Wednesday, February 24, 2010

A Tale of Two (Spin) Cities: 4Q2009 GDP

The GDP numbers look pretty good don”t they? Real GDP growth hit 4.6%, which by Malaysia’s standards means we’re officially out of recession. This hit right in the middle of the upper half of my forecast range. Rather than comment directly on the numbers, I’m going to try a little experiment and show what happens when you change the basis of calculating growth.

If you recall, for this purpose Malaysia uses quarter on the same quarter last year. Most other countries use this methodology, but very few use it as their reported growth number (with the notable exception of China). To illustrate, such a calculation would look like this:

GDP growth=GDP(4q09)/GDP(4q08)-1

The internationally accepted standard on the other hand is to use quarter on the previous quarter, using seasonally adjusted data, with growth number later annualised:

GDP growth=((GDP(4q09)/GDP(3q09))^4)-1

As an alternative, you can equivalently use a natural log formulation:

GDP growth=(1+(LN(GDP(4q09)-LN(GDP(3q09))))^4)-1

…which is a little easier computationally, as well as being level neutral.

So, here we go…

Case 1: Malaysian standard
After three straight quarters of contraction, Malaysia’s 4Q09 GDP growth numbers confirmed that the country was out of recession, posting a growth of 4.5% against the same quarter last year:


On the demand side, growth was driven by a strong recovery in trade and an 8.2% jump in domestic investment. On the supply side, growth was largely broad based and supported by fiscal stimulus measures which helped drive construction growth to 9.2%, while the manufacturing sector posted its first positive growth in four quarters. The only sector yet to recover is mining, albeit the fall in output improved to –2.8%:


Case 2:International standard
Malaysia’s 4Q09 GDP growth zoomed an impressive 14.2% despite a sharp pullback in government spending, which fell 11.1%:


Output has now recovered to the level prior to the recession, which began in 2Q08 and lasted four consecutive quarters, to 1Q09. On the demand side, growth was driven by a 28.2% jump in exports and a 21.3% increase in domestic investment.

On the supply side, growth was uneven, with agriculture and manufacturing leading the way. Services growth was sustained, but mining sector output dropped 8.50% while construction output growth fell from 9.1% in 3Q to just 1.9% in 4Q2009.


A real study in contrasts isn’t it? While the broader picture is similar, the details show significant differences:
  1. The timing of the recession changes from beginning in 1Q2009 under the official standard to 2Q2008 under the international standard –a full three quarters earlier.
  2. The length of the recession also changes from 3 quarters to 4 quarters.
  3. Positive domestic demand growth in 4Q2009 turns negative under the international standard.
  4. The biggest differences are in the sector breakdowns, with sharp construction and services growth turning into consolidation, and first manufacturing recovery has actually been evident since 2Q2009.
The billion dollar question is: Would a difference in calculation affect the planning and expectations of economic agents (consumers and businesses), and would it affect the timing and scope of government policy? I don’t necessarily think so.

If you recall, the first stimulus package was introduced in November 2008, which coincided with the GDP report for 3Q2008. At that point, under the international calculation the recession would have already entered its second quarter, but was still relatively shallow. The signs were already there however for the sharp drop in 4Q2008, and the even steeper drop in 1Q2009.

However, fiscal consolidation might have happened earlier with the emergence of the economy out of recession in 2Q2009 – that might not have affected for instance the credit support the government provided which formed the bulk of the 2nd stimulus package, but we might not have had as much of the direct spending that the government committed to (nor the pace of borrowing).

So…which way is better? From a pure macro-perspective, neither is necessarily good or bad. I don’t think there would be any changes in BNM’s policy moves for instance. But from a micro-perspective, the international standard offers a better picture of what’s happening on the ground and which areas need government support (if at all). Ideally, both methods should be reported and used for policy adjustments – as is the general practice elsewhere.

Technical Notes:
4Q2009 GDP report from the Department of Statistics


  1. I don't understand why the international stardard take Y1 Q4 on Y0 Q3. Why is that? Further, what's up with the exponent of 4?

  2. Not sure what you mean? To clarify, it's always the quarter being reported over the preceding quarter, i.e. 1Q2010 over 4Q2009, or 4Q2009 over 3Q2009 etc

    The exponent is for annualisation (4 quarters in a year). In other words, what would be the growth level for the year if you extrapolate the present quarter's performance over a full year? You can't just times 4 (which ignores increases in output between quarters), it has to be exponential.