Business Cycle Dating Committee of the Centre for Economic Policy Research Michael Artis Fabio Canova Jordi Gali Francesco Giavazzi Richard Portes (President, CEPR) Lucrezia Reichlin (Chair) Harald Uhlig Philippe Weil Embargo:.00 GMT, 2.00 BST,.00 CET, Monday 22nd Sept 200. The CEPR Business Cycle Dating Committee The Centre for Economic Policy Research has formed a committee to set the dates of the euro area business cycle. Its mission is to establish the chronology of recessions and expansions of the original euro area member countries plus Greece for 970-998 and of the euro area as a whole from 999 onwards. The euro area currently includes Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain. In determining the chronology of the euro area business cycle, the CEPR Business Cycle Dating Committee has decided to adopt a definition of a recession similar to that used by the National Bureau of Economic Research (NBER), which has for many years dated the US business cycle. We have had to adapt the NBER definition, however, to reflect specific features of the euro area. Thus the Committee defines a recession as a significant decline in the level of economic activity, spread across the economy of the euro area, usually visible in two or more consecutive quarters of negative growth in, employment and other measures of aggregate economic activity for the euro area as a whole, and reflecting similar developments in most countries. A recession begins just after the economy reaches a peak of activity and ends when the economy reaches its trough. Between trough and peak, the economy is formally in an expansion; between peak and trough it is in a recession. In both cases, growth rates may be very low. The CEPR Committee s task is significantly different from that of the NBER. The euro area groups together a set of different countries. Although subject to a common monetary policy since 999, they even now have heterogeneous institutions and policies. Moreover, European statistics are of uneven quality, long time series are not available, and data definitions differ across countries and sources.
These differences explain why some of the CEPR criteria for dating business cycles differ from those used by the NBER: Unlike the NBER for the US, the CEPR committee dates in terms of quarters rather than months. Currently, the most reliable European data for our purposes, and those around which a reasonable consensus can be achieved, are the quarterly series. The Committee analyses euro area aggregate statistics, but it also monitors country statistics to make sure that expansions or recessions are widespread over the countries of the area. There is no fixed rule by which country information is weighted. The Committee views real (euro area aggregate, as well as national) as the main measure of macroeconomic activity, but it also looks at additional macroeconomic variables, for several reasons. First, euro area series constructed for the pre-emu era reflect not only movements in economic activity but also changes in exchange rates, which are problematic. Second, statistics are sometimes subject to large subsequent revisions, and this makes them an imperfect indicator of current business cycle conditions. Third, measured does not always move in parallel with its individual major components (which may indeed be moving in different directions) or other macroeconomic aggregates such as employment. Fourth, these variables are known to display more cyclicality than and are useful in strengthening opinions when the data do not seem very decisive. They are also available (with the exception of investment) earlier and at a higher frequency than. For recent euro area data (since the end of the 990s) we use, where possible, official Eurostat statistics and focus primarily but not exclusively on () quarterly (Eurostat source); (2) quarterly employment (OECD); () monthly industrial production (Eurostat); (4) quarterly business investment (Eurostat); (5) consumption and its main components (Eurostat and ECB). For country data, we use Eurostat and OECD sources and monitor Germany, France and Italy systematically. Historical euro area data since the 970s are provided by the OECD and the ECB. We mainly use the ECB source. For national data, we have used the OECD and the IMF. We do not use a fixed rule to weight different data series, although we give primary emphasis to. The committee informally assesses the depth, duration and severity of a recession. Hence although recessions are usually characterized by at least two consecutive quarters of declining, this is not a fixed rule.
2. The euro area business cycle since 970 The Committee has identified the following cyclical episodes since 970, with peaks and troughs dated as follows: PEAK 974q 980q 992q TROUGH 975q 982q 99q Table. Chronology of euro area business cycles Thus the Committee has identified three recessions: 974q to 975q, 980q to 982q, and 992q to 99q. Figure shows the level of real (seasonally adjusted) from 970 to 998 according to two alternative series constructed by the OECD and the ECB (denoted by FHM see below). Shaded bands indicate recessions. For the same period, we also report key series plotted together with : investment (Figure 2) and Employment (Figure ). Figures 4a-4c, 5a-5c, and 6a-6c plot, investment, employment and industrial production for France, Germany and Italy. These data underpin the following analysis of the nature of the recessions identified by the Committee: Two of the three recessions we have declared in the 970s and 990s are pronounced and also synchronized across countries and variables. In both periods employment, investment and industrial production declined with. The third recession, in the 980s, exhibits different and specific characteristics. The recession in terms of aggregate output is milder but longer. does not decline sharply but rather stagnates for almost three years. Our dating is thus based on the behaviour of employment and investment which, unlike, declined sharply during the period. In this episode, we also observe more heterogeneity in output dynamics across the three large economies than in the other two recessions. That affects our designation of the date of the trough, in particular. As to recent years, the Committee judges that, based on currently available data, it would be premature to declare a peak (and hence a subsequent recession). There is still much uncertainty, as we discuss in the next section.
. The euro area since 200 The role of the Committee is to establish the chronology of recessions and expansions, not to forecast, nor even to characterize the current conjuncture. This means that it may be necessary to wait until well after the event to declare a peak or trough in economic activity. The recent period is a case in which caution is clearly required. Euro area has slowed down since the first quarter of 200. A weak resurgence of positive growth at the beginning of 2002 seems to have come to a new halt. Employment has grown somewhat, while industrial production, after having fallen sharply in 200, shows weak signs of recovery (see Figures 7, 8 and 9). Investment has been declining for more than two years, but government consumption rose 2.2% in 200 and 2.7% in 2002 (ECB Statistics Pocketbook, September 200). Figure 0 compares euro in recent years with the 980s recession by normalizing in 200q and euro in 980q (beginning of the 980s recession) to. Qualitatively, there are some resemblances between the behaviour of now and then, although no sharp fall has been recorded in the recent period, nor have we seen recently more than a single quarter s decline in sequence. And unlike the 980s episode, employment is not declining. Figures, 2, and provide information about France, Germany and Italy. Based on currently available data, our current judgment is therefore that the euro area has been experiencing a prolonged pause in the growth of economic activity, rather than a full-fledged recession as we have defined it. The picture may of course change as revised statistics appear. 4. Europe and the US A comparison between the recent US and euro-area experiences shows some interesting differences and illustrates further why we cannot declare a euro-area recession for the recent period. Current NIPA data show a peak in US in 2000q4 (compare the NBER date of March 200). According to these data, declined for three consecutive quarters (though 200q). The cumulative decline over that period was -2%. Cumulative growth from 2000q4 through 200 was 4.2%. The corresponding numbers for employment (over age 6, measured month middle of quarter) have been -5% and +0.4%. Now consider the euro area, taking 200q as the base period (growth was positive and robust in that quarter, unlike the US). The cumulative change of euro area from 200q through 200q4 was +0.%. Cumulative growth from 200q through 200q2 was +.%. The corresponding numbers for employment (based on data in ECB Monthly Bulletin) are +0.5% and +6% (the latter through 200q). Thus the euro-area economy has essentially stagnated since 200q, and we have
observed neither the sharp (though short) decline in that the US experienced nor the US recovery. This appears to repeat the pattern seen in the 980s: euro-area is less volatile than that of the US. Note also that employment in the euro area has continued to grow somewhat (although more slowly than in the 990s), whereas that in the US has not. Looking back at historical business cycle data, it should be noticed that there is no clear pattern of lead-lag relation between the US and the Euro area. The 970s recessions are synchronized, while in the 990s the US recession led the European recession. In the 980s, the euro area output behaviour was smoother than in the US: the Euro area experienced a long mild recession, while the US had two short sharper recessions. Finally, as discussed above, while in recent years the US experienced a recession and is now showing a recovery (the NBER has dated both a peak and a trough for the US in 200), current signals from the Euro area are mixed (see Figure 4). Figures 5 to 7 report US and some other key macroeconomic variables. Peaks and troughs are more clearly synchronized across variables in the US than in the euro area.
5. Data sources 5.. Historical data (970-998) 5... Euro area, investment and employment are from the data set constructed at the ECB by Fagan et al. (FHM). This data set includes national account data and labor market data. We also use statistics from OECD Quarterly National Accounts Database (Gross domestic product / by Expenditure,Comparable Table /Constant prices,constant PPPs,OECD base,annual rates SA). 5..2. Large countries The sources are OECD Quarterly National Accounts (QNA), Main Economic Indicators (MEI) and Quarterly Labor Force Statistics (QLF): Gross domestic product / by Expenditure,Comparable Table /Volume index,oecd base,sa /995=00. Source: QNA. Gross fixed capital formation / by Expenditure,Comparable Table /Volume index,oecd base,sa /995=00. Source: QNA. Civilian employment index sa/index publication base SA /Civilian Employment (survey)all persons/levels/all ages. Source: QLF. Industrial production ISIC C-E sa/index publication base SA /Production/Production by economic activity /995Y. Source: MEI. The only exception is for Germany. Since the OECD series starts only after reunification, we use statistics from the IMF ( Gdp Vol. (995=00) /Index Number /Base year: 995 /averages /constant prices). 5.2. Period 999 to present Euro area (constant price and seasonally adjusted), investment (gross fixed capital formation, constant price and seasonally adjusted) and industrial production (seasonally adjusted) series are from Eurostat. Euro area employment is the index published by OECD (Main Economic Indicators; calculation from labour force sample surveys ) Large countries s are from Eurostat (constant price and seasonally adjusted). 5.. US data Real Gross Domestic Product [Billions of chained (996) dollars] Seasonally adjusted at annual rates, Table.2., Line, Bureau of Economic Analysis. Gross private domestic investment [Billions of chained (996) dollars] Seasonally adjusted at annual rates, Table.2., Line 6, Bureau of Economic Analysis. Employment: Total Nonfarm Employment - Seasonally Adjusted, Bureau of Labor Statistics Industrial Production, NBER website. http://www.ecb.int/pub/wp/ecbwp042.zip
6. Figures Figure Euro area, 970-98 0 billions of dollars 6. 5.6 Q /992 Q /99 5. 4.6 Q /980 FHM OECD 4..6. Q /970 Q /97 Q /972 Q /974 Q /975 Q /97 Q /974 Q /975 Q /976 Q /977 Q /978 Q /982 Figure 2 Euro area vs investment, 970-98 0 billions of dollars.2. Q /979 Q /980 Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 Q /987 Q /988 Q /989 Q /990 Q /99 Q /992 992/Q Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 0.25 0.2 0 billions of dollars 980/Q 99/Q 0.2 Investment 970Q 97Q 974/Q 972Q 97Q 974Q 975Q 975/Q 976Q 977Q 978Q 979Q 982/Q Figure Euro area vs emplyment, 970-98 0 billions of dollars.2. 980Q 98Q 982Q 98Q 984Q 985Q 986Q 987Q 988Q 989Q 990Q 66 992/Q 99Q 992Q 99Q 994Q 995Q 996Q 997Q 998Q 99/Q 0.5 6 4 2 millions of workers Employment 980/Q 970Q 97Q 974/Q 972Q 97Q 974Q 975Q 976Q 977Q 975/Q 978Q 979Q 980Q 98Q 982Q 98Q 984Q 985Q 986Q 987Q 988Q 982/Q 989Q 990Q 99Q 992Q 99Q 994Q 995Q 996Q 997Q 998Q 0 08 06
Figure 4a vs investment, France, 970-98..2 992/Q.07.0.02 index 995=.00 99/Q 7 2 Investment 980/Q 7 2 982/Q 7 2 Q /970 Q 2/97 Q /972 Q 4/97 Q /975 Q 2/976 Q /977 Q 4/978 Q /980 Q 2/98 Q /982 Q 4/98 Q /985 Q 2/986 Q /987 Q 4/988 Q /990 Q 2/99 Q /992 Q 4/99 Q /995 Q 2/996 Q /997 Q 4/998 Figure 4b vs employment, France, 970-98..04 992/Q.0.02 index 995=.00 980/Q 99/Q.0 9 Employment 8 7 982/Q 6 5 Q /970 Q 2/97 Q /972 Q 4/97 Q /975 Q 2/976 Q /977 Q 4/978 Q /980 Q 2/98 Q /982 Q 4/98 Q /985 Q 2/986 Q /987 Q 4/988 Q /990 Q 2/99 Q /992 Q 4/99 Q /995 Q 2/996 Q /997 Q 4/998 Figure 4c vs industrial production, France, 970-98. 992/Q index 995=.00 974/Q 980/Q 99/Q IP Q /970 Q /97 Q /972 975/Q Q /97 Q /974 Q /975 Q /976 Q /977 Q /978 Q /979 Q /980 982/Q Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 Q /987 Q /988 Q /989 Q /990 Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998
Figure 5a vs investment, Germany, 970-98.04.05 992/Q.02 5 index 995=.00 5 5 5 0.55 0.45 Q /970 Q /97 974/Q 975/Q Q /972 Q /97 Q /974 Q /975 Q /976 Q /977 980/Q 982/Q Q /978 Q /979 Q /980 Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 Q /987 Figure 5b vs employment, Germany, 970-98 Q /988 Q /989 Q /990 Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 99/Q 8 6 4 2 Investment.05.05 992/Q.00 5 99/Q 5 index 995=.00 5 5 980/Q 0 5 0 Employment 5 0.55 0.45 Q /970 Q /97 974/Q 975/Q Q /972 Q /97 Q /974 Q /975 Q /976 Q /977 982/Q Q /978 Q /979 Q /980 Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 Q /987 Figure 5c vs industrial production, Germany, 970-98 Q /988 Q /989 Q /990 Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 5 0 5.0.05 5 992/Q.05.00 index 995=.00 5 5 980/Q 99/Q 5 0 IP 5 5 0.55 0.45 Q /970 Q /97 Q /972 974/Q Q /97 Q /974 Q /975 975/Q Q /976 Q /977 Q /978 Q /979 Q /980 Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 982/Q Q /987 Q /988 Q /989 Q /990 Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 0 5 0
Figure 6a vs investment, Italy, 970-98. 992/Q.2.07.02 index 995=.00 974/Q 980/Q 99/Q 7 2 7 2 Investment 0.5 Q /970 Q /97 Q /972 Q /97 Q /974 Q /975 Q /976 975/Q Q /977 Q /978 Q /979 Q /980 982/Q Figure 6b vs employment, Italy, 970-98 Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 Q /987 Q /988 Q /989 Q /990 Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 7 2. 992/Q.09.07.05 index 995=.00 980/Q 99/Q.0.0 Employment 0.5 Q /970 Q /97 974/Q 975/Q Q /972 Q /97 Q /974 Q /975 Q /976 Q /977 982/Q Q /978 Q /979 Q /980 Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 Q /987 Q /988 Figure 6c vs industrial production, Italy, 970-98 Q /989 Q /990 Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 9 7 5. 992/Q index 995=.00 0.5 Q /970 Q /97 Q /972 974/Q Q /97 Q /974 Q /975 Q /976 Q /977 Q /978 975/Q 980/Q 982/Q Q /979 Q /980 Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 Q /987 Q /988 Q /989 Q /990 99/Q Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 IP
Figure 7 Euro area vs investment, 999-200 0 billions of do.58.56 0.40 0.5 0 billions of do.54 0.0.52 0.25 Investment.50 0.20.48 0.5.46 0.0 999. 999.2 999. 999.4 2000. 2000.2 2000. 2000.4 200. 200.2 200. 200.4 2002. 2002.2 2002. 2002.4 200. 200.2 Figure 8 Euro area vs employment, 999-200 0 billions of dollars.58.56.0.09 index 995=.54.08.52.07 Employment.06.50.05.48.04.46.0 999. 999.2 999. 999.4 2000. 2000.2 2000. 2000.4 200. 200.2 200. 200.4 2002. 2002.2 2002. 2002.4 200. 200.2 Figure 9 Euro area vs industrial production, 999-200 0 billions of dollars.58.56.04.02.54.52 8 IP.50 6.48 4.46 2 999. 999.2 999. 999.4 2000. 2000.2 2000. 2000.4 200. 200.2 200. 200.4 2002. 2002.2 2002. 2002.4 200. 200.2
Figure 0 Euro area, 980s vs 200 onwards.02.05.0.005 : 80s Recession : since 200 Q 95 9 2 4 5 6 7 8 9 0 Figure, France vs euro area, 999-200 0 billions of dollar.59.57 0.55 0.50 0.45 0 billions of dollar.55 0.40.5.5 0.5 0.0 Euro Area France.49 0.25 0.20.47 0.5.45 0.0 999. 999.2 999. 999.4 2000. 2000.2 2000. 2000.4 200. 200.2 200. 200.4 2002. 2002.2 2002. 2002.4 200. 200.2 Figure 2, Germany vs euro area, 999-200 0 billions of dollar.59.57.55 0.500 0.495 0.490 0 billions of dollar 0.485.5.5 0.480 0.475 Euro Area Germany.49 0.470.47 0.465.45 0.460 999. 999.2 999. 999.4 2000. 2000.2 2000. 2000.4 200. 200.2 200. 200.4 2002. 2002.2 2002. 2002.4 200. 200.2
Figure, Italy vs euro area, 999-200 0 billions of dollar.59.57 0.265 0.260 0 billions of dollar.55.5.5 0.255 0.250 Euro Area Italy.49.47 0.245.45 0.240 999. 999.2 999. 999.4 2000. 2000.2 2000. 2000.4 200. 200.2 200. 200.4 2002. 2002.2 2002. 2002.4 200. 200.2 Figure 4 Euro and US recessions, 970-2002 0 billions of dollars 9. 8. 7. Euro 6. 5. 4.. 2. 9.5 8.5 7.5 Q /970 Q 2/97 Q /972 Figure 5 US vs investment, 970-200 0 billions of dollars Q 4/97 Q /975 Q 2/976 Q /977 Q 4/978 Q /980 Q 2/98 Q /982 Q 4/98 Q /985 Q 2/986 Q /987 Q 4/988 Q /990 Q 2/99 Q /992 Q 4/99 Q /995 Q 2/996 Q /997 Q 4/998 Q /2000 Q 2/200 Q /2002 200/Q 200/Q 4.8.6.4 0 billions of dollars US Euro+US US Euro 6.5 5.5 4.5.5 97/Q 4 Q /970 Q /97 Q /972 Q /97 Q /974 Q /975 Q /976 975/Q 980/Q Q /977 Q /978 Q /979 980/Q 98/Q Q /980 Q /98 Q /982 Q /98 Q /984 982/Q 4 990/Q 99/Q Q /985 Q /986 Q /987 Q /988 Q /989 Q /990 Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 Q /999 Q /2000 Q /200 Q /2002 Q /200.2.0 0.4 Investiment
Figure 6 US vs employment, 970-200 0 billions of dollars 9.5 8.5 990/Q 5 200/Q 25 200/Q4 5 millions of workers 7.5 6.5 98/Q 99/Q 05 95 Employment 5.5 4.5.5 Q /970 Q /97 Q /972 97/Q4 975/Q Q /97 Q /974 Q /975 Q /976 Q /977 Q /978 980/Q 980/Q 982/Q4 Q /979 Q /980 Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 Q /987 Q /988 Figure 7 US vs industrial production, 970-200 Q /989 Q /990 Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 Q /999 Q /2000 Q /200 Q /2002 Q /200 85 75 65 0 billions of dollars 9.5 8.5 200/Q 200/Q4.25.5.05 7.5 5 6.5 990/Q 99/Q 5 IP 98/Q 5 5.5 4.5.5 Q /970 Q /97 Q /972 97/Q4 Q /97 Q /974 Q /975 Q /976 Q /977 Q /978 975/Q 980/Q 980/Q Q /979 Q /980 Q /98 Q /982 Q /98 Q /984 Q /985 Q /986 Q /987 Q /988 982/Q4 Q /989 Q /990 Q /99 Q /992 Q /99 Q /994 Q /995 Q /996 Q /997 Q /998 Q /999 Q /2000 Q /200 Q /2002 Q /200 5 0.55 0.45