Konstantīns Beņkovskis, Julia Wörz. Evaluation of Non-Price Competitiveness of Exports from the Central, E Eastern and South-Eastern European Countries in the EU market

Evaluation of Non-Price Competitiveness of Exports from the Central, Eastern and South-Eastern Central, Eastern and South-Eastern

European Countries in the EU market

Konstantīns Beņkovskis (Bank of Latvia)Julia Wörz (Österreichische Nationalbank)

27 April 2012

Outline

• Drawbacks of traditional REER indicators

• How to assess non-price competitiveness?o Theoretical framework

o Elasticities of substitutiono Elasticities of substitution

o Dynamics in price and non-price competitiveness in CESEE

o Contribution of non-price factors in some product sectors of Estonia’s exports

• Conclusions

MotivationREER signal losses in price competitiveness – is it the whole story?

Real effective exchange rate (36 partner countries, 1999=100)

190

210CPI based

180

200

220ULC based

Bulgaria

Czech

RepublicEstonia

Latvia

Source: Eurostat3

90

110

130

150

170

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

80

100

120

140

160

180

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Latvia

Lithuania

Hungary

Poland

Romania

Slovenia

Slovakia

MotivationMarkets shares signal improving competitiveness

The share of exports in the World trade (2002=100)

190

210

230

250

Latvia

Lithuania

Estonia

Source: WTO4

90

110

130

150

170

190

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Q1-Q3*

Bulgaria

Czech Republic

Hungary

Poland

Romania

Slovakia

Slovenia

MotivationREER indicators have list of drawbacks

• Bad approximation for export priceso Whole economy, no distinction between domestic and external

markets

o Profit martins are ignored

• Structural issues are not captured:Structural issues are not captured:o Differences in export structure are not taken into account

o Need to analyse competitiveness at disaggregated level

• Focusing on price competitivenesso Some measures are adjusted for quality (e.g. CPI-based)

o However, many other important factors left aside (e.g. taste, image of brands)

o Need to consider non-price competitiveness issues

Goal of the projectPrice and non-price competitiveness of CESEE countries

• Evaluate the price and non-price competitiveness of CESEE countries– Latvia, Estonia, Lithuania, Poland, Czech Republic,

Hungary, Slovakia, Slovenia, Romania, Bulgaria

• Due to data constraints we are limiting our analysis to the exports to the EUo Still analysing the most part of CESEE exports

Short literature reviewSeveral papers dealing with quality and variety issues

• Feenstra, R.C.(1994) “New Product Varieties and the Measurement of International Prices”, American Economic Review, 84(1), pp.157-177.

• Hummels, D. and Klenow, P.J. (2005) “The Variety and Quality of Nation’s Exports”, American Economic Review, 95(3), pp.704-723.

• Broda, C. and Weinstein, D.E. (2006) “Globalization and the Gains from Variety”, Quarterly Journal of Economics, 121(2), pp.541-585.from Variety”, Quarterly Journal of Economics, 121(2), pp.541-585.

• Benkovskis, K. and Rimgailaite, R. (2011) “The Quality and Variety of Exports from the New EU Member States”, Economics of Transition, 19(4), 2011, 723-747.

• Benkovskis, K. and Wörz, J. (2011) “How Does Quality Impact on Import Prices?”, OeNB Working Papers, 175/2011.

• Benkovskis, K. and Wörz, J. (2012) “Evaluation of Non-Price Competitiveness of Exports from Central, Eastern and South-Eastern European Countries at the EU market”, Bank of Latvia WorkingPapers, 1/2012.

How to evaluate non-price competitiveness?

• We have trade data on a very disaggregated level:o prices (unit values, euro/kg)

o volumes (kg)

• Why shouldn’t we combine both sources instead of • Why shouldn’t we combine both sources instead of focusing just on one?o If real market share improves when relative export price is

increasing, it gives us some clue about non-price factors

• Consistent theoretical framework needed

Theoretical frameworkConsumer’s utility function

• First-level CES utility function (imports and domestic good)

• Second-level CES utility function (different imported goods)

1;111

>

+=

−−−

κκκ

κκ

κκ

ttt MDU

• Second-level CES utility function (different imported goods)

• Third-level CES utility function (different varieties of a good)

1;111

>

=

−

∈

−

∑ g

Cc

gctgctgt

g

g

g

g

g mdM σσ

σ

σ

σ

σ

1;11

>

=

−

∈

−

∑ γγγ

γγ

Gg

gtt MMelasticity of substitution

between products

elasticity of substitution

between varieties

quality or taste parameter

set of goods

set of countries

Theoretical frameworkMinimum unit-cost function

• After solving the utility maximization problem

o minimum unit-cost depend on price, quality or taste

gg

Cc

gctgctgt pdσσφ−

∈

−

= ∑

1

1

1

o minimum unit-cost depend on price, quality or taste parameter and set of partner countries (variety)

• The exact import price index for good g is defined as:

1−

=gt

gt

gPφ

φ

Theoretical frameworkRelative price on a single market

• Our goal, however, is to evaluate quality-adjusted relative export price indexo We can interpret xgct as country’s c exports of a product g

o We propose to define changes of relative export price as follows

( )( ) 1

where• φgt

k – minimum unit-cost of good g in case it is exported only by country k

• φgt-k – minimum unit-cost of good g in case it is exported by all

countries except k

( )( )k

gt

k

gt

gktgktgktgkt

k

gt

k

gt

k

gt

k

gt

gkt

gddppRXP −

−−

−−−

−−

−

− ==1

1

1

11

1

1

φφφφ

φφ σ

Theoretical frameworkChange of a relative price on a single market

• Relative price of a good g imported from country k relative to other origins:

∏∏−

−

−

−

∈

−

−

−−

−−

−

∈ −

−

=

kg

g

kgct

g

kg

kgct

Cc

w

gkt

gct

gct

gkt

k

gt

k

gt

Cc

w

gkt

gct

gct

gkt

gctd

d

d

d

p

p

p

pRXP

σσ

λ

λ 1

1

11

1

11

1

whereo Cg

-k – set of countries exporting to a particular market in both periods, excluding country k

o w-kgct – Sato-Vartia weights of exporters, excluding county k

o

gg

∑

∑

−

−

∈

∈− =

kgt

kg

Cc

gctgct

Cc

gctgct

k

gtxp

xp

λ∑

∑

−−

−

∈−−

∈−−

−− =

kgt

kg

Cc

gctgct

Cc

gctgct

k

gtxp

xp

1

11

11

1λ

Theoretical frameworkChange of a relative price in a single market

• Relative price of a good g imported from country k relative to other origins:

∏∏−

−

−

−

∈

−

−

−−

−−

−

∈ −

−

=

kg

g

kgct

g

kg

kgct

Cc

w

gkt

gct

gct

gkt

k

gt

k

gt

Cc

w

gkt

gct

gct

gkt

gctd

d

d

d

p

p

p

pRXP

σσ

λ

λ 1

1

11

1

11

1

1. Traditional relative price index – increase denotes worsening price competitiveness

2. Adjustment for changes in monopoly power of exporters. If set of partner countries is increasing, relative price index increase as well

3. Adjustment for changes in quality or taste. Rise in relative quality or taste decrease relative price index and improve competitiveness

1 2 3

Theoretical frameworkHow to estimate quality/taste parameter?

• After solving utility maximization problem:

+

=

gctgctgct

x

x

p

p

d

dln

1lnln

1

σσ benchmark country

relative prices (UVX) relative quantities (kg)

• Relative quality or taste depends on relative prices and relative volumes of sales

• It also depends on elasticity of substitutiono relative quantities are not important for perfect competition

gktggktgktg xpd σσ benchmark country

Theoretical frameworkAggregated relative export price

• Relative price indices on a particular market (RXP(i)gkt) should now be aggregated by products (g) and by export markets (i)

• Weights from exporter side should be used

( )∏∏= WigtiRXPRXP

whereo RXPkt – aggregated relative export price index

o Wigt is weight of a product g exported to country i in total exports to country k

( )∏∏∈ ∈

=Ii Gg

W

gktktigtiRXPRXP

Estimation of elasticitiesSystem of demand and supply equations

• We need to estimate elasticities of substitution

• Elasticity of substitution between varieties estimated from the systemo Relative demand equation:

( ) ps lnln ∆∆

o Relative supply equation:

• Absence of exogenous variables to identify the system and estimate elasticities

( ) gctgctgct

gkt

gct

g

gkt

gctd

p

p

s

sln;

ln

ln1

ln

ln∆=+

∆

∆−−=

∆

∆εεσ

gct

gkt

gct

g

g

gkt

gct

s

s

p

pδ

ω

ω+

∆

∆

+=

∆

∆

ln

ln

1ln

ln

• In order to take advantage of the independence of an

, these two equations are multiplied together to obtain (Leamer’s, 1981, approach):

gctε

gctδ

gctgctgctgctu

spsp+

∆

∆

+

∆

=

∆ lnlnlnln

22

θθ

Estimation of elasticitiesTransformation of the system

• Relative price x market shares are correlated with error term, estimates will be biased

• However, we can obtain consistent estimates by exploiting the panel nature of the data

gct

gkt

gct

gkt

gct

gkt

gct

gkt

gctu

spsp+

∆

∆

+

∆

=

∆ lnlnlnln

21 θθ

( )( );111 −+=

gg

g

σω

ωθ

( )( )( );11

212 −+

−−=

gg

gg

σω

σωθ gctgctgctu δε=

• Broda and Weinstein (2006) argue that one needs to define a set of moment conditions for each good g by using the independence of unobserved demand and supply disturbances for each country over time:

Estimation of elasticitiesGMM estimates

( ) ( )( ) cuEG ∀== 0ββ

• For each good g the following GMM estimator is obtained

o solved as constrained minimization problem. B is a set of economically feasible values of β (σg>1, ωg≥0)

( ) ( )( ) cuEG ggcttg ∀== 0ββ

( )ggg ωσβ ,=

( ) ( )gg

Bg WGG βββ

β

**minargˆ ′=∈

Estimation of elasticitiesResults for big EU countries

800

1000

1200Germany

France

UK

Medianelasticity

Median mark-up

Germany 6.19 19.3%

France 5.39 22.8%

Distribution of elasticities of

substitution

0

200

400

600

800

1.0

-1.

2

1.8

-2.

2

3.3

-4.

1

6.0

-7.

4

11.0

-13

.5

20.1

-24

.5

36.6

-44

.7

66.7

-81

.5

121.

5 -

148.

4

221.

4 -

270.

4

403.

4 -

492.

7

735.

1 -

897.

8

UK

Italy

France 5.39 22.8%

Italy 5.76 21.0%

UK 4.91 25.6%

… … …

Database and coverageEurostat Comext data

• Eurostat Comexto Import data for all 27 EU countries

o 8-digit CN classification level

• approx. 10 000 products

o 1999 to 2010, annual datao 1999 to 2010, annual data

o 50 main partner countries

• All EU countries, US, Japan, China, India, Brazil, Canada, Russia etc.

• Therefore, we are assessing competitiveness of CESEE countries on EU marketo Still analysing the most part of CESEE exports

Relative export price indicesBaltic countries and Bulgaria

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Estonia

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Latvia

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Conventional Adjusted by non-price factors

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Bulgaria

Conventional Adjusted by non-price factors

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Lithuania

Conventional Adjusted by non-price factors

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Conventional Adjusted by non-price factors

Relative export price indicesVyshegrad group countries

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Poland

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Czech Republic

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Conventional Adjusted by non-price factors

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Hungary

Conventional Adjusted by non-price factors

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Slovakia

Conventional Adjusted by non-price factors

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Conventional Adjusted by non-price factors

Relative export price indicesRomania and Slovenia

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Slovenia

708090

100110120130140

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Romania

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Conventional Adjusted by non-price factors

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Conventional Adjusted by non-price factors

Relative quality/taste of Estonia’s exportsBy main export sectors

Cumulated contribution of relative quality to export competitiveness (1999=100)

220

260

300

340

160

180

200

220

24

20

60

100

140

180

60

80

100

120

140

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Machinery and mechanical appliances Wood and wood products

Base metals and articles thereof Prepared foodstuffs

Chemical products Mineral products (RHS)

Relative quality/taste of Estonia’s exportsBy main export markets

Cumulated contribution of relative quality to export competitiveness (1999=100)

400

500

600

190

220

250

25

0

100

200

300

70

100

130

160

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Finland Sweden Latvia

Lithuania Denmark Germany (RHS)

ConclusionsQuality/taste of CESEE exports is increasing

• The traditional competitiveness measures based on unit values show that CESEE export prices increased faster than export prices of competitorso Is it creating a problem for export competitiveness?

• Our analysis indicates that it was driven by faster increase in quality/taste of CESEE exportso In fact, aggregate competitiveness (including price and

non-price factors) of CESEE even improved

Further plansNeeds for improvement in methodology, other database

• No effect on relative price index from changes in set of export products/marketso Gains from diversification?

o Theoretical framework should be more developed for supply sidesupply side

• separate taste from quality?

• Due to data constraint we analyse only competitiveness on the EU marketo Competitiveness on the World market

o UN Comtrade database