Another teachable moment in currency management

After spending $28 bn over two years propping up the tenge, late last week, Kazakhstan's central bank announced a regime shift from exchange rate targeting to inflation targeting. The decision, prompted by the Chinese and Russian devaluations and falling oil prices, resulted in a steep plunge in tenge's valuation by nearly a quarter in a single day. 

As the graphic shows, even by Kazakhstan's usual standards, with bouts of sharp devaluations, this one has been exceptionally steep.

Especially when compared to its oil exporting peers.

Oil dominates the country's export basket and contributes the lions share of the government's revenues, as seen from this 2013 export products graphic. 

For such countries, a floating exchange rate regime can be a useful automatic stabilizer. As commodity price falls and current account balance weakens, a gradually depreciating currency would serve to both keep exports competitive and squeeze imports, thereby restoring the external balance. Similar forces work in the reverse with rising commodity price and appreciating currency so as to moderate capital inflows during the commodity upturns. Further, this dynamic would also serve to promote manufacturing diversification and mitigate the "resource curse" effect during good times, which is characterized by a hollowing out of the non-commodity tradeables sector, especially manufacturing. Most importantly, it would leave the country with the monetary policy autonomy to combat the headwinds arising from external shocks. 

In this respect Kazakhstan is no different than many commodity exporting emerging market peers. In order to maximize revenues during good times, these countries play the foreign exchange markets to keep their currencies over-valued, thereby fueling asset price bubbles, only to be forced into depreciating sharply when the tide has turned and the central bank can no longer hold on. Such unpredictable reversals erode the country's macroeconomic policy credibility and markets punish it with violent swings. Inflation and economic stagnation, accompanied by banking crises, soon follow. Monetary policy discretion goes out of the window as the country is forced to hike rates to retain capital and rein in inflation. It is not devaluations per se that erode credibility and disrupts the markets, but sharp and unpredictable devaluations.

Nigeria, which has stubbornly kept its currency pegged to the US dollar, thereby suffering a steep drop in its oil revenues, looks most likely to soon burn out its reserves and be forced to let the over-valued naira depreciate steeply. 

An example of such macroeconomic prudence, with the attendant credibility and stability that anchors inflation expectations and limits cross-capital flows volatility, which enables countries to weather the storm, is Colombia, another country dependent on oil exports. The FT writes,

Under successive presidents, Colombia put together a framework that aimed for structural fiscal balance, instituted inflation targeting and achieved relatively broad-based growth... Over the past year, the economy’s exposure to the oil price (more than 50 per cent of exports) has resulted in the Colombian peso being one of the weakest freely-traded currencies in the world, by falling 36 per cent over the past 12 months. Like most EM currencies, it received another kick downwards from China’s renminbi devaluation last week. The rapid depreciation alongside falls in dollar-denominated oil prices means that in recent months, Colombia’s oil revenues in domestic currency terms have fallen relatively little, even compared with other oil economies. Economists at Barclays calculate that oil priced in Colombian pesos fell just 1.3 per cent from early May to late July, compared with falls of 6.2 per cent in Russian rouble terms and 14.4 per cent if priced in Nigerian naira...

(unlike its peers) the Colombian central bank has left monetary policy on hold since last August. Colombian consumer price inflation has recently risen to 4.5 per cent, just above the central bank’s 2-4 per cent target band, but inflationary expectations have remained well anchored... the IMF forecast that the economy would expand 3.4 per cent this year, below last year’s 4.6 per cent but far better than the recessions forecast in Russia and Brazil, and that inflation would drop back below target.

Kazakhstan may well have now embraced an inflation-targeting regime. But on the issue of which monetary policy regime best suits such countries, Jeffrey Frankel has written here about the possible superiority of NGDP targeting. Writing in the aftermath of a shift in the nominal anchor from US Dollar to a basket of currencies in September 2013, Prof Frankel had cautioned against any adoption of inflation targeting,

An example illustrates the point.  If a truly serious CPI target had been in place five years ago at the time of the global financial crisis, then Kazakhstan would have faced a difficult and unnecessary dilemma when it was hit by adverse shocks in oil prices, the housing sector, and the banking system. The country would have had either to forego the necessary February 2009 depreciation of the tenge or else to violate strongly the CPI target as the devaluation pushed up import prices.  The former choice would have been dangerous for the economy, while the latter choice would have largely defeated the purpose of having announced IT in the first place (that purpose being long-term monetary credibility).

Instead, he suggested an NGDP targeting nominal anchor for monetary policy as the best automatic stabilizer for oil exporters like Kazakhstan,

Kazakhstan is vulnerable to a variety of possible shocks, such as a fall in the world oil price, which would be better accommodated by a more flexible exchange rate regime... An alternative anchor for monetary policy, in place of either the dollar exchange rate or any version of the CPI, is nominal GDP... the innovation would in fact be better suited to middle-income commodity-exporting countries like Kazakhstan. The reason is that supply shocks and trade shocks are much larger in such countries. In the event of a fall in dollar oil prices, neither an exchange rate target nor a CPI target would let the tenge depreciate. An exchange rate target would not allow the depreciation by definition, while a CPI target would work against it because of the implications for import prices. In both cases sticking with the announced regime in the aftermath of an adverse trade shock would likely yield an excessively tight monetary policy. A nominal GDP target would allow accommodation of the adverse terms of trade shock: it would call for a monetary policy loose enough to depreciate the tenge against the dollar.

The case for NGDP Targeting examined

As I blogged earlier, Inflation Targeting (IT), which underpinned the monetary policy consensus since the nineties, looks set to be another casualty of the global financial crisis. Nominal Gross Domestic Product (NGDP) targeting has become the most discussed alternative to IT.

As the name suggests, NGDP targeting seeks to fix a trend nominal GDP growth rate as the nominal anchor for setting interest rates and other monetary policy actions. This target is more effective than a pure inflation target at boosting output, especially when the economy is facing the zero-lower bound (ZLB) in interest rate. As Simon Wren-Lewis explains, an NGDP target works by relaxing monetary policy tomorrow in order to raise tomorrow's output and inflation. Assuming rational expectations, this response in turn immediately raises inflation today (since today's inflation depends on expected inflation tomorrow) and therefore reduces real interest rates today, which in turn raises output today and again inflation today. A simple inflation target cannot generate this effect on output today or tomorrow when the economy is facing ZLB.

Further, its supporters claim that by directly targeting the level of output growth, it avoids getting entangled with the intermediate objective of inflation, and focuses on the ultimate objective of stable economic growth. Central Bankers too are loath to give up their hard-won inflation fighting credibility which has helped firmly anchor inflation expectations for nearly two decades. Also, as Scott Sumner points out, it is politically easier to mobilize support since it re-frames the debate around output and avoids the contentious topic of inflation.

At a time when the developed economies are facing deflation and liquidity trap, a generous dose of inflation can be helpful in generating growth. But popular and ideological opposition to the idea of stoking inflation, borne out of a generation of inflation targeting, comes in the way of any attempt to promote growth, even by generating inflation consistent with the defined inflation target. Furthermore, monetary policy, in particular expansionary policy, has come to be intimately associated with inflation. NGDP targeting would replace inflation with output as the nominal anchor.

This distinction is a bit of sophistry, but enough to get political traction for expansionary monetary policy. Fundamentally, the effect of monetary and fiscal policies get distributed between output and price level changes - expansion causes both growth and inflation, while contraction lowers both. In other words, inflation and growth are two sides of the same monetary policy coin, though their relative magnitudes is determined by the nature of supply-side shocks and cannot be influenced directly through monetary policy.

So my concerns with the alleged superiority of NGDP over IT is as follows

1. It may be possible that in general NGDP cycles till now have been more closely correlated with asset inflation cycles than inflation cycles have been with asset prices. But there is limited theoretical basis for claiming that business cycles correlate strongly with asset price cycles, any more so than inflation cycles. Or do we, as before, avoid addressing asset prices, and deal with them through micro- and macro-prudential regulations?

2. Given that an NGDP anchor is the sum of potential output and optimal inflation target, any volatility in potential output is likely to introduce uncertainties into an NGDP targeting framework. If the potential output is itself not stable, then NGDP targeting becomes inconsistent with IT.

3. Further, a reliable assessment of output gap is critical to setting an NGDP anchor. In conditions of zero-lower bound, accuracy of the output gap is less important since even if the central bank over-estimates its magnitude, it can very easily raise rates to correct the situation. But if it under-estimates the gap, there is no possibility of going down in the opposite direction with interest rates. But this does not hold once we are faced with different economic conditions, like an overheating economy facing inflationary pressures.

4. This brings us to the most important flaw with NGDP targeting. A simple NGDP target would reveal little about the distribution of NGDP between real output and inflation. Consider two scenarios. In the first, high inflation caused by a supply-constrained and over-heating economy keeps real output growth low and NGDP below the target. In the second, deflation co-exists with demand-constrained economic conditions and keeps NGDP below the target. The policy prescriptions for the two conditions are different. Expansionary monetary policy would be inflationary in the former while growth stimulating in the second case.

The NGDP level per se would not provide enough information to guide us on the right policy. In fact, in the former case, we can reach the target NGDP level by monetary expansion which would only raise inflation further without creating jobs or boosting real output. But in a deflationary demand-constrained economy, monetary expansion is more likely to increase NGDP by boosting real output, by skirting around the issue of inflation. Since this is exactly the problem facing developed economies today, it is natural that any policy which is likely to promote recovery attract attention.

Fortunately, we now have real world examples of these two conditions being played out simultaneously. Even as many developed economies are struggling with deflationary recessions, India is grappling with a supply-constrained inflationary economic slowdown. In fact, just as NGDP targeting, by skirting around inflation, offers a politically feasible cover for monetary expansion in developed economies, inflation targeting, by focusing directly on persistent high inflation, provides a politically convenient excuse for the Reserve Bank of India (RBI) in not relaxing its monetary tightening.

The contrast between the two conditions is striking and representative of the difficulty of having a uniform policy suitable for all conditions. One practical approach would be to stick with IT, but use nominal GDP as an anchor to help restore growth in economies facing deflationary recessions and the  ZLB in interest rate. In this context, the strategy suggested by Jeffrey Frankel for central banks to shape expectations by introducing a long-run NGDP target and then dynamic short-run targets till growth is restored, without junking the long-run inflation target, looks appropriate.

NGDP Targeting re-frames the Inflation Targeting debate

One of the biggest casualties of the global financial crisis may be the use of Inflation Targeting (IT) as the dominant monetary policy strategy. With IT having failed to avoid the crisis and now not being able to help economies out from the depths of deflation, governments and central banks look set to give IT a honorable burial. As the graphic below shows, prevailing inflation targets provide no guidance for monetary policy for countries seeking help to exit deflation.

Nominal Gross Domestic Product (NGDP) targeting is emerging as one of the strongest alternatives to IT as a monetary policy framework. As the name suggests, it projects the trend growth rate of nominal GDP as the monetary policy anchor. By subsuming inflation within the NGDP target, it avoids getting entangled in the inflation debate.

Scott Sumner has an excellent article where he lays out the defence of NGDP targeting. In making out the case for NGDP, Scott Sumner points to its greater popular and political acceptability, apart from its greater inherent effectiveness. In particular, he points to the difficulty with getting support for policies that explicitly seek out inflation in order to recover from a deep deflationary environment due to the entrenched belief that all inflation is bad. In this context, NGDP targeting provides a nice behavioral psychology sleight of hand by re-framing the debate in terms of raising nominal GDP and job creation instead of generating inflation.

Paradoxically, cognitive biases arising from our aversion to inflation and affinity for output growth, causes us to oppose expansionary policies which cause inflation while supporting those that promote growth, despite both outcomes being two sides of the same monetary policy coin (the effect of expansionary monetary and fiscal policies get distributed between inflation and growth, the relative magnitudes of each being dependent on supply-side factors). This is a classic example of how framing the terms of a debate can increase political acceptability of the same policy instrument. As Scott Sumner writes,

If we stopped talking about inflation targeting and started talking about NGDP targeting, we could greatly simplify the policy debate. Do we want more demand, or not? Most Americans surely think that more demand would be a good thing right now, but very few people want to see more inflation. To the Federal Reserve, these two effects are simply two sides of the same coin. But because the Fed expresses its aims in terms of inflation, its work is understood as a matter of managing inflation, and therefore Fed policies aimed at boosting inflation are politically problematic.

NGDP targeting therefore provides a cover for expansionary monetary policy, which has been stigmatized by its close association with inflation, to play its full role in growth by reducing the focus on inflation. It frames the terms of the debate as between growth and stagnation, not higher and lower inflation.