Sweden: 2013 Article IV Consultation

International Monetary Fund. European Dept.

doi:10.5089/9781484382554.002.A001

Jump to Content Jump to Main Navigation

Sweden: 2013 Article IV Consultation

Author: International Monetary Fund. European Dept.

Publication Date:: 05 Sep 2013

eISBN:: 9781484382554

Language:: English

Keywords:: ISCR; CR; bank; borrower; output gap; depositor; fund staff calculation; consumer price inflation; HICP rates of inflation; services inflation; growth rate; Potential output; Inflation; Consumer price indexes; Global

Download PDF (879.9 KB)

This 2013 Article IV Consultation examines the performance of Sweden’s fiscal policies to counter effects of global financial crisis. Economic growth in Sweden has been moderate since global financial crisis of 2008–2009. The IMF report posits that with potential growth moderately weaker and the natural rate of unemployment to remain elevated, policies should focus on growth-enhancing reforms, especially in the labor market. It suggests that good policies that secure the soundness of Swedish international banking groups are expected to benefit borrowers not only in Sweden but across the region.

Abstract

I. Potential Output and Unemployment¹

Sweden has suffered a permanent output loss from the crisis and potential output growth will likely moderate going forward relative to pre-crisis rates. At the same time, the natural rate of unemployment has increased and is set to decline only gradually over the medium term.

A. The Great Recession in Sweden

1. Sweden was hit strongly after the global financial crisis of 2008–09, and growth has moderated after its initial bounce back. After a contraction of 5 percent in 2009, Swedish growth rebounded to above 3½ percent in 2010 and 2011 before slowing down together with European and global trade (see Figure 1). Meanwhile, the unemployment rate increased to 8 percent by 2012 (see Figure 2).² The manufacturing sector, already on a downward trend, was also hard-hit, with a close to 10 percent decline in employment in 2009.

2. Against this background, this paper assesses the effects of the crisis on potential output. In 2012, output was 10 percent below the level that would have been reached had output continued to grow at the average 2000–07 growth rate of 3¼ percent. Under current projections, the output loss relative to the 2007 trend continues to widen over the forecast horizon, implying a permanent output loss (see Figure 3).

3. Simply extending the pre-crisis trend might not be a good indicator of Sweden’s growth potential. Table 1 shows estimates from the Swedish Ministry of Finance, the Riksbank, the National Institute of Economic Research (NIER), the European Commission, and Fund staff as reflected in the World Economic Outlook, implying output gaps between −3½ and −1 percent of potential output in 2013. The wide range of estimates indicates the uncertainty of the underlying estimates and suggests a closer look at the differences behind them. In what follows, we estimate potential output in several different ways to inform the discussion about the current cyclical stance of the Swedish economy and longer-term effects on potential growth.

Table 1.

Output Gap Estimates

(Percent of potential output)

Sources: Ameco database (June 2013), National Institute of Economic Research (June 2013b), Riksbank (April 2013), Swedish Ministry of Finance (April 2013), World Economic Outlook (July 2013).

^{^1/}Average of quarterly estimates.

^{^2/}World Economic Outlook.

Table 1.

Output Gap Estimates

(Percent of potential output)

	2010	2011	2012	2013	2014
European Commission	−1.6	−0.1	−1.3	−1.6	−1.1
Ministry of Finance	−3.1	−1.3	−2.2	−3.5	−3.3
National Institute of Economic Research	−3.6	−1.5	−1.9	−2.1	−1.7
Riksbank ^1/	−2.8	−0.5	−0.9	−1.3	−0.6
Fund staff ^2/	−1.0	0.8	−0.4	−1.0	−0.7

Sources: Ameco database (June 2013), National Institute of Economic Research (June 2013b), Riksbank (April 2013), Swedish Ministry of Finance (April 2013), World Economic Outlook (July 2013).

^{^1/}Average of quarterly estimates.

^{^2/}World Economic Outlook.

Table 1.

Output Gap Estimates

(Percent of potential output)

	2010	2011	2012	2013	2014
European Commission	−1.6	−0.1	−1.3	−1.6	−1.1
Ministry of Finance	−3.1	−1.3	−2.2	−3.5	−3.3
National Institute of Economic Research	−3.6	−1.5	−1.9	−2.1	−1.7
Riksbank ^1/	−2.8	−0.5	−0.9	−1.3	−0.6
Fund staff ^2/	−1.0	0.8	−0.4	−1.0	−0.7

Sources: Ameco database (June 2013), National Institute of Economic Research (June 2013b), Riksbank (April 2013), Swedish Ministry of Finance (April 2013), World Economic Outlook (July 2013).

^{^1/}Average of quarterly estimates.

^{^2/}World Economic Outlook.

B. Potential Output

4. Three standard methods to estimate potential output are deployed to identify general trends while avoiding a false sense of precision: a simple HP filter, a production function, and a multivariate approach as in Benes at al. (2010) (see the Appendix for details). For the univariate HP filter, the smoothing parameter is varied to cover ranges suggested in the literature.³ For the production function approach, potential levels of total factor productivity and the unemployment rate are determined by the HP filter, again obtained from a range of smoothing parameters, and combined with the actual levels of the labor force and the capital stock. To alleviate the end-point problem from the HP-filter, forecasts are extended until 2025. The multivariate filter builds on a macroeconomic model, which jointly determines potential output and the natural rate of unemployment. Note that output forecasts are determined within the model and, hence, may differ from desk projections.

5. Results point to a slightly negative output gap in 2013 after widening from 2012. Based on the range of output gap estimates in 2012, output was likely close to potential. However, as growth is projected to remain weak in 2013, the output gap is projected to widen, with all methods pointing to a negative output gap in the range of -1.6 (multivariate filter) to -0.5 (HP filter and production function with low smoothing) percent of potential output (see Figure 4). The average measure stands at a negative gap of 1 percent, broadly in line with developments in a number of peer economies. For example, Germany’s output gap is projected to widen to around -½ percent of potential output (see Figure 5).

6. Potential output growth is currently below pre-crisis rates. The various approaches point to 2012 potential growth ranging from just below 2 to 2¼ percent, relative to 2000–07 average potential growth of close to 3 percent (based on the average estimate) (see Figure 6). A decomposition of potential growth using the production function approach suggests that variations in employment and total factor productivity⁴ are important contributing factors to the current slowdown as their contributions to growth declined from 0.7 and 0.2 percentage points in 2005 to 0.5 and -0.1 percentage points, respectively, in 2012 (see Figure 7).

7. Over the medium term, the output gap is set to close while potential growth will remain somewhat muted. Results suggest the output gap will close in two to five years, depending on the model. However, all approaches suggest a decline in potential growth over the medium term relative to the pre-crisis 2000–07 average, with the average measure pointing to a ½ percentage point decline. The production function points to a slowdown in potential growth from an average of close to 3 percent during 2000–07 to just above 2 percent in the medium term as the contribution to growth from employment remains subdued over the forecast horizon (see below for further analysis of the structural rate of unemployment).

C. The Natural Rate of Unemployment

8. Headline inflation and the output gap provide seemingly conflicting views on the level of the underlying, equilibrium, rate of unemployment. Year-on-year consumer price (CPI) inflation below zero seems to suggest the economy is operating well below potential capacity and, hence, that a substantial share of the elevated unemployment rate is explained by cyclical factors. In contrast, a small output gap, estimated at around -1 percent of potential GDP in 2013, suggests a large part of the actual level of unemployment is structural in nature. This is consistent with the NIER’s (2013a) estimate of an equilibrium level of unemployment at just below 7 percent for 2013, gradually decreasing to 6.5 percent in 2017 as previously implemented reforms will have effect.

Explaining the Low Rate of Inflation

9. However, the low rate of inflation is driven mostly by exchange rate and interest rate developments. While there is some slack in the economy with a moderating impact on price developments, the direct effects of low interest rates (in particular, mortgage rates, which play a significant role for household spending in Sweden) and low imported inflation are the main factors in explaining current year-on-year deflation in consumer prices.

Low interest rates directly impact Swedish consumer price inflation. While the Harmonized Index of Consumer Prices (HICP) excludes certain interest costs, these are included in the Swedish Consumer Price Index (CPI),⁵ which the Riksbank is monitoring against its 2 percent inflation target. Hence, current low interest rates have contributed to substantial movements in the CPI (see Figure 8). To support monetary policy decision making, Statistics Sweden also computes two underlying consumer price indices: the CPIF, which captures the CPI at fixed interest rates, and the CPIX, which modifies the CPI to exclude all interest costs for owner-occupied dwellings.⁶ The CPIF and the HICP rates of inflation evolve similarly, underlining the importance of the effect of interest rate changes.

Low imported inflation is another key factor in explaining the low rate of inflation. Figure 9 shows a decomposition of HICP inflation, while Figure 10 shows the rate of inflation in components with high import content and high domestic content. Inflation in services, a mainly domestic component, has been relatively constant at just below 2 percent over the last few years, only declining recently. Conversely, non-energy industrial goods, which traditionally have high import content, have experienced deflation, contributing to lower inflation.

Estimating the Natural Rate of Unemployment

10. Mirroring the results on potential output, there are indications that the natural rate of unemployment is elevated. Based on the multivariate filter approach, which computes an estimate of the natural rate of unemployment along with potential growth, the natural rate of unemployment started to increase from the early 2000s (see Figure 11). Along with a predicted improvement in potential growth, likely reflecting past labor market reforms, the model suggests that the natural rate will begin to decline over the forecast horizon. However, with potential growth remaining moderately below pre-crisis rates, the natural rate of unemployment is projected to remain elevated at around 6½ percent over the medium term, a result also confirmed from HP-filter approaches.⁷

11. Okun’s law can be used to generate an alternative estimate of the natural rate of unemployment. Okun (1962) estimated a relationship between real GDP and unemployment, which can be expressed in two forms: the gap form and the change form. The gap form expresses a linear negative relationship between the output gap, Y^gap, and the excess unemployment, u, over the natural rate, u, during 1993–2012. The change form relates the growth rate in GDP, $\frac{Δ Y}{Y},$ , to percentage point changes in the unemployment rate, Δu. Equations (1) and (2) show both forms, where c and k are parameters. k indicates the average growth rate of potential output, and the intercept in a regression of the unemployment rate on the output gap can be interpreted as the (time-invariant) natural rate of unemployment.

\begin{matrix} (1) & Y^{g a p} = - c (u - \bar{u}) & or & u = \bar{u} - \frac{1}{c} Y^{g a p} \end{matrix}

\begin{matrix} (2) & \frac{Δ Y}{Y} = K - c Δ u & or & Δ u = d - \frac{1}{c} \frac{Δ Y}{Y} \end{matrix}

12. The results suggest a natural rate of unemployment above 7 percent. The linear relationship provides an estimate of the parameter c of around 4, meaning that a 1 percentage point increase in the unemployment rate above the natural rate (or a 1 percentage point increase in the actual rate) is associated with output, which is 4 percent below potential (or output growth, which is 4 percentage points lower). In addition, from the gap form,⁸ we find a natural rate of unemployment of about 7½ percent on average over the sample (see Figures 12 and 13)—similar to the smallest value for 2012 that the filtering methods provided but above the low rates of the early 2000s. However, any linear relationship among these variables likely changes over time, as is also visible from the HP-filter and the multivariate approach—both projecting a declining natural rate of unemployment. In addition, the ¾ percentage point decline in the unemployment rate in 2011 was somewhat larger than that suggested by Okun’s law.

13. Rolling regressions of Okun’s law illustrate the time variation in the natural rate of unemployment. We use quarterly data during 1996–2012, including quarterly output gap estimates from the multivariate filter approach, to run regressions of equation (1) (the gap form) over varying time periods with the actual unemployment rate as the left-hand-side variable. The results confirm the increase in the natural rate of unemployment in the second half of the 2000s, with the most recent period pointing to a natural rate of just above 7 percent (see Figure 14).

D. Conclusion

14. With potential growth moderately weaker and the natural rate of unemployment to remain elevated, policies should focus on growth-enhancing reforms, especially in the labor market. While Sweden ranks high in most comparisons of structural indicators, owing to extensive reforms in the 1990s, tackling the still existing areas of weakness could help raise the medium-term growth and employment outlook. For example, lifting labor market frictions and alleviating housing rigidities could help raise growth in the medium to long term.⁹

References

Benes et al., 2010, “Estimating Potential Output with a Multivariate Filter,” IMF Working Paper, WP/10/285.
- Search Google Scholar
- Export Citation
International Monetary Fund (IMF), 2013, “Sweden: Staff Report for the 2013 Article IV Consultation.”
- Search Google Scholar
- Export Citation
Ministry of Finance, Sweden, 2013, “Economic and Budget Policy Guidelines. Summary,” Spring Fiscal Policy Bill, April.
- Search Google Scholar
- Export Citation
National Institute of Economic Research, 2013a, “The Swedish Economy. Summary,” March.
- Search Google Scholar
- Export Citation
National Institute of Economic Research, 2013b, “The Swedish Economy. Summary,” June.
- Search Google Scholar
- Export Citation
OECD, 2012, “Economic Survey: Sweden 2012,” OECD Publishing. http://dx.doi.org/10.1787/eco_surveys-swe-2012-en.
- Search Google Scholar
- Export Citation
OECD, 2013, “Economic Policy Reforms 2013. Going for Growth,” OECD Publishing.
- Search Google Scholar
- Export Citation
Okun, Arthur M., 1962, “Potential GNP: Its Measurement and Significance“, Proceedings of the Business and Economic Statistics Section of the American Statistical Association. Reprinted in Cowles Foundation Paper 190.
- Search Google Scholar
- Export Citation
Ravn, Morten O. and Harald Uhlig, 2002, “On Adjusting the Hodrick-Prescott Filter for the Frequency of Observations,” The Review of Economics and Statistics, Vol. 84, No. 2 (May), pp. 371–376, 2002.
- Search Google Scholar
- Export Citation
Sveriges Riksbank, 2013, “Monetary Policy Update,” April.
- Search Google Scholar
- Export Citation

Appendix. Methodology

Production function

The production function approach assumes that output can be described by the following Cobb-Douglas production function:

Y_{t} = A_{t} {K_{t}}^{1 - α} {L_{t}}^{α}

To estimate potential output, total factor productivity, A_t, is backed out from the equation and HP-filtered. The actual stock of capital is assumed to be equal to the potential value. Labor’s share in income, α, is set at 0.55. Potential labor input is determined through an HP-filtered unemployment rate and the actual labor force.

In order to alleviate the end-point bias inherent to the HP filter, the variables are expended with projections until 2025.

Multivariate filter

The multivariate filter is based on the approach in Benes et al. (2010). Their model can be described as follows:

The output gap, ygap_t, can be described as a function of actual GDP, Y_t, and potential GDP, Y_t:

y g a p_{t} = 100 \cdot (\log (Y_{t}) - \log ({\bar{Y}}_{t}))

The unemployment gap, ugap_t, is described as the difference between the natural rate of unemployment, U_t, and the actual unemployment rate U_t:

u g a p_{t} = {\bar{U}}_{t} - U_{t}

Similar assumptions are made for the capacity utilization gap, cgap_t, computed as the difference between the actual manufacturing capacity utilization rate, C_t, and its equilibrium level, C_t:

c g a p_{t} = {\bar{C}}_{t} - C_{t} .

An inflation equation links the output gap with year-on-year core inflation, π⁴_t:

π 4_{t} = π 4_{t - 1} + β y g a p_{t} + Ω (y g a p_{t} - y g a p_{t - 1}) + ε_{t}^{π 4}

and Okun’s law, extended to include a lag effect, relates the unemployment and output gaps:

u g a p_{t} = φ_{1} u g a p_{t - 1} + φ_{2} y g a p_{t} + ε_{t}^{u g a p}

A corresponding equation is assumed for the capacity utilization gap:

c g a p_{t} = κ_{1} c g a p_{t - 1} + κ_{2} y g a p_{t} + ε_{t}^{c g a p}

Laws of motions for the equilibrium variables are then assumed, including with transitory and more persistent shocks to the system. First, the NAIRU U_t is assumed to be affected by transitory level shocks, $ε_{t}^{\bar{U}}$ , and more persistent shocks, $G_{t}^{\bar{U}}$ , such that:

{\bar{U}}_{t} = {\bar{U}}_{t - 1} + G_{t}^{\bar{U}} - \frac{ω}{100} y g a p_{t - 1} - \frac{λ}{100} ({\bar{U}}_{t - 1} - U^{s}) + ε_{t}^{\bar{U}}

where

G_{t}^{\bar{U}} = (1 - α) G_{t - 1}^{\bar{U}} + ε_{t}^{G^{\bar{U}}}

Potential output depends on the trend growth rate of potential GDP and on changes in the NAIRU:

{\bar{Y}}_{t} = {\bar{Y}}_{t - 1} - θ ({\bar{U}}_{t} - {\bar{U}}_{t - 1}) - (1 - θ) ({\bar{U}}_{t - 1} - {\bar{U}}_{t - 20}) / 19 + G_{t}^{\bar{Y}} / 4 + ε_{t}^{\bar{Y}}

where

G_{t}^{\bar{Y}} = τ G_{S S}^{\bar{Y}} + (1 - τ) G_{t - 1}^{\bar{Y}} + ε_{t}^{G^{\bar{Y}}}

Similarly for capacity utilization:

{\bar{C}}_{t} = {\bar{C}}_{t - 1} + G_{t}^{\bar{C}} + ε_{t}^{\bar{C}}

where

G_{t}^{\bar{C}} = (1 - δ) G_{t - 1}^{\bar{C}} + ε_{t}^{G^{\bar{C}}}

The effect of monetary policy is captured through the output gap, which responds negatively when inflation is running above long-term inflation expectations, or positively otherwise:

y g a p_{t} = ρ_{1} y g a p_{t - 1} - \frac{ρ_{2}}{100} (π 4_{t - 1} - π 4_{t - 1}^{L T E}) + ε_{t}^{y g a p}

The long-term inflation expectation is also subject to shocks:

π 4_{t}^{L T E} = π 4_{t - 1}^{L T E} + ε_{t}^{π 4^{L T E}}

The model is then estimated with Bayesian methods such that prior distributions help ensure reasonable parameter values. For example, potential output growth is influenced by the prior that it does not deviate too far from its steady state. Specifically, let ε_t be a measurement error that reflects the prior belief about the volatility of potential output growth around its steady state, $G_{S S}^{\bar{Y}} .$ . Then

4. ({\bar{Y}}_{t} - {\bar{Y}}_{t - 1}) = G_{S S}^{\bar{Y}} + ε_{t}

In addition to prior distributions of the parameters, the steady state unemployment rate is set at 6 percent, the steady state real GDP growth rate at 2.4 percent (coinciding with the 2018 WEO projection), and labor’s share in income is set at 0.55.

Prepared by Lone Christiansen (EUR).

The level reflects, among other things, the classification of some full-time students as unemployed, accounting for about 2 percentage points of overall unemployment.

Many authors suggest using a smoothing parameter of 100 (high smoothing) for annual data, while Ravn and Uhlig (2002) suggest a parameter of 6.25 (low smoothing), which will allow for more fluctuations in the filtered series.

Total factor productivity is determined as the residual in a Cobb-Douglas production function with real GDP, labor, and the capital stock.

In particular, interest costs for owner-occupied housing and certain costs of owner-occupied homes such as repairs, real-estate taxes, write-offs and insurance, as well as fees for tenant-owned flats are excluded from the HICP but are included in the CPI.

The interest cost in the CPI is composed of two parts: one index for interest rates and one index for house prices. The effect of the former is removed in the CPIF, while both parts are removed in the CPIX as well as the effects of changes in indirect taxes and subsidies.

The multivariate filter was also run using services inflation in order to capture domestic inflation. This did not materially alter the results.

Based on the production function approach (smoothing parameter of 100 for HP-filtered input variables).

See Policy Agenda section of IMF (2013) and OECD (2012, 2013) for further discussion.

Other Publishers

Save
Cite
Email this content

Share Link

Copy this link, or click below to email it to a friend
Email this content
or copy the link directly:

https://www.elibrary.imf.org/view/journals/002/2013/277/article-A001-en.xml
The link was not copied. Your current browser may not support copying via this button.

Link copied successfully

Sweden: Selected Issues

Author: International Monetary Fund. European Dept.

Volume/Issue:: Volume 2013: Issue 277

Publisher:: International Monetary Fund

ISBN:: 9781484382554

ISSN:: 1934-7685

Pages:: 65

DOI:: https://doi.org/10.5089/9781484382554.002

Keywords
I. Potential Output and Unemployment1
References
- Appendix. Methodology
  - Production function
  - Multivariate filter

View raw image

Contribution to Real GDP Growth

(Percentage points, unless otherwise indicated)

View raw image

Unemployment Rate

(Percent of labor force in given age group)

View raw image

Real GDP

(Bilions of chained Kronor)

View raw image

Output Gap Estimates

(Percent of potential output)

View raw image

Output Gap

(Percent of potential output)

View raw image

Potential Growth Estimates

(Percent)

View raw image

Contributions to Potential Growth

(Percentage points, unless otherwise indicated)

View raw image

Inflation

(Year-on-year, percent)

View raw image

Harmonized Consumer Price Inflation

(Percentage point contribution to overall inflation)

View raw image

Inflation and the Exchange Rate

(Year-on-year harmonized inflation in percent)

View raw image

Unemployment

(Percent)

View raw image

Okun’s Law: Change Form, 1993–2012

View raw image

Okun’s Law: Gap Form, 1993–2012

View raw image

Okun’s Law: Natural Rate of Unemployment ^1/

(Percent)

Table 1.

Output Gap Estimates

(Percent of potential output)

	2010	2011	2012	2013	2014
European Commission	−1.6	−0.1	−1.3	−1.6	−1.1
Ministry of Finance	−3.1	−1.3	−2.2	−3.5	−3.3
National Institute of Economic Research	−3.6	−1.5	−1.9	−2.1	−1.7
Riksbank ^1/	−2.8	−0.5	−0.9	−1.3	−0.6
Fund staff ^2/	−1.0	0.8	−0.4	−1.0	−0.7

Sources: Ameco database (June 2013), National Institute of Economic Research (June 2013b), Riksbank (April 2013), Swedish Ministry of Finance (April 2013), World Economic Outlook (July 2013).

^{^1/}Average of quarterly estimates.

^{^2/}World Economic Outlook.

Topics

Countries

Series

About

Resources

Abstract

I. Potential Output and Unemployment1

A. The Great Recession in Sweden

B. Potential Output

C. The Natural Rate of Unemployment

Explaining the Low Rate of Inflation

Estimating the Natural Rate of Unemployment

D. Conclusion

References

Appendix. Methodology

Production function

Multivariate filter

Same Series

Other IMF Content

Other Publishers

Inter-American Development Bank

International Labour Organization

Nordic Council of Ministers

The World Bank

Share Link

Table of Contents

Headings

Figures

Metrics

I. Potential Output and Unemployment¹