Exchange Rate Targeting: "Weak" and "Strong" are not equal

Etheorist has issued a follow-up of his thoughts in his original post outlining his support for a stronger Ringgit level. My critique of his ideas are contained in this post, where my position is essentially that a stronger MYR level is fine provided it is consistent with changes in the economy, but not otherwise.

I find myself having problems with his new post as well. My main contentions are two-fold: first with reference to the central bank "engineering" an appreciation of the exchange rate, which has monetary implications, and second the likely impact on the economy of an artificial appreciation of the exchange rate from the Balassa-Samuelson based model I described before, which has real economy implications. I also have some different interpretations of economic developments of the past decade, but since this post is going to be long and a bit technical, I will deal with those in a following post.

First, to deal with the problem of engineering a non-secular appreciation of the exchange rate. To understand the issues here, one must first have an understanding of how a central bank influences the exchange rate above or below the market determined rate. Let's first examine the case of weakening the currency, since everyone appears to believe the stylized fact that the MYR is below its long term equilibrium level.

To achieve a lower exchange rate requires the central bank to sell domestic currency for foreign currency. On the central bank's balance sheet, this appears as an increase in international reserves (+assets), in exchange for depositing money in the banking system's accounts with the central bank (+liabilities). This results in an expansion of the money supply, which if unsterilised could trigger an undesired credit expansion and demand-fed inflation, exacerbated by the lower short term nominal interest rates from the monetary injection.

Sterilisation refers to central bank open market operations designed to keep interest rates and money supply stable in conjunction with central bank forex transactions. It’s similar to normal liquidity management operations in the sense that the mechanism is the same.

Since in this example a monetary injection was the result of the forex transaction, the central bank must drain liquidity from the interbank market. This is done by selling the central bank's holdings of securities to the banking system, enough to fully or partially offset the monetary injection. Since the central bank in our example is interested in weakening the exchange rate of the domestic currency, unsterilised intervention is also an option at the risk of higher inflation, but also with the potential benefit of higher economic growth.

Thus the practical limits of weakening the exchange rate are: what constitutes an acceptable level of higher inflation; or if sterilisation is resorted to, the borrowing capacity of the central bank. At least that's the case if you don't want to create another Zimbabwe – the central bank’s credibility in keeping the soundness of the domestic currency is at stake.

Strengthening a currency’s exchange rate over its market-determined rate operates in the opposite fashion: the central bank buys domestic currency in return for foreign exchange. This appears as a drop in international reserves on the asset side of the central bank’s balance sheet, and as a drop in the banking system’s cash balances with the central bank. This is also ipso facto a contraction in the money supply, and pushes up short term interest rates.

An engineered appreciation of the exchange rate is thus inherently deflationary, both in terms of the domestic money supply as well as in terms of import prices. If this is undesirable, for instance when inflation and growth are already too low or negative, the central bank can again resort to full or partial sterilization via a purchase of securities from the banking system, thereby injecting funds into the banking system.

From the above it’s easy to see the limits of artificially boosting the exchange rate level beyond the market clearing level: it is limited by the amount of international reserves at the central bank, as well as the supply of available securities in the banking system. This further implies that a strengthening operation is unlikely to succeed for long without a fairly deep supply of public debt.

In both weakening and strengthening operations, if the degree of misalignment in the target exchange rate is far enough away from the fundamental equilibrium rate, the central bank must continually intervene to maintain its target parity which puts it up against the limits of intervention that much faster. Note also that increasing or decreasing the amount of foreign currency in the system does not have any impact on the banking system’s capacity to lend, since loans can only be made to residents in domestic currency.

A second implication is that by virtue of targeting the exchange rate, volatility of either or both money supply and interest rates will become an order of magnitude greater, because the central bank cannot target all three. This is actually easy to see from the pre- and post-2005 Malaysian experience, where interest rates were higher and spreads far wider pre-2005.

A third implication is that artificially weakening the exchange rate is far easier than strengthening it, because of the nature of the limits imposed by the mechanisms employed. In extremis, a central bank can print money to fund forex purchases, which is not an option the other way around – the level of international reserves is a hard limit.

Are there ways around these limits though? For strengthening the exchange rate, yes there is – you can partially or fully close the capital account. But since this entails a general withdrawal from global financial integration, it also means closing access to foreign investment (particularly portfolio) as well as turning our backs on development. I’ll return to the topic of intervention and sterilisation in a future blog post.

A second less painful way around this is to simultaneously engineer an increase in the equilibrium rate. That ensures that misalignments between the equilibrium rate and the nominal rate aren’t too far apart. What are the elements that need to be pushed to achieve this objective? This leads me back to the Balassa-Samuelson Hypothesis model I described before.

To summarise, an economy can be divided into two producing sectors – tradables and non-tradables. Domestic prices (adjusted for the exchange rate) in the tradable sector are everywhere equalized internationally, since demand and supply are determined on a global basis. Domestic prices in the non-tradable sector however are determined purely by local conditions. Labour and capital are the production inputs, where the former is immobile but the latter fully mobile across international borders.

The effect on exchange rates can be described as follows: say for instance that there is an exogenous expansion in the non-tradable sector. This bids up wages across the whole economy, which reduces the relative productivity in the tradable sector. There is thus upward pressure on prices in the tradable sector. However, since prices in the tradable sector are set internationally, the end result is ceterus paribus an appreciation in the real exchange rate of the domestic currency to equate the international prices of tradables. Note that this effect arises from structural changes in the domestic economy rather than changes in international relative prices.

Now that we have that in mind, what’s the impact of an artificially raised exchange rate without these structural changes? An appreciation in the exchange rate creates an identical reduction in relative productivity for the tradable sector as in my example above. However, since there has been no corresponding increase in total domestic demand within the economy, the result is a reduction in the total demand for labour thus putting downward pressure on wages and prices in the tradable sector. This in turn will tend to depreciate the exchange rate. The net result is a tendency to revert to the equilibrium exchange rate, but with higher unemployment.

The opposite is true of an artificially weakened exchange rate. You get an increase in relative productivity of the tradable sector, which has the effect of putting upward pressure on prices and wages in the tradable sector since total demand is also higher. This will have the effect of putting upward pressure on the exchange rate, with the net result a return towards the equilibrium rate but with higher employment.

In short, exchange rate appreciation above the equilibrium level, as a policy tool, won’t result in a rise in incomes but rather the reverse. If I can resort to that hoary old chestnut phrase, “correlation does not imply causality”. In this case, one must distinguish between productivity in the non-tradable sector which is not subject to international competition, and productivity in the tradable sector which is subject to international price pressures. An increase in the former results in a depreciating exchange rate, while for the latter the opposite is true.

More to come!