# Growth in a dollarized economy: a conundrum?

One of the indispensable elements an “analytical narrative” on the Turkish economy must definitely include is the full story of asset (currency) substitution. Since currency substitution is similar to inflation and is largely inertial, it would and should have been impossible to expect a “reversal of fortune” in short notice unless drastic shocks were to be fed in. Nonetheless, for the time being such an element seems to have been fed in: Currency substitution is reversing its course, which is bad indeed. Back in early 2004, currency substitution had reversed course, and that was one of the key factors determining the success of the Derviş-IMF program of 2001. Had it not reversed course, neither inflation nor TRY interest rates could have fallen between 2004 and 2008, and the AKP possibly wouldn’t have won the 2004, 2007, 2009 and 2011 elections in the first place. Clearly, the economic program of Kemal Derviş as such might not have offered one such beneficial shock. The outside world also helped a lot, due to the famous’ savings glut’ and all that. Now, currency substitution, in fact asset and liability substitution, could be back with a vengeance. If it is, we will truly be back to 2001, if not in terms of the depth of recession in terms of the behavioral pattern of residents and non-resi

dents alike, financially-speaking. The last thing we need is an acceleration of dollarization that reaches a point of no return. If it does, then it will take a lot of time to change that. Because currency (asset) substitution is inertial, it feeds into habits and develops long-memory dependence.

Beyond the Hurst coefficient

That the currency substitution phenomenon is subject to inertia can be conceptualized through the Hurst exponent. Macroeconomists of the Massachusetts variety had resorted to path-dependence or hysteresis in order to explain European unemployment – or Eurosclerosis - in the late 1980s. Benoit Mandelbrot had already pioneered this approach in the early 1960s, both in economics and in finance. Mandelbrot wasn’t truly well received in those disciplines and fame came to him through other channels. But what he wrote between 1960 and 1972 has had nonetheless some impact and was echoed by others afterwards. It would be unfair – to both Mandelbrot and to the economics profession - to claim that he had no effect whatsoever.

Technically, the story goes like this: The Hurst exponent, H, is defined as H=log(R/S)/log(N) where N is the duration of the sample of data, and R/S the corresponding value of rescaled range. In this way, Hurst generalized an equation valid for the Brownian motion in order to include a broader class of time series. In fact, the distance R covered by a particle undergoing random collisions is directly proportional to the squareroot of time N: R=k*N0.5 where k is a constant which depends on the time-series. The generalization proposed by Hurst was: R/S=k*NH where H is the Hurst exponent. If H=0.5, the behavior of the time-series is similar to a random walk; if H<0.5, the time-series covers less “distance” than a random walk (i.e., if the time-series increases, it is more probable that then it will decrease, and vice-versa); if H>0.5, the time-series covers more “distance” than a random walk (if the time-series increases, it is more probable that it will continue to increase). Given a time series x(n), n=1....N, H can be estimated by taking the slope of (R/S) plotted vs. n in a log-log scale. In order to see if there is hysteresis in the series, we may compute the Hurst coefficient. It is embedded in many statistical packages as a widely used technique. Turning the currency substitution into an index beginning in 1986 and processing the rate of change of the index by means of the Hurst procedure, we obtain 0.69 in 2003 Q2. This is the peak level around the 2001 crisis. All in all, path-dependence had increased before the crisis because data between 1996 and 2001 pointed to a very clear uptrend. Now, Hurst is going up again.

Memory-dependence is key

This suggests that asset (currency) substitution is not random and that once reversed it may follow the exact opposite path, building renewed inertia as the new path continues to unfold. That poses an interesting question: Is it true that the ongoing and forthcoming current account deficit reduction will suffice to generate a sufficient movement downward in the path of the nominal exchange rate? From the behavior of both firms

and households, it seems not. It all hangs in a balance and the name of the beginning of a solution is credibility. However, this would only be the beginning. Just as it was the case in late 2003/early 2004, even the existence of a consistent and credible program run under the auspices of the IMF and carrying the name of such a reputable economist as Kemal Dervis didn’t do trick at once. It took more than two years for currency substitution to reverse course, and the shock ingredient was a massive capital inflow from the outside world. The other side of the coin is: Should the bandwagon effect reverse sign once more, would it be possible to get reverse asset (currency) substitution back any time soon? After all, we really want dollarization to come to a halt, right? Or do we need another two years so things get back to their older path, even in the presence of a credible commitment?

Could a higher real interest rate help?

Nonetheless, this is not the whole story, as a quick guided tour of historical time series reveals. In a Bayesian VAR model with the standard Litterman priors, we observe that in whatever way we constrain the future path of key variables, the leading role of the exchange rate is kept constant. We also see that if we don’t constrain the series, the nominal exchange rate increases, growth turns negative whereas the current account balances. This is pretty much what we are experiencing today. If we keep the future expected interest rate low enough, then the exchange rate increases, resulting in high inflation and negative growth each and every time. This is a practical certainty. The stress on the exchange rate is inherited the next year and, consequently, currency substitution goes on and on, with growth turning negative at each jump and the current account deficit, if it ever widens, beginning to shrink sizably. Admittedly, we should take the Euro-USD parity into account, but replacing the USD with the Euro or by a combination of the two will not change the effect of the exchange rate. Note that the situation isn’t symmetric. That is, if the interest rate is kept high, let’s say 200 basis points higher as the effective funding CBRT rate goes, it may not suffice to reverse currency substitution. It may only help keep it under control, in other words, endure that it doesn’t jump-start. However, after 0.6 Hurst, it is difficult to render substitution tame. We may be at a critical juncture because both asset and debt substitution, including government debt, have been edging up for some years and substitution is already high enough to generate inertia.

Can a re-dollarized economy grow at par?

This is indeed a tricky question. On this, I offer two observations. The first goes like this: Let’s try to capture dollarization by means of GARCH class models instead of simply using the Hurst coefficient as a landmark. What does this tell us? First, in the beginning of the substitution process, exchange rate volatility matters as well as inertia. Second, as time goes by and substitution takes root, volatility effects wither away, and in the end, there is only inertia left. Third, people don’t consider future conditional variances (volatility) much, but mostly the current volatilities. There is kind of myopic attitude here. Hence, a timely positive shock killing current exchange rate volatility can have an immediate impact. One shouldn’t allow time to pass because this kind of dollarization behavior takes time to form but it accelerates rapidly after a threshold. The second is a more structural observation. Is there any way to accurately forecast the potential (trend) growth rate in a highly dollarized economy? It is definitely down now, possibly to below 4 percent. The upshot is, should substitution endure, how would we compute an ever-decreasing potential? As such, the remarks claiming ‘technical recession is over’ are meaningless because we risk losing the very metric against which growth can be rendered measurable.

Will there be growth?

Not in a meaningful way, although on paper some figures will appear due to the business cycle frequency. Nevertheless, we are beyond that point now. That quarter-on-quarter seasonally- and calendar-adjusted GDP turned positive is meaningless because of three reasons. Primo, it is due to very high public expenditures ahead of elections, and can’t be repeated. Because there is no investment, we can only gather that these expenditures served in lieu of transfers. Secondo, inventory depletion has been running its course for the last three quarters, and we don’t know what will happen next. Terzo, it is because imports collapsed that net exports (external demand) helps GDP to keep its head a bit. It isn’t export-driven growth; it is in fact only a statistical artefact of domestic demand collapsing.

Last but not least, dollarization is ongoing. As such, I would rather expect negative head-on GDP growth in Q2, rather than the positive sign quarter-on-quarter adjusted datum suggests. Of course, the enigmatic behavior of inventory accumulation may have a bearing therein, but in which direction I can hardly guess. Note that in 2002, in the aftermath of the crisis, inventory behavior was so problematic that GDP data were put under scrutiny. No, there won’t be any meaningful growth in whatever sense GDP growth refers to something important.

The more fundamental question remains: Can we grow again despite dollarization – and fast? Can we even ‘measure’ the economy’s potential as long as both debts and assets continue to be dollarized?