In this article we focus on the performance of one specific indicator which can play a crucial role in these adjustment processes: unit labour costs (ULCs). We emphasise the importance of understanding the different patterns of ULC developments across the different sectors of the economy, particularly in so-called ‘tradables’ and ‘non-tradables’.
As a preliminary, we consider ULC developments across all European Union countries and focus on differences between the economy as a whole and the manufacturing sector in particular. We single out the manufacturing sector as the principal tradable sector in most economies; the differences in developments between the manufacturing sector and the economy as a whole are then interpreted as an indication of the degree to which cost-competitiveness in the tradable sector has deteriorated or improved relative to all sectors of the economy. In Figure 1 we look at developments in two periods: the pre-crisis period 2005-2008 (Figure 1a) and the crisis period 2009-20113 (Figure 1b).
The main patterns that the figures show are as follows:
There are much stronger movements in ULCs in many low- and medium-income economies than in advanced economies; this was the case both in ULC growth in the pre-crisis period (interpreted at the time partly as the workings of the Balassa-Samuelson process leading to price level convergence between advanced and catching-up economies, but partly also reflecting an overshooting in real exchange rate appreciation caused by strong capital inflows) as well as in ULC declines during the crisis period.
In general, ULCs rise more moderately in the manufacturing sector than in the economy as a whole across most economies which is natural as manufacturing is also a sector with generally higher relative labour productivity growth.
None the less, exceptions to that general trend can also be observed: a number of economies either experienced very similar developments in ULC growth in manufacturing compared to the economy as a whole or the relative ULC position of their manufacturing sectors deteriorated in the pre-crisis period. Latvia, Romania, Croatia, Spain, Greece and Cyprus fall into this category.
Furthermore, some economies can be seen to have undergone major shifts during the crisis period, with ULCs in manufacturing dropping significantly (and competitiveness thus improving). Poland, Estonia, Lithuania, Latvia, Romania, and Ireland are among those countries.
Moreover, one group of economies displays a persistent and significant differential in terms of ULC developments in both periods (favouring the relative competitiveness of the manufacturing sector). This group includes the Czech Republic, Poland, Slovenia, Slovakia and Ireland. Yet another group of economies is characterised by low differentials or ‘perverse’ ULC developments in the manufacturing sector relative to the economy as a whole (i.e. a deterioration in the relative ULC position of manufacturing). This group comprises Hungary, Spain, Italy, Portugal and Croatia. We consider developments in this latter group a problematic issue, unlike the pattern in the former group that we regard as a sign of healthy developments in the competitiveness of these economies.
Let us now discuss in greater detail those factors which drive relative ULC developments across sectors and time-periods.
The following decomposition formula is applied:
? ULC = – ? Output + ? Employment + ? Compensation Rate (in NCU) –
? Exch. Rate
Exchange rate is defined as NCU/EUR. It is clear that for those countries that adopted the euro at a particular juncture or maintained a fixed currency regime in relation to the euro, changes in the exchange rate play no role in driving ULCs. Output is identified with gross value added (GVA).
Using the above formula, we first present an overview across the entire range of economies and we then select a few country examples in order to point out diverse patterns of ULC developments that occurred in the pre-crisis and the crisis periods.
Figure 2 (a and b) shows the decomposition of unit labour cost developments into the various components (changes in output, employment, labour compensation, exchange rate) for the periods 2005-08 and 2009-2011 respectively. Without going over these developments in any detail, we want to point out the following features:
Firstly, ULC developments are in general much more dramatic in the NMS than in the GPS or in the advanced EU economies; this likely reflects the stronger catching-up gap which still had to be covered by the new member countries in price level convergence compared to the GPS economies which had been EU (and thus Single Market) members for a much longer period and where the nominal convergence process had already taken place earlier.
Secondly, we can see a marked difference between the pre-crisis period and the crisis period in that in the pre-crisis period labour compensation growth played a much bigger role in determining overall ULC developments than during the crisis years. During the crisis years, it is relative output and employment growth/contraction (and hence implicitly labour productivity) which had a dominating role to play in determining ULC developments. There are exceptions to these patterns: e.g. fast output growth was important for manufacturing productivity growth in the Czech Republic, Slovakia and Poland in the pre-crisis period, and wage growth was high in Bulgaria during the crisis years. Thirdly, exchange rate developments played a significant role only in very few economies as most economies were either members of the EMU or had opted for fixed (or quasi-fixed) exchange rate regimes. The economies where exchange rate flexibility still played a significant role in the pre-crisis period were the Czech Republic, Slovakia, Poland and Romania (all appreciating vis-a-vis the Euro) and Poland, Hungary and Romania (depreciating vis-a-vis the Euro) during the crisis years.
We shall now discuss a few country examples regarding the factors which played important roles in ULC developments and point to different developments in this regard in different (tradable and non-tradable) sectors (see Figures 3). The examples chosen should demonstrate both differences in the weights which different factors have in driving ULC developments in different economies and the degree to which (particularly the vulnerable) economies manage to re-equilibrate real exchange rates in the course of the crisis.
We start with Latvia (Figure 3a): this country went through rather dramatic structural adjustment processes, which are apparent in the growth and employment adjustments across sectors. Over the period 2009-2011, Latvia experienced a sharp contraction in output particularly in construction and in non-tradable services, whereas the impact on manufacturing and tradable services was far less negative over the same period.
In terms of ULCs and their components, the country registered a steep rise in ULCs in the period prior to the crisis (2005-2008) in the construction sector relative to the other sectors of the economy. Wage compensation per worker in the construction sector rose by nearly 40% p.a. while in the economy as a whole it grew by some 22%. In that period, ULCs grew by 33.9% in the construction sector mainly driven by wage growth, while ULCs grew by 18.2% in the economy as a whole. Once the crisis struck, ULCs fell by -4.7% in the economy as a whole, while decreasing by -10.1% in the manufacturing sector and by -4.7% in the construction sector. The crucially important component in the construction sector that drove ULCs down was a dramatic contraction in employment (-22.7%) accompanied by a decline in output of -19.4%. In the manufacturing sector, on the other hand, output increased over the period 2009-2011 by 3.0% p.a. and manufacturing was the only sector where output did not decline (the output decline in the economy as a whole was -4.7% p.a.). The developments favouring the tradable sector during the crisis period are also apparent when one compares the tradable and non-tradable market services sectors. The data show that the decline in output (and employment) was more substantial in the non-tradable services sector than in the tradable services sector. Hence overall there was a clear shift during the crisis period towards tradable activities (manufacturing and tradable market services) and away from non-tradables (construction and other non-tradable market services). Furthermore, the Latvian case clearly shows – and this finding applies to all economies – that over the crisis period, relative ULCs across sectors are driven far less by differential movements in wage compensation, but much more by the differentiated movements in output and employment (and hence in labour productivity).
In the case of Slovenia (Figure 3b), as in the case of Latvia, exchange rate adjustments only played a role in ULC developments in the first period (2001-2004). Thereafter, in the run-up to joining the eurozone and then having acquired EMU-membership in 2007, devaluations could no longer contribute to improving the competitiveness of the Slovene economy. From that point on the two other variables, labour productivity and labour compensation, determined ULC developments. Moving straight to the period 2009-2011, the period of adjustment, quite striking differences between the Slovene and the Latvian economies can be observed. The difference lie mostly in the productivity growth figures. In Latvia over the period 2009-11, productivity growth rates in the total economy, manufacturing and the construction sector were -3.8%, 14.2% and 4.3%, respectively, whereas the figures for the corresponding sectors in Slovenia were substantially lower -0.2%, 2.8% and -7.6%. If we take those figures together with the growth rates in compensation rates per worker, we obtain the corresponding ULC growth figures in Slovenia: for the economy as a whole +2.7% (Latvia -4.7%), manufacturing +1.1% (Latvia -10.1%) and the construction sector +8.4% (Latvia -4.7%). ULC developments in favour of manufacturing were corrected to a far greater degree in Latvia than in Slovenia. If we examine the factors behind the productivity growth figures, we can see that these productivity ‘improvements’ were due mostly to employment contraction in Latvia being much starker than in Slovenia.
Romania (Figure 3c) also offers evidence (from the standpoint of ULC developments) of comparatively pronounced adjustments favouring the tradable sector. Furthermore, given the country’s flexible exchange rate regime, exchange rate adjustments still play a role in contrast to the two economies discussed above. Concentrating on the adjustment process during the crisis period, we can see that ULC developments are strongly differentiated across sectors. For the period 2009-2011, we find that ULCs fell for the economy as a whole by -0.7% p.a., but dropped in manufacturing by -10.5%; they rose in the construction sector by 10.9%, while tradable services also developed differently (+2.2%) in comparison to the non-tradable services sector (+5.4%). Hence, overall the tradable sectors (manufacturing and tradable services) improved their relative positions in terms of ULCs compared to the non-tradable sectors. Over and above that, Figure 3c also shows that devaluation contributed to a decline in ULCs (expressed in EUR) by 4.8% per annum; this devaluation, of course, only bears relevance for the tradable sectors as it contributes to improving their competitiveness. Hence taking the differential impact of exchange rate devaluation into account, the difference in the impact of adjustments favouring the tradable sector as against the non-tradable sector over the crisis period is even more pronounced.
If we look in greater detail at the different components which explain ULC developments across the different sectors in Romania, we can see that manufacturing whose relative ULC position was greatly improved (a) benefited from a far more moderate increase in wages (growth in employee compensation rose by only 2% p.a. as against 3.1% in the economy as a whole); and (b) underwent a much more pronounced decrease in employment (-5.4%) as compared to the other sectors (-1.0% for the economy as a whole). Moreover, output developments were distinctly more positive (+2.7%) as against negative growth rates in the other private sector activities. Furthermore, the different ULC patterns between tradable and non-tradable services sectors were mainly due to the far more moderate wage growth in the former; that effect was further bolstered by the exchange rate devaluation benefiting the tradable sectors.
Let us now shortly discuss developments in the GPS countries (Greece, Portugal, Spain): Figures 4a-4c show unit labour cost developments prior to and following the impact of the crisis for these economies. We observe the following:
Greece experienced rather unfavourable developments in ULCs in manufacturing relative to the economy as a whole prior to the crisis: while ULC grew on average by 2.7% per year in the economy as a whole in the pre-crisis period (2005-08), they grew by 9.2% p.a. in manufacturing; the main reason was particularly fast wage growth and negative output growth. The situation was better in tradable services (ULCs fell by -0.9% p.a. in that period driven by a relatively favourable output performance). When we come to the crisis period (2009-2011), we see a pattern of ‘internal devaluation’: ULCs in manufacturing decline by -5.3% p.a. while they increase in the economy as a whole by 1.4% p.a.; there is output contraction in the economy as a whole while there is slight output growth in the manufacturing sector. However, most of the decline in ULCs in the manufacturing sector is due to a dramatic fall in employment (- 6.0% p.a.). In the tradable services sector there is a sharp decline in output by close to 10.0% per annum and the collapse in output and employment in the construction sector is dramatic. Hence we can see that the main drivers behind ULC developments in the different sectors during the crisis are output and employment developments.
In Spain, we can similarly witness an employment and output driven process of adjustment of relative ULCs in the different sectors of the economy during the crisis: again, employment contraction in the manufacturing sector was very strong (-6.8% p.a. in the crisis period) outstripping output contraction, so that ULCs fell by -3.2% compared to -0.7% in the economy as a whole. The fall of employment in the services sectors and of ULCs was more moderate. Wage growth fell substantially compared to the pre-crisis periods, but remained in positive territory. Following a sustained boom of construction activity in the pre-crisis period, this sector experienced – like in Greece – a sharp contraction in employment and output (more in the former than in the latter) during the crisis.
Finally, the pattern of relative adjustment in ULCs in manufacturing compared to the economy is also visible in Portugal during the crisis years (with ULCs falling by -1.0% p.a. over the years 2009-2011 in manufacturing with a slight rise of 0.3% in the economy as a whole), again driven by a much stronger contraction of employment levels in manufacturing than in the other sectors of the economy with the exception of construction.
The findings of this analysis can be summarised as follows:
The decomposition of relative ULC developments across sectors into employment, output, wage and exchange rate effects is of importance to understanding the manner in which the relative cost position of the tradable sectors improves or deteriorates (relative ULCs are one of the indicators of ‘real exchange rates’).
Furthermore, an analysis by sector is important: drawing on information solely on ULCs for the economy as a whole and then comparing those costs across countries can be quite misleading to assess developments regarding different economies’ competitiveness (which should be based on an assessment of competitiveness of the tradable sectors).
Although we have instances of differential developments in compensation rates across sectors in the short to medium term, differential developments in output and employment (and hence in productivity) play – in most instances – a much more important role in driving relative ULCs across sectors. Two issues follow from this: First, although ‘wage flexibility’ (across sectors) might be an important determinant of competitiveness in the longer run, in the medium and short term, the relative development of output and employment are a far more decisive factor in determining whether the tradable sector regains competitiveness. Thus, a sharp drop in output (and hence utilisation levels), if not matched by an even greater drop in employment, would be detrimental to this particular indicator of competitiveness. Secondly, it is important to assess the extent to which, in the course of a crisis, productivity developments might be long- or short-term in nature (e.g. whether reductions in staffing levels are temporary or long-term).
The example of Slovenia and its comparison with Latvia show that Slovenia failed to make a successful transition adjusting to firmly fixed exchange rates (by virtue of its being a member of the eurozone). Once exchange rate flexibility was lost, Slovenia did not manage to maintain (or restore in the crisis period) competitiveness in its tradable sector. In Latvia, on the other hand, the adjustment processes during the crisis period were dramatic (in terms of both output and employment in the non-tradable sector), thus supporting a shift towards competitiveness.
In economies with flexible exchange rates, exchange rate adjustments – as demonstrated in the case of Romania – continue to play a role in supporting a return to competitiveness. They can further accentuate the differential impact that ULC developments have on tradable as distinct from non-tradable activities.
Conclusions
Whether adjustments in ULCs across sectors (and thus in real exchange rates) which are mainly based on sharp relative employment and output adjustments during the crisis years, will support the tradable sector in vulnerable economies in the longer-term remains an open question. Real exchange rate adjustments could be short-term or lasting, and gains made in ULC developments which might have involved substantial capacity contractions might keep such economies ‘trade-balance constrained’ for a long time to come. Hence there is a relative price and a capacity effect to such adjustments which both have to be considered. The monitoring of these issues will remain a vital issue to understand the future course of ‘North-South’ gaps and ‘external imbalances’ in the European Union.
References
European Commission (2010), Competitiveness Report 2010, Enterprise and Industry DG, Luxembourg.
European Commission (2012), Competitiveness Report 2012, Enterprise and Industry DG, Luxembourg.
European Commission (2012), ‘Quarterly Report to the Euro Area’, Volume 11, No 2 (2012), Directorate - General Economic and Financial Affairs of the European Commission.
