Documenting Income Inequality: Malaysian Household Incomes 1970-2009

[Warning: this post is very-long and graphics-intensive]

We had an interesting debate going on in the comments to this post regarding inequality, but very specifically on racial inequality in Malaysia. I’d been meaning to post on the data underlying my arguments in that debate, so here it is even though a couple of months late.

It’s been a central pillar of public policy since 1970 to embed affirmative action in the economy along many dimensions, from education to corporate ownership to public procurement. What success any of these policies have had, has to be tempered with the inefficiencies caused by tampering with market forces. There’s certainly a case to be made on both sides of the fence, but much of these tend to descend into emotional and value laden arguments, and not really economic ones.

In any event, I don’t propose to rehash any of these debates, except to lay out what the data on incomes actually says in chart form, and what I think it might all mean. All the data is taken from the EPU website (see the link at the end of the post). The only transformation I’ve done is to inflation adjust the data for some of the charts (clearly marked).

The data covers mean household income from 1970 to 2009, and is sourced from household income and expenditure surveys, which are conducted periodically but not, unfortunately, on an annual basis. Clicking on any of the charts should take you to a larger version.

So without further ado:

Except for the obvious recession periods (1985-1987, 1997-98), incomes of Malaysian households have generally risen (first chart above, in RM). The second chart above shows the ratios (by ethnicity) against the average Malaysian household, while the third chart is the same as the second, but omits foreigners for a closer look at the data trends.

The big variation in incomes of foreign workers can be largely explained as changing demographics – in the post-independence era and before the drive towards industrialisation, foreigners were primarily high income expatriates running or working for foreign owned companies, or in government. As manufacturing and estate activities increased, there was a dramatic shift towards low wage foreign labour, hence the drop in average foreign incomes.

One thing that’s clear though is that income by ethnicity is converging on the Malaysian norm, with two distinctive periods – 1970-1987 and from 2002-onwards. In the period in between, income ratios by ethnicity were largely stable, with a tendency towards divergence. This is true not only of average households across the spectrum, but also true within income stratas as well.

For example, here are the ratios for the middle 40% of income earners by ethnicity relative to the whole:

For all intents and purposes, the chart above is very nearly identical to the one before using aggregated incomes.

More interestingly still, we can see where the idea that inflation has eaten into take-home pay in recent years has come from (inflation adjusted incomes; 2000=100):

It’s not true of Malaysian households generally and definitely not for Bumiputera households, but it is specifically true for Chinese households and somewhat so for Indian households. In fact real incomes have fallen from 2002 onwards for Chinese households and from 2004 are effectively flat for Indian households.

Take away the impact of the 1997-98 recession, and real incomes for Chinese and Indian households have barely increased since 1997. Real incomes in 2009 for Bumiputera households on the other hand are as much as a third higher than in 1997.

Again this is across the board, though these trends appear to be stronger for high and middle income households than they are for poor households, where incomes even for Chinese households have shown some increase, at least up to the advent of the Great recession (inflation adjusted incomes; 2000=100):

So the conclusion that can be reached is that inequality by the race dimension is gradually being reduced and is at a historical low.

Unfortunately, we’re not so lucky with overall income inequality:

Again across the board – income inequality is a Malaysian phenomenon, and not wholly a cross-ethnic phenomenon. While income inequality has fallen generally from an average around 0.50 in the 1970s to around 0.45 since 1987, there has been little to no significant improvement since.

There does however continue to remain inter-ethnic income disparities (here, Buimputera households against Chinese households by income strata):

From around 40%-45% of Chinese household incomes in 1970, Bumiputera incomes have closed to 68%-74% in 2009. The gap has been closed to the smallest level historically, but its still there.

Nevertheless, questions remain:

1. How have Bumiputera incomes gained so much, and just as importantly, why were the gains confined to pre-1987, and post 1997?
2. What explains the stagnation in inter-ethnic income ratios between 1987-1997?
3. Why, despite poverty alleviation schemes which appear to have been successful, has this not resulted in reductions in income inequality measures post-1987?
4. Why did Bumiputera incomes continue increasing in the 2000s, yet Indian and Chinese incomes stagnate or retreat?

There’s obviously the impact of NEP and post-NEP policies, but that explains only some part of the convergence in the inter-ethnic income ratios, but not fully. These policies continued to operate throughout the sample period, but the biggest impact appears to have been pre-1987, and less so afterwards.

The fact that income growth and ratios behaved so differently across time suggests that affirmative action policies aren’t the main driver for income convergence – quite a few measures were rolled back in the aftermath of 1997-98, yet this period marks the beginning of the sharpest income gains for Bumiputera households.

My working hypothesis revolves primarily around commodity prices (particularly rubber and CPO), but even that’s an imperfect measure going by these next two charts:

Up to 1987, rural incomes and Bumiputera incomes were virtually synonymous. But after the 1987 recession, Bumi incomes also began to have a growing urban component (first chart). The second chart suggests much the same story – if agriculture growth was driving Bumiputera income gains, the biggest impact was pre-1987 and less so afterwards.

Here at least we have a partial explanation of why Bumi income ratios stopped climbing after 1987 – urban income growth on average outstripped rural income growth from 1987-1997, helped along by weak commodity prices in US dollar terms after the coordinated forex intervention following the Plaza Accord. Commodity prices only began climbing again (in inflation adjusted terms) after about 2000.

Nevertheless, we’re still left with the mystery of why Chinese and Indian incomes (presumably more concentrated in urban communities) stagnated while Bumiputera income gains began climbing again post-1997. Commodity prices certainly were a factor here, but not the whole story given the increasing urban component of Bumiputera incomes.

Affirmative action policies were also a factor, but also nowhere near the whole story, especially considering the rollback of some of the equity ownership criteria after 1997 and liberalisation of some economic sectors.

One way to maybe resolve these puzzles is to look at the ethnic makeup of the labour force in different economic sectors, but I don’t have that data. What data I do have suggest somewhat perversely that Chinese incomes are more strongly affected by agriculture than by manufacturing, which doesn’t make any sense to me.

So we’re left with more than a few conundrums for people to continue arguing over…and fertile research grounds for aspiring PhDs.

Technical Notes

• Data on household incomes from the Economic Planning Unit
• Consumer price index data from DOS (spliced series rebased to 2000=100)