Worrying About an Unobservable
Recessions associated with a financial crisis, such as the severe contraction in output that many European economies are currently experiencing, often involve large and highly persistent—sometimes permanent—output losses (Cerra and Saxena, 2008; and IMF, 2009e). Financial crises that include credit crunches and housing busts tend to be particularly severe (Claessens, Kose, and Terrones, 2008) and, on average, last for two years with deep and persistent effects on asset prices, output, and employment, accompanied by massive increases in government debt (Reinhart and Rogoff, 2009). The severity of a recession described by changes in actual output, however, is often insufficient for guiding the response of policymakers. More information is needed about those changes in output—their nature, sources, and long-term consequences—which are often expressed collectively as “potential output” or the “output gap” (that is, the percentage difference between potential and actual output).
Because potential output cannot be observed directly, using it as a basis for policymaking is inherently difficult and forces policymakers to rely on imperfect estimates that can cause serious policy errors. Several years of strong growth, for example, may be mistakenly interpreted as a new long-term, sustainable trend, which could subsequently turn out to be unsustainable. Emerging economies are particularly prone to this type of uncertainty because the process of catching up with the output levels of advanced economies is rarely a steady process but often involves much variation. Fiscal policymakers in advanced and emerging economies alike run the risk of interpreting short-lived booms as a new steady state and then fail to adjust polices quickly enough. For instance, some believe that uncertainty about potential output in the euro area and its economies contributed to insufficient fiscal consolidation and structural reform during relatively good times and put both the Stability and Growth Pact (SGP) and the Lisbon Strategy under great strain in the early 2000s (Cotis, Elmeskov, and Mourougane, 2004). Similarly, monetary policy errors occurred when policymakers falsely believed that the large recessions during the 1970s and 1980s had little effect on potential output and thus on the underlying price pressures; that miscalculation is another example of uncertainty about potential output that led to relatively high inflation in many economies and proved costly to correct.
The questions the current crisis poses for policymakers are not different from those raised during past contractions: what precisely is the effect of the crisis on potential output and how long will the effect last? Further, how uncertain are the answers to these questions? Over the past decades, new methods for estimating potential output have developed and broadened the underlying concepts of potential output. But even today, measurement of potential output is often as much art as science, involving strong assumptions and conceptual choices, especially in times of severe economic crisis.
Long-Term Effects
Will the current crisis affect Europe’s growth potential in the long run? The long-term effects on growth are often assessed by using the concept of trend growth or “steady-state growth.” In the theoretical literature, the assessment and forecasting of steady-state growth are based on a production function that brings together the contributions from its main inputs—labor, capital, and productivity—to determine the time path of output. Measurement of labor productivity–enhancing technological change, embodied in new capital goods, as well as the evolution of labor skills that affect productivity, is very difficult; and, together with other inputs (public sector capital, for example) that are hard to measure, they are usually summarized by the term total factor productivity (TFP), or the “residual” in explaining longer-run growth. TFP is thus a key component in the assessment of an economy’s growth potential but is also the central source of uncertainty in such estimates, since little is known about its deep structural sources.
Nevertheless, applying the concept of growth accounting to European countries yields interesting insights into the long-term drivers of growth (Table 4). Three stylized facts come out of the growth-accounting exercise:
First, gross value added (GVA) in the euro area was increasing at 2 percent per year, a much lower rate than in the United States or the United Kingdom during 1980–95. The differences arose from hours worked with respect to the United States and from TFP growth with respect to the United Kingdom.
Second, the differences between the euro area and the United States and the United Kingdom became larger in the next decade. The euro area continued to grow at 2 percent, while the United States and the United Kingdom registered even higher growth rates. During this time, the United States outperformed in TFP growth, whereas the United Kingdom accumulated capital—and both showed relatively strong growth in financial sector GVA.
Third, growth in European emerging economies far outpaced growth in the euro area. For the handful of emerging economies for which growth accounting is available, TFP growth and capital accumulation held the key to the differences with the euro area.
Sectoral growth accounting reveals large differences in the sources of growth among the advanced countries. The effect of the crisis, therefore, on steady-state growth is likely to vary among them, especially between the euro area and the United Kingdom. The large differences stem from the contribution of finance and business services to overall GVA growth (Table 4 and Figure 12). The contribution of the financial industry to GVA growth for the euro area (0.6 percentage point) was much lower than that for the United Kingdom (1.6 percentage points) or for the United States (1.2 percentage points) in the decade preceding the crisis. In fact, TFP growth in the financial industry in the euro area was negative, led by Germany, France, and Italy, in contrast to the relatively large contribution of TFP in the financial industry to total productivity growth in the United Kingdom and the United States.
Gross Value-Added Growth and Contributions, 1980-95 and 1995-2005
(Annual average volume growth rates, percent)
Due to data limitations, the euro area only includes the following countries: Austria, Belgium, Finland, France, Germany, Italy, Netherlands, and Spain. Slovenia and Slovak Republic have been added to New Member States since they are very recent entrants to the euro area.
Gross Value-Added Growth and Contributions, 1980-95 and 1995-2005
(Annual average volume growth rates, percent)
VA | L | H | LC | K | KIT | KNIT | TFP | LP | |
---|---|---|---|---|---|---|---|---|---|
(1)=(2)+(5)+(8) | (2)=(3)+(4) | (3) | (4) | (5)=(6)+(7) | (6) | (7) | (8) | (9)=(1)-(2) | |
1980-1995 | |||||||||
MARKET ECONOMY | |||||||||
Euro area 1/ | 1.9 | 0.0 | -0.2 | 0.3 | 1.0 | 0.3 | 0.7 | 0.9 | 1.9 |
France | 1.8 | -0.1 | -0.5 | 0.4 | 0.7 | 0.2 | 0.4 | 1.2 | 1.9 |
Germany | 1.9 | -0.2 | -0.4 | 0.2 | 1.2 | 0.3 | 0.9 | 0.8 | 2.1 |
Italy | 1.9 | 0.3 | 0.2 | 0.1 | 0.8 | 0.3 | 0.6 | 0.8 | 1.6 |
United Kingdom | 2.5 | -0.3 | -0.6 | 0.3 | 1.2 | 0.6 | 0.7 | 1.5 | 2.7 |
United States | 3.0 | 1.2 | 1.0 | 0.2 | 1.1 | 0.5 | 0.6 | 0.7 | 1.8 |
1995-2005 | |||||||||
MARKET ECONOMY | |||||||||
Euro area 1/ | 1.9 | 0.6 | 0.4 | 0.1 | 1.1 | 0.5 | 0.6 | 0.2 | 1.3 |
France | 2.4 | 0.7 | 0.3 | 0.4 | 0.9 | 0.4 | 0.5 | 0.8 | 1.7 |
Germany | 1.0 | -0.4 | -0.5 | 0.1 | 1.0 | 0.5 | 0.5 | 0.4 | 1.4 |
Italy | 1.2 | 0.8 | 0.7 | 0.2 | 1.0 | 0.3 | 0.8 | -0.7 | 0.4 |
United Kingdom | 3.2 | 0.8 | 0.4 | 0.5 | 1.5 | 0.9 | 0.6 | 0.9 | 2.4 |
New Member States | |||||||||
Czech Republic | 2.7 | -0.1 | -0.3 | 0.2 | 2.3 | 0.7 | 1.6 | 0.6 | 2.9 |
Hungary | 4.4 | 1.0 | 0.6 | 0.4 | 0.8 | 0.3 | 0.5 | 2.6 | 3.4 |
Poland | 3.6 | 4.1 | |||||||
Estonia | 7.1 | 7.5 | |||||||
Latvia | 6.9 | 6.6 | |||||||
Lithuania | 6.2 | 6.0 | |||||||
Slovak Republic | 3.1 | 3.8 | |||||||
Slovenia | 4.1 | 0.4 | 0.0 | 0.4 | 2.3 | 0.5 | 1.8 | 1.4 | 3.7 |
United States | 3.7 | 0.7 | 0.4 | 0.3 | 1.3 | 0.8 | 0.6 | 1.7 | 3.0 |
FINANCE AND BUSINESS SERVICES | |||||||||
Euro area 1/ | 2.8 | 2.1 | 1.8 | 0.2 | 2.0 | 1.1 | 0.9 | -1.3 | 0.7 |
France | 2.8 | 1.9 | 1.6 | 0.4 | 1.7 | 0.8 | 0.9 | -0.8 | |
Germany | 1.4 | 1.4 | 1.4 | 0.0 | 3.2 | 1.6 | 1.6 | -3.3 | -0.1 |
Italy | 2.9 | 2.6 | 2.4 | 0.2 | 0.7 | 0.6 | 0.2 | -0.4 | |
United Kingdom | 6.4 | 2.6 | 2.0 | 0.5 | 2.7 | 1.7 | 1.0 | 1.1 | 3.8 |
New Member States | |||||||||
Czech Republic | 3.5 | 1.4 | 1.1 | 0.3 | 1.9 | 1.3 | 0.6 | 0.2 | 2.1 |
Estonia | 2.3 | 0.1 | |||||||
Hungary | 5.1 | 3.9 | 3.3 | 0.6 | 0.7 | 0.5 | 0.2 | 0.5 | 1.2 |
Latvia | 8.5 | 3.9 | |||||||
Lithuania | 6.5 | 10.5 | |||||||
Poland | 7.7 | 0.1 | |||||||
Slovak Republic | -0.2 | -0.2 | |||||||
Slovenia | 5.8 | 2.5 | 2.0 | 0.6 | 3.4 | 1.3 | 2.0 | -0.1 | 3.2 |
United States | 4.3 | 1.9 | 1.5 | 0.4 | 1.9 | 1.2 | 0.7 | 0.4 | 2.4 |
Due to data limitations, the euro area only includes the following countries: Austria, Belgium, Finland, France, Germany, Italy, Netherlands, and Spain. Slovenia and Slovak Republic have been added to New Member States since they are very recent entrants to the euro area.
Gross Value-Added Growth and Contributions, 1980-95 and 1995-2005
(Annual average volume growth rates, percent)
VA | L | H | LC | K | KIT | KNIT | TFP | LP | |
---|---|---|---|---|---|---|---|---|---|
(1)=(2)+(5)+(8) | (2)=(3)+(4) | (3) | (4) | (5)=(6)+(7) | (6) | (7) | (8) | (9)=(1)-(2) | |
1980-1995 | |||||||||
MARKET ECONOMY | |||||||||
Euro area 1/ | 1.9 | 0.0 | -0.2 | 0.3 | 1.0 | 0.3 | 0.7 | 0.9 | 1.9 |
France | 1.8 | -0.1 | -0.5 | 0.4 | 0.7 | 0.2 | 0.4 | 1.2 | 1.9 |
Germany | 1.9 | -0.2 | -0.4 | 0.2 | 1.2 | 0.3 | 0.9 | 0.8 | 2.1 |
Italy | 1.9 | 0.3 | 0.2 | 0.1 | 0.8 | 0.3 | 0.6 | 0.8 | 1.6 |
United Kingdom | 2.5 | -0.3 | -0.6 | 0.3 | 1.2 | 0.6 | 0.7 | 1.5 | 2.7 |
United States | 3.0 | 1.2 | 1.0 | 0.2 | 1.1 | 0.5 | 0.6 | 0.7 | 1.8 |
1995-2005 | |||||||||
MARKET ECONOMY | |||||||||
Euro area 1/ | 1.9 | 0.6 | 0.4 | 0.1 | 1.1 | 0.5 | 0.6 | 0.2 | 1.3 |
France | 2.4 | 0.7 | 0.3 | 0.4 | 0.9 | 0.4 | 0.5 | 0.8 | 1.7 |
Germany | 1.0 | -0.4 | -0.5 | 0.1 | 1.0 | 0.5 | 0.5 | 0.4 | 1.4 |
Italy | 1.2 | 0.8 | 0.7 | 0.2 | 1.0 | 0.3 | 0.8 | -0.7 | 0.4 |
United Kingdom | 3.2 | 0.8 | 0.4 | 0.5 | 1.5 | 0.9 | 0.6 | 0.9 | 2.4 |
New Member States | |||||||||
Czech Republic | 2.7 | -0.1 | -0.3 | 0.2 | 2.3 | 0.7 | 1.6 | 0.6 | 2.9 |
Hungary | 4.4 | 1.0 | 0.6 | 0.4 | 0.8 | 0.3 | 0.5 | 2.6 | 3.4 |
Poland | 3.6 | 4.1 | |||||||
Estonia | 7.1 | 7.5 | |||||||
Latvia | 6.9 | 6.6 | |||||||
Lithuania | 6.2 | 6.0 | |||||||
Slovak Republic | 3.1 | 3.8 | |||||||
Slovenia | 4.1 | 0.4 | 0.0 | 0.4 | 2.3 | 0.5 | 1.8 | 1.4 | 3.7 |
United States | 3.7 | 0.7 | 0.4 | 0.3 | 1.3 | 0.8 | 0.6 | 1.7 | 3.0 |
FINANCE AND BUSINESS SERVICES | |||||||||
Euro area 1/ | 2.8 | 2.1 | 1.8 | 0.2 | 2.0 | 1.1 | 0.9 | -1.3 | 0.7 |
France | 2.8 | 1.9 | 1.6 | 0.4 | 1.7 | 0.8 | 0.9 | -0.8 | |
Germany | 1.4 | 1.4 | 1.4 | 0.0 | 3.2 | 1.6 | 1.6 | -3.3 | -0.1 |
Italy | 2.9 | 2.6 | 2.4 | 0.2 | 0.7 | 0.6 | 0.2 | -0.4 | |
United Kingdom | 6.4 | 2.6 | 2.0 | 0.5 | 2.7 | 1.7 | 1.0 | 1.1 | 3.8 |
New Member States | |||||||||
Czech Republic | 3.5 | 1.4 | 1.1 | 0.3 | 1.9 | 1.3 | 0.6 | 0.2 | 2.1 |
Estonia | 2.3 | 0.1 | |||||||
Hungary | 5.1 | 3.9 | 3.3 | 0.6 | 0.7 | 0.5 | 0.2 | 0.5 | 1.2 |
Latvia | 8.5 | 3.9 | |||||||
Lithuania | 6.5 | 10.5 | |||||||
Poland | 7.7 | 0.1 | |||||||
Slovak Republic | -0.2 | -0.2 | |||||||
Slovenia | 5.8 | 2.5 | 2.0 | 0.6 | 3.4 | 1.3 | 2.0 | -0.1 | 3.2 |
United States | 4.3 | 1.9 | 1.5 | 0.4 | 1.9 | 1.2 | 0.7 | 0.4 | 2.4 |
Due to data limitations, the euro area only includes the following countries: Austria, Belgium, Finland, France, Germany, Italy, Netherlands, and Spain. Slovenia and Slovak Republic have been added to New Member States since they are very recent entrants to the euro area.
While the differences indicate some unused growth potential in the euro area financial industry, they are also likely to signal a certain unsustainability of financial sector growth in the Anglo-Saxon countries: large and extraordinary profits might have distorted TFP growth in the United States and the United Kingdom. With the disappearance of these exaggerated profits in the financial sector, long-term growth in the United Kingdom will probably suffer more than that in the euro area. Most likely, the euro area will return to its steady-state growth of about 2 percent in the long term following the crisis, with some upside potential—for instance, if financial sector TFP growth returns to positive levels in the long run.
In general, applying the concept of steady-state growth to emerging economies is far from straightforward. Clearly, TFP growth tends to be higher in emerging than in advanced Europe, where investment in new technologies and capital equipment can still significantly accelerate the convergence process. That very process, however, complicates the analysis, and, unlike in the advanced countries, past growth in overall GVA is not easily interpreted as steady-state growth. Indeed, countries with lower incomes at the beginning of the transition tended to grow faster. For instance, the Czech Republic, which has recently been reclassified as an advanced country, grew at an average of 2.7 percent per year before the current crisis, whereas Estonia, which had a much lower initial income than the Czech Republic, was growing at close to 7 percent (Table 4).
Growth in GVA in the financial sector in the decade preceding the crisis stands out in the emerging economies. In general, a higher growth contribution from the services sector is not surprising during a period still influenced by the transition from a planned to a market economy. However, there are also notable differences among countries. In Latvia, for instance, the contribution of GVA in the finance and business services to overall GVA growth is relatively large. The financial industry in Estonia, though, grew less than overall industry, while in the Slovak Republic it did not grow at all. In other emerging markets for which data are available, the contribution of this sector was moderate. As in the advanced countries, the emerging economies in which the role of the financial sector in overall growth is relatively larger are likely to suffer more from the crisis. It is now clear, through hindsight, that growth in this sector was unsustainable and largely dependent on rapid growth in real estate services and other nontradable sectors.
The crisis, however, could slow the catching-up process in emerging Europe and lower its growth potential in the future, especially if capital inflows dwindle (Box 5). The slowdown will be more intense in poorer countries, which, over the past decade have been profiting from “downhill” capital flows—that is, from rich to poor countries—reinforcing the process of income convergence (Abiad, Leigh, and Mody, 2009). Relatively higher degrees of financial integration seem to have attracted capital from richer countries in particular in the financial and real estate sectors.13
Risks to Medium-Term Growth and Convergence in Emerging Europe
It is well established that greater financial integration helped European countries converge to higher income levels faster than the rest of the world (Abiad, Leigh, and Mody, 2009). Greater financial linkages within Europe and prevalence of foreign-owned banks in emerging Europe are reasons behind the larger volumes of bank-intermediated flows in Europe compared to the rest of the world. Thus, medium-term risks to potential growth from the financial crisis in the region mainly arise from reduced capital inflows. Both FDI and bank-related inflows are instrumental in helping these countries converge to higher income levels of their Western European neighbors, facilitated by higher financial and trade integration. The medium-term growth effects of capital inflows and other factors can be gauged by estimating a growth model with the usual determinants found in the literature (Vamvakidis, 2008; Abiad, Leigh, and Mody, 2009; and Schadler and others, 2006). Although growth regressions of this type have a number of well-known shortcomings, they provide a useful simple framework for analyzing the effects of the crisis. Applying the coefficients of the estimated model, a scenario analysis focuses on the variables most likely to be adversely affected in this crisis, based on the World Economic Outlook (WEO) forecasts of these variables over 2008–14 for the emerging economies.
Growth Regression 1/
(Dependent variable = annual growth of PPP adjusted per capita income (4-year averages) 1996-99 to 2004-07)
The equation is estimated by generalized least squares with random effects and time dummies for 2000-03 and 2004-07. A ***, **, and * represent significance at 1, 5, and 10 percent respectively. Initial income is log (real per capita PPP-adjusted GDP) at the beginning of the 4-year periods. Bank-related inflows comprise non-FDI, nonportfolio inflows into the banking and corporate sectors. Wald tests show that the three FDI-related coefficients are jointly significant at 2 percent, and the three bank-related flow coefficients are significant at 3 percent.
Growth Regression 1/
(Dependent variable = annual growth of PPP adjusted per capita income (4-year averages) 1996-99 to 2004-07)
Coefficient | Std error | ||
---|---|---|---|
Constant | 7.951 | 2.824 | *** |
Dummy for Europe (EUROPE) | 3.711 | 2.188 | * |
Initial income | -0.691 | 0.187 | *** |
Age dependency rate | -0.056 | 0.015 | *** |
Primary school enrollment ratio | 0.016 | 0.011 | |
Institutional development (Index of Economic Freedom 2006) | 0.156 | 0.296 | |
Gross fixed investment/GDP | 0.028 | 0.030 | |
Gross fixed investment/GDP * EUROPE | -0.084 | 0.083 | |
FDI/GDP | 0.435 | 0.190 | ** |
FDI/GDP * Initial income | -0.033 | 0.026 | |
FDI/GDP * Initial income * EUROPE | -0.003 | 0.009 | |
Bank-related inflows | 0.478 | 0.208 | ** |
Bank-related inflows * Initial income | -0.048 | 0.021 | ** |
Bank-related inflows * Initial income * EUROPE | -0.003 | 0.003 | |
Government debt/GDP | -0.011 | 0.004 | ** |
Government debt/GDP * EUROPE | -0.024 | 0.012 | * |
Number of observations | 273 | ||
Number of countries | 98 | ||
R-square | 0.28 |
The equation is estimated by generalized least squares with random effects and time dummies for 2000-03 and 2004-07. A ***, **, and * represent significance at 1, 5, and 10 percent respectively. Initial income is log (real per capita PPP-adjusted GDP) at the beginning of the 4-year periods. Bank-related inflows comprise non-FDI, nonportfolio inflows into the banking and corporate sectors. Wald tests show that the three FDI-related coefficients are jointly significant at 2 percent, and the three bank-related flow coefficients are significant at 3 percent.
Growth Regression 1/
(Dependent variable = annual growth of PPP adjusted per capita income (4-year averages) 1996-99 to 2004-07)
Coefficient | Std error | ||
---|---|---|---|
Constant | 7.951 | 2.824 | *** |
Dummy for Europe (EUROPE) | 3.711 | 2.188 | * |
Initial income | -0.691 | 0.187 | *** |
Age dependency rate | -0.056 | 0.015 | *** |
Primary school enrollment ratio | 0.016 | 0.011 | |
Institutional development (Index of Economic Freedom 2006) | 0.156 | 0.296 | |
Gross fixed investment/GDP | 0.028 | 0.030 | |
Gross fixed investment/GDP * EUROPE | -0.084 | 0.083 | |
FDI/GDP | 0.435 | 0.190 | ** |
FDI/GDP * Initial income | -0.033 | 0.026 | |
FDI/GDP * Initial income * EUROPE | -0.003 | 0.009 | |
Bank-related inflows | 0.478 | 0.208 | ** |
Bank-related inflows * Initial income | -0.048 | 0.021 | ** |
Bank-related inflows * Initial income * EUROPE | -0.003 | 0.003 | |
Government debt/GDP | -0.011 | 0.004 | ** |
Government debt/GDP * EUROPE | -0.024 | 0.012 | * |
Number of observations | 273 | ||
Number of countries | 98 | ||
R-square | 0.28 |
The equation is estimated by generalized least squares with random effects and time dummies for 2000-03 and 2004-07. A ***, **, and * represent significance at 1, 5, and 10 percent respectively. Initial income is log (real per capita PPP-adjusted GDP) at the beginning of the 4-year periods. Bank-related inflows comprise non-FDI, nonportfolio inflows into the banking and corporate sectors. Wald tests show that the three FDI-related coefficients are jointly significant at 2 percent, and the three bank-related flow coefficients are significant at 3 percent.
Gross fixed investment, inflows of foreign direct investment (FDI) and other bank-related capital inflows are among the growth determinants most likely to be adversely affected in this crisis (see table), for instance if foreign investors readjust their risk perception or domestic demand declines, including due to credit constraints. The large increases in government debt due to cyclical factors and public-intervention policies during the crisis will also weigh on overall growth via crowding-out effects. The growth regression is based on 4-year average panel data (1996–99, 2000–03, 2004–07) for 98 countries and a number of additional standard growth determinants unlikely to have been affected by the crisis. The “European” effects are captured by level and interactions of the dummy for Europe (EUROPE).
In particular, the following model is estimated:
Yit=α0 + α1Initial incomeit + β EUROPE + γ0FDIit + γ1 FDI * Initial incomeit + γ2FDIit * Initial incomeit * EUROPE + ŋ0Bankflowsit + ŋ1Bankflowsit * Initial incomeit + ŋ2Bankflowsit* Initial incomeit * EUROPE + δ0 GGDebtit + δ1GDebtit * EUROPE+ Ψ0Xit + εit
Yit is annual growth of (PPP-adjusted) real GDP per capita. Poorer countries grow faster by virtue of a lower starting level of income (α1). The effect of capital inflows on growth works through two channels in Europe: the direct level effect (γ0 and ŋ0) and the convergence effects (γ1 + γ2 for FDI in percent of GDP, and for ŋ1 + ŋ2 bank-related capital inflows in percent of GDP) based on the interaction of FDI and bankflows with initial per capita income and the EUROPE dummy. The results suggest that capital inflows have helped poorer countries grow faster, with a higher-than-average effect in Europe. Government debt has weighed down on growth, again with a larger European effect. Other growth determinants captured in the vector Xit include gross fixed investment in percent of GDP, which is highly correlated with FDI and bank flows and does not have a significant additional effect.
The regression can be used for scenario analysis. Based on the estimated model, the effect on average growth is computed by calculating the average fall in FDI and the average increase in government debt/GDP from WEO projections for 2008–14. For instance, the drop in average FDI between 2004–2007 and 2008–14 is 49 percent, and the increase in government debt/GDP is 35 percent. In the absence of WEO projections on bank-related capital inflows, the percentage for FDI drop is applied to bank-flows as well. These ratios are then applied to the 2004–07 average ratios of the three variables for each of the countries, and the effect on average growth for 2008–14 is computed.
Impact of the Crisis on Changes to Medium-Run Growth Estimates in Europe (WEO Average Scenario) 1/
Source: IMF staff calculations.1/ Based on growth regression (table). Scenario: 49 percent fall in FDI/GDP (FDI) and bank-related inflows and a 35 percent rise in government debt/GDP (DEBT) from the average in 2004–07 over 2008–14. These numbers were chosen by taking the average percent fall in FDI (WEO) projections, and applying the same percent to bank-flows and the average increase in DEBT over the same period and applying them to the other countries.Impact of the Crisis on Changes to Medium-Run Growth Estimates in Europe (WEO Average Scenario) 1/
Source: IMF staff calculations.1/ Based on growth regression (table). Scenario: 49 percent fall in FDI/GDP (FDI) and bank-related inflows and a 35 percent rise in government debt/GDP (DEBT) from the average in 2004–07 over 2008–14. These numbers were chosen by taking the average percent fall in FDI (WEO) projections, and applying the same percent to bank-flows and the average increase in DEBT over the same period and applying them to the other countries.Impact of the Crisis on Changes to Medium-Run Growth Estimates in Europe (WEO Average Scenario) 1/
Source: IMF staff calculations.1/ Based on growth regression (table). Scenario: 49 percent fall in FDI/GDP (FDI) and bank-related inflows and a 35 percent rise in government debt/GDP (DEBT) from the average in 2004–07 over 2008–14. These numbers were chosen by taking the average percent fall in FDI (WEO) projections, and applying the same percent to bank-flows and the average increase in DEBT over the same period and applying them to the other countries.Based on this simulation, the crisis could reduce medium-term growth by 0.6–2.5 percentage points in the NMS and by 0.4–2.2 percentage points in the other emerging economies (figure). By design, countries in which capital inflows had a larger role in GDP and are relatively poor are likely to experience a stronger adverse effect on average growth in 2008–14, a result consistent with the finding that countries with high precrisis investment-to-GDP ratios tend to have large output losses (IMF, 2009e).
Note: The main author of this box is Srobona Mitra.
Medium-Term Effects
For several reasons, the short- and medium-term effects of the current crisis might well exceed their longer-term impact. All components of the production function are subject to medium-term swings around their steady-state levels, because investment declines and structural unemployment often increases as recessions hit, with a temporary yet persistent effect on potential output. Recessions associated with financial crises have especially severe impacts on medium-term growth, if accompanied by credit crunches and real estate bubble busts, and have a persistently adverse effect on output (Cerra and Saxena, 2008; Claessens, Kose and Terrones, 2008; and IMF, 2009e). Thus, the next five to seven years are likely to see large drops in growth rates followed by a recovery driven by time-varying potential growth around its long-run steady state. The medium term is also the relevant horizon for monetary policymakers, and most of the fiscal costs of the crisis (from financial sector intervention, stimulus expenditure, and cyclical adjustments) will be determined during this time.
Although European experts and policymakers agree that potential output in advanced countries is likely to fall as a result of the financial crisis and the consequent deep global recession, the degree of its decline is far from certain. A recent study published by the OECD (Furceri and Mourougane, 2009), for instance, estimates that the permanent reduction in potential output due to typical financial crises is between 1.5 percent and 2.5 percent, with even higher losses for particularly deep crises. While these estimates are high, they also indicate substantial variation among countries, a point also stressed in IMF (2009e). According to the forecasts by the European Commission (2009c), the growth of potential output in the euro area could decline to 0.7 percent during 2009–10, from 1.8 percent during 2000–06, indicating a loss of more than 1 percent per year relative to the precrisis developments. Recent estimates based on a European Commission simulation model put the estimated loss at anywhere from 0.5 percent to about 4.5 percent, depending on the scenario (Koopman and Székely, 2009). One reason for such divergence is that the current economic and financial crisis is a unique event. Another is that it is not easy to distinguish medium-term movements in potential from other, shorter-term fluctuations in output.
With a focus on the euro area as an example, standard statistical methodologies for calculating medium-term potential output, such as the HP filter, generally confirm that potential growth turned negative in 2008:Q4–2009:Q1, but the extent of the drop varies widely (Figure 13). Despite falling potential growth, the euro area output gap (that is, the percentage difference between estimated potential and observed output) opened widely in early 2009 with the continued sharp drop in actual output (Table 5). The HP filter, however, renders real-time estimates of potential output at the end of the sample period imprecise if a persistent shock has recently occurred. A simple method for avoiding such end-of-sample problems is to extend the sample using forecasts of GDP. The forecast-adjusted HP filter tracks the sharply falling potential growth since 2007, taking into account the expected persistence of the dramatic fall of output growth in 2009 and the sluggish recovery from 2010.
Euro Area: Potential (Medium-Term) Growth with Different Methodologies, 1993:Q2–2009:Q1
(Percent)
Sources: IMF, World Economic Outlook; and IMF staff estimates.Euro Area: Potential (Medium-Term) Growth with Different Methodologies, 1993:Q2–2009:Q1
(Percent)
Sources: IMF, World Economic Outlook; and IMF staff estimates.Euro Area: Potential (Medium-Term) Growth with Different Methodologies, 1993:Q2–2009:Q1
(Percent)
Sources: IMF, World Economic Outlook; and IMF staff estimates.Potential Output and Output Gaps in the Euro Area
Percentage points of potential output.
AR(1) coefficient.
Potential Output and Output Gaps in the Euro Area
Output Gap | Potential Growth | |||||
---|---|---|---|---|---|---|
Methodology | 2009:Q1 | Change since 2007:Q4 1/ | 2009:Q1 | Change in the level of potential output since 2007:Q4, percent | Persistence 2/ | Variance |
HP filter | -4.38 | -5.87 | 0.32 | -0.04 | 1.00 | 0.01 |
HP filter with WEO forecast | -1.24 | -4.66 | -0.13 | -0.21 | 1.05 | 0.03 |
New Keynesian Model | -3.25 | -4.57 | -0.25 | -0.74 | 0.37 | 0.06 |
Percentage points of potential output.
AR(1) coefficient.
Potential Output and Output Gaps in the Euro Area
Output Gap | Potential Growth | |||||
---|---|---|---|---|---|---|
Methodology | 2009:Q1 | Change since 2007:Q4 1/ | 2009:Q1 | Change in the level of potential output since 2007:Q4, percent | Persistence 2/ | Variance |
HP filter | -4.38 | -5.87 | 0.32 | -0.04 | 1.00 | 0.01 |
HP filter with WEO forecast | -1.24 | -4.66 | -0.13 | -0.21 | 1.05 | 0.03 |
New Keynesian Model | -3.25 | -4.57 | -0.25 | -0.74 | 0.37 | 0.06 |
Percentage points of potential output.
AR(1) coefficient.
Monetary policymakers are primarily interested in the level of output that avoids pressure for inflation to either increase or decrease. Conceptually, this calculation requires an approach to potential output different from that captured by simple statistical trends (see also Chapter 3). For example, following a standard methodology derived from New Keynesian Models (NKM), potential output is defined as the hypothetical, flexible-price equilibrium output that would be produced in the absence of price and wage rigidities. NKM potential output is more volatile than an HP-filtered trend because it also reflects the reaction of potential output to short-term, transitory aggregate supply shocks, depending, among other things, on the elasticity of labor supply (Basu and Fernald, 2009).14 Estimating potential growth for the euro area using these different methods illustrates this point (Figure 13).15
Although the output gap in 2009:Q1 (the most recent reading) from all estimates is large and negative, its size varies widely (Table 5). For instance, the simple HP-filtered estimates point to a rapidly widening output gap at the end of the sample. In contrast, the 2009:Q1 gap for the forecast-adjusted HP filter is much smaller, reflecting the sharp drop in potential growth (Figure 13). The latter is picking up the drastic fall in GDP forecast for 2009, followed by a very modest recovery in 2010. The NKM-based output gap for 2009:Q1 is large compared to the forecast-based. HP-filtered estimate, which reflects the smaller and less persistent drop in the estimate of potential growth.
Differences in country-specific and sometimes in methodological factors lead to large variations in estimates of how the crisis will affect potential growth for individual countries. As for the aggregate euro area, however, potential growth is projected to fall over the medium term in most of the large, advanced European countries as well (Box 6). The projected potential losses in output vary between ½ percent and 5½ percent in 2009, with larger drops in countries suffering from deflated booms in the financial or construction industry (or in both)—for instance, Ireland, Portugal, Spain, and the United Kingdom.
Effect of the Financial Crisis on Potential Growth in Western Europe
While there is little doubt that the financial crisis will adversely affect the growth potential of Europe, it is difficult to estimate its impact precisely. Nor will applying standardized or uniform methodologies (for example, HP filters) likely be able to capture the differing elements that weigh on the growth potential of individual countries. Hence, IMF staff uses an approach that allows level and growth rate effects on potential output estimates to reflect country-specific circumstances. In addition, the extent to which potential output losses will be recovered in the longer term is quite uncertain. With that caveat in mind, illustrating the diverse impact of the financial crisis on the near-term growth potential of countries in Western Europe nonetheless provides insight.
IMF staff currently projects a slowdown in the growth of potential output across the region, including some contractions in the level of potential output (first figure). The extent of the slowdown varies widely, ranging from ½ percent to 5½ percent between 2007 and 2009. The sharpest slowdowns are expected in countries where the crisis is likely to require large structural adjustment, for instance, when precrisis growth reflected unsustainable contributions from certain sectors. In other countries, continued weakness in investment and in the labor market is expected to reduce potential output.
Selected European Countries and the United States: Potential Output Growth, 2007–09
(Percent)
Sources: IMF, World Economic Outlook; and IMF staff estimates.Selected European Countries and the United States: Potential Output Growth, 2007–09
(Percent)
Sources: IMF, World Economic Outlook; and IMF staff estimates.Selected European Countries and the United States: Potential Output Growth, 2007–09
(Percent)
Sources: IMF, World Economic Outlook; and IMF staff estimates.Extreme expansions of the financial and construction sectors will partly unwind, subtracting from potential output (second figure). In Ireland, in particular, both the financial and the construction sectors expanded rapidly going into the crisis. Growth in Iceland, the United Kingdom, and Portugal was founded on a booming financial sector; growth in Spain, on a construction boom. These sectors are unlikely to recover to their precrisis strength in the medium term.
Selected European Countries and the United States: Financial Intermediation and Construction, 2000–07
(Share of Gross Value Added)
Selected European Countries and the United States: Financial Intermediation and Construction, 2000–07
(Share of Gross Value Added)
Selected European Countries and the United States: Financial Intermediation and Construction, 2000–07
(Share of Gross Value Added)
A structural adjustment to a more balanced growth path might also reduce potential output. For example, heavy reliance on export-driven growth has expanded the German export sector, including car production, at the expense of the nontradables sector. With global imbalances unwinding, a downward correction in world trade—and, with it, German exports—is likely for the medium term (third figure). Potential output will fall while resources are reallocated.
Selected European Countries and the United States: Change in Balance of Real Growth
(Foreign balance minus total domestic demand, percent of real GDP)
Sources: IMF, World Economic Outlook; and IMF staff calculations.Selected European Countries and the United States: Change in Balance of Real Growth
(Foreign balance minus total domestic demand, percent of real GDP)
Sources: IMF, World Economic Outlook; and IMF staff calculations.Selected European Countries and the United States: Change in Balance of Real Growth
(Foreign balance minus total domestic demand, percent of real GDP)
Sources: IMF, World Economic Outlook; and IMF staff calculations.Note: The main author of this box is Franziska Ohnsorge.
For Europe’s emerging economies that are still converging toward some steady state, risks to potential growth mainly stems from reduced capital inflows that the region has, so far, depended upon (Box 5). The relevant question here is: what could go wrong in the crisis that could also affect medium-term growth and convergence in emerging Europe? Simulations from an econometric growth model suggest that lower foreign direct investment (FDI) and bank-related inflows (for example, those from foreign parent banks to eastern European subsidiaries) and higher government debt stemming from the crisis could shave medium-term growth by 0.6–2.5 percentage points in the New Member States and by 0.4–2.2 in the other emerging economies. Countries in which capital inflows had a larger role in GDP and are relatively poor are likely to experience a stronger adverse effect on average growth in 2008–14. Still, average growth in the emerging economies is likely to be higher than in the advanced economies, mainly because convergence will probably continue at some level.
Policy Implications
Further structural reforms remain a priority, particularly in the euro area. Beyond the medium term, the large advanced economies in the euro area should return to their historical levels of potential (or steady-state) growth of about 2 percent on average, with some variation reflecting country-specific demographics. The historical difference between TFP growth in the euro area and in the United States, however, suggests that there are incentives and scope for structural reform, even though some TFP growth in the United States during 1995–2005 could have been distorted by exaggerated profits in the financial and the information and technology sectors.
The EU’s Lisbon Agenda is a useful template for such reforms (European Commission, 2005, 2007), with policy implications for all countries. A number of important items are on the broader reform agenda:
promoting education and training of the labor force;
modernizing social protection systems, including pensions and health care, to ensure their social adequacy, financial sustainability, and responsiveness to changing needs and to support labor force participation and better retention in employment;
strengthening private incentives for research and development;
improving productivity-enhancing infrastructure;
far-reaching and timely implementation of the Services Directive;16 and
unwinding the state aids and job subsidies implemented to boost demand-side policies during this crisis.
Next to their longer-term effects, some of these measures will also favorably affect the medium term. For instance, labor force training would prevent the loss of skills that usually accompanies people who leave the labor force during a recession. Job training is particularly welcome, as it can foster labor mobility and prevent skill erosion during a period where longer unemployment spells threaten to turn cyclical into structural unemployment, with negative repercussions for potential output.
In addition, sound policies that maintain the momentum toward convergence are essential for emerging economies. To secure the capital inflows required for a smooth continuation of the convergence process, governments will need to ensure a business-friendly environment, strengthen policy frameworks that reduce political uncertainty, further support integration of financial markets and provide macroeconomic stability (Abiad, Leigh, and Mody, 2009). In many emerging economies, policymakers will face the additional burden of very volatile exchange rates and financing costs against the backdrop of high government debt and fiscal deficits. Improving frameworks for fiscal policy and boosting their transparency could help reduce the volatility of the business cycle and improve growth prospects (see Chapter 4).
Mirroring the argument for advanced European economies, not all of these inflows will have been translated into sustainable productivity increases. In many emerging European countries, the value added in the finance and business services area far outpaced overall value added (Figure 12).
For discussions of the NKM, see, for instance, Coenen, Smets, and Vetlov (2009); Woodford (2003); Justiniano and Primiceri (2008); and Weidner and Williams (2009). Kuttner (1994) and Laubach and Williams (2003) discuss the volatility of NKM potential output. In the long run, flexible-price potential output should converge to the long-term trend or steady-state output, but over shorter horizons they could be very different, even if the latter included some time-varying component captured, for instance, by statistical filters.
The estimates are based on a simple NKM model for the euro area estimated using Bayesian methods and the Kalman filter. The model allows for persistent shocks to output, inflation, and exchange rates.
Available via the Internet: http://ec.europa.eu/internal_market/services/services-dir/index_en.htm.