Thursday, October 17, 2013

Forex Fallacy

[This post is loooong. For those without patience, you can skip to the end without missing too much]

Tong Kooi Ong talks forex policy and promptly makes a meal of it:

How the middle class is subsidizing the Corporate Elites and why it has to stop

There is a feeling that Malaysia’s middle class are generally not a happy lot. Many moan about the rising cost of living, education and healthcare, their relatively low wages and rising debt as they borrow more to buy homes and cars…

…A major problem lies in the weak ringgit, which results in different purchasing power for the two “middle classes”…

…In US dollar terms, Malaysian salaries are 22% of the level in the US, 16% of Australia and 24% of Singapore. It is true that the cost of living in these countries is much higher, and one should not compare apples with oranges. But if we compare purchasing power based on McDonald’s Big Mac index, which is widely used, we are still very much worse off…

…The weak ringgit is clearly a major problem for the urban middle class as local salaries can’t keep up with an increasingly US dollar-denominated and international cost base…

…The main reason was actually our own industrialisation policy, which was started by former prime minister Tun Dr Mahathir Mohamad…

…However, there were also unintended consequences of that policy on the ringgit.

To make Malaysia more export competitive, the ringgit had to be kept low. With export sales denominated in US dollars and costs in ringgit, a lower ringgit helped to make products cheaper in US dollar terms and, at the same time, boosted the profits of Malaysian corporates and industries.

This was further assisted by a policy that kept interest rates low and reduced the funding costs for big businesses to grow through debt.

In fact, real interest rates were very low and negative in some years, even though Malaysia’s Consumer Price Index understates real inflation for urban dwellers. The ringgit has also been on a depreciating trend since 1981, even before the big decline during the 1997/98 Asian financial crisis (see Charts 2 and 7).

The weakening ringgit was also a consequence of the fact that many of these new industries were not globally competitive. In other words, we had import costs from a weak ringgit but no export gains for some of our huge industrial projects…

…Today, however, all our cheap labour consist of foreigners. FDI is falling. The old low-cost manufacturing hub model no longer works, and indeed it now creates major social costs and problems.

A new strategic policy is needed to bring the country to the next stage.

It is time to have a policy of positive real interest rates and to promote a gradual appreciation of the ringgit.

Corporates and industries must be forced to become more efficient and make products that are more globally competitive, and capital must be more efficiently utilised. We can learn from countries that have successfully moved out of the middle-income trap.

There are so many things wrong with this “analysis” that it’s hard to figure out where to start. It begins with a false premise, compounds it with flawed data analysis, and concludes with a Frankenstein proposal. My critique is so long, I’m going to have to split this post into three parts.

A False Premise

Is the Ringgit weak? Figuring things out like that is harder than it looks. Talking about this always reminds me of Nigel Tufnell in that infamous “11” scene in Spinal Tap:

Exchange rates are relative prices, not absolute ones. Turning the knob up to “11” doesn’t necessarily make the amp louder. That’s fundamental to the amp itself, and the numbers are just for reference.

Looking at things purely from the lens of the USD exchange rate tells almost nothing about where the equilibrium rate is or should be. You can have a currency depreciating against the USD, but progressively getting overvalued over time. You could also have a currency appreciating against the USD, but getting more competitive at the same time.

Couching things purely in terms of a currency’s USD value also takes the unspoken assumption that the USD has some fundamental absolute value, which is obviously not true. In truth, the purchasing power of any given currency can and does fluctuate. In modern times, this is typically one way i.e. a declining purchasing power through inflation. Not even gold, which has traditionally been seen as inflation resistant, is a really good store of value.

Two charts to illustrate all these points:


The chart above shows the USDMYR rate (USD per 1 Ringgit) from 1990-1998. Between 1990 and 1997, the Ringgit tended to trade – or to be more precise, was constrained – between 2.50 and 2.70 Ringgit per USD. No obvious changes here, except for after the Asian Financial Crisis when it crashed and was finally pegged to the USD at 3.80. This fools a lot of people into thinking that RM2.50-2.70 is some magical number that defines the Ringgit’s “true” value.

Here’s the flip side:


Again, the exchange rate between 1990-1998, but this time between the Yuan and the Ringgit. China devalued the Yuan by a third on Jan 1, 1994. At a stroke, the Ringgit went from competitive to seriously overvalued overnight. Can you tell that from the USD chart? No, you can’t.

There’s actually other ways to look at whether a someone has a thumb on the scales of exchange rates. The first, as the article notes, is to look at the real interest rate, but this is actually wrong. What matters is not the real interest rate, but the real interest rate differential.

Since I’ve already covered that recently, I won’t get any further into that except to note that a negative real interest rate isn’t exchange rate negative, if everyone else has negative real interest rates at the same time. The converse is true – you should not expect a currency appreciation with a positive real interest rate, if everyone else is also offering positive real interest rates.

Second is to look at international reserves. If a central bank is supressing exchange rate appreciation directly, international reserves should be increasing:


Looking at the chart, there are three clear periods of forex intervention. The first in the early 1990s, the second during the period of the USD peg and immediately after, and the third in 2010. The first and third coincide with heavy inflows of portfolio capital, while the second was in relation to maintaining the USD peg. Only in that middle period could it be said that BNM was actively involved in supressing exchange rate movements.

Contrast the above with this:


Here, except for the middle period between 1997 and 2002, reserve accumulation was near constant. If an exchange rate is deliberately prevented from appreciating, central bank intervention needs to be more or less continuous, as the factors that underlie the appreciation need to be constantly offset by purchasing foreign currency and selling domestic currency. Outside the USD peg period, BNM’s intervention has been episodic, not continuous. For the curious, that second chart is Singapore’s MAS.

Last, you might want to pay attention to the current account surplus/deficit. In fact, this is how most economists would define currency weakness or strength. The equilibrium exchange rate should be the point where the current account is exactly in balance – that’s one of the major criteria used by the IMF for example:


On that basis, the Ringgit appears undervalued. The problem is that the current account approach doesn’t factor in global supply chains.

One thing the article gets seriously wrong is the claim that an undervalued exchange rate makes local industries more competitive. But that’s only true to the extent that costs are wholly local currency denominated, and that exports and imports are completely independent of each other, which is something commonly assumed in basic Econ 101.

Trouble is, for Malaysia and for many East Asian economies trade doesn’t actually behave that way. In a disaggregated supply chain, the impact of exchange rate movements are muted – a weaker exchange rate makes your domestic input more competitive, but makes your imported inputs more expensive. And vice versa – a stronger exchange rate makes domestic inputs less competitive, but makes imported inputs much cheaper.

The end result is that current account surpluses and deficits greatly exaggerate the extent of currency under- or over-valuation. 40% of China’s electronic exports is imported content; the figure is about a third for cars “made” in Germany. Even commodity exports are not free of imported inputs – Malaysian palm oil for instance depends a great deal on imported fertiliser (about a quarter of costs).

Bottom line: there’s little evidence that since the unpegging of the Ringgit from the USD in 2005, there was a deliberate policy choice to maintain a weak Ringgit exchange rate. Neither is there much evidence for the period before the Asian Financial Crisis.

Flawed Data Analysis

The Big Mac index is popular. That doesn’t mean it’s correct. While it largely does capture differences in purchasing power, that actually has surprisingly little to say about equilibrium exchange rates. I’ve written about the Big Mac index before and I’ve little to add to what I wrote 4 years ago apart from saying that it no more than proves that the Penn Effect is alive and real – higher income economies have higher price levels, and thus stronger exchange rates.

To prove it, just plot prices against GDP per capita – in fact that’s just what the Economist magazine has done in it’s latest iteration of the index. Even the new adjusted index they use shows the same pattern.

Moreover, such simple approaches tend to ignore other fundamental factors that affect exchange rates. The article correctly notes the depreciation of the Ringgit in the 1980s, and attributes this to the “Look East” policy and the government’s policy of industrialisation which supposedly requires a weak exchange rate policy. But I see three other factors that should be considered, and these were largely echoed by Jomo in his presentation last month at UM.

First was the Volcker moment, when Fed chair Paul Volcker raised short term interest rates in 1981 to double digit levels in a successful attempt to tame inflation (the real interest rate differential that year with Malaysia averaged a staggering 8.8%).

Incidentally, this plunged the US into its deepest recession between the Great Depression of the 1930s and the recent Great Recession of the 2000s. The interest rate differential alone was enough to explain the beginnings of the Ringgit’s depreciation during that period.

The combination of a high USD and a sharp US recession also turned a decade-long boom in commodity prices into a two decade-long bear market. Inflation adjusted commodity prices have yet to regain their peaks from 1980-81, despite a runup in prices over the 2000s. Weaker commodity prices result in lower foreign exchange earnings for any given level of exports, which puts downward pressure on the exchange rate.

Third, the Plaza Accord of 1985 saw a coordinated attempt by major central banks to sell down the USD and buy up JPY, which seriously compounded the problem. CPO prices fell 80% between the end of 1984 to mid-1986. Average crude oil prices lost almost 2/3rds during the same period. Crashing commodity prices and a depreciating USD combined to further lower export income in MYR terms, which obviously put even more downward pressure on the USDMYR exchange rate. Again, something I’ve covered before recently in more detail.

Take these three factors, and you don’t need policy induced weakness or a lack of competitiveness argument to explain the lower USDMYR exchange rate. Given that commodity prices have yet to recover to 1980 levels, it’s not hard to see why USDMYR hasn’t gone back to RM2.70 per USD either.

But beyond that, the whole narrative that industrialisation requires low real interest rates is at odds with policy imperatives and the canonical approach to industrialisation – the general rule is precisely the opposite.

Industrialisation from an agrarian economy requires importation of fixed capital, which would be cheaper if the exchange rate is stronger. High real interest rates are then required to maintain the exchange rate, and to redirect savings from other projects to those the government prefers. In turn, these policy moves result in a current account deficit. If you look at the pattern of most developing countries over time, that’s what generally happens.

The East Asian economies, Malaysia among them, were unusual in having built up a large pool of domestic savings which could be used to finance investment. Low real interest rates (to be more precise, real interest rate differentials) were thus a consequence of policy and circumstance, not a policy goal. Although it should be said that Malaysia in the pre-AFC period was also characterised more by CA deficits than surpluses, partly from higher imports but also from volatility of commodity export income.

A Frankenstein Policy

If you want to see what the impact is of a strong currency policy, look up Britain, 1925, gold standard. Then look up higher unemployment, general strike, and uncompetitive exports. JK Galbraith called it, “The 1925 return to gold standard was perhaps the most decisively damaging action involving money in modern times.” Actually we don’t even have to go back that far – Malaysia circa 1994-1996 is a pretty close model too.

The evidence suggests that far from forcing industries to become more competitive, strong currencies cull out weak companies and replace them with nothing else. The productive capacity either shuts down or simply shifts offshore. That’s what happened to Japan in the wake of the 1985 Plaza Accord, incidentally much to Malaysia’s benefit. The US does not have an exchange rate policy per se, but the Dollar’s “exorbitant privilege” as the world’s major reserve currency means that it is generally overvalued, with predictable consequences for US industry.

The UK suffered a two-year recession in 1990-1992, because the Pound was linked to the Deutschemark, and the cost of German reunification prompted the Bundesbank to tighten monetary policy, a stance that was transmitted to the rest of Europe via the fixed peg European Exchange Rate mechanism (ERM), whether it suited them or not. The GBP became overvalued relative to the DMK. Relief only came when GBP was unceremoniously bundled out of the ERM in 1992.

Does an overvalued currency improve your purchasing power? Yes, but only for the lucky sods who keep their jobs.

Monetary policy considerations also apply – the policy trilemma rears its ugly head. If, as implied by the article, an exchange rate level or pace of appreciation should be targeted, that means you must either forego capital mobility or an independent monetary policy.

If you choose an independent monetary policy (i.e. the ability to choose interest rates, again as implied in the article), then you MUST have strict capital controls. If on the other hand you choose free flow of capital, then you MUST give up monetary independence. In essence, if the target is the USD exchange rate, your monetary policy will be run by Ben Bernanke, or Janet Yellen if and when she takes over.

At this stage, somebody’s bound to bring up Singapore. But Singapore is in fact a perfect case study for all I’ve talked about here. Singapore uses an exchange rate target as its primary monetary policy instrument – the slope and width of SGD appreciation against the currencies of its trade partners. In other words, MAS runs a policy of deliberately appreciating the SGD against everybody else.

But look at the details and the consequences. MAS is constantly increasing its international reserves; Singapore has an enormous public debt because they use government securities to lock up excess liquidity i.e. reserve accumulation through sale of SGD is simultaneously sterilised. So what actually happens in practice is that MAS is constantly slowing the natural appreciation of the SGD. Real interest rate differentials between Singapore and the US are constantly negative, not positive.

Singapore also allows free flow of capital, which means domestic liquidity and interest rates are inherently tied to foreign exchange liquidity.

In sum, Ben Bernanke sets interest rates for Singapore.

And there’s another, more important lesson from this. If you want a strong currency and better purchasing power, it’s better and more sustainable to work at the underlying factors behind currency appreciation rather than trying to artificially boost it through policy means. At heart, this means paying attention to the “internal” real exchange rate i.e. the exchange rate between tradables and nontradables. There are no short-cuts here.

We want “10” to be louder. Trying to make it louder by changing the number on the dial to “11” doesn’t work.


  1. Thanks for this, Hisham. A long read, but I always suspected there was something wrong with Tong's article.

  2. Very good and detail work. You obviously remember our monetary history, which many people unfortunately never bother to read, or simply forgets.

    Weakening the Ringgit argument is indeed a Frankenstein claim made by Tong. When Malaysia unpeg the Ringgit, BNM and the whole country wanted it to appreciate, not depreciate, for fear of being labelled as not so successful (in graduating from the Asian Crisis of 98).
    Until today, the whole country is still in the 'apreciate' mood, and as of now, BNM seems to be on the neutral side, which is very good.
    You've clearly spent a good deal of time researching data and articulate them in articles, kudos...

    1. Sharif,

      Thank you for the kind words, though I don't know if I deserve it.

      My background is monetary economics, and my Masters thesis was actually on this subject (exchange rate valuation), so I am already very familiar with both the theory and empirical data relating to the Ringgit.