Lunes 22 de junio de 2009

Is Latvia the new Argentina?

Autor: Eduardo Levy Yeyati

The strategy of engineering an “internal” depreciation under a peg in Latvia (via contractionary fiscal policy, wage cuts and price deflation) implicit in the IMF program (IMF 2009) is proving too painful, if not self-defeating as in the 2001 collapse of Argentina’s currency board (De La Torre et al 2003). While the economy (and the fiscal accounts) worsened more than expected, nominal depreciation in most trading partners led the Lat real effective exchange rate to appreciate 6% in the first quarter of 2009.

Figure 1. The agonising dynamics of the “internal” devaluation strategy…

As a result of so much ineffective economic agony, the political consensus seems to be crumbling. The Swedish Riksbank, by arguing that the spillover on Swedish banks would be the same regardless of whether the devaluation is internal or nominal, is recognising the latter as a real possibility, whereas the IMF, which went along reluctantly under European pressure, is suggesting early euro adoption to place the burden of crisis resolution back where it belongs – the EU.

However, a disorderly abandonment of the peg would raise questions about the currency board arrangements in the region (Lithuania, Estonia, Bulgaria), bring the issue of Eastern exposure of Western European banks back into the focus, and test currency stability in Eastern Europe. Given that, the EU commitment to support an orderly exit is high, as are Latvia’s incentives to cooperate to preserve its participation in the European Exchange Rate Mechanism (ERM II).

In this light, while the analogy with Argentina 2001 may be useful to understand the dynamics of the crisis and the fragility of the peg, Latvia’s options exceed those available to Argentina at the time.

Three pressure points

In a nutshell, Latvia’s current currency predicament has broadly three fronts:

Deposits: Depositors typically hedge the mounting currency risk first by “dollarising” bank deposits and, when devaluation is perceived as a bank solvency problem, by withdrawing to purchase foreign exchange. It is the second, more drastic avenue that can ultimately break the bank (if the central bank limits the provision of liquidity to face the deposit run) or the peg (if the central bank assists the private bank, injecting liquidity that ultimately goes against central bank reserves). Local deposits stabilised with the IMF deal (slowed down, in the case of nationalised domestic bank Parex, by monthly withdrawal limits similar to the “corralito” imposed in Argentina 2001) but continued to switch into foreign exchange deposits. By contrast, foreign deposits continue to fall.

Foreign currency borrowing: Loan non-performance – already increasing across currencies due to the economic malaise – should spike within foreign-denominated debtors after a devaluation.¹ Foreign exchanged-related insolvency risk is at the root of rollover risk: from end-November 2008 to end-April 2009, banks’ foreign liabilities fell by €2.4 billion, suggesting that Swedish banks are reducing their exposure to their own Latvian branches – the main driver of reserve losses, which fell over the same period by €0.4 billion, despite an IMF disbursement of €0.2 billion. By taking over domestic Parex and requiring capital injections as a condition for liquidity assistance to foreign banks (another similarity with the Argentine crisis), the Latvian central bank has signalled that, while domestic banks will be rescued, foreign banks will likely be left on their own. Indeed, bank internationalisation allows Latvia to partially “export” the fiscal cost of the bailout, bringing the case closer to the Uruguayan 2002 episode.²

Liquidity: It is the contractionary effect of unsterilised foreign exchange intervention that is pressing the authorities to reconsider the original plan, even moving them to consider issuing “vouchers” for certain transactions, a version of the Argentina’s quasi-moneys issued by the federal and local Treasuries in 2001.

Figure 2. …and the central role of banks

Three exit strategies

As the defence of the peg becomes increasingly untenable, the focus is shifting to a few alternative avenues to avert a disorderly currency collapse.

Float

Judging from past experience, even if backed by an augmented EU-IMF program, a devaluation would overshoot the ex ante real exchange rate misalignment (which, based on the recent evolution of the Latvia’s REER, would place the needed correction already at a sizable 50%) fuelled by the run to dollarise savings before the new, higher exchange rate materialises. Argentina is a case in point – after the discrete 40% devaluation of January 2002 succumbed within a month to reserve drainage and parallel market pressures, the exchange rate overshot from 1 to 4 before coming down to 3 by end-2002.

However, given the current depth of the crisis and the fact that the inevitable debt write-downs that would benefit Latvian debtors at the expense of Scandinavian banks could boost the post-crisis rebound, devaluation may ultimately deliver the faster road to economic recovery.³

But, in the particular case of Latvia, a good old devaluation would reset the clock for euro adoption and, given it potential implications for other ERM II countries, may draw little EU support. While this option remains the exit of last resort, it is unlikely to be the route chosen in the first place.

Euroise

The Argentine analogy is, again, illuminating. Faced with concerns about its peg’s sustainability after Brazil abandoned its crawling peg in 1999, an Argentine mission to Washington to secure the endorsement of the US to de jure dollarisation – and the lender of last resort services of the Fed – was given a sympathetic but discouraging message – even if Argentina presented an attractive payoff for such a contingent liability (which it did not), a treaty would open the door for other financially dollarised countries to request similar treatment and would never pass Congress.

By contrast, Latvia’s early euro adoption would only test the EU commitment to euro convergence – a key driver of the Eastern European leveraging story. However, politics are more complicated. In Europe, there are many countries currently warming up to adopt the euro that could see Latvia’s case as a useful shortcut to avoid improbable but still possible currency stress down the road. Moreover, Latvia’s main regional exposure is vis-à-vis Scandinavian banks; the fact that euro adoption would grant Swedish banks a euro lender of last resort (the ECB) should certainly make euro countries uneasy.

Ultimately, while euroisation seems to be the solution favoured by the IMF, it would only make sense provided that the Euro zone approves, an unlikely event.

Realign

A strategy halfway between euroisation and floating – a contained devaluation that preserves Latvia’s ERM II status – falls short by most accounts, but it is nonetheless the most likely to broker a compromise between all relevant players (the Lats, the EU, the IMF, Sweden). The natural way to implement this would be a negotiated one-off 15%-30% realignment of the central parity preserving the ECB commitment to intervene at the bounds, and the time table for euro adoption, accompanied by the widening of the current +-1% band to the ERM standard +-15%. True, it’s hard to find successful contained devaluations under a currency run in recent economic history. But there are more things at stake in the Latvian peg – even a devalued one – than just a nominal anchor, which makes this strategy, if not a sure cure, at least a viable therapy.

Crucially, the plan requires a “Uruguay 2002”-type solution to the banking problem – limiting or suspending emergency assistance to foreign branches, thereby eliminating about 60% of the foreign exchange bank liabilities (a scenario that the Riksbank is regarding as increasingly likely). As for domestic banks, full deposit insurance plus explicit government support should counter the deposit run and keep dollarisation within the banks. Should the run continue, banks would be nationalised, reprogramming deposits to reduce the pressure on the Lat and restructuring non-deposit liabilities. The other critical aspect of this strategy is, or course, IMF-EU-EBRD money (as in EBRD’s recent involvement in large local bank Parex) to defend the new band ceiling (since the inadequately adjusted exchange rate will likely be under stress), thus providing assurance that Latvia will not be the next Argentina.

Such a scheme should receive support from all quarters. The EU minimises contagion (despite some predictable near-term ripples), the IMF avoids another embarrassing collapse and ensures that some of the foreign exchange pressure is transfered elsewhere, and Latvia stays the ERM course and shares the losses with Sweden – a fitting epilogue to a crisis that was in part rooted in reckless lending by foreign banks.

For all the obvious similarities with Argentina 2001, Latvia presents a more complex case that is poised to become the acid test of euro commitment.

References

De la Torre, Augusto, Eduardo Levy-Yeyati, and Sergio L. Schmuckler (2003), “Living and Dying with Hard Pegs: The Rise and Fall of Argentina’s Currency Board”, Economia, vol. 5

Levy-Yeyati, Eduardo (2006), “Financial dollarization: Evaluating the consequences”, Economic Policy, January.

Levy-Yeyati, Eduardo, Maria Soledad Martinez Peria, and Sergio L. Schmukler (2004). "Market discipline under systemic risk - evidence from bank runs in emerging economies," World Bank Policy Research Working Paper 3440.

IMF (2009) Country Report No. 09/3

1 Typically in financially dollarised economies, a small currency mismatch at the bank hides a considerable large exchange rate-related credit risk (see Levy-Yeyati 2006).

2 For a comparative analysis of the crises in Uruguay and Argentina, see Levy-Yeyati, Martínez Pería and Schmukler (2004).

3 In Argentina, where balance-sheet effects were diluted through forceful conversion of foreign exchange liabilities at the peg exchange rate, growth picked up right after the devaluation was complete.

This article may be reproduced with appropriate attribution.