Annex. Econometric Analysis
Econometric analysis builds on Cechetti et al. (2011), which uses a new dataset on debt levels for a group of 18 OECD countries, based primarily on flow of funds data. The paper uses data over the period 1980-2006, but since the authors had compiled data through 2009, this analysis uses the full sample.
The empirical specification is derived from the neoclassical growth model of Solow, where per capita income growth depends on the initial level of physical and human capital, savings rate, population rate, and technology. In addition to these standard regressors in the growth literature, measures of public and private sector debt are added to the specification to see whether they have an impact on growth independent of other determinants. Panel data regressions are estimated using country-specific and time-specific time effects. More specifically:
gi,t+1,t+k is the k-year forward average of annual real GDP per capita growth between years t+1 and t+k. The analysis uses k=5
yi,t is the log of real per capital GDP at time t;
μit and γt are country-specific and time-specific dummies;
Xi,t includes gross saving as a share of GDP; population growth; number of years spent in secondary education, as a proxy for the level of human capital; the dependency ratio; openness to trade measured by the sum of exports and imports to GDP; CPI inflation as a measure to macroeconomic stability; the ratio of liquid liabilities to GDP, as a measure of financial development, and a dummy to control for banking crises.
Di,t includes, depending on the specification, the ratio of debt to GDP of public and/or private sector (household and corporate sector) as well as interactions with dummy variables indicating whether the debt ratios are above a threshold level.
Least squares (LSDV) estimation is used. The presence of a lagged dependent variable in the right hand side (dynamic panel) implies that the estimates may be biased. However, it has not been proved that generalized methods of moments (GMM) or instrumental variables (IV) outperforms LSDV in small size panels, like the one this analysis uses (N = 18).
The analysis tries to assess whether the growth impact of high debt in one sector depends on the level of indebtedness in other sectors. Debt is considered to be “high” if it’s above a certain threshold identified as the sample mean. The thresholds are 73 percent of GDP for public debt, 98 percent of GDP for corporate debt and 48 percent of GDP for household debt. For instance, in the specification to estimate the impact of public debt on growth and its differential impact depending on the level of indebtedness in the private sector, the regressor α′i,t becomes:
Buiter, Willem and Ebrahim Rahbari, 2012, Debt of Nations: Mr. Miawber’s Vindication: Causes and Consequences of Excessive Debt, Citigoup.
Baum, Anja, Cristina Checherita-Westphal and Philipp Rother, 2012, Debt and growth: new evidence for the euro area, ECB Working Paper Nr. 1450 (Frankfurt).
Cuerpo, Carlos, Inês Drumond, Julia Lendvai, Peter Pontuch and Rafal Raciborski, 2013, Indebtedness, Deleveraging Dynamics and Macroeconomic Adjustment, European Economy, Economic Papers 477.
Cussen, Mary and Bridin O’Lear y, 2013, Why are Irish Non-Financial Corporations so Indebted? In Quarterly Bulletin 01 /13, Central Bank of Ireland (Dublin).
Cussen, Mary, Bridin O’Lear y and Donal Smith, 2012, The Impact of the Financial Turmoil on Households: A Cross Country Comparison In Quarterly Bulletin 02/12, Central Bank of Ireland (Dublin).
European Central Bank, 2013a, Integrated Euro Area Accounts for the Fourth Quarter of 2012, Box 4 in Monthly Bulletin, May (Frankfurt), pp 46–50.
European Central Bank, 2013b, The Eurosystem Household Finance and Consumption Survey: Results from the First Wave, Statistics Paper Series 2/13 (Frankfurt).
Herdon, Thomas, Michael Ash and Robert Pollin, 2013, Does High Public Debt Consistently Stifle Economic Growth? A Critique of Reinhardt and Rogoff, PERI Working Paper Series 322 (Amherst)
International Monetary Fund, 2012a, Access to Credit, Debt Overhang, and Economic Recovery: The Irish Case, Chapter 2 Ireland: Selected Issues Paper, IMF Country Report No. 12/265 (Washington).
International Monetary Fund, 2012c, Household Consumption, Wealth and Saving, Chapter 1 Ireland: Selected Issues Paper, IMF Country Report No. 12/265 (Washington).
International Monetary Fund, 2012d, Spain: Vulnerabilities of Private Sector Balance Sheets and Risks to the financial Sector: Technical Note, Financial Sector Assessment Program (Washington).
International Monetary Fund, 2013a, France: Housing Prices and Financial Stability: Technical Note, Financial Sector Assessment Program (Washington).
International Monetary Fund, 2013b, Italy: The Financial Situation of Italian Households and non-Financial Corporations and Risks to the Banking System: Technical Note, Financial Sector Assessment Program (Washington).
International Monetary Fund, 2013d, Portugal’s Corporate (De) Leveraging, Chapter 3 Portugal: Selected Issues Paper, IMF Country Report No. 13/19 (Washington).
International Monetary Fund, 2013e, Slovenia: Staff Report of the 2013 Article IV Consultation, IMF Country Report No. (Washington).
International Monetary Fund, 2013f, The Kingdom of the Netherlands: Staff Report of the 2013 Article IV Consultation, IMF Country Report No. 13/48 (Washington).
International Monetary Fund, 2013g, European Union; Financial System Stability Assessment, IMF Country Report No. 13/75 (Washington).
International Monetary Fund, 2013h, Household Savings Ratio in Spain—How Low Can it Go? In Spain: Selected Issues Paper (forthcoming).
Kumar, Manmohan S. and Jaejoon Woo, 2010, Public Debt and Growth, International Monetary Fund Working Paper 10/174 (Washington).
Laeven, Luc and Thomas Laryea, 2009, Principles of Household Debt Restructuring, International Monetary Fund Staff Discussion Note 09/15 (Washington).
Laryea, Thomas, 2010, Approaches to Corporate Debt Restructuring in the Wake of the Financial Crisis, International Monetary Fund Staff Discussion Note 10/02 (Washington).
Liu, Yan and Christoph B. Rosenberg, 2013, Dealing with Private Debt Distress in the Wake of the European Financial Crisis: A Review of the Economics and Legal Toolbox. International Monetary Fund Working Paper 13/44 (Washington).
Panizza, Ugo and Andrea Filippo Presbitero, 2012, Public Debt and Economic Growth: Is There a Causal Effect? Money and Finance Research group working paper No. 65 (Ancona).
Reinhart, Carmen M. and Kenneth Rogoff, 2010, Growth in a Time of Debt, American Economic Review: Papers and Proceedings 100, pp. 573–578.
Reinhart, Carmen M., Vincent R. Reinhart, and Kenneth Rogoff, 2012, Debt Overhangs Past and Present, NBER Working Paper 18015 (Cambridge).
Ruscher, Eric and Guntram B. Wolff, 2012, Corporate Balance sheet Adjustment: Stylized Facts, Causes and Consequences, Bruegel Working Paper 2012/03 (Brussels).
Prepared by Fabian Bornhorst and Marta Ruiz Arranz.
See Cussen and O’Leary (2013) for a discussion of consolidated vs. non-consolidated corporate debt in the euro area.
See also Cuerpo et al. (2013) for an identification of countries currently facing private sector deleveraging pressures based on various indebtedness indicators. For an overview, see also Buiter and Rahbari (2012) and McKinsey (2012).
A full assessment of house prices would have to go beyond affordability ratios (price-to-income and price-to-rent ratios) and include other fundamentals, including supply constraints. See IMF 2012d, IMF 2013a, IMF 2013f.
In the euro area, high corporate debt is also associated with lower per capita GDP growth during the period from 1999-2011 (ECB 2012).
Historical episodes include: Canada (1979-1984), Denmark (1987-1994), Germany (2000-11), UK (1990-96), Finland (1989-1997), Japan (2001-11), Norway (1988-1995), and Sweden (1989-95). In the last four, household deleveraging was associated with a banking crisis. These episodes were selected among advanced economies that experienced a reduction in the household debt-to-disposable income ratio of more than 10 percentage points.
Historical experience offers one possible benchmark. Model based approaches can also be employed to derive optimal levels of leverage or indebtedness to gauge deleveraging needs, see e.g., Cuerpo et al., 2013.
For a discussion of the role of inflation in assisting the deleveraging process, including its costs, see IMF Fiscal Monitor (April 2013).
Identification of historic corporate deleveraging episodes is based on Ruscher and Wolff (2012), who use the sector’s net lending/borrowing data as a marker, combined with indebtedness data from Cecchetti et al. (2011). It comprises of episodes with a significant debt reduction (10 percent of GDP or more), which, on average, lasted 6 years. A number of shorter episodes of corporate deleveraging identified by Ruscher and Wolff (2012) did not result in a significant debt reduction.
See Annex 1 for details on the econometric analysis. Debt is considered to be “high” if it is above the mean value in the sample. The mean values are 73 percent of GDP for government debt, 48 percent of GDP for household debt, and 98 percent of GDP for corporate debt. The thresholds identified in Cechetti et al. (2011) are also used as a robustness test. The main results hold but the higher thresholds relative to the mean, particularly for household debt imply that there are very few observations when debt is high in all sectors at the same time.