In the early 2000s, both Brazil and Turkey experienced severe financial crises. Currency depreciation and public sector solvency concerns in the context of significant liability dollarisation pushed both countries to the verge of default. Luckily, both countries embarked upon successful IMF-supervised economic adjustment programme under new governments and experienced a decade of solid economic growth. Coincidentally (maybe) both countries have recently been experiencing wide-spread popular protests, albeit for a variety of different reasons. Arguably, the economic stabilisation of the past decade and the emergence of a politically vocal middle class have played a role here. But this is a separate topic. Let’s focus on economics.
The combination of fiscal adjustment and growth acceleration underpinned improving government solvency. In Brazil, net public sector debt has fallen from more than 60% of GDP in 2002 to 35% of GDP today. Gross general government debt - the more widely used indicator for purposes of cross-country comparisons - remains relatively elevated at 69% of GDP. The more limited decline of gross debt is to some extent due to central bank’s sterilised FX intervention and accumulation policy as well as sizeable off-balance financial transactions in the guise of lending to public sector banks. In Turkey, net general government debt has fallen a little faster than in Brazil, decreasing from 70% of GDP in 2002 to less than 30% of GDP today. Gross general government debt has fallen by much more than in Brazil and today stands at around 36% of GDP.
Both sovereigns have successfully reduced their (net) foreign-currency (FCY) exposure. At end-2012, Brazil’s sovereign net FCY creditor position was 14.2% of GDP, compared to Turkey’s 1.7% of GDP. Both countries have largely eliminated domestically-issued FCY(-linked), while a decade ago, 30-40% of domestically-issued government debt in both countries was linked to the exchange rate. In Brazil, the general government domestic debt is 55.8% of GDP, compared to total debt of 58.7% in 2012. By comparison, Turkey’s domestic debt amounts to 28.4% of GDP, compared to a total of 37.6% of GDP. The Brazilian central bank holds FCY assets worth 14% of GDP, roughly the same as in Turkey, translating into net FCY creditor position in both cases.
A sovereign net FCY creditor position means that, unlike a decade ago, exchange depreciation leads to a decline – rather than a significant rise – in the net debt-to-GDP ratio. Accumulating large FX reserves does come at a direct financial cost, however. First of all, the central bank typically sells government bonds from its portfolio to absorb the additional liquidity created by FX purchases, thus effectively financing lower-yielding FCY-denominated assets by way of higher-yielding local-currency (LCY) government debt. Moreover, to the extent that a currency is undervalued in real terms, reducing net FCY debt deprives the sovereign of the opportunity to reduce the LCY value of FCY liabilities by way of real currency appreciation. Last but not least, the resulting larger stock of domestic LCY debt will tend to help keep domestic interest rates high.
On the flipside, a large net FCY creditor position (aka large FX reserves) provides the authorities with more ammunition to intervene in the FX market and provide FCY funding to domestic financial institutions and corporates in case of a “sudden shock” (e.g. Brazil in 2008-09). Brazil certainly has a much a more favourable external liquidity position than Turkey. It may be worth noting that the CBRT’s net international FX reserves amount to less than half its reported reserves after adjusting for domestic banks’ FX deposits with the central bank. In practice, this may not make a significant difference, but in this respect Brazil’s position is more favourable, too.
While both the Turkish government and the banking sector can sustain even a large exchange rate shock, limited FCY liquidity means that the impact in terms of higher interest rates, lower local liquidity and economic growth will be much more significant than in Brazil. It is no coincidence that Turkish real GDP declined dramatically in late 2008 and early 2009, while the Brazilian economy only registered a very mild decline in output. Turkey is no doubt more exposed to a sudden stop, and while the government and the banking sector seem well-positioned to withstand significant exchange rate depreciation, the authorities have significantly less scope to soften the impact of such a shock as far as external liquidity is concerned.
Interestingly, Brazil’s significantly more favourable FCY position does not translate into lower sovereign risk as reflected in CDS spreads. Curiously, as of early July Brazilian 5Y CDS spreads traded at 195 bp versus 180 bps in Turkey. Naturally, Turkey’s gross and, less so, net public sector debt is lower than Brazil’s. But its external position, if not vulnerable, is undoubtedly more sensitive to an external shock than Brazil’s. It looks as if the benefits Brazil derives from a large sovereign FCY creditor position in terms of markets’ perception of sovereign risk are at best very limited compared to Turkey, while it undoubtedly translates into higher (quasi-) fiscal costs. Nominal and real interest rates in Brazil remain significantly higher than in Turkey. That said, large FCY holdings do make Brazil less susceptible to a sudden stop than Turkey. Think of greater quasi-fiscal costs as an insurance premium.