Europe: constraining wages or constraining flows of capital? (graphs)
from Merijn Knibbe
Summary. Mickey Levy, chief economist of ‘Monetary Financial Institution (MFI) ‘Bank of America’, recently stated that European nations have to constrain wages to get rid of current account deficits. A more thorough look at the statistics shows that Levy uses flawed reasoning. Also, as wage growth in most (not all) countries was moderate post 1999 (Spanish manufacturing wages did for instance not increase faster than German manufacturing wages) while there is ample evidence that ‘easy credit’ led to extra ordinary booms and busts, it might very well be that it are “MFI’s gone wild” which have to be constrained instead of wages.
This and/or next week I will publish a number of posts which take issue with Mickey Levy. He recently published an article on VOXEU, ‘Diverging competitiveness in the EU: constraining wages is the key”. Levy looks at Unit Labor Costs (ULC) in the EU Germany, Spain, Italy, Greece, Portugal and France, sees that ULC have not increased in Germany but did increase in the other countries and: presto! Higher ULC means lower competitiveness so ULC has to fall which means ‘wage constraint’ in the entire EU! But he should have taken a better look at the information. And he should not have used ULC. And he should have looked at more countries. Using the wrong information and concepts leads him to misunderstand the problem. He’s paid high wages to understand such problems – but he doesn’t. To avoid misunderstandings: there can be times when wages increase too fast. But at this time, that’s not the problem. And after the introduction of the Euro this was not the root of the problem, either.
First, there are at least three reasons why ULC are not an apt indicator of competitiveness:
• ULC are a macro metric, based upon the accounting constraint (here somewhat simplified) that (total wages + total profits = total income), while ULC is defined as (wages/(total income)). It’s just the share of labor in total income, and not anything like a ‘cost’! But this metric, though useful for macro-analysis of income distribution, is not fit for micro analysis. It can, for instance, decrease when wages increase but when profits increase faster – an increase in profits which might be due to increased competitiveness but also to price developments (oil in Norway…). And the other way around. Therewith the spike of ULC in the EU in 2008 and especially 2009 – profits declined. Levy however does not investigate if something like this happened in Germany (higher profits explaining low ULC growth).
• When capital pours into a nation, financing for instance a building boom, ULC will typically increase, as the work force and therewith total wages increase – while an increase in profits has to wait until the new malls and airports and luxury resorts are finished (if ever). Levy however does not investigate if this happened in Spain and Ireland.
• ULC is a composite variable. As always, this can lead to aggregation problems when two periods are compared. An example. Hourly labor costs in construction are generally lower than average hourly labor costs in the entire economy, while ULC in labor intensive construction are generally above average. Extreme construction booms like those in Spain and Ireland and Latvia and Lithuania and Estonia which lead to an upswing in construction will therefore lead to an decrease of the economy wide average wages and an increase of economy wide ULC, while the opposite occurs when the boom turns to bust. But – unless the boom creates a tight labor market which drives up the general wage level and crowds out other sectors of the economy, as in the Baltics – this does not affect the export sector of the economy. Levy did not investigate if something like this happened. But Derry O’Brien (2011) did, for Ireland (p. 22). He shows that such arithmetical effects were non trivial (understatement) in the Irish case, where economy wide ULC fell dramatically – but where (sub-) sectoral ULC fell hardly at all and hourly wages in manufacturing kept increasing faster than in its main competitors up to and including 2009. Somewhat comparable criticism has been voiced by Felipe and Kumar (2011). The chief economist of the Bank of America should know such studies, when he writes about ULC. He even gets paid to know this. But he doesn’t. And if he does he’s just pleading to increase profits at the expense of wages.
With this in mind, we can take a fresh look at the problem which, according to Levy, is:
“Since unification (he means: the introduction of the Euro, M.K.), unit labour costs – wage compensation adjusted for labour productivity – in troubled Eurozone nations have risen dramatically faster than in Germany and other high-performing nations. The sources of these unit-labour-cost divergences are very instructive. Contrary to the common view, the largest source of diverging labour competitiveness in many Eurozone nations has been wage increases that exceeded productivity gains.”
And indeed. Wage increases in some countries have been high. As short ago as the first quarter of 2010 nominal Greek manufacturing wages increased with 10% – while some years earlier wages in the Baltic states increased with up to 54% in two years. But this did not happen everywhere. Developments in Italy and Portugal were much more moderate. Spanish hourly labor costs in ‘industry and services’ in fact declined from 59% of the German level in 1999 to 53% in 2003, to rise again to 59% again in 2007 (wages in Spanish manufacturing showed the same development, though at a slightly lower level, 57%). Let’s states this the other way around: during this entire period wages in German manufacturing were about 75% higher than in Spain and this hardly changed (all data: Eurostat). After 2007 there is a break in the series and it’s difficult to compare developments after 2007 with developments before 2007 – but there is no sign of any kind of irresponsible wage increases in Spain. If Germany is an example of moderate increases of wages, Spain is too. And Spanish productivity in fact increased, relative to German productivity, between 1999 and 2010… Looking at more detailed data the whole story of Levy just evaporates. And if moderate wage increases did not keep Spain out of trouble before, why should they do so now?
It’s time for another story. To get an understanding of what really happened it helps to look at some graphs. In some blogposts which, I hope, will publish this or next week I will analyze this information more thoroughly, at the moment the graphs are meant to show that the situation is indeed more complex than indicated by the standard narrative.
The first graph shows hourly wage costs in EU in 2009 as well if the country had a deficit or a surplus on the current account between 2004 and 2010 (EU countries with more than 1 million inhabitants).
Wage differences in the EU are very large, while low manufacturing wages are associated with deficits on the current account. I would not have predicted this, ten years ago. Remarkable: the low level of wages in the UK. Levy states that there is little room to improve relative productivity of the troubled countries in Europe. But the graph indicates that countries like Portugal, Greece, the UK and Spain either have low wage costs per product or have ample scope to increase productivity. Levy’s idea that constraining wages is the only option for such countries (let alone the transition countries) is plain wrong. The question stays, however, why European low wage countries – all of them, and not only the EU countries! – had deficits on the current account, the opposite of what was to be expected. Wages were often very low but countries like Bulgaria did have access to the same trucks and computers and and supermarket systems and airports as countries like Belgium or Austria – why couldn’t they compete? This brings us to the next graph, from Kash Mansori, which gives a hint about what might have happened: credit fuelled booms. More on these problems next time.
In the 1970s, German works councils in firms had already learned that one had to be careful about the expertise of economic experts. Accordingly, each interested party in a negotiation had to bring their own “experts” with them to the table in negotiations over wages and benefits, not necessarily because people were dishonest but because the science is inexact and with an inexact science then interests express themselves as “science.” So we can’t rely on the scientific opinion of people who work for investment institutions when considering investor logic.
This begins to sound like economists are varieties of lawyer, some bringing accusations and others needed to cross-examine the purported evidence. But who are the complainants and why is there no preliminary hearing to examine the evidence and establish there is a case to be made before the accusation is allowed to be made public?
A minor quibble: ULC is defined as total labor cost divided by real output, not nominal total income. It’s not simply the share of labor cost in total income, you have to divide by the price level. The Eurostat data tells me that Germany has had the lowest inflation in the whole Euro area from 1990 to 2011. The price level (what Eurostat calls “harmonised consumer price index”) has risen by 22% in Germany, 41% in Spain, 48% in Greece and 29% in the Euro area as a whole. How does this affect your conclusions if you account for changes in the price level?
@RRR,
On the definition of ULC: I use the Eurostat definition, which divides nominal wages by nominal GDP and therewith implicitly deflates wages by the price index of GDP:
“Real unit labour cost growth
This derived (sic, M.K.) indicator compares remuneration (compensation per employee in current prices) and productivity (gross domestic product (GDP) in current prices per employment) to show how the remuneration of employees is related to the productivity of their labour. It is the relationship between how much each “worker” is paid and the value he/she produces by their work. Its growth rate is intended to give an impression of the dynamics of the participation of the production factor labour in output value created. Please note that the variables used in the numerator (compensation, employees) refer to employed labour only, while those in the denominator (GDP, employment) refer to all labour, including self-employed”
To me, this is the labor share (including ‘other cost’ like money paid by employers to pension funds of health insurance!) in GDP (but the caveat about self employed is of course important, if there is a shift of self employed to wage labor, ULC will rise).
I’ll return to the price level, an important point. Do we agree that, aside from such quibbles, it is important to look at sectoral developments (for instance differences between government wages and manufacturing wages), too?
But if ULC is defined in this way it’s a completely useless as an indicator for competitiveness because, as you point out, it’s simply equal to the income share of labour. Are you sure this is what Levy uses?
That aside, I agree completely with the importane of studying sectoral development.