United States: Selected Issues
Author: International Monetary Fund
Publication Date:
02 Aug 2012
eISBN:
9781475504910
Language:
English
Keywords:
ISCR; CR; recession; mortgage; WEO study; crisis trend; l crisis; post recession window; WEO database; Banking crises; Global financial crisis of 2008-2009; Economic recession
Download PDF (420.5 KB)

This Selected Issues paper for the United States discusses the microeconomics of the country—household wealth and savings. Households’ consumption-saving decisions have an important bearing on the U.S. economic outlook. This paper demonstrates how households with consistently lower income, which have shown growth in the years prior to the crisis, experienced larger declines in their saving rates and a larger rise in their indebtedness before the crisis, contributing significantly to the dynamics of the mean saving rate.

Abstract

This Selected Issues paper for the United States discusses the microeconomics of the country—household wealth and savings. Households’ consumption-saving decisions have an important bearing on the U.S. economic outlook. This paper demonstrates how households with consistently lower income, which have shown growth in the years prior to the crisis, experienced larger declines in their saving rates and a larger rise in their indebtedness before the crisis, contributing significantly to the dynamics of the mean saving rate.

III. Output Losses Following Financial Crises—A Sensitivity Analysis1

Almost four years after the crisis, it is still much debated whether U.S. output will return to its pre-crisis trend. We carry out a sensitivity analysis of the methodology used in Chapter 4 of the October 2009 WEO for assessing the output losses following financial crises. In particular, we check whether the main results of that chapter are robust to changes in the definition of the pre-crisis output trend, the events used for the analysis, and the window over which the output losses are being assessed. Using historical episodes, we also discuss the possibility of a full catch-up in the long run. Our results confirm the WEO 2009 empirical finding that financial crises are followed by significant output losses.

A. Introduction

1. The economic recovery from the Great Recession has been sluggish by most standards, in particular considering the depth of the output loss during the crisis. Whether U.S. output will eventually return to its pre-recession trend or remain on a permanently lower trajectory has important implications for policymaking. A number of research papers have documented that output remains below its pre-recession trend following financial crises, including IMF (2009), Reinhart and Rogoff (2009), and Cerra and Saxena (2008). By contrast, Papell and Prodan (2011) find that GDP tends to return to its pre-crisis trend, but only after a long period (an average of nine years). To gain further insights into this question, this chapter carries out a sensitivity analysis of the methodology used in Chapter 4 of the 2009 World Economic Outlook (IMF, 2009).

2. We examine whether the choice of countries, trend definitions, and the length of the post-crisis window makes a difference to the assessment of output losses in the aftermath of financial crises. A first question is whether estimates of long-lasting output losses in the WEO study are driven by the inclusion of severe crises in emerging market economies, some of which exhibited unusual growth patterns ahead of the crises. Another question is whether the estimated pre-crisis trends were too optimistic, as some of the financial crises in the WEO study were preceded by credit booms and asset price bubbles, giving the appearance of strong potential growth ahead of the crisis and a loss thereafter. Finally, whether the 7-year post-crisis window in the WEO study is long enough to capture an eventual return of output to its potential has been questioned—it is thought, for example, that after the Great Depression the US output eventually returned to pre-crisis trend, although this catch-up took over a decade.

B. Analysis

Revised Data

3. First, we replicate the analysis of the WEO 2009 study using newly available data. For the 86 banking crises identified in the sample,2 we estimate linear trends using real GDP per capita from 10 years to 3 years before the crisis event. As in the WEO study, when the trend is negative, we re-calculate it over a longer period—from 20 years to 3 years prior to the crisis, and use the larger of the two estimated trends. These trends most frequently involve countries with multiple spells of financial crisis over short periods or other significant disruptions (no advanced economy episodes were adjusted using this technique). Once a trend has been estimated, it is extrapolated over the 7 years following the crisis. We then calculate differences between actual real per-capita output and the extrapolated trend over a 7-year window to produce time-series of output loss or gain. The data underlying these updated trend estimates are from the latest World Development Indicators database, and supplemented with country authorities’ data and the WEO database where necessary. As such, it incorporates any revisions to historical data that may have occurred after the WEO study.

4. Qualitatively, our update confirms the results of the WEO study that financial crises are followed by permanent output losses. However, our estimate of the average output losses is somewhat smaller than in the WEO. Under the new data, the mean output gap 7 years after the crisis is 8.6 percent of the pre-crisis trend, compared with the 12.5 percent identified by IMF (2009). Likewise, the median output loss comes to 6.3 percent of the pre-crisis trend using the updated dataset, compared to 10.1 percent in the WEO (2009) study (Figures 1 and 2).

Figure 1.
Figure 1.

Output Evolution after Banking Crises

Citation: IMF Staff Country Reports 2012, 214; 10.5089/9781475504910.002.A003

Source: Fund staff estimates on data from the World Development Indicators, World Economic Outlook, and country sources.
Figure 2.
Figure 2.

Output Evolution after Banking Crises

Citation: IMF Staff Country Reports 2012, 214; 10.5089/9781475504910.002.A003

Source: Fund staff estimates on data from the World Development Indicators, World Economic Outlook, and country sources.

Omitting Negative Trends

5. These initial results could potentially be influenced by outliers. In particular, there are 26 episodes in the sample for which we estimate a negative output trend over the 10-year window and therefore consider a revised trend based on a 20-year timeframe. All of these cases involve emerging economies, and many of these episodes involve countries with repeated contractions reflecting more than one spell of financial instability or other major disruptions over the estimation period. The direction of the potential bias introduced by the treatment of these events is ambiguous. On one hand, the replacement of a negative trend with a larger one yields more negative gaps between actual and trend output for any level of observed real GDP per capita, possibly bloating the estimated output losses. However, fitting curves over multiple steep contractions could produce trends—and output loss estimates—that are too weak. In any case, it may be inappropriate to include these cases in estimates of “typical” recoveries, since a proper upturn did not take place. Dropping these episodes yields estimated output losses that are larger: the average 7-year loss for the narrow set of 58 episodes3 is 14.1 percent of pre-crisis trend, while the median stands at 10.9 percent (Figure 3).

Figure 3.
Figure 3.

Output Evolution after Banking Crises

Citation: IMF Staff Country Reports 2012, 214; 10.5089/9781475504910.002.A003

Source: Fund staff estimates on data from the World Development Indicators, World Economic Outlook, and country sources.

Limiting the Analysis to Advanced Economies

6. The experiences of emerging and developing countries following banking crises may also not be informative for an advanced economy such as the United States. As advanced economies typically have greater financial development and more space for forceful interventions during banking crises, the shape of the recovery there could be quite different than in emerging market economies. Dropping all events in emerging economies, we are left with 9 cases of banking crises in advanced economies starting in 1997 or earlier. These episodes include the U.S. Savings & Loan Crisis in the late 1980s, the Nordic crises in the early 1990s, and Japan in the mid-1990s, among others as shown in column (III) of Table 1. Results for these cases show even steeper 7-year output losses than in the full sample (Figure 4), averaging 19.2 percent of the pre-crisis trend; the median loss is 16.7 percent.

Table 1.

Financial Crises in the Sample

article image
Sources: Laven and Valencia (2009)Notes: Columns I and V (baseline) refer to the episodes included in IMF (2009). Columns II and VI (no trends <0) exclude all countries for which we estimated a negative trend leading up to the crisis. Columns III and VII (advanced only) include only those episodes in WEO (2009) that are classified as advanced economies. Columns IV and VII (with extra cases) include several other cases added by the authors. These additional cases were chosen either to update the Laeven-Valencia dataset, which ended in 2007, or in the case of the United Kingdom, because it was included in Reinhart and Rogoff (2008).
Figure 4.
Figure 4.

Advanced Economies’ Output after Banking Crises

Citation: IMF Staff Country Reports 2012, 214; 10.5089/9781475504910.002.A003

Source: Fund staff estimates on data from the World Development Indicators, World Economic Outlook, and country sources.

Using Different Trend Definitions and Using a Longer Post Recession Window

7. Finally, since the period over which these trends are estimated could influence the results, we replicate the analysis for advanced economies using six different methods for estimating trends, while extending the post recession window to 10 years. We further augment the sample with some additional episodes that were not considered in the WEO 2009 study, as shown in column (IV) of Table 1. For the purposes of comparability, we retain the 10-year to 3-year trend. However, we add three alternative measures of trend, estimated from 20-years prior to the crisis to 1, 3, and 5 years before the onset. We also add two other estimates beginning 10-years before the crisis and extending to 1 and 5 years prior. The longer-trends and the ones ending 5 years prior to the event are intended to be less sensitive to excessively large growth associated with the bubbles that frequently grew in the years immediately preceding the crisis.

8. Under all estimates of trend, we find that in the aftermath of advanced economies’ banking crises, per-capita output typically lags behind trend for at least 10 years after the crisis (Figure 5). Estimated output losses vary depending on the definition of trend but remain large, with the average 10-year loss in a range of 15.6 percent to 24.4 percent. Moreover, in the 7 cases for which it is possible to observe real GDP per capita for 20-years following the onset of the crisis, in only three instances—after the 1991 Sweden and Norway crises and the 1973 United Kingdom crisis—does real per-capita output rise above any of the six trend definitions at any point along the 20-year horizon.

Figure 5.
Figure 5.

Evolution of Advanced Economies’ Output Following Financial Crises Under Alternate Trend Definitions

(percent of pre-crisis trend)

Citation: IMF Staff Country Reports 2012, 214; 10.5089/9781475504910.002.A003

Source: Fund staff estimates.

C. U.S. Episodes: The Great Depression, the S&L Crisis, and the Great Recession

9. Some analysts point to the behavior of output in aftermath of the Great Depression as evidence that U.S. output is likely to catch up to its pre-crisis path over time. Papell and Prodan (2011) argue that the long-term effects of the Great Depression—the most severe crisis and recession in U. S. history—could provide an upper bound for the long term effects of the Great Recession, given the advances in economic policymaking. Other analysts have highlighted the stability of the long-term trend in U.S. per capita real GDP since the late 19th century, and the fact that output has always returned to the long-term trend in the aftermath of recessions. We examine the behavior of output in the aftermath of the Savings and Loan (S&L) crisis in the 1980s, and the Great Depression using various trend definitions. After the S&L output lagged slightly behind trend over the next decade (Figure 6). In the Great Depression, however, we find that output did surpass its pre-recession trend after 11–13 years, depending on the trend definition used (Figure 6).

Figure 6.
Figure 6.

The Great Depression, S&L Crisis, and the Great Recession

Citation: IMF Staff Country Reports 2012, 214; 10.5089/9781475504910.002.A003

Sources: Bureau of Economic Analysis; Historical Statistics of the United States; Haver Analytics and Fund staff estimates.

However, the recovery of output coincided with a surge in federal spending as the United States entered the WWII (Figure 6). A strong rise in government spending is unlikely to be repeated at the current juncture; in fact, federal consumption and investment expenditures have been trending down since late 2010.

D. Conclusion

10. The WEO 2009 finding of large output losses following banking crises is robust for advanced economies. For the U.S., the prediction of a long-lasting deviation from the pre-crisis trend, as incorporated into staff’s economic projections, underscores the importance of policies designed to limit medium-term output losses, in particular through targeted measures aimed at boosting housing and labor markets. The findings also imply that budget revenues may not fully recover to their pre-crisis trend, highlighting the need for a medium term fiscal consolidation plan that places the public debt ratio on a sustainable path.

References

  • Cerra, Valerie and Sweta Chaman Saxena, 2008, “Growth Dynamics: The Myth of Economic Recovery,” American Economic Review, 2008, 98:1, p. 439–457.

    • Search Google Scholar
    • Export Citation

  • International Monetary Fund, 2009, World Economic Outlook, October 2009: What’s the Damage? Medium-Term Output Dynamics After Financial Crises (Washington).

    • Search Google Scholar
    • Export Citation

  • Papell, David H. and Prodan, Ruxandra, “The Statistical Behavior of GDP after Financial Crises and Severe Recessions,” paper prepared for the Federal Reserve Bank of Boston conference on “Long-Term Effects of the Great Recession,” October 2011.

    • Search Google Scholar
    • Export Citation

  • Reinhart, Carmen M. and Kenneth S. Rogoff, 2009 “The Aftermath of Financial Crises,” NBER Working Paper No. 14656 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

  • Reinhart, Carmen M. and Kenneth S. Rogoff, 2008, “This Time is Different: A Panoramic View of Eight Centuries of Financial Crises,” NBER Working Paper No. 13882 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
1

Prepared by Geoffrey Keim, Oya Celasun, and Martin Sommer.

2

As identified in Laeven and Valencia (2010). For a list of events, please refer to Table 1, columns (I) and (V).

3

Please refer to Table 1, columns (II) and (VI) for a list of these 58 events.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
United States: Selected Issues
Author: International Monetary Fund
Volume/Issue:
Volume 2012: Issue 214
Publisher:
International Monetary Fund
ISBN:
9781475504910
ISSN:
1934-7685
Pages:
119
DOI:
https://doi.org/10.5089/9781475504910.002

Output Evolution after Banking Crises

Output Evolution after Banking Crises

Output Evolution after Banking Crises

Advanced Economies’ Output after Banking Crises

Evolution of Advanced Economies’ Output Following Financial Crises Under Alternate Trend Definitions

(percent of pre-crisis trend)

The Great Depression, S&L Crisis, and the Great Recession