The GDP growth for the second quarter decelerated further to 4.0% YoY from 1Q16’s 4.2% YoY. But the most interesting GDP component ever since the GST was implemented in April 2015 is private consumption.
And there was a huge jump in that part of the equation. The 2Q16 private consumption expanded 6.3% YoY from 5.3% YoY in the previous quarter. It suggests things are normalizing.
Is growth normalizing?
I did a bootstrap model comparing actual growth with what it would have been without GST. It does show some kind of normalization.
The modelling is very naive with just a bit of seasonality sprinkled in it. The blue line is the actual YoY GDP growth while red is the counterfactual if there was no GST imposed:
But it is important to say it is not so straightforward to claim consumption growth is normalizing. The fact is 2Q15-1Q16 growth are incomparable to the 2Q16 rate. The latter period is the first time the GST effect has been controlled on the year-on-year basis since the tax was implemented while the earlier ones are polluted by base effect. Perhaps, it is better to compare 2Q16 rate with those before 2015 to get a feeling how the GST has impacted economic growth as a whole.
But we have only one point so far. Maybe it is wise to be patience and wait for more data point to be available.
Difference-in-difference
Nevertheless, the nature of 2Q16 makes YoY difference-in-difference analysis across time possible for the first time. Diff-in-diff is done to compare how a certain thing (in this case, YoY GDP growth) behave in two different situations with respect to one factor, while controlling for everything else. Controlling for everything else is tough, but in our case, we are interested the impact of the GST on GDP growth only.
More specifically, what we want to know is whether consumption GDP growth is weaker with GST than without, post-2Q15-1Q16 transition period. Or to put it simply, is the GST-drag on consumption GDP growth a one-off thing?
In the spirit of stylized facts, we want to determine whether it is Case 1 (which is bad):
Or Case 2 (which is okay):
Case 1 offers a much bigger output loss than Case 2. There is a Case 3 where output loss happened only during a few periods, but I do not think it is realistic since the GST is an ongoing concern.
YoY chart above suggests we are closer to Case 1.
Quarterly growth suggests case 2
But keeping in mind the issues about base effect for YoY method, maybe quarter-on-quarter calculation would be of a better help.
And QoQ suggests Case 2 is in play. There is a persistent negative effect on growth rates. From the Department of Statistics seasonally adjusted data, 2Q consumption grew only 0.7% QoQ. In 2011-2014, growth that quarter averaged close to 2% QoQ but in 2Q16, it was only 0.6%:
Indeed, QoQ growth rates since 2Q15 has been weak compared to previous years. QoQ in an way does suggest some kind of a growth slowdown.
How much output loss we suffered?
The easier question is whether the GST has adversely affected the GDP levels. It is easier because base effect is pretty much irrelevant to levels. The answer is, it has.
In the first chart, you can see the GST roughly took 1 to 2 percentage points off quarterly consumption GDP growth. That is equivalent to MYR47 billion output loss in real terms (2010 prices) in the 1Q15-2Q16 period. This includes the abnormal spending increase in 1Q15.
MYR47 billion sounds large. So is Najib’s billion ringgit donation. But to put the number is the proper context, the total size of real consumption GDP during that period was MYR843 billion. So, that is about 6% output loss in that period.
But I have not done the same diff-in-diff for other GDP components. I would speculate the overall impact is bigger than MYR47 billion. But it is hard to imagine it in my head since the expected impact is all over the place.
But I am certain the overall economy lost some economic output. That is probably Captain Obvious speaking.
Implications
If Case 1 is true, then the government has less room to mess around with its GST revenue and start encouraging investment to raise the GDP potential so that the loss is recovered as soon as possible.
