Monday, March 21, 2022

Minimum Wage Revised Part II: Some Evidence on Wages

Part I dealt with some of the criticisms of the minimum wage. Part II here will look at some of the data around the impact on wages and prices. For a look at the impact on employment, see this old post.

How do wages react to an increase in the minimum wage? To try to answer this question, I'll use some estimates from the EPF on wages at various parts of the distribution. To my knowledge, I've only shown this particular data in public just once before, and that was many years ago. Since then there has been multiple revisions of Malaysia's minimum wage, with varying results at least in terms of boosting wages, as you shall see.

This first graph is from the initial implementation in January 2013, when the minimum wage was set at RM900 for Peninsular Malaysia and RM800 for Sabah and Sarawak:

Reading from top left to bottom right, each graph represents a decile in the wage distribution, ie from bottom 10% to the 60th percentile. S1 is the bottom 10%, S2 is the bottom 20%, and so on. S5 would then be the median, right in the middle of the distribution.

One can immediately see the impact that the setting of the minimum wage had on wages across the distribution:

1.    There was an immediate and significant increase in wages at the outset, all the way from the bottom to at least the median (hard to tell with S6); and

2.    For S1 and S2, there was also an acceleration of growth, which lasted a little over 2 years for S1 and about a year for S2. 

Side note on the data: S1 and S2 tend to be workers with irregular and inconsistent contributions - the median is just one contribution a year. I suspect these are mainly part-time or shift workers paid by the hour, hence why the indicated salary levels remain below the minimum wage.

Coming back to the data, not only did the 2013 minimum wage set a floor under the labour market, it also pushed wages up for the entire bottom half of it. As evidence of a policy having its intended effect, this is pretty solid.

For the subsequent revisions, it's a lot less exciting. These happened in July 2016 (RM1,000), January 2019 (RM1,100), and 2020 (RM1,200). I'll forbear from showing 2016 because the data is a little weird, and I think there's an issue with how the it was loaded in the database. Any effects for 2020 of course got wiped out by Covid. Only the January 2019 data is worth showing, because of the interesting contrast it makes against the 2013 data:

We can only say with certainty that there was an effect for the 2nd decile (ie S2, the bottom 20%). The rest show the mostly normal seasonal pattern of bonus payments, and not much else. Also note that the chart covers the start of Covid as well, which swallowed up any effect from the revision to RM1,200. 

Moving on, another interesting question to ask given my previous post, is did all this cause prices to increase? That's tougher to see (index numbers, 2010=100):

The only categories in the CPI where price increases coincided with the setting of the minimum wage are Education and Furnishings, Household Equipment, & Household Maintenance. And for Education, that's actually a common seasonal pattern, as prices appear to rise every January anyway, though you could argue this particular time was a little bigger. As I argued previously, minimum wage hikes should boost services inflation, but that doesn't appear to have happened here. Restaurants & hotels and Recreation services & Culture saw price increases only a year afterward. Healthcare barely deviated at all. I checked for price movements around the revision dates, and didn't find much else either. I wouldn't necessarily take this as proof positive that minimum wages never cause inflation however.

My takeaway from all this is that context matters. We have a situation where a change in the minimum wage caused a significant shift in half the income distribution, and a situation where there was only a very localised impact. Changes in the price level were also very localised. As such, generalised statements aren't useful, as any impact from changes to a minimum wage depend on the particular situation, as I argued in my previous post.

I'll delve a little further into that in the next post.

No comments:

Post a Comment