Lies, Damn Lies, and Then There Are Statistics

UBS Securities Asia Ltd has issued a report that has generated a bit of a buzz in the online media titled "Malaysia - Another Bizarre Story" (try here, alternate here). The headlines in the online media however are less timid:

“Money leaving Malaysia in massive amounts and bizarre fashion”

“Malaysia's Disastrous Capital Flight”

Ye Gods.

I’ve covered the issue of capital flight before in this post (read the comments for an interesting discussion). There’s no doubt that capital has been leaving the country in the aftermath of 1997-98, which in my opinion and from anecdotal evidence is largely due to domestic investment abroad. However, this UBS report is highly misleading as it purports to show that large amounts of capital left the country in 2009, which isn’t the case at all - at least, not any worse than it usually has.

Here’s my view (point by point, in blue) of where this report went wrong.

“Question: which Asian country had the biggest FX reserve losses in 2009? The answer is Malaysia, and by a very wide margin; we estimate that official reserves fell by well more than one-quarter on a valuation-adjusted basis.”

This is the main evidence that the report uses to justify its conclusion. But note that the basis of this estimate is from peak 2008 levels against current levels, which is more than a little disingenuous (look at notes for the first chart in the report). The assessment of “more than a quarter” loss suffers from what’s called the “base” effect – if you use a period with a high denominator, you can get negative growth even if the actual levels are increasing (international reserves, RM millions):

I put in a longer sample period so it’s easier to see what’s going on. Using a more conventional year on year calculation, this is what you get (log annual changes):

See what I mean? For most of 2009 growth in reserves was negative, despite the fact that the actual level of reserves was largely flat to increasing. Malaysia suffered technical deflation for much of 2009 for much the same reason, but I don’t hear anyone crowing over consumer prices falling.


Here's what monthly growth looked like (log monthly changes):

Through the whole of 2009, we've had exactly three months where reserve growth was negative, while the rest of the year saw reserves rising. This type of problem is the reason why I've consistently advocated looking at levels rather than growth metrics during times of violent change. You simply miss a lot of what's going on by using a period-to-period percentage change approach.


I’m also a little suspicious of the claim that the estimates were valuation adjusted – BNM already books revaluation losses and gains every quarter. If you revalue again based on published statistics, that would constitute double counting and would exacerbate movements in reserves.

“Other structural surplus neighbors like China, Hong Kong, Singapore, Taiwan and Thailand have all seen sizeable increases in FX reserves over the past 12 months … and yet Malaysian reserves nearly collapsed.”

Sloppy analysis – reserves at end 2008 stood at RM317.5 billion, versus RM331.3 billion at end 2009, including a revaluation loss of over RM10 billion for 2009 as a whole. That means a net increase of RM13.8 billion over a 12 month period – sure doesn’t look like a collapse to me.

More important, the first three countries listed all have an explicit exchange rate target. China has effectively halted Yuan appreciation against the USD in the last year and a half, Hong Kong has a currency board arrangement with the USD which effectively means forex reserves back the money supply, and Singapore explicitly uses the exchange rate as the instrument for implementing monetary policy unlike the more conventional interest rate target as used in Malaysia. I don’t know off-hand how Taiwan manages monetary policy (Taiwan is not an IMF member, and aren’t categorized within the IMF’s exchange regime framework), but Thailand uses a managed float with an inflation target, also implying intervention in foreign exchange markets.

Why this is important is that the countries listed all accumulate or lose reserves due to implicit or explicit exchange rate targets. Malaysia does not fall under that category under either a de facto or de jure basis, and I’ve gone on record to state that I don’t think BNM has an exchange rate target for the Ringgit at all. Hence, if there is no intervention, there should be effectively little to no change in reserves – which is what happened in 2009.

So what’s the story for 2008, because there was a massive loss in reserves back then? Looking at the reserve levels, you see a mild run-up from 2005 to 2006, and then acceleration from end-2006 to about mid-2008. Not coincidentally, reserve accumulation happened simultaneously with the commodity bubble of those years:

That's the main difference between Malaysia and the other countries we were compared to in the report - we are a commodity exporting country, none of the others are. Given that exports of commodities didn’t change much in volume but export receipts did, we had an abnormal influx of income that was more nominal than real. In the aftermath of the collapse in prices post July 2008, we had a drop in income that was also more nominal than real. Hence, from BNM’s perspective, the appreciation of the exchange rate (as well as the money supply impact of the inflow of trade receipts) warranted intervention to mitigate the Ringgit’s (and the money supply's) rise during the boom, and limit the Ringgit’s (sharp) fall (and the potential contraction in the money supply) in its aftermath. In addition, BNM had to meet the sudden demand for USD from the banking system as depositors and investors pulled out:

If you look at newspaper reports at the time (4Q 2008), BNM as good as admitted buying MYR against USD to support the currency:

Now if these factors are taken away – receipts from a commodity bubble boosting forex supply within the banking system and forcing an overshooting appreciation of the exchange rate, we would not have had reserve accumulation in 2007-2008 in the first place, and reserve accumulation would simply be on the same trend as it was from 2005-2006. And we wouldn't be arguing about a loss of reserves we probably shouldn't have had in the first place.

3.       “And this despite a massive, unprecedented decline in high-powered “base” money, as shown in Chart 4. Indeed, over the past 12 months Malaysia recorded one of the biggest base money contractions in the entire EM world, matched only by the Baltic states (Chart 5). This is in part because the Malaysian central bank responded with a sharp drop in reserve requirements to keep banks liquid … but still, we can’t help but note that the domestic financial system seems uniquely unaffected by apparent capital outflows.”

This one’s a bit funny – the writer was obviously not referring to M1 (currency + demand deposits) (RM millions; and log annual changes):

He’s referring to base money which is a bit different – currency + bank reserve deposits. This is a rather funny metric to use, because under a modern regulatory system, bank reserve deposits are barely relevant. As long as the statutory reserve ratio is below the risk-weighted capital ratio (8% under the original Basle requirements, somewhat more nebulous under Basle II), then the money multiplier is effectively limited by the capital ratio and not the reserve ratio.


Hence BNM’s cut in the reserve ratio was a token and not an effective policy change, unlike the situation in China for instance where the reserve ratio is one of the primary instruments for managing the supply of credit (because it's double the capital ratio). BNM hasn’t seriously used the reserve ratio as a policy instrument since before the 1997 crisis.

What’s even funnier is the description of the cut as “sharp” – it was a 50% cut, which sounds big until you realize that it was from 1% to 0.5%. And the banks have more or less ignored the reserve requirement anyway – the banking system has been flush with cash for years, and they haven’t bothered to lend out the excess (banking system deposits with BNM in RM millions; loan-deposit ratio):

Hence there should be no surprise that interest rates have trended lower in 2009 – there hasn’t actually been a large outflow of capital (and hence a contraction in the money supply), there hasn’t been a massive loss of reserves, and...there isn’t really a story here except maybe UBS wanting to sell something.


(H/T Hafiz Noor Shams - free plug for you, mate!)