Critiquing The Critique; Or This Is NOT How You Calculate Inflation Part II

I got tipped off about an analysis of the ETP last week, which makes some of the same points I made two years ago (excerpt):

A Critique of the ETP (Part 2)
We won’t really be twice as rich in 2020

RM48,000 in 2020 is not real income.
The ETP promises to double gross national income GNI) per capita to RM48,000 by 2020 from RM23,700 in 2009. However, RM48,000 in 2020 will be worth a lot less than RM48,000 today, just like RM100 today buys a lot less than RM100 eight years ago, thanks to ever-rising prices. If Malaysians are really to be twice better off, nominal income must be RM64,000 by then, to compensate for the 2.8% per year inflation that PEMANDU expects.

Nothing transformational in the RM48,000 target.
This target is for nominal$ income, which includes inflation, and not real income, which strips out inflation. Because of inflation, nominal GNI per capita growth averaged 8.2% from 2001-2010, whereas real GNI grew only 3.2%. At the historical average 8.2% per year growth rate, nominal incomes will exceed RM48,000 by 2018 anyway, with or without the ETP or PEMANDU.

PEMANDU and its expensive consultants cannot even get basic mathematics correct.
If the income target is RM48,000, PEMANDU’s 6% real GNI growth rate and 2.8% inflation forecasts are wrong. If its growth and inflation forecasts are right, then the RM48,000 target is wrong - it should be RM54,145 in 2020, not RM48,000. Furthermore, key metrics of some EPPs – the investment value, GNI contribution and jobs created – are unavailable.

Grade ‘D’ for data transparency.
In this series, we evaluate the ETP on its own terms based on the goals and plans outlined in the ETP Roadmap. PEMANDU scores a ‘D’ for data transparency. Like us, Malaysia’s top research house finds it impossible to get the numbers to add up.

Now, I absolutely agree that there’s a strong disconnect between the nominal target embedded in the ETP and the real target we’re aiming for. In fact I wrote about in July 2010 (excerpt):

The NEM In Numbers: Nominal Versus Real

…One of the things that has been bothering me about the New Economic Model is that the public discourse revolves around trying to reach over 6% real growth. That’s largely false, or to be more precise, largely misleading.

The goal of the NEM is for Malaysia to reach high income status, which is defined as between US$15,000-US$20,000 in per capita income by 2020, compared to our reported level of US$6,896 as of 2009 (source: IMF World Economic Outlook Database April 2010). Per capita income here is defined as nominal GDP as the numerator and the total population as the denominator.

Two problems immediately come to mind here, relative to what’s being bandied about. First the income target is in US$, not in local currency terms. Secondly, we’re talking about a real (inflation adjusted) target, when we’re actually aiming for a nominal (current Ringgit) target.

My calculations pretty much jives with the REFSA paper – if the real projected numbers come anywhere close to actual, we’re going to hit the high income target in the 2017-18 time frame. It’s also nice to have an independent assessment of the progress in the ETP.

But…

It might help if the authors did a bit more homework.

For example, in point number 3 from the summary above, the authors accuse Pemandu and it’s consultants of not being able to add the numbers up correctly. That’s true – if Pemandu were the ones doing the calculating.

But the truth of the matter is that the numbers weren’t generated for the ETP, they came from the New Economic Model (NEM) Part I document and the source is the NEAC. Note that my own critique was written in July 2010, four months before the ETP was launched, and about the same time as the ETP “labs” were being conducted. The NEM document at that point had been available to the public for four months.

Credit (or discredit) where it’s due – the reason why Pemandu’s expensive consultants didn’t get the numbers right was because the numbers were actually done by a bunch of economists (the list of NEAC council members is available here) working in conjunction with the EPU and the World Bank, and not by Pemandu or its consultants. The inflation numbers were actually sourced from the IMF’s forecasts for the next decade.

Economists can’t count, obviously, LOL.

That aside, the focus in the article on the RM48,000 nominal income target is also misplaced. The actual target is really US$15,000-US$20,000, which is where the high income threshold is expected to be. There’s nothing firmer than that, for a very simple reason – it’s arbitrary and third party determined.

Let me explain: the determination of who’s high income and who’s not is made by the World Bank (details here), not by Malaysia. And their methodology is based on nominal GNI per capita in US$ through the World Bank Atlas method (which adjusts for exchange rate movements).

Hence, trying to fine tune the 2020 RM GNI target to get exactly double the real income from 2009 is a bit of a fool’s errand, because despite the talk from the government, that was never the real goal being contemplated. It’s always been about Malaysia graduating from the World Bank’s middle income category to it’s high income category.

Never mind if the ETP document stresses on the numbers – I’ve never viewed the targets as hard numbers to aim at, but more aspirational in nature. A guide to where we ought to be if you will, as opposed to where we must be.

There are just too many variables with too high a degree of uncertainty involved to forecast economic numbers eight years into the future with any kind of accuracy. Anybody involved in economic forecasting can tell you it’s hard enough to make an decent forecast 8 months ahead, never mind 8 years ahead. That’s why I’ve never bothered critiquing the overall numbers, or the EPP breakdowns, as they stand.

So why have the target numbers at all? Because you need some sort of guide to work towards. Would it make a difference if the target was approximately 15% higher, as the authors demand? Not really. These types of forecasts are most useful as a guide to policy action – if changing the targets don’t entail a change in policy action, then the exercise is pretty pointless.

Moreover, whether we achieve the ETP/NEM targets is predicated on two things outside the government’s effective control – the World Bank’s high income threshold in 2020, and the MYRUSD exchange rate. Since we don’t control the former, and BNM is adamant that the latter be market determined, we’re actually shooting at an arbitrary RM-denominated GNI target that’s probably as good as any other.

More importantly, my view has always been that since quite frankly we’re likely to hit the high income threshold sooner or later (we’re NOT in a middle income trap *cough*), the focus should be on structural aspects of the economy and quality of life issues e.g. the distribution of income and wealth, education, and R&D, among others. Focusing the debate on the medium term putative NEM numbers distracts from more pressing long-term issues.

However, since we are discussing numbers, I do have a critique of the critique to make, which explains the subheading of this blogpost. Just like Dato’ Seri Anwar, the authors made two elementary errors in calculating inflation (oh, the irony!!).

From page 3:

At the national level, this is reflected in the GNI per capita growth rates. Nominal GNI per capita grew by 8.2% per year on average from 2001-10. However, because of inflation, real GNI per capita grew by just 3.2% per year.

And from page 4:

The economy is of course much more complex. The GDP deflator is the official estimate of average inflation across the economy. The GDP deflator is then applied to the nominal numbers, to deflate them down to an estimate of the real numbers.

From which they derive (on Pg 5):

Sources: Data is as per sources quoted in this report. Note that PEMANDU’s 2.8% inflation forecast is well below the 5.0% average from 2001-09, as measured by the GDP deflator.

Two problems here. First, the GDP deflator is indeed a measure of inflation across the whole economy – that’s why it’s NOT used as a measure of consumer inflation, and also why it’s wrong to use it to measure changes in actual consumer purchasing power.

Breaking down the GDP numbers, on the expenditure side we have consumption (both public and private), investment (also both public and private), and net exports. The problem here is that since domestic consumers don’t buy exports (otherwise, duh, they’re not labelled exports), deflating nominal gross incomes using the total deflator will not give an accurate picture of domestic inflation.

If the exchange rate is appreciating, as has been the case with the Ringgit generally since it was floated in July 2005, the nominal value of exports will rise faster than the domestic rate of inflation (by the same token, import price changes would understate inflation). Using the GDP deflator as a measure of inflation would therefore overestimate the rate of consumer inflation in Malaysia, especially since the volume of exports is around 100% of Malaysia’s GDP.

Now the NEM does indeed use the GDP deflator as its measure of inflation – but they don’t apply it in terms of consumer purchasing power, unlike this report. Whether the loss in consumer purchasing power (and hence incomes) through inflation would be the same as the deflator relies on a number of different variables, particularly on domestic inflation but just as importantly on the Ringgit’s external purchasing power. Appreciation of the exchange rate, which is empirically likely with higher per capita incomes, would partially offset losses in purchasing power from domestic inflation sources.

Second problem, the authors use per capita differences in nominal and real GNI as their measure of inflation i.e. 8.2% nominal growth less 3.2% real growth yields 5% average inflation.

Since they’re using per capita numbers, this would only be accurate if there was no population growth.

But if you take away population growth (which averaged 1.8% during that decade) from that inflation calculation, we get the actual average annual change in the GDP deflator of 3.2% – in fact that’s the number you get when you do the calculation the right way, by looking at the average change in the nominal against real numbers on a non-per-capita basis. But again, that 3.2% deflator number still overstates domestic inflation, because it still includes changes in export prices.

So, do we have here a case of the pot calling the kettle black?