REER as rupee's fair value signal is hazy

During the brief respite from the sharply rising volatility in most emerging market (EM) exchange rates (including the rupee), key policy decisions on exchange rate management — as those on Friday — must be based upon a sense of a “fair value” of the currency. The recent sharp depreciation of the rupee has re-ignited the debate about a fair value of the currency. The metric widely used for this is the Real Effective Exchange Rate (REER). The REER, while probably the most parsimonious, is a very ambiguous measure of fair value, and needs to be interpreted with due caution. The RBI’s 2017 Annual Report has an excellent analytical overview on aspects of REER methodology.

The first problem in interpreting the REERs is that the indices are based only on the basket of merchandise trade, services are not included. The share of global services trade relative to merchandise is rising, from 21 per cent of merchandise exports in 2011 to 58 per cent in 2017 (although this is partially due to the fall in commodities prices). The IMF has a set of REER indices, which incorporate travel (but not all) services, available for some countries (not India). For countries with significant services exports (like India), the exclusion of services might render a REER problematic.

Second, the use of the consumer price index (CPI) as a deflator for the weighted trade basket in the REER is a problem, since the CPI weights a large number of non-traded commodities and services, whose price dynamics might be very different from traded items.

Third, as a corollary of the second, the effects of productivity changes might be very different for traded and non-traded baskets of goods and services. The former, in India’s case, are likely be more “price takers”, while the latter will have (at least) a labour cost component less influenced by global prices. OECD has a set of country REERs derived from unit labour cost (ULC, a proxy for productivity). The contrast from a comparison of the Rupee’s REER (computed by OECD) using both the standard trade weights (like RBI’s) and ULC could not be more startling. The former rises from 100 in 1994 to 136 in December ’16; the latter falls to 43 in this period.

Reflecting these features, the chart summarises the consequences of different construction methodologies on the path of the Rupee’s REER, comparing the RBI Index with one developed by the Bank for International Settlements (BIS). Based on the chart, the RBI Rupee as a basket looks at first sight to be overvalued, having risen for the most part since 2000, and about 36 per cent up as of April 2018. This raises another important issue in interpreting the extent of overvaluation, namely, the reference time point used for comparison. Which year should be considered as the base, presumably being a “normal” or “neutral” year?

The BIS REER index moves, in comparison, are more damped, with periodic reversions to mean, and the overvaluation, if any, is moderate, about 5 per cent as of April 2018. This discrepancy arises from the differences between the RBI and BIS REER methodologies, as follows.

Based on the above, the fundamental problem is the choice of weights for the individual currencies in the basket. RBI weights for individual currencies are based on the share of bilateral merchandise trade (for example, between India and the US) whereas BIS uses “double weights” incorporating the exchange rates of countries which also trade with the bilateral pair (for example, say Vietnam and Sri Lanka’s trade with the US). This accounts for both bilateral trade and third market competition. The simple intuition of the underlying “elasticities of substitution” is that if one country’s currency depreciation is similar to those of its trade competitors, there is no comparative advantage in exports. The BIS trade weights are also time varying versus RBI periodic re-indexing.

An advantage of the BIS indices is their availability for most countries using a uniform methodology, allowing cross country comparisons of REER changes a BIS REER indexes gives a feel for relative exchange rate valuations. The first notable is that generally since early 2013, the group of advanced economy (AE) currencies has progressively become undervalued and EMs, overvalued (note that the reference date for this is January 2010). The US was the not-able exception.

A weakness in the BIS methodology (vs an “ideal” EER) is that, inter alia, adjustments for modern value chains are not considered; exports of country X to destination Y might be components for an onward destination Z (with a large value add) with very different economic (and currency) characteristics. So the substance of trade might largely be between countries X and Z, while the weights would not reflect this. The rising value addition in third world countries in the global trade chain potentially increases the distortion of valuations signals embedded in the REER weights.

Returning to the IMF indexes, while the inclusion of travel services doesn’t seem to create wide differences in the REER trends relative to BIS, the metrics are not comparable since IMF does not incorporate the BIS “double weighting”.

This article only touches upon the sources of ambiguity in interpreting REER. An informed and data driven debate on estimating a currency’s valuation should now follow.
The writer is senior vice-president, business and economic research, Axis Bank. Tanay Dalal contributed to the article. Views are personal.

Business Standard is now on Telegram.
For insightful reports and views on business, markets, politics and other issues, subscribe to our official Telegram channel