Wage-productivity gap: where is it?

The problem with exports is not wages being too high in tradables. On the contrary, the evidence is that wages have evolved quite in shape with productivity, especially in manufactures. The problem is about to convince workers and unemployed to move from services, where wages are higher, to manufactures, where wages are lower.

INTRO

Competitiveness is a rather complex concept. In a broad sense, it refers to the extent to which a nation provides economic agents with (socially-aligned) incentives to produce and invest. For a moment, however, let’s focus on a narrow concept of competitiveness, related to wage costs. In particular, let’s stick with a definition proposed by Olivier Blanchard (2007), who refers to competitiveness as the inverse of unit labour cost (ULC) in tradable goods sectors relative to the corresponding world value.
In its influential paper, Blanchard (2007) argues that competitiveness in Portugal deteriorated significantly since the mid-nineties. According to the author, this reflected a misalignment between wages and productivity growth, which caused profitability in tradable goods sectors to shrink. However, Blanchard did not provide evidence supporting the claimed loss of competitiveness. The author showed that economy-wide ULC increased in Portugal faster than in the EU15, but he did not distinguished tradable goods (T) from non-tradable goods (N). The European Commission (2011) goes along with the same argument: “Since the introduction of the euro, Portugal has experienced significant real exchange rate (REER) appreciation vis-à-vis its trading partners, due to wage growth largely outstripping productivity advances (Graph 3)”. However, Graph 3 in the document only displays economy-wide real exchange rate indexes…
In a contrasting view, Campos e Cunha (2008) argues that “there is no room to claim a loss of competitiveness (...)” (p.158). The author points out that, in a small open economy, the real exchange rate and aggregate demand are two sides of the same coin. As it is well known, in a well functioning economy, an aggregate demand expansion translates into higher N-prices, while T-prices are bounded to remain unchanged. This causes a real exchange rate appreciation that has nothing to do with wage-productivity gaps. Fagan and Gaspar (2007) illustrate the argument is in the context of an endowment economy where, by definition, there is no such a thing as competitiveness.
Competitiveness and the real exchange rate do not necessarily go along.

FORMAL

To illustrate this, let’s look at the real exchange rate index based on nominal unit labour costs. By definition, ULC=W/a, where W refers to the compensation per employee (“nominal wage”) and a refers to Gross Value Added at constant prices per worker (“productivity”). Now, assume that the production function is a Cob-Douglas with labour-output elasticity equal to b and let Z be a variable measuring the wage-productivity gap: Z=ba/(W/P)=bP/ULC. When computed in terms of a base year (the constant b disappears) the index 1/Z is labelled Real Unit Labour Cost (RULC) or “real wage gap”. In a frictionless economy, Z=1. In a world with frictions, in face of a wage push, firms may opt to maintain a higher level of employment than that implied by the textbook wage-productivity rule. When this is so, the producer margin shrinks (Z<1).
Using the definitions above, the real exchange rate index based on nominal unit labour costs (RER-ULC) becomes ULC/ULC*=(P/P*)(Z*/Z). This means that RER-ULC accounts for two effects: wage-productivity misalignments (Z*/Z) and the increase in the relative price of non-tradable goods (P/P*). In a well functioning economy there are no wage-productivity misalignments, so Z=Z*=1. Still, unit labour costs may increase relative to abroad, whenever non-tradable good prices increase relative to abroad. This means that we can hardly rely on RER-ULC as an indicator of competitiveness. Competitiveness a la Blanchard is accounted for by the component (Z*/Z) and in the proportion corresponding to tradable goods, only.

EVIDENCE

Returning to Portugal, we now look at the data. In Table 1, se see that between 1995 and 2010 there was a
significant real exchange rate appreciation, both measured in terms of RER-ULC and in terms of relative price deflators (RER-P). As we just showed, the difference between these two indexes captures the real wage gap in Portugal relative to the real wage gap abroad (relative RULC, line c). As shown in the table, most of the RER-ULC appreciation along 1995-2010 was accounted for the relative price effect, RER-P. The distance between RER-ULC and RER-P peaked up at 6% in 2005, to decline to 3% in 2010. Definitely, this is too little to support the usual narrative, that Portugal faces a huge problem of wage competitiveness.

Table 1: Real exchange rate indexes and adjusted wage share

1995 2000 2005 2010
(a) RER - ULC 100 107 116 115
(b) RER- P 100 104 109 112
(c) Relative RULC 100 103 106 103
(d) Labour share (Port) 66.0 67.2 68.2 66.3
(e) Labour share (EU 15) 66.9 66.8 65.5 66.0

It should be noted that the measure Z*/Z accounts for both domestic developments and developments abroad. As for developments abroad, during this period Germany was recovering from excessive wages that came along with unification. This helps explain the slight decline in the EU15 labour share (e) between 1995 and 2005.
As far as Portugal is concerned, a branch level inspection reveals that most of the (small) increase in the labour share along 1995-2005 is accounted for weird developments in the sectors of building construction and agriculture. In these two sectors, indeed, wage bills increased well beyond productivity. I suspect, however, that this is more a statistical artefact – mirroring the decreasing informality in these sectors - than evidence of dysfunctional wage setting. In any case, in the sectors of manufactures and services, real wages have evolved quite in shape with productivity along the entire period. Because these two sectors account for the bulk of gross value added in Portugal, with no surprise the share of labour on domestic income (d) remained pretty constant: at the aggregate level, there is no evidence of wage productivity mismatches. If there is any, I’m missing something.

WHAT HAPPENED?

All in all, the evidence suggests that so far Campos e Cunha was right: relative price effects, rather than dysfunctional wage setting explain the recent process of real exchange rate appreciation. This accord to the widely accepted narrative that at the origin of the current external imbalance is the once-and-for-all shock consisting in the liberalization of capital inflows followed by EMU accession: this triggered the aggregate demand and pushed the relative prices of non-tradable goods above the level that is consistent with the internal and external balance.
The story has to be however completed with the oil shock that marked the dump phase: Portugal is highly vulnerable to oil prices and after 2000 the relative price of oil increased sharply. As it is well known, in a small open economy a terms of trade deterioration is equivalent to a “negative” Dutch Disease: the relative price of non-tradable goods has to decrease to restore the external balance. In an ongoing work with Miguel Faria e Castro, we find that the recent oil price shock accounts for roughly one half of the current real exchange rate misalignment in Portugal (either measured in terms of RER-P or RER-ULC).

SO WHAT?

It may be argued that this discussion is irrelevant: after all, a real exchange rate appreciation impacts negatively in the current account, irrespectively as to whether it comes through an increase in the relative price of non-tradable goods or through wage-productivity misalignments. For policy purposes, however, the distinction is very important.
Indeed, if the real exchange rate appreciation was related to dysfunctional wage setting institutions, one would be tempted to question whether Portugal would be able to sustain its membership in a low inflation club like the euro-area.
If, as it appears to be the case, the real exchange rate over-valuation is wage-gap free, then there are no reasons to conclude - at least by now - that nominal wage rigidities will impair the upcoming macroeconomic adjustment: after all, the return to external balance involves the re-allocation of labour across industries, which necessarily entails the celebration of new labour contracts.
This does not mean that the average wage in the economy shall not decline. As a rule, a “white collar” to “blue collar” metamorphosis will involve the acceptance of lower wages. Not only because wages are on average lower in manufactures than in services, but also because recycled workers will lose the value of whatever specific knowledge they accumulated in the job they had before. Hence, we shall expect workers to resist the move until it becomes a matter of personal urgency.
This discussion drives us back to the beginning of this post. Eventually a policy of reducing wages in existing contracts will not help workers to move from non-tradable good sectors to tradable good sectors. What we really need is to remove those barriers lying on the path of entrepreneurs and job-seekers that impair investment and are delaying the re-allocation of labour from N to T. On the policy dimension, this includes the provision of a stable macroeconomic environment, fiscal sustainability, lower costs on dismissals and less incentive for workers to remain unemployed while searching for a perfect match. In a word, we need more focus on competitiveness, broad sense.

REFERENCES

Blanchard, O., 2007. Adjustment within the euro: the difficult case of Portugal, Portuguese Economic Journal 6(1): 1-21.
Campos e Cunha, L. , 2008. Is the Dutch disease pandemic in the South? In Francesco Franco (ed.), Challenges Ahead for the Portuguese Economy, Imprensa de Ciencias Sociais, Lisbon.
European Commission, 2011. The Economic Adjustment Programme for Portugal, European Economy: Occasional papers 79, June.
Fagan, G., Gaspar, V., 2007. Adjusting to the euro. ECB Working Paper Series, Nº716, January.