d^.