Tuesday, November 24, 2009

3Q GDP: Good But Not Great

The headline numbers in last Friday's GDP report show seemingly good progress: -1.2% yoy, compared with a consensus estimate of -2.0%. It's even better than the forecast of 1.7% I did a couple of weeks back:

Looking at the levels, it certainly seems like Malaysia is on pace for a "V" shaped recovery:

Still there are some signs that all is not going smoothly. We have of course the big boost (finally) from excess government spending - not (seasonally adjusted levels and log annualised quarterly changes):

It sure doesn't look like third quarter government expenditure has seen any higher growth than it normally does historically. And export growth is er, iffy while private consumption growth has been sustained - except it's slowed. The upshot of it all is that GDP growth has also slowed, suggesting the recovery is already beginning to lose some steam (log quarterly annualised changes, seasonally adjusted):

To be fair, it was always unlikely for output to fully recover in a couple of quarters. And given the external situation, exports were never going to be of any real support to growth. More positively, the inventory changes number also came in negative, which suggests continuing spare capacity and the potential for a sustained boost in production to rebuild stocks - this recovery still has some legs then.

So this is probably the best to be expected, though I could have wished for better.

Monday, November 23, 2009

The China Factor

If as in my last post, fiscal stimulus is contraindicated as a way to full recovery for Malaysia, then we need to look at external demand.

The Asian Development Bank has just published research that looks into the impact of China on the rest of the region (Warning, pdf link). Specifically, the research asks the question: Can China pull East Asia out of recession on its own:

"Developing Asia has traditionally relied on exports to the United States (US) and other industrialized countries for demand and growth. As a result, the collapse of exports to the US and other industrialized countries during the global financial and economic crisis has sharply curtailed gross domestic product (GDP) growth across the region. The emergence of the People’s Republic of China (PRC) as a globally influential economic force is fueling hopes that it can supplement the US as an additional source of demand and growth. The central objective of this paper is to use vector autoregression (VAR) models to empirically investigate whether exports to the PRC have a significant and positive effect on the GDP of nine developing Asian countries. The study’s results from a three-variable VAR model indicate that PRC’s imports have a significant positive effect on the GDP of regional countries. However, the study’s results from a four-variable VAR model indicate that the PRC’s apparently positive impact reflects the US’ demand for Asian goods, rather than independent demand from the PRC. Therefore, overall, the study’s evidence suggests that the PRC is not yet an engine of growth for the rest of the region."

The answer is taken as a unit, China does exercise considerable influence over the economies of the region. However, once you add external factors, then we are still looking at final demand from the US underlying even China's import demand.

So looking at current events, I would say that China's stimulus spending has helped support the region during the past year, but full recovery (if it will occur at all) depends on recovery in US consumer demand. With global rebalancing in full swing, that's by no means a given.

I suspect what we'll see going forward is a permanently lower path of global growth, which I referred to here. In other words, there will be no full closure of the global output gap but rather destruction of over-capacity instead, which suggests a slow, hesitant recovery path.

On a side note: as with all VAR studies, we're looking at historical data and the framework used is agnostic of structure. Nothing says that a secular shift towards consumer consumption in China won't change future interrelationships. A global crisis like the one we're experiencing is a natural structural break.

Technical Notes:
D. Park and K. Shin, "The People's Republic of China as an Engine of Growth for Developing Asia?: Evidence from Vector Autoregression Models", ADB Economics Working Paper Series No. 175

Friday, November 20, 2009

Talking Stimulus; Or Why Fiscal Deficits Might Fail To Boost Economies

This article on VoxEU caused a stir in the economics blogosphere went it first appeared. I’d originally intended to post on this much earlier, but for a number of reasons I’ve sat on it for a while. What’s interesting is that the underlying research attempts to measure the impact of government fiscal deficits on GDP – that amorphous quantity known as the fiscal multiplier.

There are two contending views on the fiscal multiplier – Keynesians of course contend that the multiplier is positive and above unity i.e. for every dollar spent by the government, GDP expands by more than a dollar. The opposite viewpoint, typically associated with monetarists and Neo-classicals, is that the multiplier is below unity, and at the extreme can actually be negative. Empirical investigations into the question have yielded a range of estimates from zero to well over 2, which means the question is far from settled.

The paper explains why there’s such a diversity of estimates – fiscal policy is equally diverse in its application, the structure of the economy matters, monetary policy reaction matters, as does the position of the economy relative to full employment output. The methodology at arriving at the estimates also matters enormously (check the paper’s critical appraisal of the Obama government’s estimates). So the real answer is: it depends.

Here are some of the key issues (not limited to the discussion in the paper):

1. The output gap – the greater the gap between the current level of output and full employment output, the greater the likely impact of fiscal spending. This is the prototypical Keynesian argument, and one Neo-classicals have a hard time with. Why? Because under the assumption of complete and fully clearing markets, by construction unemployment cannot exist. You then get howlers like, “people must be demanding more leisure time,” in an attempt to explain the drop in employment.

2. Crowding out – the effectiveness of fiscal spending depends on its use of available resources. The Keynesian argument is that since output is below capacity, idle resources can be used by the government to boost economic activity back to full employment level. However if resources (both real and financial) are not idle, then government spending partially or fully replaces private sector spending, and the multiplier must be less than one and at the limit, zero.

3. Central bank independence – the more optimistic estimates of fiscal multipliers generally have one thing in common, which is that they assume no change in monetary policy. But deficit spending requires raising government debt, which in turn pushes up interest rates. How a central bank would react to that depends on how independent it is and the outlook for inflation. If inflation is low or there is deflation, monetary expansion is called for which would tend to exaggerate the effect of fiscal expenditure. But over the medium term to long term, as output picks up, inflation should rise and monetary tightening will reduce the stimulus impact. Over time then, the multiplier should therefore be declining. Under full employment of course, independent central banks would act to offset the inflationary impact of increased government spending, and it’s more than possible that under such circumstances the fiscal multiplier can actually be negative.

4. Trade openness – another key assumption of the Keynesian argument is that government spending falls mainly on local production of goods and services. If it doesn’t, then the impact of government expenditure is actually lost to the economy, as it “leaks” to other countries. This was explicitly tested in the paper and found to be a significant factor – more “open” economies tended to have lower multipliers.

5. Structural fiscal spending – policy choices matter. Tax cuts and rebates will have different effects on the economy then construction projects, although as a rule, direct spending will have the higher multiplier.

In addition, the paper finds that multipliers in developing countries as well as highly indebted countries vare lower than in developed countries, but higher for countries with fixed exchange rates than for countries with flexible exchange rates.

Do these estimates apply to Malaysia? I’ve off and on been monitoring whether any research has been conducted on the fiscal multiplier(s) for Malaysia – believe it or not, I have yet to find any empirical research into this question. If anybody knows of any, please let me know.

The authors of the paper unfortunately don’t list the countries included in their sample, although it’s a mix of 45 developed and undeveloped economies (it’s an unbalanced panel VAR regression – I presume Malaysia is included). But even so you can’t directly take their estimates as applying to Malaysia, as panel studies lose country specific information. It’s quite possible (probable) that you get unbiased statistically significant aggregate estimates, yet have one or more countries in the sample with wildly different experiences.

But looking at the list of factors above (as well as the estimates from the paper), it’s hard to be confident that fiscal stimulus would have much impact on the Malaysian economy. While I accept the basic Keynesian premise that in the presence of an output gap and idle resources (little or no crowding out effect) expanded fiscal expenditure can be desirable, the mitigating factors mean that any impact will leak from the economy resulting in a low multiplier.

Here’s why I think so: first is the relative openness of the economy. The authors of the paper use a threshold of total trade exceeding 60% of GDP to define open and non-open economies; the ratio for Malaysia exceeds 200%. Also, the evidence on the Balassa-Samuelson effect on Malaysia is inconclusive, where studies usually take government consumption as a proxy for spending on non-tradables. What that means to me is that there’s a high import element in government spending.

Think about the structure of the March 2009 stimulus package for instance. Only about RM20b+ can properly be called stimulus spending, out of which most are construction projects. The reason for this policy choice is that construction typically (from previous studies in developed countries) has a high multiplier effect. But in Malaysia’s case, construction labour is mainly foreign and a lot of raw materials are sourced overseas. While we’d get a long term benefit from the infrastructure created, the short to medium term impact on GDP is already half-lost at the outset. Given the import content of some of our domestically produced and consumed goods as well, I’m not sure that tax cuts or rebates would be any better.

Second, Malaysia is definitely not a high income economy, so if we take the estimates from the paper, the impact is indistinguishable from zero. And third, the Ringgit is relatively flexible (well I think so anyway), which should also means a low multiplier.

So why bother with deficit spending? Under the current circumstances I can think of only one, albeit very good, reason. While taken individually, Malaysia and its trade partners would be “open” economies, in aggregate the world is a “closed” economy. That means aggregate excess fiscal spending will have an impact somewhere, even if the benefits do not fall wholly on the country spending the money. There’s no doubt that China’s massive stimulus spending over the past year helped pull East Asia very quickly out of recession. Malaysia’s fiscal spending therefore is in effect our contribution to reflating the global economy, not just Malaysia’s alone.

The risk here is we beget a “free rider” problem – if only one country is doing the spending, they bear the cost while gaining less of the benefit. Fiscal rectitude on the part of their trade partners under those circumstances would amount to being bad global citizens. Also pertinent is that the US stimulus spending could have the paradoxical effect of exacerbating global trade and capital imbalances that helped start this whole mess in the first place. Hence the calls for coordinated international action in stimulus spending earlier this year; and the talk about the need for coordination when withdrawing stimulus spending, which should presumably happen later next year. So under the circumstances, where we have had a synchronous global downturn, there are solid grounds for fiscal deficit spending.

Just let’s not get the politicians too comfortable spending taxpayers’ money when we get out of this.

Ethan Ilzetzki, Enrique G. Mendoza & Carlos A. Vegh, “How big are fiscal multipliers? New evidence from new data”. Link to full paper here.

Thursday, November 12, 2009

State By State GDP Data: About Time

The Department of Statistics has released state-by-state GDP data for 2005-2006, with 2007 and 2008 data to be released soon.

Two years isn't nearly enough data to look at trends or to do any kind of substantial analysis, but there are some nice nuggets in there. Check out especially Jadual 3 (warning: PDF link), which lists per capita GDP by state. Even though Selangor is the largest in terms of contribution to national GDP, it's only third in terms of income per capita, and on par with Sarawak, Labuan and Negeri Sembilan(!). So much for being developed.

September Industrial Production: Down But Not Out

September's IPI report shows production falling in most sectors, even on a seasonally adjusted basis - no biggie, since we're still looking at a Hari Raya effect here (log monthly changes):

I'll only admit to being worried if the numbers don't turn up strongly next month.

Here's an interesting question to ponder while we wait: can we use the IPI to provide an advance forecast for current quarter GDP? The Q3 numbers will be coming out very soon, so this is an exercise with a quick payoff, and since we're interested in a forecast rather than in the structural relationships there's no need to do a full scale model. The following specification works pretty well:

log(rgdp) = constant + D2 + D3 + D4 + log(ipi) + AR(1)

The IPI figure used here is the quarterly average, the D* terms are seasonal dummy variables, while the AR(1) term models the sample error as a 1 period autoregressive scheme. Sample period is 2001:1 to 2008:4:

The R2 is a suspiciously high 99.7%, which really warrants some cointegration testing, especially with a DW stat close to 2. But since I'm not interested in structure I'm ignoring those for now, especially since all the diagnostics check out ok. The full sample forecast against realization shows a fairly close relationship:

...but note that the estimated forecast is overstating GDP over the last three quarters. There lies the weakness of most forecast models when you're trying to foresee trouble: because of the techniques used, models like this will show a rosier picture than the reality during a downturn, and be more pessimistic when the economy is in a bubble. In any case, I'm only interested in the 3Q2009 forecast, so using a 1-step forecast yields:

Point Forecast:133572.56, Interval forecast: 135834.37-131310.76

Which is about -1.7% growth for the point forecast and 0.0% to -3.3% growth for the interval forecast (95% confidence level), using a year-on-year comparison. For quarter-on-quarter annualised, the equivalent numbers are 22.7% (!!!!), and 31.3% to 14.6%.

Technical Notes:
1. All data from DOS

Monday, November 9, 2009

Big Mac Index Yet Again

Someone else responded to Mr Gunasegaram's article and thinks a "strong" currency is not a good idea:

"First of all, I believe this is a crazy and short-sighted idea. It likely leads to a recession. When our currency is strong, everything we import seems to be at a discount and this encourages people to spend more on foreign goods and services through imports.

Strong currency also discourages our usual foreign trade partners to buy goods from us, thus export value will drop.

It also causes a country to lose its competitive advantage in attracting foreign investments.

At a time when inflow of money is significantly less than outflow, a serious current account deficit occurs in a broader sense. Malaysia will then be in deep trouble, after a short period of delusive wealth."


"If the ringgit were to be strengthen to a level of one for one US dollar, I would say good luck to everyone in Malaysia, welcome to the “burgernomics club”.

We will be staring with mouths watering at the Big Mac burger with an attractive price of RM3.54 in Malaysia just to find that our wallet is empty. And conveniently, if you want some money, IMF out there is willingly to lend you plenty.

There is no free lunch. Permanent and sustainable wealth does not come from currency rate delusion."

He's a bit less polite than I was though.

Ye Gods, Not Again: The Big Mac Index And The Ringgit

What is it about The Star and a "weak" Ringgit?

I thought I'd done with the Ringgit for a while, but P Gunasegaram repeats that tired old mantra that Purchasing Power Parity is a valid theory, again quoting The Economist magazine's Big Mac Index. I've covered the weaknesses of the BMI here and here, and I'm not about to go into that again except to note that PPP as a guide to currency value is at best a hypothesis, not a theory. A theory requires that the underlying hypothesis is supported by empirical data, which conspicuously is lacking for PPP.

Mr Gunasegaram also repeats that other tired mantra that an increase in reserves indicates a de facto attempt to devalue the currency. Again, this is not necessarily true as it completely ignores the potential impact of trade and capital flows on the domestic monetary aggregates (and thus inflation), which would necessitate central bank intervention. Since this intervention is functionally equivalent to currency intervention, what looks like an attempt to weaken the currency is actually something quite different.

We're also ignoring recent history here. Here's the log annual change in reserves and the MYRUSD exchange rate (where a negative change indicates appreciation) since 2000:

Two things to note here - first during the fixed rate period, reserve accumulation was volatile. If PPP was in fact operative and the Ringgit is substantially undervalued against the USD then reserve accumulation according to Mr Gunasegaram's hypothesis should have been consistently positive - it was not.

Second during the floating rate period, reserve accumulation had the opposite sign to what his hypothesis suggests. In other words, reserves increased when the Ringgit appreciated, and decreased when the Ringgit depreciated. That's a rather big hole in the "weak" currency meme, but is consistent with BNM's stated policy of ameliorating currency volatility (and not aiming for a target level) when necessary.

If BNM was intervening to create a "weak" Ringgit, then there would be no call to intervene when the Ringgit started weakening last year - which in fact actually happened. And the intervention conducted was more a factor of demand for USD (a response to domestic monetary conditions), rather than to achieve some target rate for the exchange rate.

You can find my extended critique of the "weak" currency meme here and especially here.

On another note, while a "strong" currency policy would indeed have the effect of increasing de facto individual incomes, there's an underlying assumption that nothing else will change. I roughly calculated trade elasticities with respect to the exchange rate in an earlier post - a 1% appreciation in the exchange rate resulted in approximately a 1.55% to 1.57% drop in exports.

Think about that for a minute - if my estimates are correct (and to be fair I'm not all that certain), since the drop in exports exceeds the rise in the exchange rate that implies a drop in export volumes. Since volumes will drop and there is no endogenous change in the supply for labour in the tradables sector, you'll get your higher income at the cost of higher unemployment (and since we're talking about low-cost production, also higher income inequality).

Let's just say I don't think pushing for a "strong", as opposed to a fairly valued exchange rate, is a good policy.

Technical Notes:
1. Data for reserves and MYRUSD exchange rate from BNM's Monthly Statistical Bulletin.

Friday, November 6, 2009

September Trade: I'm Not Worried - Yet

It's an unfortunate fact that most point forecasts are nowhere close to actual realization, a truism that applies to practically any endeavour not just economics (stocks, weather, sports, you name it). Last month's trade numbers are making me look like a genius though - which means I'm probably due for a run of bad luck!

I've already commented on the impact that Ramadhan would have on trade, especially with nearly a full week of production stops due to Eid il Fitri. It's no surprise then that both exports and imports appear to have regressed (log annual and monthly changes):

But the export number was presaged by the drop in imports in August - the point forecasts for my two models were just RM300 million off the actual (the 95% confidence interval was over RM10-13 billion), which is about as close to perfect as you can get. Since both models are driven by imports as predictors of export performance, no surprise that they caught the potential drop in exports.

Seasonally adjusted month-on-month growth moreover was essentially flat (log monthly changes, seasonally adjusted):

I'm taking this development positively, given the underlying cause. What I'm hoping and expecting for now is that my models underperform the next couple of months - we're coming up to the end-of-year shopping season in developed countries, and there should be a bounce in output and exports in October and November.

I'm discounting the fact that imports were flat in September - capital goods imports accounted for most of the poor showing, while there was a marginal increase in intermediate goods imports. October exports should therefore range closer to the upper bounds of the interval forecasts.

Next month's predictions:

Seasonally Adjusted Model

Point forecast:RM43,133m, Range forecast:RM48,479m-RM37,788m

Seasonal Effect Model

Point forecast:RM46,735m, Range forecast:RM53,251m-RM40,219m

Technical Notes
1. September Trade data from Matrade

Thursday, November 5, 2009

September 2009 Monetary Policy Update

Banks have begun to raise their lending rates on home loans and other financing – is this justified? From a consumer perspective this is a blow against disposable income, albeit a rather small one (for the moment). But from the banks’ perspective, the rate hike has been probably overdue. The Base Lending Rate (BLR) is supposedly the rate at which banks would lend to their best customers, and essentially provides enough of a margin over their cost of funds (COF) to cover loan defaults, overhead, and reserve requirements. But stiff competition in the banking sector has rendered BLR irrelevant as the benchmark lending rate:

Average lending rates have been consistently below BLR since 2004, and below 3% above overnight money since early 2007:

With prospects of economic recovery now clearer and loan demand sustained, there’s a feeling that interest margins have been overly compressed and banks are mispricing default risks. I have some sympathy for this view – when it comes to loan supply, it’s probably a little better to err on the higher side for pricing. If there is one lesson that we’ve learned from this past crisis, it’s that it’s all too easy to misjudge risk in the financial sector, more so since we have an environment of very low loan defaults and high domestic liquidity. But for those reasons, I don’t expect too much in terms of rate hikes in the next couple of months though – really about 20-30bp, 50bp on the outside.

Speaking of liquidity, there’s been slight movement on the monetary front (log annual changes):

I expect monetary growth to fall back in October-November, but to pick up later in December.

There’s not a whole lot of movement on the interest rate front either. BNM has kept the OPR at 2.00%, and the government only borrowed in September to redeem RM4 billion in MGS that had come due. MGS yields as a result stayed pat:

RM2 bilion in Khazanah bond redemptions also partially offset a year high RM4.5 billion in PDS issuance, so bond supply only marginally expanded.

The only big movement on the monetary front has been the Ringgit, but that's largely a US dollar story and not a Ringgit story, so I'll leave that for another blog post.

Tuesday, November 3, 2009

What The KLCI Says About Economic Policy

Usually it doesn't.

Teoh Kok Lin makes that point today (emphasis mine):

"BUDGET 2010 was, shall we say, not that warmly received by the man in the street. Some pointed out that the budget gave few goodies to the rakyat or businesses compared with previous years, hence it was not as good and that’s why the stock market was down 0.5% last Monday.

I believe a slight disappointment with the budget played a small role, if any, in the stock market. However, I am perplexed how one can infer from “one-day” stock market movements whether the budget is good or not."

To be more rigorous about it, here are the stats on the log difference in daily closing for the KL Composite Index (sample: January 2000-September 2009):

The two stats of interest are the mean (0.0001) and the standard deviation (0.0096). The observations show that daily changes are approximately normally distributed somewhat more leptokurtic than a normal distribution (from the Jarque-Bera stat "peakier" with a kurtosis stat > 3), with a mean of zero:

Using a 95% confidence interval and the standard deviation of 0.0096, we get:

(1.96 x 0.0096) x 100 = 1.88%

Which means that any movement plus/minus 1.88% is just the normal daily trading range for the KLCI, and doesn't say anything at all about the budget, RPGT, credit card taxes, fuel prices, commodity prices, the US dollar, Tun Mahathir's comments, corruption, politics, or Samy Vellu's hair (or lack thereof).

Edited (changes in blue): The actual distribution of changes has a greater central tendency than a normal distribution. Thanks Hafiz!

Monday, November 2, 2009

Credit Cards: Much Ado About Nothing

Reading through the Sunday Star yesterday, I'm puzzled by the amount of vitriol attracted by the new credit card levy, with a RM50 tax on each principal card, and RM25 on each supplementary card.

The ostensible idea was to encourage "prudent spending", by effectively increasing the costs of access, which had dropped due to the proliferation of "free for life" cards. I'm not sure this is the right policy, or to be more precise, the most effective policy to implement based on the stated goal. On the other hand, a more effective policy might involve heavy handed intervention in the workings of the financial sector that would contravene the spirit of the past ten years of financial liberalization. It’s certainly generated a lot of angst among card-holders, but judging from the public response, the tax appears to be effective in achieving the government’s goal of reducing card “collections”.

But before going into a discussion of right or wrong, it's probably best to proceed from a basis of firm facts - what is the current situation in Malaysia now?

With the explosion of consumer banking that began after the 1997-98 crisis, credit cards and mortgages have been a key battleground for the banking industry. The number of cards have grown exponentially (in millions):

...as have outstanding balances:

This has been coupled with a fairly steady drop in bad debts, although you should note that the default rate is far from typical of other bank business lines (which average under 2%):

On the face of it, this look likes a fairly attractive market for the banks. From the consumer point of view, it doesn't look too bad either. On per card basis, it looks like consumers are acting relatively responsibly:

Balances per card are dropping, as are transactions. The percentage of balances not paid-off at the end of the reporting period is dropping as well even as credit limits are rising, indicating we're in no danger of turning into another Korea just yet.

What complicates matters is that number of cards ≠ number of cardholders. How would recasting the numbers on the basis of population look like? More specifically, based on labour force numbers, we see a very different picture - the ratio of the number of cards to the work force has risen from about 1 in 5 in 1998 to near parity in 2008:

Rollover balances have tripled in the last ten years, indicating card debt is increasing as a percentage of income, and transactions velocity has increased from about once a month to twice a month. The latter is not necessarily bad if balances were stable, but they are in fact increasing which is a little worrying.

So there’s a solid basis for the government’s concerns over easy access to high-cost personal credit. Is a flat tax the optimum effective solution to reduce card usage? Not hardly – ideally there would be some (fairly high) minimum threshold of income to qualify, and perhaps a graduated tax based on credit limits similar to the annual commitment fees banks charge businesses for revolving credit.

I’d also consider putting in a minimum interest rate floor, not just an interest rate ceiling as currently practiced. We are after all talking about unsecured revolving credit where even with bad debts at an historical low is seeing defaults average in the mid-teens (comparable to junk-bond default levels). Some banks are offering under 10% p.a. interest rate charges - I'm not sure if this is not mispricing the financial risk involved.

But such measures are a little harder to enforce on individuals – if income is the barometer, what about the self-employed or those who earn on a commission basis? If access to this segment is curtailed for what is arguably a legitimate portion of the working public, then hard rules based on income are inequitable – and we have enough inequity in this country already, thank you. It’s also more than possible to set limits on the number of cards any one person can have (through CCRIS), though this may involve some loss of privacy. On second thoughts, maybe not such a good idea.

On that basis, a flat tax is a second best solution that despite the fact that it is not yet in place, already appears to be working by all accounts - people are already talking about cutting up their excess, unused cards. I feel the tax measure was really put in place to replace the annual fees typically charged for credit cards, but too often waived by banks in the interests of gaining and retaining customers. Maybe if banks had committed to charging annual fees we might not be having this very public debate.

Technical Notes:
1. Credit card data from BNM's Monthly Statistical Bulletin.
2. Population data from DOS and EPU