There is a very good article published in the Malay Mail over the weekend.[1] It is about the brisk sales of high-end cars like BMW and Mercedes in Malaysia while entry-level, compact cars like Perodua are doing badly, and the consumption pattern in the economy by income levels.

[The divergence car sales by prices are part of the weird macroeconomic statistics that have been coming out of Malaysia since the GST was introduced back in April 2015, along with the collapse of oil prices. Weirdness like the consumption growth is rising nicely while big ticket items like cars and homes are not doing so well in terms of growth. In fact for cars as a whole, it has been declining. It may be some non-business cycle explanation behind this, like the widening of train network in Malaysia but for now (a person driving BMW rarely if ever take the train, for instance), let’s put that aside and maybe discuss that on another day.]

But I have one, tiny issue with the article, way, way down. The author questions the accuracy of the 2014 Gini coefficient, which suggests inequality, decreased compared to 2012, while pointing to the latest car statistics (2015-2016), which may say otherwise:

While BMW Malaysia posted its highest ever sales growth last year, its rival Mercedes posted record sales for the second straight year in 2016. A total of 9,047 Mercedes vehicles reported to have left its showrooms in the first nine months of 2016, marking a 10 per cent growth compared to the same period last year, according to Malaysia’s leading automotive online magazine

In 2015, when the local economy appeared to be slowing, Mercedes sold a total of 10,845 vehicles, a record increase of 56 per cent from 6,932 units in 2014.


Based on the sales data, slower economic growth was not affecting all segments of the country equally. While those in the lower income brackets are complaining of rising costs, their more well-off counterparts have been splurging.

”What it indicates is that while the low and middle income earners are experiencing fiscal constraints, it is business as usual for the higher income group,” Muhammed told Malay Mail Online in a phone interview.

Global trend

Putrajaya has so far shrugged off the idea. According to official statistics, the country’s Gini coefficient series shows a downward trend in household income inequality from 2004 to 2012, after which it falls off drastically — the Gini coefficient was 0.46 in 2004 and only dropped by 0.3 percentage point after eight years. But in 2014, it dropped to 0.40.

I find the final paragraph (especially the final sentence) in this excerpt as slightly confusing. After laying out the situation in 2015 and 2016, the article goes on talk about the 2014 Gini data, which might not describe 2015 and 2016 very well.

Confusing, because the Gini coefficient is a low-frequency data done, at the moment, every 2-3 years together with the household income survey. Not only that, it lags severely, published only after the survey was done months earlier: the 2014 data was released in June 2015 while the 2012 data was released in July 2013. You can see its low frequency in the chart below:[2]

In contrast, car sales statistics are high-frequency data available monthly.[3] 

High-frequency data do have a lot to say about the present time. But I feel it is unfair to cast aspersion on low-frequency data from all the way back by using more recent information (2015-2016 car sales) the way the article does. It is unfair because the two datasets describe two different points of time. They are not contemporaneous data and not comparable, at least in the way it is being done.

I am aware of the paper by Lee Hwok Aun and Muhammed Abdul Khalid which the Malay Mail article cites. But the paper utilizes car sales data that is contemporary to the Gini coefficients it uses (before that, I would like to say it is a shame that Lee had to leave University of Malaya because of the short-sighted government spending cuts). You can read the working paper here.

In contrast, the Malay Mail article is taking post-2014 car sales data to question a 2014 Gini coefficient.

So, I think the attack on the 2014 Gini coefficient from the Malay Mail angel might be overdone, and different from the Lee-Muhammad criticism. The 2014 coefficient is just unable to describe the 2015 and the 2016 situation, especially since the GST was introduced in April 2015. There was a significant structural break since the 2014 coefficient was released.

[Also, I feel it is important to note that the Gini we have here measures income inequality, not consumption inequality. We have the Household Expenditure Survey though. Just saying]

In any case, the next Gini coefficient may come out this year or the next. The Household Income Survey is supposed to be done twice in five years (last done in 2014). The latest coefficient may yet go up.

Mohd Hafiz Noor Shams. Some rights reservedMohd Hafiz Noor Shams. Some rights reservedMohd Hafiz Noor Shams. Some rights reserved

[1] — Syed Jaymal Zahiid. What BMW and Perodua sales data says about the economy. Malay Mail. January 21 2017.

[2] — 2014 Household Income Survey. Department of Statistics. Accessed Jan 23 2017.

[3] — Malaysian Automotive Association. Accessed Jan 23 2017.

Mohd Hafiz Noor Shams. Some rights reservedMohd Hafiz Noor Shams. Some rights reservedMohd Hafiz Noor Shams. Some rights reserved

p/s — I am not defending the Gini coefficient per se. I am saying the criticism mounted in that particular way is off the mark.

Trackback URI | Comments RSS

Leave a Reply