Vivek Moorthy: Time to take a 'random walk'

The RBI can declare an annual target and intervene only if the rupee moves outside a moderate range.

In a recent article, former Chief Economic Adviser Shankar Acharya has staunchly defended the policy of forex intervention to prevent a rising rupee, as India has been intermittently doing since 1995, informally based on a real exchange rate target. To justify this policy he lists a few "textbook propositions" whose applicability for India he denounces as rubbish (August 9). His strong dismissals of these textbook propositions are generally on weak grounds; space constraints prevent elaboration here.

Nevertheless, one vital statement "The exchange rate is the single most important (relative) price in the economic domain" warrants rebuttal. Wrong. The exchange rate is a crucial relative price, but so is the price of oil and the price of wheat. For a large, domestic demand-centred economy, there is no justification to pursue an exchange rate target at the expense of a price (i.e. inflation, say CPI) target.

Overall his article provides a good point of departure for evaluating desirable and feasible combinations of rupee policy now. To begin with, he has outlined the data which show that the rupee rise this summer was mainly due to relaxing the cap on external commercial borrowings (ECBs). Of the $ 22 billion rise in the capital account surplus in 2006-07, about two-thirds ($13 billion) was from ECBs. I am very much in sync with him and many others on this, and had argued for curbs on inflows and easing of some outflows way back in early 1997.

Where Acharya goes astray, in my opinion, is in defending a (real) exchange target. Curbs on inflows are useful to reduce volatility, and to prevent a huge rupee rise that will hurt exports. But such curbs can be combined with a flexible exchange rate which, under current circumstances, is desirable in my opinion. With stiff curbs on capital inflows, it is unlikely that a flexible exchange rate regime can hurt exports much. Far from being mutually exclusive, capital controls and a flexible rupee are highly compatible policies. The reader of his article might have come away thinking or concluding otherwise.

In contrast to his advocacy of pegging, the Economic Advisory Council under Dr Rangarajan has made careful and balanced recommendations in July 2007. The EAC has recommended a combination of (i) curbs on debt (but not equity inflows) (ii) allowing some rupee rise and (iii) sterilizing inflow of forex, resulting from intervention, by hiking the cash reserve ratio. It should be mentioned that Dr Rangarajan, our greatest academic and policy economist, had recommended curbs on debt inflows as part of the recommendations of the 1993 BOP Committee which he headed.

Where the EAC falls short, in my opinion, is in not tackling a crucial policy choice: should forex intervention be fully discetionary, or transparently conducted. If the latter, should it be done as and when the RBI decides to, or should it be based on an explicitly announced target or exchange rate band?

Due to the explosive growth in global and Indian financial markets and capital inflows this is not a minor matter. We cannot muddle along as usual. In my opinion, it is dangerous for the RBI to buy and sell dollars at its discretion and semi-secretly (due to the roughly two-month delay in reporting intervention) as at present. The current policy might suddenly lead to a huge loss in reserves to Soros-type speculators and/or damage to the economy in trying to maintain a particular exchange rate.

There are grounds for believing that systematic and transparent intervention at specific levels is preferrable to secretive ad hoc intervention. An explicitly announced REER target (what the RBI has implicilty been doing for years) is one such transparent policy. Unfortunately it is the wrong one since it conflicts with the more fundamental goal of inflation control. And in the long run a central bank cannot determine the real exchange rate any more than it can determine the real interest rate or real GDP growth. All that it can determine in the long run is a nominal variable "" the price level or exchange rate as a final goal.

All across the world, it is becoming clear that economic fundamentals favour rising currencies in emerging economies, as they grow richer and their productivity rises. In the Indian press, Ruchir Sharma provided valuable evidence that accordingly policy is changing for many countries (Economic Times, June 5). The market meltdown and related repatriation by FIIs, hedge funds and others weakened the rupee (week ending August 18), but over time the rupee will tend to be under pressure to rise due to 'rising' (although not zooming, to be realistic) India.

The Internet has made the world flatter. India's previously non-tradable services, starting with English speaking voices are increasingly being exported. The 1993-1994 goods based REER index and/or weak rupee policy that Rajwade, Bhalla, Acharya and others are defending is badly outdated. At the other extreme, a pure float under high capital convertibiity (not even 'fuller', let alone full) leads to sharp swings in the real exchange rate, causing huge dislocations to exports, and lives of those working in export industries.

Both these extremes can be avoided by limiting inflows and intervening to limit volatility and nothing else. Even equity inflows will need to be limited. It is evident from August's gyrations that some of the Yen carry trade is not a pure uncovered interest arbitrage debt play. Rather hedge funds and other punters have borrowed in Yen and invested in, among other assets, emerging market equities.

How can a policy of intervening only to limit volatility be operationalised? I have suggested that the RBI can declare an annual target and intervene only if the rupee moves outside a moderate range, say between plus 5 per cent and minus 8 per cent, allowing for our 3 per cent higher inflation. Once a year the target should be adjusted to some average rate that prevailed in the previous year. In academic parlance, this is a 'random walk band.'

With fairly stiff curbs on inflows, the demand and supply of rupee forex will come mainly from the current account (i.e exports and imports), which does not gyrate hugely. As a result, it should be easy for the RBI to manage the target band. And it will benefit the economy to allow the exchange rate to respond mainly to current account pressures.

Free market purists would say that all target zones are bound to fail "" either firmly fix or freely float, but avoid the precarious 'unstable middle'. There are umpteen instances of exchange rate tunnels, snakes and zones failing, as Milton Friedman, the most vocal proponent of a pure float, never failed to point out. From this perspective, the 'crawling peg' most often ends up being a falling peg!

However, in retrospect, most such failures occurred because either there were huge capital flows, or the central bank was not independent enough to ensure low inflation in sync with the weighted inflation rate of the target zone partners. The polar extremes of freely float or firmly fix are better than the 'pragmatic middle' only if the modern way of ensuring macroeconomic discipline "" an independent central bank "" is not politically feasible. Further, a 'random walk' target, unlike a simple crawling peg based only on inflation differences allows the real exchange rate to vary in response to fundamentals.

If the RBI can deliver 4-5 per cent inflation steadily as a final goal, then a 5 per cent annually adjusted random walk real exchange rate target band, under moderate to stiff capital controls, is a consistent, feasible and desirable policy.

vivekmoorthy@iimb.ernet.in