THERE IS NO ONE-TO-ONE correspondence between the rate of interest and the exchange rate. Still, more often than not the rate of interest is the only variable that can control the exchange rate. Thus, it is possible to defend the currency, both under a peg and under the float, up to a certain threshold beyond which the relationship is reversed. The main example is when the degree of fiscal dominance over monetary policy determines the threshold. Still there are many other situations when non-monotonic relationships lead to regime-switching, whereby the relationship between the exchange rate and the interest rate becomes non-linear. I stick to the synthesis provided by Lance Taylor some 15 years ago that exchange rate expectations and the conditioning interest rate – actually the difference between the U.S. and Turkish rates, but never mind – are the only variables of interest when it comes to equilibrium exchange rate determination.
AN EXAMPLE OF PERVERSE MONETARY POLICY EFFECT
A widely acclaimed conjecture in developmental macroeconomics maintains that a central bank-engineered rise in the real rate of interest will lead to a real currency appreciation and will make government debt attractive. We have indeed seen many instances in Turkey whereby this conjecture proved true. But we have also witnessed that, if real interest rates increase the probability of default, government debt may become less attractive – at the going rates of interest – and currency depreciation may ensue. The latter outcome is more likely the higher the initial debt stock, the higher the proportion of FX-denominated debt, and the higher the price of risk. If an interest rate spike raises the doubt of default, inflation targeting – be it implicit or explicit – under the float and via a Taylor rule is not immune to a perverse effect. Even an interest rate increase in response to higher inflation can lead to a real depreciation. Assuming structural change is incomplete, and therefore the pass-through is still high, a real depreciation automatically implies higher inflation – such as what has always been happening in Turkey.
The situation is symmetric. Just as interest rate decreases in response to lower inflation can lead to real currency appreciation, rate hikes could well imply real depreciation. We may call it a “vicious cycle”, generated by a perverse monetary policy effect. It is fiscal policy that makes all the difference, and in such a case, it may be advisable not to change the policy interest rate, but increase the primary surplus by cutting government expenditures more. It is the fiscal policy, not monetary policy, which should become the main defense line under this outcome. But there are other scenarios as well. Let’s look at the bond market.
BONDS AND ARBITRAGE
Now, consider U.S. Treasury bills and Turkish TRY – and also FX-denominated bonds. There are obviously functional forms that depict how capital flows to Turkey are determined. I will pick one. In such an equation, assuming international investors are risk-averse, there will be a term that represents the adjustment for risk on Turkish dollar-denominated bonds. There will also be a term that denotes the risk aversion of international investors and a variable that is a proxy for the variance (volatility) of returns. Thus, the higher the expected real return on Turkish dollar-denominated bonds or the lower the expected real return on U.S. Treasury bills, or the lower the risk on Turkish dollar-denominated bonds, the larger the (portfolio) capital inflows to Turkey. This is exactly, and in a very simple set-up, what we, along with many an analyst, routinely keep saying in the ordinary vernacular. The capital flow equation states that if the risk aversion coefficient of foreign investors is higher than the average risk aversion, then an increase in risk leads both to capital outflows and to a rise in the stated rate of interest on government debt, via the arbitrage equation. It is possible to show – at the risk of filling these lines with some mathematical notation – that along with an increasing probability of default, rising interest rates would go in tandem with currency depreciation. Also any risk that emanates from another source, i.e. political, etc. would be conducive to the same result.
ALL ANCHORS ARE LOST, ONE BY ONE
Let’s see the main parameters of Turkish political economics post-2001. Because every single factor that helped it become after 2001 a moderately growing economy (among the EM peer group, Turkey’s growth was about average even in the 2003-2010 eriod, total factor productivity improving well for a couple of years only, and garden variety inflation, i.e. single digit) have been reversed. Inflation fell to single digits. Obviously, the exchange rate played a big role there. Inflation could not have been reduced to low levels in the absence of an explicit or implicit nominal anchor, and this is what happened. We knew theoretically from the papers of Michael Kumhoff, a student of Alan Drazen, that should an exchange rate argument be explicitly recognized in the Taylor Rule, the resulting dynamics would be qualitatively similar to that of a pegged exchange rate regime. This is what happened between 2002 and 2008.
This is not the end of the story. Inflation fell as a result of a vector of factors; it was not caused by a single variable. Domestic demand was weak and, for other than necessary goods, food and suchlike, it became more price and income-elastic after the 2001 crisis. That affected pricing behavior, drove profit margins down – until 2005, mark-ups tended to fade and pricing power was lost. Firms were forced to experience a “learning process” which they have possibly undone. Productivity increased, though most of it came from lay-offs at the beginning but TFP also rose. Unit labor costs dropped, since real wages and employment both fell after 2001 and stayed there. Adding the productivity effect and the euro/dollar parity change into account, the ‘overvaluation’ of the local currency exporters lamented about back in 2004-2006 wasn’t as significant as it appeared to be. That was a good equilibrium. But it didn’t last.
The IMF had helped hugely also. Front-loaded advances and postponement of back payments put aside, the signaling effect of the IMF support as a full commitment device to the Turkish cause facilitated reatly our sovereign bond issuance in the early 2000s. True, emerging bonds in general had performed admirably throughout and spread compression was a pervasive emerging market phenomenon, but Turkish Eurobonds did particularly well at that time. Then things began to worsen after 2010.
WHERE HAS ALL THE ‘PRUDENCE’ GONE?
From the political economics standpoint, there seemed to be a process of reconstructing, reshaping of (market and political) power, redistribution of assets and income and politically structural change. This is why both the U.S. and the EU applauded the performance of the first decade after 2001. All of the above pluses are now minuses. Everything has been reversed in toto. All post-2001 gains have been lost. The first and foremost omen that pointed to change in the wrong direction was the overreliance on construction.
POLICY REVERSALS DON’T HELP EITHER
They definitely don’t but it is no longer a matter of short-run expediency. It is a matter of long-run strategy. The change is now long overdue. Structural reform is to be taken very seriously. Therein lies the future. And it provides food for thought. What kind of equilibrium does the thing called structural transformation purport to lead to? Equilibration is a process, it takes time in the real calendar, and it is learned and has surely to be social. Communication needs to be concrete and explicit and virtual communication won’t wash. Certainly, sociology and psychology should not be banished from the constitution of the equilibrium solution of the structural reform game. Structural reform can be a panacea for many variables evolving in opposite directions, but it nonetheless signals that the whole parameter space has to shift. Change should be specified, well-elaborated, and the use of such blanket terms shouldn’t cause boilerplate hyperbole to crowd out analysis.