The Global Financial Crisis and the double recession in Spain Background Much of the western world experienced a slowdown of economic activity sometime between the latter part of 2007 and the beginning of 2008. Six months to a year into the recession, most of these advanced economies would then experience a full blown financial crisis. The crash of the property bubble and the subsequent evaporation of collateral, paired with mounting defaults by households that were losing jobs in increasing numbers, led to a rapid deterioration of bank balance sheets. The scale and the international reach of the ensuing collapse of the financial system were reminiscent of the Great Depression. The crisis in Spain included additional complications, many of them shared across the euro area. First, the regulatory architecture of the Eurozone was not set up to deal with European-level resolution of banking system fragility. Second, as tax revenues dwindled and public expenditures exploded to make-up for the collapse in demand, public debt reached levels last seen in World War 2. Public debt default became a very real possibility for several countries in the periphery of the Eurozone. As a result, spreads on sovereign debt over German bunds for many of these countries spiked. In the case of Spain, the 10-year sovereign spread with respect to Germany peaked over 600 basis points in the summer of 2012, further restricting the flow of credit. Additionally, Spain was also experiencing a painful process of deleveraging and internal devaluation as the imbalances that had accumulated over the period prior to the crisis started to be corrected, most notably the large external deficit. These complications make dating the 2008 recession challenging. While there is little ambiguity in dating the peak of economic activity following a long stretch of rapid growth in excess of 4%, it is harder to date the end of the recession. In part this is due to the slow tick-up 1
in economic activity at the beginning of 2010, which was not shared as widely across economic sectors. Summary We date the peak of economic activity in 2008Q2, about two quarters later than in the U.S. and one quarter after the peak dated by the CEPR for the euro area. We then date the subsequent trough in 2009Q4 for a peak-to-trough output loss of about 4.6%. Economic activity picked up haltingly and unevenly. Exports responded well and there were visible improvements in industrial output. The service sector endured the recession better but recovered less briskly. We date the next peak in economic activity in 2010Q4. Over this brief one-year expansion, output would grow by a meager 0.5 percentage point. Finally, we date the trough of the second recession in 2013Q2, when GDP, private consumption and investment reached their minimum levels and a quarter after the highest record of unemployed. During this second recession GDP lost about 5.65% peak-to-trough, the largest loss and longest recession we have dated since the start of our sample in 1970Q1. With output essentially moving sideways between 2009Q4 and 2010Q4, one could argue that this period merely reflected a pause within a more prolonged period of recession. However, we considered three factors in dating another peak of economic activity in 2010Q4. First, lumping together the lackluster expansion with the recession that began in 2008Q2 would have made for a recession lasting over 5 years. This is considerably longer than virtually any of the recessions dated in other countries. Second, recoveries from financial crises are notoriously slow due to the ongoing process of deleveraging and concomitant restrictions on the flow of credit. In that context, the pace of the recovery in 2010 was not at odds with the international historical experience during financial crises. Third, disaggregated data by sector indicate that industrial output experienced a significant upsurge, and that the service sector was growing, albeit moderately. 2
Figure 1 presents a summary of the behavior of GDP ( PIB in Spanish) as well as the behavior of the industrial sector (denoted VAB Industria ) and the service sector ( VAB Servicios ). Figure 1. The double recession PIB and VAB: Industrial sector and service sector 2008Q2 = 100 110 2008.Q2 2010.Q4 VAB Service 105 100 2007.Q4 95 PIB 90 VAB Industrial 85 2009.Q4 2013.Q2 80 mar-05 oct-06 may-08 dic-09 jul-11 feb-13 sep-14 abr-16 Notes: GDP (denoted PIB in Spanish) is the solid thick line in blue whose units are shown in the left axis by normalizing 2008Q2 to 100. The dashed lines shoe the value added of the industrial sector (in red) and the services sector (in green), also normalized to 100 in 2008Q2. Hence, any point can be read as a percent change with respect to 2008Q2. The vertical line indicates the peak of industrial output, which preceded the dated peak of general economic activity by two quarters and coincides more closely with the start of the recession in the U.S. and then in Europe. The shaded grey regions correspond to the recession periods associated with the peak and trough dates, also indicated in the graph. We note that the peak of economic activity in the industrial sector precedes that of the overall economy by nearly 2 quarters and coincides more closely with the start of the recession in the U.S. and a quarter later, in Europe. However, for reasons that will become clear 3
momentarily, we date the overall peak of economic activity in 2008Q2. The figure illustrates that the recession was felt unevenly (we have omitted the agricultural sector, which tends to be particularly volatile). The industrial sector (which includes construction) collapsed by over 12% in the next quarters after the industrial peak in 2007Q4. Meanwhile, the service sector saw barely any losses. By mid-2009, industrial output began a rapid recovery until the next peak of economic activity that we have dated. Industrial output improved by about 5% until 2010Q4. The recovery in services ticked up by 1.5%. Although the overall change in GDP was more muted (about 0.5%), we feel confident that it is appropriate to differentiate the period from 2009Q4 to 2010Q4 from the two recessionary periods that bracket this brief interlude of economic recovery. Figure 2 provides additional evidence in support of our dating. It shows the monthly series of the annual rate of growth in employment. As we have discussed earlier, employment tends to recover quite slowly relative to the recovery in economic activity. Nevertheless, the peak of 2008Q2 coincides with the first time the annual growth rate of employment turns negative. Employment losses continued all the way to the end of 2013. However, the hemorrhaging in employment slows down at the end of 2009 before worsening again in 2011, after the dating of the second peak shown in Figure 2, 2010Q4. 4
Figure 2. Employment during the crisis Employment (annual growth) 8 2008.Q2 2010.Q4 4 0-4 2009.Q4 2013.Q2-8 mar-05 nov-06 jul-08 mar-10 nov-11 jul-13 mar-15 nov-16 Notes: the figure displays the monthly series of the annual growth rate in employment (afiliados). The peak of 2008Q2 coincides with the start of employment losses. These continued all the way to the end of 2013. Grey shaded regions indicate recessions. As a way to bring in more timely information and to provide the public with real time indication of the state of the business cycle, the committee publishes a composite index of economic activity displayed in Figure 3. The way to interpret the indicator is not as a rate of change of any particular variable. Rather, it should be compared with the zero line, which roughly flashes recession whenever the indicator falls below zero, indicating economic expansion otherwise. The construction of the indicator and its interpretation are discussed in more detail on our website. 5
Figure 3. Economic activity in real time CF Indicator (November 2016) 15 10 2008.Q2 2010.Q4 5 0-5 -10-15 -20 2009.Q4 2013.Q2-25 mar-05 nov-06 jul-08 mar-10 nov-11 jul-13 mar-15 nov-16 Note: Grey shaded regions indicate recessions. The indicator turns negative a month before the peak of 2008Q2 although it aligns almost perfectly with the subsequent trough. The exit from the second recession would also be, according to our indicator, within a couple of months of the trough of 2013Q2. The indicator consolidates information from a variety of data series observed at monthly and quarterly frequencies but is not explicitly designed to match the recession dating we provide. 6
Conclusion The economic landscape since 2008 has been grim both at the domestic and international levels. The recession that started in 2008Q2 was followed by a full blown financial crisis at the international level. Within the Eurozone, things were further complicated by a slowly evolving institutional structure and a sovereign debt crisis. These factors complicate the dating of the end of the recession that began in 2008Q2. However, a closer look at the behavior of the Spanish economy during this period reveals that the economy had begun to improve by the end of 2009. The subsequent recovery, while very gradual, was visible in the two largest sectors of the economy: the industrial and the services sectors. We therefore feel comfortable dating the trough of 2009Q4. We then date a peak of 2010Q4 which marks the start of the second period of recession until the trough of 2013Q2. Our dating is consistent with the dating of the CEPR for the Euro area. The CEPR dates the first peak in 2008Q1 a quarter after we do; the trough in 2009Q2 is two quarters earlier than in Spain; and the following peak in 2011Q3, three quarters after the peak we date for Spain; and the trough that closes the double recession in 2013Q1, a quarter before we do. It has to be taken into account that the CEPR dates the business cycle of the whole Euro area, while our interest is limited to Spain. 7