Journal Issue
Share
Article

Bulgaria

Author(s):
International Monetary Fund
Published Date:
June 2010
Share
  • ShareShare
Show Summary Details

I. Estimating Potential Growth and Output Gaps in an Emerging Economy: the Case of Bulgaria1

A. Introduction

1. The current economic crisis raises the question whether potential output growth in Bulgaria in the coming years could be markedly lower than that during the boom years. The current recession was preceded by an investment boom in construction, real estate and financial sectors. Now that the boom in these sectors has ended, the significant decline in investment could have large negative effects on potential output. Moreover, it will take considerable time for the excess labor and resources to be absorbed by other sectors, which suggests that the natural rate of unemployment may rise and remain higher.

2. Economic crisis aside, the estimates of potential output based on conventional methods (detrending methods, production function approach) are often subject to various statistical problems. They are sensitive to methods as well as sample selections used for estimations. Estimates based on the commonly used HP filters and production function approach can differ significantly, depending on sample period, detrending parameters, and assumptions about the initial capital stock and the capital depreciation rate (Cerra and Saxena, 2000; Zhou, 2003).

3. For emerging economies including Bulgaria, the estimation of potential output is further hampered by short data periods and measurement problems with capital stocks. 2 Linear time series methods (e.g. VAR model, multivariate model) restrict parameters to be constant, and are therefore also not appropriate for the emerging Eastern European countries that have gone through significant structural changes as well as boom and bust cycles since 1992, most recently during 2005-09.

4. To overcome some of these problems this chapter presents a simple method for estimating Bulgaria’s potential growth and output gaps. Assessments of potential growth and output gaps are particularly important, since they serve as crucial inputs for assessing the stance of fiscal policies. Instead of estimating the level of potential output, we estimate potential growth, based on an estimated capital-output ratio and assumptions about future employment and productivity growth. We then identify a base year when the economy is at its potential (i.e., the output gap is zero) by assessing inflation and current account developments (i.e., the internal and external balances). The level of potential output and output gaps are calculated with the estimated potential growth and the selected base year. Unlike the conventional methods discussed above, our estimated output gaps depend only on potential growth but not on sample selections.

B. The Method

Estimating Potential GDP Growth

5. The estimation of potential output growth is based on the trend growth of (i) the capital stock; (ii) the labor input based on an assumed labor participation rate and working-age population growth; and (iii) TFP estimates based on a simple two-factor Cobb-Douglas production function.

6. The capital stock is estimated based on a capital/output ratio that is derived from a simple profit optimization solution. Since the National Statistics Institute (NSI) does not publish data on the capital stock, they are estimated based on the commonly used permanent inventory method (for example, Ganev, 2005). But we differ from many including Ganev (2005) in estimating the initial capital stock. Instead of following the commonly used approach which assumes that the initial capital stock is equal to the gross investment in the initial year (or a multiple of that, as in Ganev, 2005), we estimate the initial capital stock based on a simply analytical framework, in which a profit maximization based on a two-factor Cobb-Douglas production function

determines the capital-output ratio as follows:

where α is output elasticity of labor, δ is capital depreciation rate, and r represents the real interest rate. Based on the assumptions of α = 0.6, r = 0.06, and δ = 0.1, the estimated capital-output ratio is about 2.5 in 2000, in line with the existing estimates for OECD countries and for Bulgaria (Ganev, 2005). The assumed depreciation rate of 10 percent is slight higher than the usual assumption for OECD countries (about 8 percent), but is assumed to capture the significant economic transformation that took place in Bulgaria when new investment may have simply replaced the old capital stocks (for example, inactive production facilities).

7. The estimated values for capital stock are presented in Table 1. The implied capital output ratios range from 2 to 2.5, and show a rising trend during the recent investment boom.

Table 1.Estimates of Capital Stock, 2000-2009 1/(In billions of leva at 2001 prices, unless otherwise indicated)
YearGross capital formationGDPCapital stockCapital-output ratio
20005.428.972.12.5
20016.430.071.32.4
20026.431.470.62.3
20037.632.971.12.2
20048.735.172.82.1
200511.237.376.72.1
200613.439.782.42.1
200716.342.190.42.1
200818.544.799.82.2
2009 (est.)12.342.4102.22.4
Sources: NSI; IMF staff estimates.

Assuming capital-output ratio equals to 2.5 in 2000, and capital depreciation rate of 10 percent for 2001-09.

Sources: NSI; IMF staff estimates.

Assuming capital-output ratio equals to 2.5 in 2000, and capital depreciation rate of 10 percent for 2001-09.

8. The decomposition of growth suggests that growth has been increasingly driven by capital and labor inputs while the contribution of TFP growth declined steadily since 2002. The average TFP growth for the period of 2001-08 is about 2.5 percent (see Table 2). Our estimates measure the labor inputs with hours worked (national account based), though the results based on employed workers turned out to be very similar (2.6 percent). Based on the trend growth in employment and in TFP, the potential growth is about 5 percent during this period, in line with the estimate by the Bulgarian National Bank (BNB, 2009).

Table 2.Sources of Growth, 2001-2009.(Annual percentage change)
YearGDP growthContribution of labor 1/Contribution of capitalGrowth of TFP 2/
20014.10.0−0.54.5
20024.50.1−0.44.8
20035.01.40.33.3
20046.62.40.93.3
20056.21.42.12.7
20066.31.83.01.5
20076.22.03.90.3
20086.01.94.2−0.1
2009 (est.)−5.0−1.70.9−4.3
2001-20085.61.41.72.5
2001-20094.41.01.61.8
Sources: NSI; IMF staff estimates.

Labor inputs are measured with hours worked.

Solow’s residual, using labor share of 0.6 and capital share of 0.4.

Sources: NSI; IMF staff estimates.

Labor inputs are measured with hours worked.

Solow’s residual, using labor share of 0.6 and capital share of 0.4.

Potential Growth 2009-15

9. Growth declined sharply in 2009 after the economy was hit hard by the global economic and financial crisis. Capital inflows dropped from a peak of 44 percent of GDP in 2007 to less than 10 percent of GDP in 2009. As a result, investment fell by nearly 30 percent, after rising more than 20 percent annually during the previous two years. Employment also fell and while the unemployment rate rose rapidly.

10. Can Bulgaria return to the high potential growth rates of 2001-08? Growth during the boom years was driven by large capital inflows that fueled strong growth in the non-tradable sector. As capital inflows are likely to stabilize at a level well below that of the boom years, and growth in the non-tradable sector is likely to remain weak at best, growth would only be high if the tradable sector takes over as an engine of growth. More specifically:

  • The end of the investment boom and sharply reduced capital inflows would imply much lower investment and hence lower potential growth in the coming years.

  • With lower investment, the robust employment growth during 2001-08 would be difficult to achieve. Much of the strong employment growth was driven by strong growth in the non-tradable sector. Total employment rose by 20 percent during this period, of which 15 percent was the contribution from the construction, real estate, wholesales and financial service sectors (Figure 1).

  • Of course, lower investment could be offset by higher TFP growth. But improving TFP growth would require supply side reforms.

Figure 1.Bulgaria: Employment Growth and Sectoral Contributions

(In percent)

Source: NSI; IMF Staff calculations.

1/ The number of hours worked, national account based.

11. The potential output growth rate in the coming years could be markedly lower than that during 2001–08. The government’s 2009-12 Convergence Programme envisages a potential growth of about 3 percent in 2009-12. Staff estimates indicate that achieving this would require a rapid reversal of the trending decline in TFP growth, specifically, a TFP growth of 2.5 percent and a potential employment growth of about 2 percent, which is close to that during 2001-08 and may be difficult to achieve for the reasons discussed earlier. Staff’s current scenario assumes a gradually increased trend growth in both full employment (and a return to the NAIRU of 5 percent) and TFP growth, with potential growth rising to 3.5 percent in 2014-15. Potential growth for the period 2009-15 is about 2.6 percent (Table 3), slightly lower than the official estimates (BNB, 2009).

Table 3.Estimates of Potential Output Growth, 2009-2015(Annual percentage change)
YearPotential GDP growth 1/Contribution of labor 2/Contribution of capitalTrend growth of TFP
20091.50.50.70.2
20101.50.80.10.6
20112.00.70.40.9
20123.00.90.51.6
20133.20.90.71.6
20143.40.90.91.6
20153.60.91.11.6
2001-20085.01.51.02.5
2009-20152.60.80.61.2
Sources: NSI; IMF staff estimates.

Based on labor share of 0.6 and capital share of 0.4.

Full employment measured by hours worked.

Sources: NSI; IMF staff estimates.

Based on labor share of 0.6 and capital share of 0.4.

Full employment measured by hours worked.

12. potential growth is set to decline further in the longer term. Bulgaria faces a serious problem of aging population. Its population is projected to decline by 28 percent between 2008 and 2060, while the old age dependency ratio would exceed 60 percent in 2060.3 This is expected to reduce its labor force significantly by more than 25 percent in the next 50 years. The 2009-12 Convergence Programme foresees a steady decline in potential growth to 0.3 percent in 2050, even with the labor participation ratio raised to 70 percent.

Estimating Output Gaps

13. As the first step, we choose 2005 as the base year when the economy is at its potential and the output gap is closed. This is based on the developments in both internal and external imbalances. Specifically, in 2004 the inflation in the non-tradable sector reached its lowest since 2001 and the current account deficit was at a reasonable 6.6 percent of GDP. Both inflation and current account deficit began to rise in 2005 and large imbalances emerged after 2005. Output gap estimates using the HP filter also suggest that 2004 or 2005 was the turning point of the recent economic cycle in Bulgaria, and results were not sensitive to the detrending parameters or the end-points of the sample (Figure 2).

Figure 2.Bulgaria: Estimates of Output Gaps based on the HP filter

Source: Staff Estimates.

14. Output gaps are calculated based on the potential growth rates and assuming potential output equals actual output in 2005. The estimates suggest the economy was about 3.5 percent above potential in 2008 and 3.2 percent below potential in 2009 (Figure 3). The authorities’ output gap estimates in the Convergence Program were 5.8 percent for 2008 and -2.7 percent for 2009. Output gap estimates assuming a zero gap in 2004 are also presented (Figure 4) and the results are similar, although in this case the output will return to its potential in 2014 instead of 2015.

Figure 3.Bulgaria: Potential Output, Growth, and Output Gaps

(Based on output gap = 0 in 2005)

Source: NSI; and IMF staff estimates.

Figure 4.Bulgaria: Potential Output, Growth, and Output Gaps

(Based on output gap = 0 in 2004)

Source: NSI; and IMF staff estimates.

C. Conclusions

15. Potential output growth in the coming years could be markedly lower than that during the boom years. The current downturn was preceded by an investment boom in construction, real estate and financial sectors. Now that the boom has ended, it may take considerable time for the excess labor and resources to be absorbed by other sectors, in particular by the export sector. This suggests that in the next few years, the natural level of rate of unemployment will rise and remain higher, and the full employment level is likely to decline. To sustain potential growth at about 3 percent, as set in the government’s recent convergence program, would require a rapid reversal of the recent trend decline in TFP growth and a steady increase in employment growth to close to 2 percent. This would be challenging and require significant improvements in labor productivity and competitiveness, as well as reforms to further improve labor mobility and participation.

16. Our output gap estimates suggest that it could take about five years before the output in Bulgaria returns to its potential. Large output gaps emerged in 2007-08 as a result of the domestic demand boom. The end of the boom in end-2008 and the large negative output gaps in the next 3 years implies a slack economy and downward pressures on core inflation.

References:

Prepared by Jianping Zhou.

For example, annual real GDP series are available for 1996-2009, while quarterly GDP data are available only after 2002 Q1. Similar data problems also exist in other emerging EU countries, and for countries with data going back further, the quality of the data for early years is often questionable.

Eurostat News Release, 119/2008.

Other Resources Citing This Publication