NEW OECD WORKING PAPER REPORTS ON DECLINING LABOUR SHARE AND INCREASING WAGE INEQUALITY
Under the title of ‘Decoupling of wages from productivity: Macro –Level facts”, OECD Economics Department recently published a paper describing how medium wages have been lagging behind productivity developments since 1995 in most OECD countries.
OECD methodology
The paper splits up the total divergence between the evolution of medium wages and productivity into, on the one hand, a gap that measures the decline in the labour share (which is equal to the difference between the evolution of productivity and the change in average wages) and, on the other hand, a second gap that is due to the different dynamics between the average wage and the medium wage. The latter gap points to rising inequalities within the group of workers, in particular top wage earners racing far ahead of the rest of workers.
The graph below provides a picture of these developments for an unweighted average of 24 OECD countries.
It is useful to note that the OECD is using a particular methodology in this paper to calculate labour shares as it does not cover the total economy but excludes sectors such as housing and non-market sectors. According to the OECD, excluding these sectors is warranted because these may distort the numbers for the total economy. For example, since the GDP of the housing sector mainly consists of (imputed) rents which then go under the nominator of ‘capital income’, the capital share of the total economy will automatically increase, whereas the labour share will fall in case the economy is experiencing a housing price boom with related hikes in housing rents.
In practice, using this methodology has the effect of understating the decline of the labour share with the OECD stating (see paragraph 21) that the decline in the gross labour share for the average of the OECD has been small. In doing so, Table 4 reports a decline for the G7-average that is as low as 0.2%.
Two critical observations on the method used by the OECD
First, as stressed, this is an unweighted average for the entire OECD/G7 countries. A closer look at the OECD/G7 average shows that the quasi stagnation of the G7 average for example is the result of the labour share increasing in France, the UK and Italy, which gets offset by a declining labour share in the US, Japan and Canada.
However, when using a weighted average, the conclusion changes dramatically. For the G7, the labour share decline becomes a full 2.3 percentage point drop (instead of an almost non-noticeable -0.2pp).
Changes in Labour Share, percentage points, 1995–2014
Second, the impact of the rise in precarious work practices, in particular of economically dependent self- employed, is not adequately reflected. Indeed, in calculating the share of labour, the labour part of the income of the independent workers is imputed in the total labour share on the basis of average wages paid to the dependent work force. However, if the self-employed worker is actually a precarious worker on rather low pay (a phenomenon that is on the rise), this boils down to artificially inflating the labour share while underestimating the capital share of the economy. For example, research from the central bank of the Netherlands finds that the traditional methodology used by the OECD is resulting in a labour share that significantly exceeds 100% of the total value created in sectors such as construction that have a high incidence of self-employed contracts. The Dutch Central Bank therefore uses an alternative measurement where the contribution of self-employed workers is measured by their ‘mixed’ income, a measure that is based on their real earnings. The graph below compares this alternative indicator (blue line) with the traditional method to calculate the labour share (grey line). The result is clear: The labour share is much lower than under the traditional calculation.
In short: the traditional labour share has gone back up, while the alternative labour share has been unable to recover from the financial crisis.
Unfortunately, the OECD does not acknowledge this statistical problem. If it did, the impact it may have on labour share trends for other OECD economies is probably not to be underestimated.
Alternative Labour Share for the Netherlands
Relevant conclusions
The conclusions drawn by the OECD working paper nonetheless are correct and relevant.
Inequalities indeed increased significantly within the group of workers, as can be seen from the next graph where the income (wage) of the top 1% of workers has been skyrocketing.
The paper also explicitly confirms that diverging trends in labour shares and wage inequality in individual OECD countries “suggest that longer term global trends such as technological change and globalisation alone cannot fully account for decoupling wages from productivity” (par 29). The OECD underlines the significant role of public policy settings in shaping the effects of these global trends on labour shares and wage inequality. In other words, mega-trends such as technological progress and globalisation should not be used as an excuse to present rising inequalities and falling labour shares as ‘regrettable but unavoidable’ consequences: Policy also matters and it matters a lot.
The OECD intends to undertake further research on public policy settings, both from a macro as well as from a micro (firm level) point of view. What that research will deliver on concrete policy conclusions, remains to be seen.
