Prices in the EU – converging or diverging?

Spoilers. The Single Market and the Euro have decreased price differences within the EU. The process has been stronger for prices of goods than for prices of services. More can be done in EU to complete the Single Market for Services in order to increase competition and lower profits for the benefits of customers. If you agree with this, you should cast your vote on a party that wants to improve the functioning of the Single Market and increase trade within EU and with countries outside EU. 

Have price differences between countries in the EU diminished? The implementation of the Single Market and the introduction of the euro in several Member States have led to  increased integration of the economies.[1] The increased intra-EU trade has led to more competition on EU markets and together with deeper integration of labour, products and financial markets reduced costs for households and firms which you can see here and here. The Euro has made it easer to compare prices in different countries and removed costs of currency conversion and eliminated exchange risk premia between a number of countries.[2]

Therefore, the Single Market and the Euro should lead to decreased price differentials between the countries. Decreased price differentials over time implies that prices of the same products in different countries converge. This process can be reinforced or hampered by other developments. Two such developments will be discussed below; the enlargement of the EU and the Global Financial Crisis (GFC).

The Enlargement and the Global Financial crisis impacted on prices in the EU

The enlargement of the EU in 2004 had two counteracting effects on convergence of prices.[3] Firstly, as income levels in the new entrants began to catch up with the old Member States, price levels in the former approached the price levels in the latter implying a “convergence from below”. Secondly, with an enlarged EU, increased trade flows strengthen the competition pressure on prices implying a “convergence from above”.

The GFC caused trade to decrease which should slow down or halt the convergence process as import competition becomes weaker in every national market. Also, since the EU countries were impacted by the GFC to varying degrees, differences in income elasticities across the EU countries could lead to halted convergence or even divergence of prices. Adding to that, some countries, e.g. Greece, raised several taxes and administrative prices during the GFC in order to increase government revenues.

Measuring convergence: beta-convergence and sigma-convergence

As hinted in the previous paragraph, convergence can mean two things; a reduction of price dispersion of price levels across countries and developments towards a mean. The first type of convergence which refers to reduced price dispersion of price levels is called sigma-converge. The second type of convergence, beta-convergence, occurs when prices or incomes in poorer economies grow faster than in rich economies.[4] Sigma-convergence can be measured by analysing the variance of the same products across countries over time. A reduction of the variance, standard deviation or coefficient of variation of prices, indicates an occurrence of sigma-convergence. Beta-convergence can be studied by comparing the relative changes of prices between certain years with the price level in the first year. I will below try to illustrate sigma-convergence by graphs showing standard deviations of prices. (See the Box in the end of this post for a justification of the choice of sigma- before beta-convergence graphs.) Note that the graphs below are not formal tests of convergence. Both beta- and sigma-convergence should be analysed by econometric methods taking other explanatory variables into account. A quick search on the Internet will provide several of such studies.

Income and prices and goods versus services

Until now I have not separated between prices of different kinds of goods nor between prices of goods and prices of services. But there are good reasons to make this distinction. Goods are more traded than services so one should expect price differences between goods to be lower than between services.[5] And the Balassa-Samuelsson effect tells us that income matters for price levels so we should expect prices in relative rich countries to be higher than in less rich countries.[6] Therefore, one can expect a stronger relationship between income levels and price levels of services than between income levels and price levels of goods, c.f. Figure 1.

Figure 1. Price and income levels. GDP per capita vs price indices of consumer goods (left) and consumer services (right) 2015

Figur 1 till inlägget

Source: Eurostat

It therefore makes sense to study convergence prices of goods and services separately. And there are of course differences between different kinds of goods as well as between different kinds of services. But taking this into account would require more space than this blog so I will ignore most of these differences below. I will only separate between consumer goods and other types of goods and consumer services and other types of services. The data presented below was extracted from here

The EU Enlargement increased the convergence of prices but the GFC stopped it

The EU enlargements in 2004 and 2007 did not increase the price dispersion of goods and services within the EU27, c.f. Figure 2.

Figure 2. Standard deviations of prices of goods and services in the EU 1999-2017.

Figur 2 till inlägget

Source: Eurostat. Note. Consumer goods and consumer services are aggregates. The aggregate Total goods consists of the two other large aggregates Consumer goods and Capital goods. Consumer goods can be further split into Durable, Non-durable and Semi-durable goods.

But the GFC did increase the dispersion of prices as the decreased sigma-convergence, as measured by standard deviations, shows. Convergence came to a halt when the GFC broke out and have levelled out since even though the latest observations indicated that convergence is happening again. The price differences of goods did not increase during the GFC contrary to the price differences of services which kept increasing until a few years ago.

The aggregate EU27 can be split in different ways. It is divided into three sub-aggregates: EU15, EU10 and EA12 below.[7] EU15 consists of the countries in the EU before 2004, EU10 contains the Member States that became members of the EU in 2004 and EA12 the original Euro Area countries. Looking at the left panel below, price differences of goods in EU15 and EA12 have decreased throughout the whole period even though the GFC slowed down the process. Especially so for EA10 where price differentials started to increase after the GFC until 2013 after which they plateaued. The effects of the GFC on the EU10 countries varied substantially. While Poland as the only country in the EU did not experience a decline in output, it took between eight and nine years for Cyprus, Latvia and Slovenia to again reach the pre-crisis levels of output.

In the aftermath of the GFC, consumer services prices are still diverging

The effect of the GFC on the dispersion of prices of services was larger and lasted longer as shown in the right panel below which also shows some marked difference between the three aggregates. In fact, price differences of services are as large or larger in 2017 as in 1999 in EU15 and EA12 while they have started to decline again in EU10, c.f. Figure 3.

Figure 3. Standard deviations of prices of consumer goods (left) and consumer services (right) in the EU 1999-2017.

Figur 3 till inlägget

Source: Eurostat. Note. Consumer goods and consumer services are aggregates.

Some of the countries in EA12 were especially hard hit by the GFC, e.g. Ireland, Italy, Greece, Portugal and Spain. The right panel above implies that a large part of the convergence in the first part of the period in EU15 occurred due to developments in the non-EA countries Denmark, Sweden and the UK. An interesting observation is that prices of services in EU10 began to converge again faster than prices of goods.

Seems it is time for a furry intervention.

“No, that’s the wrong graph”

wrong graph

Yes it is. Thank you.

In order to try to understand what is going on in EU10, I split the aggregate Consumer goods into the sub-aggregates: Durable goods, Non-durable goods and Semi-durable goods. This shows that aggregates with less traded goods, Non-durables and Semi-durables drive the convergence while the aggregate which contains more goods that are traded shows convergence, c.f. Figure 4.

Figure 4. Convergence and divergence of prices of goods within EU10 2008-2017.

Figur 4 till inlägget

Source: Eurostat. Note. Durable goods are goods like washing machines, cars, computers, kitchen eqpt and furniture. Semi-durable goods are goods like clothing, footwear, books and toys. Non-durable goods are mainly foodstuff and liquid stuff.

As Figure 3 implied, also the prices of Non- and Semi-durable goods have converged since 1999 and since the Enlargement. Judging by the latest observations it also appears that the price differences of these kinds of goods are again converging. The data on Eurostat’s web site does not allow for a decomposition of Consumer services so I don’t know what’s going on in the right panel of figure 3. Eurostat have more detailed data also for Services. These data are available for people working for international organisations or universities.

The Single Market and the Euro has increased convergence but more needs to be done

This post shows that overall the Single Market, the Enlargement and the Euro have all contributed to more integrated markets and similar prices within the EU. The GFC did temporarily hamper the process but the latest developments indicate that prices are converging again.

The above exercises also imply that goods markets are more integrated that services markets and that price differences of goods are smaller for goods that are traded more. This post also implies that more efforts are needed to improve the Single Market for Services especially in the Euro area. In this respect would a Digital Single Market help the convergence process as it would make it easier to compare and buy prices from different destinations within the EU. Digital trade has become more important over time but many restrictions still hamper it. Some markets for services appear to be more prone for actions than other.[8] Also, many professions in the EU are still heavily regulated by national authorities. These national regulations do in many cases hamper cross-border movements of services and tend to affect availability and prices of services negatively. Differences in regulation of retail shops can also affect the convergence of goods.

Differences in regulation of retail shops can also affect the convergence of goods. As shown here, there are substantial differences in regulations on establishment and operation of retail shops in the EU. These restrictions tend to hamper the competition in different ways giving consumers less choice and higher prices. This and many other studies confirm that markups (profits) tend to be higher in regulated markets. Removing regulations that hamper competition and leads to lower markups may also decrease income inequality as total income, i.e. GDP is made up of labour income (wages), profits capital income.

For you who will vote in the elections to the European Parliament, the implications of the above is to vote for parties and candidates that want to improve the functioning of the Single Market and increase the number of countries which the EU have free trade-agreements with.

The post also shows that to understand fully what is going on, one needs to go further into detail in as disaggregated categories of goods and services as possible. One should also try to look at developments in different types of categories of countries in order to for example see what happened in the countries that were most affected by the GFC.

Box. Beta- or sigma-convergence

Box

Read more.

Barro, R. J. & Sala-i-Martin, X. (2012). Economic Growth. MIT Press, 2nd edition.

in t’Veld, J. (2019). “Quantifying the Economic Effects of the Single Market in a Structural Macro Model. Discussion Paper 094. European Commission.

Kommerskollegium (2015). Economic-effects-of-the-european-single-market. Review of the empirical literature. https://www.kommers.se/Documents/dokumentarkiv/publikationer/2015/Publ-economic-effects-of-the-european-single-market.pdf

Quah, D., T. (1997) “Empirics for Growth and Distribution: Stratification, Polarization, and Convergence Clubs.” Journal of Economic Growth, 2.

Young, A., T., Higgins, M., J. & Levy, D. (2008) “Sigma Convergence versus Beta Convergence: Evidence from U.S. County-Level Data. Journal of Money, Credit and Banking, Vol. 40, No. 5

Eurostat database:

https://ec.europa.eu/eurostat/web/purchasing-power-parities/data/database

Footnotes.

[1] The Maastricht Treaty established the Single Market for goods in 1992. The euro was introduced in 1999 in eleven EU countries. The Single Market for services is not as complete as for goods. The Services Directive was adopted in 2006 and covers around 45% of EU GDP. In addition to the Directive, a number of laws provide rules for financial services, telecommunications and more: https://ec.europa.eu/growth/single-market/services_en

[2] This does not mean that country-specific risk premia have been eliminated as was particularly evident during the Global Financial Crisis. 

[3] Ten new Member States entered in 2004, two in 2007 and one in 2013.

[4] Yeah, I know, that definition was a bit sloppy, but you can find a better definition easily by searching the Internet.

[5] Other reasons for this are that more services are regulated and often provided by governments.

[6] That’s not all the Balassa-Samuelsson effect tells us, look it up!

[7] Including Bulgaria and Romania which entered the EU in 2007 does not change the developments of EU10 in Figure 2. The country aggregates are kept unchanged during the whole time period.

[8] Using anecdotal evidence, a well-informed source once told me that French notaries are doing exceptionally well driving around in Jaguars drinking champagne and eating Nutella. Some of you might take that less serious than others.

[9] This is often referred to unconditional convergence. An analysis of conditional convergence would add some explanatory variables such as income and trade.

[10] The speed of convergence is calculated as – ln(1+β) where βis the coefficient for the initial price in the equation mentioned in the text. Multiply by 100 and you get the estimated reduction of prices differences within a year.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.