Explaining Household Income

DS Wahid’s announcement over the preliminary results of the 2014 Household Income Survey (HIS) has created a minor furore (excerpt):

RM5,900 average household income? Lies, damned lies, and statistics, says MP

KUALA LUMPUR, Sept 9 ― The federal minister's claim that Malaysian households make an average of RM5,900 monthly is a farce that does not represent the actual earning capacity of most Malaysians, an opposition lawmaker said today.

DAP's Bukit Mertajam MP Steven Sim said the figure is a result of creative use of statistics in the 2014 Household income Survey (HIS) preliminary report to paint a “dishonest” picture of increased prosperity among the people….

…On Sunday, Minister in the Prime Minister's Department Datuk Seri Abdul Wahid Omar was quoted by national news wire Bernama as saying that the average household income in the country has risen to over RM5,900 a month, a significant increase from the RM5,000 monthly average recorded in the 2012 HIS.

Sim stressed that the figure does not make sense when the government itself admitted that 80 per cent of Malaysian households had benefitted from the Bantuan Rakyat 1Malaysia (BR1M) cash aid programme, whose recipients must earn less than RM3,000 a month.

He added that government statistics also showed that 82.5 per cent of Malaysians below the age of 30 earn less than RM3,000 a month as at September last year….

…Sim stressed that the median household income in Malaysia ― which stood at RM3,626 in the 2012 HIS ― gives a clearer idea of how much Malaysian households actually earn as it differentiates between low and high income earners….

Certainly it would be better to use the median rather than the mean to give a more accurate portrayal of the current state of household income.

However, that doesn’t much change the thrust of the results – while the median is much lower than the mean due to the skewness of the distribution (and nobody, I think, disputes this), I think growth will almost certainly be close to the same whichever measure you use. The growth across the two periods is really the more important point here, not the levels.

But I think there’s a more fundamental problem here than whether the results are political spin or outright lying. In the immortal words of Inigo Montoya, “I do not think it means what you think it means.”

To be more specific, the word income as used in the HIS – not just for Malaysia, but nearly the world round – is not exactly the same as the usage in every day language.

I could point out the obvious, that income ≠ wages. Most people would equate income with salaries and wages, as for most people this would be their main source of income. But household income also includes things like rent, interest and dividends, even if capital income is only really a big factor for the top 10% or so. Nevertheless, when we’re talking about aggregated data, some element of capital income is included, even if it doesn’t figure for the bottom 80%, and whether you use the median or the mean.

There’s also some element of business income. At the national level, “mixed income”, which covers income derived from everything from nasi lemak stalls to million dollar, multi-partner legal and accountancy firms, is about a third the size of wage income. Only about 2/3rds of Malaysia’s total labour force earns a salary. Everybody else is either running a registered business (about half a million people or so), or in the informal sector which is not well tracked. Put another way, you need to cut 20%-25% from household income to arrive at a figure approximating wage income.

But you could say that for the average Malaysian, none of these really make any difference. There would still be some discrepancy between the HIS data and what people are earning on the ground. So here’s what really sets apart income in the national statistics sense, and income in the everyday common usage sense. What’s that one thing?

Property.

Not in the form of rent (which is an explicit financial flow), but in the form of imputed rent (which isn’t). Home ownership in Malaysia (believe it or not) exceeds 70%, which means the flow of “income” derived from housing services goes far beyond the top 10% or even top 20% of the income distribution. In fact, for many in the bottom 40%, home ownership represents their only substantial store of wealth, and thus a major source of the disconnect between household income and wages.

Let me explain why: when you live in your own home, you are de facto receiving gratis a flow of housing services. This flow of housing services is represented in the statistics by imputed rent. In essence, you’re paying rent to yourself, and this “unearned” and non-financial income needs to be taken into account (refer here for a decent discussion on the subject). I know that sounds really strange, but the concept is embedded in the national accounts and household income data – it’s part of GDP, too. Again, I stress this is not a uniquely Malaysian practice, but is global best practice.

But there’s some method to the madness – if you live in a 5 room, 5 bathroom mansion with a lawn, the implied income flow you receive should be much, much higher than if you’re living in a 2 x 2 shoebox apartment, even if your “earned” income is the same. In accounting for this “income” flow, typically the actual rent of an equivalent property is used as a benchmark. One implication of this is that if actual house rentals go up – and they have indeed been rising in Malaysia though not quite as fast as house prices – so does the implied income from housing services, even if there are no actual money flows involved.

Imputed rent is a not inconsiderable proportion of Malaysian household income in the HIS. My crystal ball spits out an estimate of above 10% of average gross income (ok, it’s not a crystal ball, I’m just not allowed to disclose my source).

If you put all these differing strands together, you can form a rough guesstimate of average wages: take away 10%-15% of gross income for imputed rent, take away another 20%-25% for capital and mixed income, then divide by 1.8 (approximate average number of income earners per household).

For an average household income level of RM5900, that works out to an average salary per worker of between RM2,350 and RM2,100. The median is another 25%-30% below that, which gives approximately RM1,650-RM1,450. I don’t think that’s too far from where most people think Malaysian wages actually are. In fact, after accounting for growth, that’s more or less where the 2012 Salary and Wages Survey puts it.

To sum up, there’s really no inconsistency between the HIS numbers, which includes both earned and unearned income, and the numbers quoted for BR1M or from LHDN, which uses only earned income alone. Rather, the difference is really in how income is defined and computed.

If I have a criticism of our household income data, it’s that earned income is not disclosed along with gross income. Singapore for instance, reports both.