Tuesday, March 24, 2015

Wages and The CE/GDP Ratio

I’ve come across the same dilemma myself, but a box article in BNM’s 2014 Annual Report outlines the latest data (excerpt; emphasis added):

Trends in Malaysia’s Gross Domestic Product by Income

…In terms of share, capital income forms the largest component of GDPI (Chart 3). However, with the growth of labour income outpacing the growth of capital income, the share of labour income to GDP has risen steadily from 29.5% in 2005 to 33.6% in 2013. By definition, however, the labour income component in GDPI excludes income earned by self-employed individuals…With such adjustments, the share of labour income for Malaysia is higher, on average, by 8.0 ppt. throughout the period (Chart 4)….

...Despite the rising trend, the labour income share in Malaysia, even after accounting for the income of the self-employed, remains relatively low compared with other upper-middle income economies (Chart 5). Within the region, Malaysia’s adjusted labour income share (42.1%) is lower than Korea (54%) and Singapore (44%).

While this may be a cause for concern, it is important to note that a higher share of labour income does not necessarily correspond with higher average wages. As noted in both Chart 5 and 6, Brazil and India, for instance, have higher labour income shares than Malaysia, but lower average monthly wages. On the other hand, Singapore has a relatively low labour income share, but a high average wage level. Of significance, most advanced economies tend to have both high average salaries as well as higher labour income shares….

…Chart 7 provides a snapshot of the adjusted labour income shares and average wages5 across industries in Malaysia. Similar to the cross-country observations, the variations in labour income shares across industries do not necessarily correspond with the variations in average wage levels. For example, the share of labour income in the labour-intensive industries tends to be higher than average, but the average wage levels are lower than the national average….

I call this a dilemma, because targeting a higher employee compensation/GDP ratio is not the same as raising wages. The sectors with the highest wages (e.g. oil & gas) also tend to have lower COE ratios; conversely sectors with higher than average COE ratios also show average wages lower than the national average. If you don’t understand how something came to be, it’s tough to figure out what policy levers to pull.

The article suggests that higher productivity will tend to drive higher average wages. I won’t dispute that or disagree with it, but I don’t think the matter is that simple – advanced economies have both higher COE ratios AND higher wages. For that matter, what explains the different labour income shares between Singapore (which has higher “productivity”) and Korea? Higher productivity increases the size of the pie, but we still need to look at how the pie is divided as well.

I’m grasping towards the idea that these cross-sector and cross-country differences could be due to different institutional incentives and attitudes towards savings and investment versus consumption on the one hand, and labour/capital bargaining power on the other, but my modelling skills just aren’t up to figuring that out yet.


  1. En Hishamh,

    To raise productivity, one can only add resources (equipment/it/others) therefore capital. This will lead to different skill sets (normally higher) hence higher wages.

    On balance maybe the original share remains much unchanged.

    Zuo De

    1. @Zuo De

      Raising productivity is more than just adding more capital. There's more than one measure of productivity, but the one most economists refer to is total factor productivity, which is the productivity level you get over and above what you put in in labour or capital.

      I'm on an automation drive in my department. We just changed our process flow and used the software that we already had more efficiently, and cut the time of working on a report from 6-7 hours, to just 2-3 hours.

    2. I think there is a different between industries / sectors.

      In the service sector that i am in, a human can only do so much. Plus some new equipment maybe a bit more but no that much more as the human factor is controlling.

      As for manufacturing, automation should increase productivity over and above .....

      I think your reduction in report writing is more on the approval cycle rather than you can churn out reports as one need to think what to write how to write and this human factor cannot be improve by the software, please enlighten.

      Zuo De

    3. @Zuo De

      I'M the approving authority :)

      The biggest time saver was actually automating the generation of charts from data providers such as Bloomberg and Haver. What I had my staff do is to prepare Word and Excel data and chart templates, which meant that pulling the data from the systems resulted in the automatic creation of the charts required. We could spend our time solely on analysis, rather than wasting them on time consuming manual work. As I recall, most people only use 1% of the total capabilities of Office software.

      I agree that service industries are quite different and productivity is more challenging. But that's the function of an entrepreneur and managerial innovation - asking how things can be done faster and better with the same resources.

    4. My apology for my ignorance (on approving), my bad.

      Zuo De

  2. Is Zairil's call that EF test is flawed data correct?

    1. @anon 10.22

      I have no idea. My personal experience suggests English standards have indeed declined.