Hyperbolic talk of currency war masks reality of persistent US structural power (2011)

The world economy is on the brink of a “currency war”. Some policymakers have threatened “retaliation”. This is hyperbole, for not only is war is an odd term to refer to a situation where somebody sells you something at a discount, which is effectively what you do when you keep your exchange rate at “competitive” levels. But a good, old-fashioned term like “beggar-thy-neighbour” policies would do just fine to describe what is going on. International monetary relations are fraught with tensions, which the G20 summit in Seoul, unsurprisingly, left unresolved. In times of low growth, economic uncertainty and/ or high unemployment, tapping external demand through a “competitive” exchange rate becomes both an attractive policy option and a bone of contention. 

In this context, the US Federal Reserve has been pursuing a policy of quantitative easing aimed at pushing down bond yields, raise US asset prices and weaken the exchange rate. In doing so, the Fed is seeking to both fend off deflation risks and support economic activity. Other countries fear that lower interest rates and a weaker dollar will put appreciation pressure on their currencies and increase capital inflows – especially countries with flexible exchange rates, an open capital account and high interest rates. 

A country can respond to capital inflows in several ways. (1) It can absorb the capital inflows into official reserves via FX purchases. If left unsterilised, this will raise domestic inflation. If fully sterilised, this may or may not push up interest rates, but it will typically addition to the fiscal costs of carrying reserves. (2) It can let the exchange rate appreciate. While this may diminish the attractiveness of local assets in the eyes of foreign investors, it is likely to lead to an “overshooting” of the exchange rate, slowing economic growth and a deterioration of the current account. (3) It can try to limit capital inflows through capital controls or macro-prudential measures. If implemented successfully, this will reduce secondary market liquidity and push up domestic yields. All these responses carry economic, financial and/ or, in the case of controls, implementation risks. Policymakers facing surging capital inflows and an appreciating currency find themselves between the proverbial rock and a hard place. 

While Fed quantitative easing has a very significant impact on other, especially smaller economies, any “retaliatory” measures these may take have virtually no economic or financial impact on US. Several factors explain this asymmetry. (1) The dollar exchange rate and, even more so, any bilateral exchange rate is far less important to the US than for its trading partners, with the possible exception of the EU given the importance of bilateral trade. (2) Capital controls in other countries do not impose any tangible economic or financial costs on the US – other than perhaps opportunity costs for individual US investors. (3) Sanctions by other countries targeting US exports are neither credible, for, assuming the US responds in kind, these typically more trade-dependent economies would incur relatively greater economic costs than the US. 

“Currency war” therefore seems a misnomer. It may feel like war to smaller countries, as they feel compelled to take “defensive” action to fend off capital inflows. But if this is a war, the US has not noticed. The US pursues an economic policy it deems to be in its interest and however these countries respond to it is of little consequence to the US. This is a prime example of continued US “structural power”. Structural power is the power of a state to indirectly influence others by controlling the structures within which they must operate. This differs from “relational power”, or the ability of one state to influence another state's behavior directly. This describes the situation quite accurately, for Washington is not seeking to influence other countries’ behaviour. If one prefers to stick to martial metaphors, US policy may be said to cause “collateral damage”. 

US structural power is attributable to the dollar’s status as a reserve currency, the US government’s ability to issue debt in its own currency and the large size of the US economy on which other countries are – or perceive themselves to be – depend in terms of generating economic growth. Keen to maintain a competitive exchange rate, the other economies have little choice but to absorb US capital outflows into FX reserves. This is not to say that some of these countries could not inflict meaningful costs on the US by, for example, dumping US assets or imposing protectionist measures on US exports. However, both the economic and financial costs of such measures would be far greater for these countries than for the US. 

In other words, the US retains maximum policy flexibility, while its own choices have a very tangible impact on, and create significant constraints for, other countries. As long as other economies depend more on the US market for their exports than vice versa, they will not only be highly constrained in terms of policy options, but they will also wield little in the way of bargaining, let alone retaliatory power vis-à-vis the US. Even holding large amounts of US debt will provide them with very little leverage as long as they are reluctant to let their currency appreciate. True, the US has thus far failed to get other countries, notably China, to appreciate their currencies. Whether one attributes this to a lack of US relational power or a reluctance on the part of Washington to yield its power, US structural power remains the defining characteristic of today’s global monetary and financial system. 

BRIC public sector debt is very manageable in spite of rising off-balance sheet risks (2011)

Over the next five years, government debt in the G4 (US, Japan, Germany and the UK), with the exception of Germany, will rise very significantly. The G4 will therefore have to undergo a multi-year fiscal adjustment in order to put debt on a sustainable path. By contrast, the BRIC governments, having weathered the global crisis fiscally unscathed, will see their debt levels largely unchanged and face a very limited or even non-existent need for a fiscal adjustment. Even if BRIC public sectors were forced to take on contingent liabilities (e.g. in case of a banking sector bail-out), they would still be in better shape, financially speaking, than the G4.

Public debt matters – economically and politically. A high and rising level of public sector debt will, sooner or later, push up domestic interest rates, crowd out private-sector investment and eventually threaten government solvency. Rising debt levels also limit fiscal flexibility, ultimately forcing governments into an extended fiscal retrenchment. This can have consequences reaching far beyond the economic realm. A very large debt burden, for instance, forced Great Britain to liquidate its empire following WWII. Similar concerns have informed recent analyses by US foreign policy in the context of an unsustainable US fiscal policy (Bacevich, Johnson, Mandelbaum).

Over the next five years, government debt is set to decline moderately in the BRICs, while it will rise dramatically in the G4 (US, Japan, Germany and the UK), with the exception of Germany. The G4 will therefore have to undergo a multi-year fiscal adjustment in order to put debt on a sustainable path. By contrast, the BRIC governments, having weathered the global crisis fiscally unscathed, will see their debt levels largely unchanged and face a very limited or even non-existent need for a fiscal adjustment. A combination of significantly lower fiscal deficits, faster economic growth and higher inflation will ensure debt sustainability. According to IMF estimates, the fiscal deficit in the BRIC will average 3.1% of GDP versus 5.6% of GDP in the G4. The real (inflation-adjusted) deficits in the BRIC will be even smaller, of course. 

That said, the level of government debt in the BRIC countries varies substantially. Gross general government debt in Brazil and India amounts to a sizeable 67% and 75% of GDP, respectively, while in China and Russia it stands at a very low 20% and 10% of GDP, respectively. It is also noteworthy that BRIC government debt is almost exclusively held by residents. Only in Brazil do foreigners own a little more than 10% of total government debt. This contrasts sharply with Germany (50%), the UK (30%) and the US (50%), where foreigners hold much larger shares of government debt. If government and central bank balance sheets are consolidated, the public sectors in the BRICs are also all net foreign (currency) creditors. China’s public sector owns net foreign assets worth a stunning – for a country the size of China – 50% of GDP. A very low dependence on foreign financing sharply limits BRIC governments’ financial vulnerability.

Recently, many analysts have expressed concern about the size of (more broadly defined) public sector debt and contingent liabilities in the BRICs (and especially in Brazil and China). Non-financial public sector (NFPS) debt comprises, in addition to the debt of the general government sector, the liabilities of the central bank and nonfinancial public-sector-owned companies. Net NFPS debt, the more relevant indicator from a debt sustainability perspective, amounts to a very manageable 40% of GDP in Brazil, the only BRIC country providing consolidated PS figures. Russia has no doubt the lowest net NFPS debt – the general government is net creditor, after all! For both India and, even more so, China, wide-spread government ownership of non-financial (let alone, financial) companies at both the central and local government levels make it virtually impossible to estimate net public debt with any degree of accuracy.

The Chinese authorities have just released estimates of direct and explicitly-guaranteed local-government debt, a source of concern to analysts following the massive surge in bank lending to local governments in 2008-09, putting it at a manageable 27% of GDP. India has been providing more comprehensive state government debt statistics all along. Neither country provides NFPS estimates. However, unless one assumes that the liabilities owed by public sector companies vastly exceed their assets, net NFPS is bound to be sustainable in all BRICs given the combination of (strong) economic growth, (generally small) government deficits and (low) real interest rates.

What about contingent liabilities? Past experience suggests that banking sector crises are the single most important source of contingent liabilities. All BRIC economies have been experiencing strong real credit growth over the past two years, raising concerns about a future rise in non-performing loans and the potential need for the government to extend financial support to the banking sector. All other things equal: the larger the size of the banking sector and the larger the share of lending by government-owned banks, the greater the potential liabilities. Bank lending to the private sector amounts to 50% of GDP in Brazil, India and Russia and a very considerable 135% of GDP in China. Government-owned banks account for 50% of total banking sector assets in China, 40% in both Brazil and Russia and 70% in India. Ceteris paribus this suggests that China faces the potentially largest contingent liabilities as a share of GDP. Admittedly, in the event of a systemic banking crisis, governments often have little choice but to bail out banks regardless of ownership.

Naturally, credit quality, capital buffers and profitability also affect the level of contingent liabilities. Fitch estimates that the Chinese banking system might require financial support in the order of 10-30% of GDP in a moderate and severe stress scenario, respectively. Bail-out costs in the other BRIC countries would be significantly smaller given the much smaller size of their banking sectors. In short, the contingent liabilities associated with even a severe banking sector crisis would not undermine debt sustainability in any of the BRICs. China, facing potentially the largest contingent liabilities, is best placed to sustain an increase in liabilities given that it enjoys the strongest economic growth outlook, has comparatively little debt and a captive domestic investor (depositor) base. If, unrealistically, all the debt of all non-financial public-sector entities, including central and local governments, were added up and assumed by the central government, Chinese gross public debt may add up to 150% of GDP or so (before intra-public sector netting!). Even in this scenario, this would not break the “sovereign” bank – even if real GDP growth declined substantially (from 10% to 6%), real interest rates rose (by 200bp) and the primary fiscal balance worsened (by 1% of GDP).

In sum, even if BRIC public sectors were forced to take on contingent liabilities, they would still be in better shape, fiscally speaking, than the G4. This does not, however, necessarily translate into higher sovereign risk in the G4, for the latter have a number of things going for them (e.g. large, diverse and “deep” investor base, solid political and economic institutions, a strong debt service track record). Nonetheless, financially and politically, the BRICs will benefit from far greater fiscal flexibility than the advanced economies over the next decade and beyond, off-balance sheet liabilities notwithstanding. Higher underlying growth will increase BRIC governments’ fiscal resources relative to the G4. This will have wide-ranging consequences for the economic and political position of the BRIC relative to the G4.

Demographic trends in advanced and emerging economies and their potential consequences (2011)

Population growth is projected to turn negative in China (2025) and Brazil (2040). Russia’s population has been declining for over a decade. This is worth keeping in mind when analysing the “rise of the BRICs”. The BRICs (with the exception of India) will sooner rather than later be confronting significant demography-related challenges, including a shrinking labour supply, a potentially declining pace of innovation, declining domestic savings, rising pension and healthcare expenditure and, more arguably, a structural decline in inflation. Moreover, both developed and emerging economies must not discount the non-linear effects population decline is bound to trigger.

Demographic trends will diverge sharply over the coming decades. Germany and Japan, the world’s third- and fourth-largest economies, are already experiencing population decline. Less often-mentioned is the fact that most top-tier emerging markets, including the BRICs, are also rapidly approaching an inflection point. Population growth will soon turn negative in China (2025) and Brazil (2040). Russia has been experiencing population decline since 1995. Only India’s population, projected to overtake China’s by 2020 or so, will continue to grow past the middle of the century. By contrast, the UN projects the populations of France, the UK and the US to expand, albeit gradually, until the end of the century. 

Population aging, even more so population decline, is bound to have important political, economic and social consequences. First, labour supply, defined as the population aged 15-64, will shrink. Increasing labour market participation, let alone raising working hours, will only go so far. Sooner or later, man hours worked will decline. Only in Brazil, India, the UK and the US will the potential labour force be tangibly larger in 2030 than today.

Meanwhile, the dependency ratio (population aged younger than 15 and older than 65 relative to the rest) will be rising in both the BRICs and the G5. This will likely reduce savings. If investment remains unchanged, interest rates will have to rise. If investment declines, too, as is likely, a rise in interest rates will be avoided, but economic growth will diminish. Either scenario is negative for government debt sustainability.

Second, there is some evidence that innovation and technological progress is furthered by large, expanding populations (Kremer 1993), all other things being equal. The older populations become, the more reluctant they will be to adopt new technologies (demand) and the less able they will be to innovate (supply). This may, however, be offset by the greater economic incentives to innovate in the context of a declining labour supply and the resulting upward pressure on wages. A declining ability to innovate is arguably less relevant for the less advanced economies – not least because they can simply import (and adapt) existing technology. By contrast, advanced economies operating at the “technological frontier” and facing declining capital and labour supply will have to rely, even more than before, on total factor productivity growth. Just when innovation is becoming more important, the capacity to innovate may be declining!

Third, an increase in the old-age dependency ratio will typically raise pension and healthcare expenditure. Not only will this weigh on the domestic savings rate but it will also raise government outlays, putting pressure on the fiscal accounts. Entitlements are significant in all advanced economies and, if left unreformed, will put major pressure on fiscal positions at a time when household savings are declining. More likely than not, this will put downward pressure on public (and private) investment and thus negatively affect the economy’s growth potential. It may also raise interest rates.

Fourth, and admittedly more speculatively, a declining population and a rising old-age dependency ratio may drag down inflation due to decreasing demand (Ezer 2011). Empirical support for such a view is mixed. In the context of a zero (nominal) interest rate, Japan has been suffering from deflation over the past two decades. While this can be largely attributed to the “debt overhang” following the financial crisis two decades ago, it is possible that demographic factors have also contributed to deflationary pressures. If there is some truth to this, population aging may lead to low inflation. A very low level of inflation, let alone deflation, would negatively affect debt sustainability by setting a floor for real interest rates or even raising the real value of debt.

As such, this is bad news, especially for the initially more rapidly aging G5. As regards Japan and the US at least, potential growth may already have declined due to on-going post-crisis deleveraging. The structural fiscal position has deteriorated sharply and public debt has moved onto an unsustainable path. Over the medium term, the situation will become even more challenging due to a declining labour supply, declining savings and investment, diminished growth, rising age-related (government) expenditure and a more limited ability to diminish the real debt burden due to lower inflation. By contrast, a (generally) more favourable demographic outlook, a much more sustainable initial fiscal and public debt position and a significant (and undiminished) post-crisis growth potential puts the BRIC in a very different position – at least in the short- to medium-term.

It would be a mistake to underestimate the possible lateral (or non-linear) effects of population aging/ decline by disregarding economic, political and financial feedback loops. Mathematically, population growth tends to turn exponential, as does population decline – speaking of non-linearities! Negative population dynamics tend to intensify in the context of positive feedback loops and negative externalities. For instance, a rising fiscal burden, diminished growth prospects and less innovative society may drive young entrepreneurial types into emigration, while fiscal pressures and declining public investment may lead to a general deterioration of economic, financial and social conditions, which might further accelerate population decline (positive feedback). On other hand, rising labour incomes amid a declining workforce may attract greater immigration, softening the demographic downward trend (negative feedback). Last but not least, it may raise fertility against the backdrop of rising labour compensation.

Politically, the emerging “grey majority” might make reforms aimed at lifting economic growth or reforming public finances by limiting pension and health care expenditure well-nigh impossible. Domestically, this may result in political stalemate and intensify economic stagnation/ decline. In foreign affairs, the “grey majority” may diminish the tendency to pursue aggressive policies, while fiscal weakness may diminish the inclination to engage in financially costly armed conflicts. Incidentally, fewer children may make parents less inclined to lend support to armed conflict, while a smaller share of young people may make it more difficult for governments to garner domestic support for a high-stakes foreign policies, as the risk-averse, older majority prevails.

To sum up, demographic change will have important, if somewhat underdetermined economic, financial and political consequences. While virtually all top-tier emerging markets and advanced economies are aging, different countries will be confronting demography-related challenges with greater or lesser urgency. Interestingly, the dividing line is not as clear-cut as might be expected. China and Russia will be characterised by a rapidly aging demographic, while among the advanced economies the Anglo-Saxon countries will be confronted with less dramatic change – relatively speaking. This should be borne in mind when analysing the “rise of the BRICs” and the “decline of the ‘West".

Sino-US financial imbalances bound to grow further (2011)

China’s official FX reserves hit a stunning USD 3.2 tr in July. In order to prevent too rapid a nominal appreciation against the USD, the People’s Bank of China (PBoC) is forced to recycle a large share of the balance-of-payments surplus into dollar-denominated assets. Official FX reserves are typically invested in liquid, high-grade debt instruments. About half of China’s FX reserves are invested in US treasuries and agencies, making the Chinese government the single largest creditor of the US government.

Beijing will have no choice but to pile into US government debt for the foreseeable future. If the Chinese current account surplus gradually moves back to 7.5% of GDP by 2015, as the IMF – assuming an unchanged real effective exchange rate – projects, and assuming China will be running a capital account surplus of 2% of GDP a year (below its historical average!), it could see its FX holdings increase by USD 3.5 tr by 2015. Even under much less aggressive assumptions, FX reserves are bound to increase by a massive USD 2.0-2.7 tr over the next five years. Assuming China keeps the share of dollar-denominated assets constant (at an estimated 2/3), continues to invest the bulk of its dollar-denominated holdings in US treasuries and reduces its exposure to agencies, holdings of US treasuries could rise from currently USD 1.2 tr to nearly 4 tr under the admittedly aggressive IMF baseline scenario.

If, on the other hand, the Chinese authorities, who project a current account surplus of 4% of GDP, are correct, China will accumulate only USD 2.25 tr in additional FX reserves by 2015 (assuming a relatively small capital account surplus of 1.5% of GDP). This would nonetheless sufficient to push total Chinese holdings of US government debt to nearly USD 3 tr. If only 1/2 of the cumulative increase in FX reserves were invested in US government debt, Chinese holdings would turn out lower, but would still exceed USD 2.5 tr by 2015. In short, even under such a “best-best-case” scenario, Beijing’s exposure to Washington and Washington’s reliance on Chinese financing will continue to grow tangibly.

Assuming furthermore (as per Congressional Budget Office) that US treasuries (held by the public) climb to USD 12.6 tr (or 70% of US GDP) by 2015, China could end up holding close to 1/3 of the total – or 2 ½ times today’s share – under the IMF baseline scenario. Or to put it differently, Beijing would hold 37% of Chinese GDP worth of US treasuries (up from 20% of GDP today), while Washington would owe 20% of US GDP worth of treasuries to the Chinese government (up from 7.5% of GDP today). Under the other scenarios, these ratios would be somewhat lower, but not dramatically so.

A number of caveats are in order. First, the current account projections might be way off the mark. The IMF estimates appear to be at the high end. China’s real effective exchange rate will almost certainly appreciate. And if it does not, would the deficit countries be prepared to absorb the rising trade surpluses such a projection implies? That said, it is difficult to see why the surplus should fall much below 3-4% of GDP, short of Beijing allowing the RMB to rise very significantly in nominal terms. And this looks similarly unlikely.

Second, China might ease restrictions on private-sector capital outflows (while maintaining controls on inflows). This might help lower the cumulative balance-of-payments surplus, FX reserve accumulation and thus the pace of treasury accumulation. While the Chinese authorities profess a desire to move towards greater capital account openness, this will invariably be a very gradual process and will not be anywhere near completion by 2015. Besides, the capital account would have to shift into deficit to make a tangible dent in the overall balance-of-payments surplus. This does not look like a very probable scenario at the moment. Even if FDI outflows, which face the least restrictions, rise to USD 100 bn a year, this would be insufficient to push the capital account into deficit, given continued strong FDI and “other” capital inflows. Moreover, a sharp rise in outward FDI might trigger rising political resistance among recipient countries. It already has. (No wonder that Beijing put a bilateral investment treaty near the top of its agenda during last May’s SED.)

Third, China might decide to accumulate dollar-denominated assets other than US government debt and/or divest a share of its existing holdings of US government debt. The China Investment Corporation (CIC) is reportedly to receive another USD 200 bn, in addition to its initial USD 110 bn allotment, of PBoC foreign reserves. However, the new allocation would amount to less than 6% of official reserve holdings. Moreover, the traumatic 2008 experience when the CIC’s equity investments lost significant value will ensure that any larger shift out of US treasuries into higher-risk assets will be a very gradual process. The speed of diversification will inevitably lag the increase in FX reserves over the next few years. Even a deteriorating US fiscal outlook would not do much to change this given that virtually all other USD assets would presumably be similarly negatively affected by rising US sovereign risk. Purchasing non-treasury USD assets would not allow China to diversify the risk associated with its USD assets. If it did shift out of treasuries, Beijing might end up fuelling the very crisis it seeks to avoid. Chinese holdings of US treasuries are bound to grow further.

In short, even determined efforts by the Chinese authorities will be insufficient to prevent a further, significant accumulation of US treasuries. It will not be possible to shift additional reserve holdings into higher-risk instruments sufficiently quickly to avoid a further accumulation of US treasuries. Only the US treasury market offers sufficient depth to absorb Chinese USD investments. Long story short: Beijing will continue to accumulate claims on the US government, while Washington will see its liabilities vis-a-vis China continue to rise. The financial stakes in Sino-US relations will be rising inexorably for the foreseeable future – relentlessly raising the bilateral economic and political stakes.

BRIC banking systems after the crisis (2011)

State-led economic development, if successfully implemented, is appropriate during the early “catch up” phase of economic growth. However, as growth becomes more dependent on indigenous innovation and hence a dynamic private sector, a shift to more market-led rather than state-directed development becomes necessary. This also applies to the banking sector. Subject to proper regulation, banking systems that rely on private-sector banks and market-led credit allocation will eventually tend to generate superior economic outcomes. That said, we are unlikely to see a significant reduction in public-sector bank ownership in the BRIC countries anytime soon, nor, for that matter, a tangible increase in foreign ownership.

Following the global financial crisis, BRIC governments have become (even) less prepared to reduce their presence in the domestic banking system. After all, policymakers’ success in overcoming the credit crunch in 2008-09 in part relied on their ability to provide credit to the economy through public sector-owned banks. In the absence of often substantial public-sector bank lending, the decline in domestic demand in a number of countries (e.g. Brazil, China, India) would have been much more severe. Conventional monetary policy would have been and was insufficient to stimulate bank lending (aka Keynesian “liquidity trap”).

Enter the Beijing consensus, exit the Washington consensus. The Beijing consensus is committed to, among other things, the state playing an activist role in economic sectors deemed “strategic”, invariably including the banking sector. This takes the form of outright government ownership or at least significant government intervention. Instead of near-exclusively relying on private-sector, market-led processes, the state takes an activist approach going far beyond merely regulating private-sector activity. Historically, this type of successful developmentalist, state-led economic policy and development does nonetheless rely on functioning private markets – nowhere is this more evident than in today’s China, where the private sector has been the main engine of economic growth. 

History suggests that this strategy, if successfully implemented, is appropriate during the early “catch up” phase, when per-capita income is low and growth is significantly driven by large-scale investment in physical infrastructure and the introduction of “off-the-shelf” technologies. However, as per-capita income rises, growth becomes more dependent on a dynamic private sector and indigenous innovation. Eichengreen et al. (2011) identify a per-capita income of USD 17,000 (in 2005 constant international dollars) as a critical threshold where economies experience a tangible downward shift in their trend growth. This is where state-directed policies are bound to become less effective in terms of generating growth than a dynamic private sector. This suggests that – following recent PPP revisions – smart state-led policies remain broadly appropriate in low per-capita-income India (USD 3,500) and, less so, in China (USD 7,700) and Brazil (USD 10,000), while they are bound to be less effective in Russia (USD 15,200), all other things being equal. Naturally, should China continue to grow at near-double-digit rates, it would, as Eichengreen et al. (2013) point out, reach the “critical threshold” before the end of the current decade.

The greater need for a shift to more market-led rather than state-directed development also applies to the banking sector. At a stage of economic development where per-capita income is low and capital productivity is high, it is not difficult to identify profitable and economic growth-generating lending opportunities. This tends to change as economies move into middle-income territory. This is why, subject to proper regulation, banking systems that rely on private-sector banks and market-led credit allocation will eventually tend to generate superior economic outcomes.

This flies in the face of policymakers’ recent successful experience with counter-cyclical state-directed credit policies. After all, the extensive use of government-directed bank lending played an important role in sustaining domestic demand and economic growth (China, India) or may, at least, have prevented an even sharper economic contraction (Brazil, Russia). Interestingly, real bank lending grew significantly faster in the BRIC countries, where governments play an important role in the banking sector. Real credit growth averaged almost 25% in China in 2009-10, while public-sector banks in Brazil, impressively, doubled lending from 10% of GDP in 2008 to 20% of GDP in 2010. 

This contrasts sharply with the contraction in credit experienced by many developed markets and relatively anemic credit growth in those emerging markets where government ownership of banks is very limited (e.g. Mexico, Eastern Europe). Admittedly, other factors such as extensive foreign ownership and significant cross-border lending may also have contributed to differential credit growth. But the role played by public-sector banks was undoubtedly important.

It is perfectly sensible to pursue counter-cyclical state-directed credit policies if the banking system is dysfunctional and is suffering from market failure. However, time inconsistency and politicians’ desire to dish out favours risk turning counter-cyclical policies into pro-cyclical ones. Interestingly, among the BRICs, only Brazil seems to have given in to this temptation, while China, more accustomed to state-directed lending and more concerned about its inflationary consequences, has not. But unless top-notch governance regimes are in place, extensive state-directed credit allocation, especially if sustained over a longer period of time, carries the risk of capture by “rent-seekers”. And rare is the government (or the bureaucracy) that manages to privilege medium-term economic efficiency over short-term political considerations in a consistent manner. An economy that grows at double-digit rates may be able to afford this (China), whereas most economies, especially those constrained by low savings rates (Brazil), cannot. Last but not least, an extensive public-sector presence also undercuts, and if does not undercut then it certainly slows, the development of a more sophisticated banking and financial sector capable of sustaining economic development once an economy moves into middle-income territory.

That said, it is difficult to see why the BRIC governments would be willing to substantially reduce, let alone relinquish their role, in the domestic banking sector over the next few years. Some BRIC governments have sold (China) or are planning to sell minority stakes (Russia) in major state-owned banks. But none of them is seriously considering giving up control. True, Brazil did fully privatise a number of its public-sector banks in the 1990s (and even sold some of them to foreigners), but this occurred against the backdrop of severe financial pressures and an urgent need to resolve a banking crisis. Short of a major crisis, which is unlikely given solid economic fundamentals, we will not see a substantial decline in public-sector ownership and control in the BRICs over the next decade or so.

Similarly, if the history of banking sector opening since the 1990s is anything to go by, none of the BRIC economies will see a significant increase in foreign bank ownership. While opening the banking sector to foreigners has always been a politically unpopular proposition in the BRICs, economically and intellectually it seemed difficult to contest its benefits. The view that greater foreign ownership is unambiguously a good thing, bringing superior regulation, fresh capital, financial innovation and better risk management, has at least been called into question in the wake of the global financial crisis. There are also concerns among BRIC policymakers that a large foreign presence may allow external shocks to be transmitted more easily. This is debatable, however. Extensive foreign ownership may actually have helped avert a larger crisis thanks to co-ordination committing foreign banks to maintain the lending of their domestically incorporated subsidiaries, recapitalise local subsidiaries (if necessary) and, more generally, allow for an “orderly” de-leveraging (e.g. Vienna Initiative). Still, we are not likely to see either a significant reduction in public-sector ownership or a substantial increase in foreign ownership in BRIC banking sectors in the near or even medium-term future.

Why Brazil is both catching up and falling behind (2011)

Brazil’s medium-term economic prospects underpinned by solid fundamentals, favourable demographics and strategic natural resource exports are bright. Nonetheless, while per-capita income growth has picked up tangibly over the past few years, it remains way below that of Brazil’s East Asian “peers”. Per capita GDP in China and Korea, as a share of Brazil’s, increased from 7% and 62% in 1980 to 60% and 270% in 2010, respectively. Even if Chinese growth declines a couple of percentage points from current levels, its per-capita income will exceed Brazil’s by 2020.

Brazil is doing well. It weathered the 2008-09 global crisis largely unscathed (except for a significant growth slowdown in 2009). It is once more attracting record levels of capital inflows, including very significant FDI. Record-high terms-of-trade and a strong exchange rate are supporting a consumption boom. Unemployment is at all-time lows. Major oil discoveries carry the promise of turning Brazil into an important energy exporter over the next decade and generate windfall revenues for the government. Large FX reserves, a flexible macro-framework (however, imperfectly adhered to in practice), a stable political system, a favourable demographic outlook and an expanding middle class, underpinned by a solid banking system and an increasing number of internationally competitive companies, bode well for medium-term economic growth.

Brazil thus has legitimate grounds for optimism. But everything is relative. Per-capita GDP growth accelerated from less than 1% in the 1980s and 1990s to 2.4% during 2001-10 (or 3.5% since 2004). While this represents a significant improvement, it falls way short of star performers like China and Korea, both of which experienced dramatically higher per-capita growth for any given level of income. As a result, per-capita GDP in China and Korea, as a share of Brazil’s, increased from 7% and 62% in 1980 to 60% and 270% in 2010, respectively. Even if Chinese growth declines a couple of percentage points from current levels, its per-capita income will exceed Brazil’s by 2020.

East Asia’s rapid economic development can be attributed to a number of (causally) difficult-to-disentangle factors, including high investment and large domestic savings, favourable demographic developments (falling dependency ratio) and an outward-oriented, export-led industrialisation strategy, resulting in a high degree of trade openness, which, in turn, offers access to advanced technology and fuels productivity gains. Brazil shares next to none of these characteristics. It has remained a relatively closed economy with merchandise trade accounting for a mere 20% of GDP (compared with 3-4 times that share in China and Korea). The share of commodities in total exports is high (and has actually been increasing). National savings and domestic investment remain relatively low.

Investment is a major driver of economic growth, and investment is largely financed by domestic savings. This means that unless an economy runs a significant current account deficit (imports foreign savings), its investment capacity is limited by its domestic savings. Worryingly, economic stabilisation, an improved outlook and extremely favourable international conditions have not yet had a tangible impact on domestic savings. Especially the terms-of-trade shock should have had a positive, if typically only transitory, impact on savings. Brazil’s domestic savings rate averaged 16.9% of GDP in 2001-10 and a mere 17.8% during the second half of the decade, compared with 17.5% of GDP during the “lost decade” of the 1990s.

This is a problem, for not only has the savings rate barely budged (so far) but it remains a stunning 30 percentage points below China and more than 10 percentage points below Korea’s. High savings rates in EM Asia and low savings rates in Latin America are generally attributed to a varying combination of historical (hyper-inflation), economic (growth), demographic (change), (fiscal) policy and even cultural factors. The difference in investment rates is slightly less pronounced given Brazil’s greater reliance on foreign savings to finance domestic investment. It is likely that marginal capital productivity, especially in infrastructure, is higher in Brazil than in its Asian peers, given the very limited investment in Brazil in this sector over the past three decades. Brazil should therefore get more “bang for its buck”, but an investment rate of 20% of GDP will be insufficient to sustain a growth rate of more than 5% annually. Micro-economic evidence suggests much the same. Brazil’s infrastructure has come under heavy pressure.

Similarly, a number of industries are struggling to find qualified staff. In short, a relative lack of investment in physical infrastructure (e.g. getting commodities to ports, for example) and human capital (e.g. engineers capable of implementing large-scale infrastructure projects) is creating bottlenecks that are driving up prices. In short, if Brazil wants to raise its sustainable rate of economic growth, it needs to raise its investment in a sustainable manner, that is, without running too large a current account and fiscal deficit. The IMF projects the investment to rise from just below 20% of GDP today to an average of 21% of GDP during 2016, assuming an almost unchanged savings rate of around 17.5% of GDP. This compares poorly to China and Korea, which will invest more than 45% and 30% of their respective GDP during 2011-16.

The most likely scenario is one where Brazil settles onto a medium-term growth trajectory of 4.5% per year. This will be politically and economically sustainable. The government faces very limited incentives to pursue major, growth-enhancing structural reform: from an electoral point of view, the short-term political costs of reform would outweigh the political benefit of higher medium-term growth. Smaller, incremental reform less prone to encounter political opposition is much more likely. This is not a catastrophe, but it won't allow Brazil to grow anywhere near Asian levels. Hence Brasilia should not be surprised if China overtakes Brazil in terms of per-capita income soon.

Everything you always wanted to know about Brazil’s public debt (2011)

Brazil’s net public sector debt has been declining, more or less, continuously since 2002. Thanks to large primary surpluses and solid economic growth, net debt has declined from more than 63% of GDP in September 2002 to less than 40% of GDP in July 2011. Similarly, gross government debt declined by 20% of GDP, from a peak of 82% of GDP in September 2002 to 63% of GDP today. Meanwhile, gross domestic government debt has remained largely unchanged due to the sizeable domestic debt issuance necessary to finance public-sector assets accumulation.

The structure of public sector liabilities has improved tangibly over the past decade. Significantly, the public sector became a net foreign (currency) creditor in 2006. The share of both FX- and Selic-linked debt has diminished and the overall maturity structure has improved. As a result, were a financial shock equivalent to 2002 to occur today, the net debt ratio would fall by 3% of GDP rather than increase by 17% of GDP. Indeed, the debt ratio did decline during late 2008 on the back of a sharp currency depreciation. Gross financing requirements remain nonetheless large by international standards.

Net public debt will continue to decline over the short to medium-term. Under our baseline scenario, the ratio will decline by an average of 1.0-1.5% of GDP a year. Under an optimistic scenario – one where a tighter short-term fiscal policy combined with a medium-term adjustment reduces real interest rates materially – net debt could fall to as low as 20% of GDP by 2020. Such a scenario is, however, unlikely to materialise. By contrast, gross government will decline more slowly, depending on continued FX reserve accumulation and likely further lending to official banks by the Treasury.

It would be highly desirable to upgrade the present fiscal framework. Targeting primary surpluses was an appropriate policy when government solvency was at stake. With medium-term solvency not an issue anymore, it is time to “fine-tune” fiscal policy by shifting toward a structural (cyclically-adjusted) primary surplus target, or, even better, a structural nominal balance. This “upgrade” should be accompanied by broader structural reforms.

Brazil has come a very long way since 2002. Brazil suffered a series of major financial crises over the past three decades. Massive foreign borrowing in the 1970s followed by the so-called Volcker shock (read: high US interest rates and US recession) triggered a precipitous decline in commodity prices and eventually forced Brazil into external default. A great many stabilisation plans during the second half of the 1980s and early 1990s failed to bring spiralling inflation under control. After the real plan succeeded in doing so, Brazil suffered several subsequent balance-of-payments crises. In 1998-99, after a period of heightened global risk aversion following the Russian default, large-scale capital outflows forced Brazil off its crawling currency peg. A major contributor to and/or accelerator of both the 1998-99 and 2002 crisis were rising government debt and a debt structure, notably large shares of FX- and Selic-linked debt, that amplified financial shocks. Not only did the public debt structure make Brazil more vulnerable to both endogenous and exogenous shocks. But a high level of debt repeatedly forced the government into pro-cyclical fiscal and monetary tightening, exacerbating economic downturns.

Since the last major crisis in 2002, Brazil’s financial resilience has increased dramatically. The 2008 global crisis proved Brazil’s newfound financial resilience. Its solid external solvency and liquidity position (notably large FX reserves) was instrumental in allowing Brazil to weather the crisis – financially and economically – relatively unscathed. Lower public debt and a net foreign-currency creditor position at last severed the nefarious link between balance-of-payments-shocks-cum-currency-depreciation and public sector debt sustainability. It not only allowed the government to let fiscal stabilisers soften the economic downturn but it even allowed it to provide counter-cyclical fiscal and monetary stimulus. Thanks to the public sector’s and the economy’s net FCY creditor position, the government was also able to let the currency depreciate without triggering the systemic financial stress in the public, financial or corporate sectors related to FCY mismatches.

Brazil has come a long way indeed in terms of public sector debt sustainability, that is, solvency, liquidity and sensitivity to shocks. First, public debt is at very manageable levels. In net terms, public debt amounts to slightly less than 40% of GDP, compared to more than 60% in 2002. Second, its dependence on private foreign financing, whether international bond markets or foreign buying of domestically issued debt, is small to negligible. FCY-denominated/ -linked debt amounts to less than 3% of GDP. Non-resident holdings of domestic LCY debt amounts to a relatively modest 5% of GDP. Third, the aggregate (public- and private-sector) external balance sheet has removed the long-standing “fear of floating”. Thus a flexible exchange rate allows Brazil to absorb external shocks more easily than in the late 1990s and early 2000s when policy-makers were frequently forced to respond to external shocks with massive monetary tightening in an attempt to limit excessive currency depreciation, which, in turn, was threatening to further undermine debt sustainability. Today, a 10% nominal depreciation reduces the net public sector debt ratio by 1% of GDP. At the height of the global financial crisis in late 2008, net public sector debt had dropped by a full 5% of GDP, largely as a result of the public sector’s net FCY creditor position. In short, Brazil’s public debt position is quite solid.

Gross government debt vs. net public sector debt. After repeated financial crises in the 1990s and early 2000s, Brazil managed to put its fiscal house in order and reduce the public debt burden thanks to large primary surpluses and improved economic growth – and despite very, very high real interest rates of 10% or more. Brazil traditionally uses the public sector (PS) primary surplus to measure its fiscal effort. Following the return to low inflation during the second half of the 1990s, the public sector’s fiscal position deteriorated sharply due to a declining “inflation tax”. The primary balance was raised significantly in the wake of the 1998-99 financial crisis and again in the wake of the 2002 capital account crisis. The fiscal adjustment during the first year of the Lula administration was decisive in that it led to a near-continuous decline in net public debt, thus putting government debt dynamics on an unambiguously sustainable path.

Net public sector debt spiked to 63% of GDP in September 2002 at the height of the balance-of-payments crisis, mainly on account of sharply depreciated BRL and, somewhat less so, a widening fiscal deficit. Almost 50% of the public debt was linked to or denominated in foreign-currency or linked to the Selic rate (if one accounts for central bank swaps). Since then, debt has been gradually declining and today stands at just below 40% of GDP. The ratio even dipped to below 36% of GDP at the height of the 2008 financial crisis when – thanks to the public sector’s long FCY creditor position – the sharp devaluation of the BRL led to an improvement in the net public sector debt position.

Under the more widely used gross general government (GG) debt concept, which covers the federal government, states and municipalities, Brazil’s debt – under the “new” definition – has fluctuated in the 55-60% of GDP range since 2007. The “new” gross government debt definition as opposed to the conventional “old” definition only counts debt held by the public (including repos) as government debt, while federal debt securities held by the central bank are not counted as debt. Under the “old” definition, which counts central bank holdings of government debt as GG debt, this ratio declined from more than 80% of GDP in 2002 to 62% of GDP in July 2011.

Not only has gross government debt (under the new definition) remained largely unchanged, fluctuating in the 55-60% of GDP range since late-2006, but domestic debt has actually risen during this period from 51% of GDP to 54% of GDP. This happened in spite of, by Brazilian standards, strong economic growth, averaging 4.5% in real terms. The unchanged debt ratio is almost entirely attributable to the accumulation of assets financed by domestic debt issuance. Loans to the financial public sector (mainly: BNDES) and central bank foreign reserve assets account for the lion’s share of asset accumulation. Credit to public-sector banks increased from 0.5% of GDP at end-2006 to 7.3% of GDP in August 2011. During the same period, the stock of FX reserves grew from 8% to 15% of GDP. In dollar terms, the BCB accumulated more than USD 270 bn in additional FX reserves since 2007. Roughly speaking, leaving aside the question whether it has been financed by repos or Treasury issuance and excluding the “carrying costs”, this added 7% of GDP worth of debt to the gross domestic debt stock. All other things equal, gross domestic government debt would be roughly 15% of GDP lower than it is today, had the government not extended loans to official banks and not accumulated FX reserves since 2007. In other words, gross domestic GG debt would be closer to 40% of GDP had no further asset accumulation taken place.

There are drawbacks to public sector asset accumulation. The accumulation of public sector assets financed by public sector liabilities is worth having a closer look at. As regards FX reserve accumulation, the marginal economic and financial benefits of accumulation are, initially, significant in terms of lowering risk premia and foreign borrowing costs, including those of the private sector, and in terms of reducing economic and financial volatility (susceptibility to both endogenous and exogenous shocks). However, leaving aside the gains and losses resulting from currency depreciation/appreciation, the public sector incurs a financial loss. The interest rate paid on the domestic liabilities issued to finance FX reserve accumulation is greater than the interest received on its FX assets. The so-called “negative carry” is greater than 1,100 bp at the moment (roughly: Fed funds rate minus Selic). The public sector incurs losses, leaving aside currency-valuation-related gains or losses, worth 1.0-1.5% of GDP every year. Again, this can be considered an insurance premium paid to protect the economy against disruptive shocks and an indirect subsidy to the private sector in terms of cheaper external borrowing.

Similarly, as regards lending to public banks, the Treasury lends at TJLP (taxa de juros de longo prazo or long-term interest rate) but has to finance the loan at roughly Selic, resulting in a negative run rate of currently more than 500 bp. This roughly translates into a loss of more than 0.3% of GDP (roughly: 500 bp x 7% of GDP). The TCU federal watchdog has recently put the fiscal cost at 0.35% of GDP. True, to the extent that BNDES makes a profit by lending at a rate higher than TLJP and, ultimately, pays dividends to the Treasury, the effective loss (assuming zero credit losses incurred by BNDES) will be slightly smaller. Again, this holds in strictly financial terms and does not take into account the non-financial benefits in terms of broader, generally difficult-to-quantify growth and development objectives.

That said, in terms of FX reserves, the public sector looks “over-hedged”. The public-sector is a net foreign (currency) creditor to the tune of 10-12% of GDP. Foreign gross GG debt currently amounts to a mere 2.4% of GDP. This strongly suggests that the additional economic and financial benefit derived from additional FX reserve accumulation does not compensate for its fiscal costs. (This suggests that continued FX reserve accumulation is driven by other concerns.) A similar argument applies to the accumulation of Treasury claims on public-sector banks. Initially being a response to the credit crunch in 2008, public-sector-backed lending represented an economically and politically opportune way to support growth in the context of a severe credit crunch. Today the continued increase in lending comes at a tangible fiscal cost and it is arguable whether the non-financial benefits resulting from it offset the financial losses incurred by the Treasury. It would seem more sensible for official banks to raise the needed financing in the market rather than rely on the Treasury for funding.

Where one stands on this issue is largely determined by one’s assessment of the economic, non-financial benefits relative to the financial losses of both FX reserve accumulation and BNDES lending. It is incontrovertible, however, that this policy causes non-negligible financial costs and may even raise medium-term fiscal risks. The Treasury runs some (admittedly limited) credit risk on its loans. The BNDES loan portfolio is of relatively high credit quality.

But in a severe economic scenario, it would nonetheless likely suffer losses, which could indirectly impact the Treasury. While the public sector does not run credit risk on its FX holdings, currency appreciation does lead to accounting losses. On the flipside, of course, currency depreciation results in valuation gains. All told, too narrow a focus on direct financial costs and potential losses overlooks the potential “opportunity costs” of asset accumulation. By preventing a more sustained decline in domestic debt, this policy helps keep domestic interest rates at a level higher than under a scenario where (external) asset accumulation had not taken place.

Again, the stock of federal debt securities held by the public would be closer to 40% of GDP rather than the 53% of GDP, had the central bank stopped accumulating FX reserves in 2007 and the Treasury not extended financing to BNDES. A faster decline in domestic interest rates would not only have created “fiscal space”, enabling the government, if necessary, to respond to an economic slowdown forcefully. It would also have allowed for a decline in gross and net interest payments and, under the assumption that the government had refrained from increasing current expenditure, for higher public investment or more rapid deficit reduction. In addition to narrowing the “carry” differential and thus further reducing opportunity costs, it should also have led to less severe currency appreciation pressure earlier in the year.

The debt structure has improved tangibly. An improving outlook for public sector solvency and lower, but still very high (real) interest rates have also allowed the authorities to improve the debt structure and reduce the risks attached to it. As mentioned above, the vulnerability to exchange rate and interest rate shocks has diminished dramatically. In 2004, 1/4 of all federal public debt securities outstanding3 was FCY-linked or -denominated, rendering the debt vulnerable to exchange rate shocks. A very considerable 1/2 was linked to the Selic rate. By 2011, the debt structure had improved very dramatically. Today FX debt is only 4% of the total debt stock, while Selic-linked debt amounts to 1/3 of the total. This, as pointed out several times already, renders public debt significantly less susceptible to currency and interest-rate shocks, as the 2008 crisis demonstrated. Moreover, more than 40% of all domestic debt securities were maturing over the next 12 months in December 2002, or 17% of GDP. Today, less than 23% of all securities will mature over the course of the next 12 months, or less than 11% of GDP. This is still higher than the government’s deposits with the central bank and is high by emerging markets standards. But roll-over risk has undoubtedly declined over the past decade, even if gross financing requirements remain comparatively high.

In terms of the holders of government debt, a little more than 10% of total government debt securities are held by non-residents, and the vast bulk consists of holdings of LCY-denominated debt, whether domestic or external. This significantly improves the risk features from the creditor’s point of view by transferring FCY-related risks to the investor. Before 2005, there were, admittedly, next to no nonresident holdings of domestic government debt. But given that foreigners are predominantly invested in longer-term and/or fixed-rate instruments, financial risks from the Treasury’s point of view are low. By comparison, non-residents hold 50% of both German and US government debt securities (held by the public) due to their reserve currency status.

The structure of foreign government debt also improved. On the back of FX reserve accumulation, the public sector became a net foreign (-currency) creditor in 2006. Brasilia retired its IMF debt in 2005, repaid its Paris Club debt in 2006 and it pre-paid World Bank debt in 2009. Today the bulk of government debt is owed by the federal government in the form of international bonds. (It is worth mentioning that 10% of international bonds are denominated in BRL.) Starting in 2006, the Treasury has also been repurchasing international bonds.

Long story short. Not only has net public sector and, to a lesser extent, gross GG debt declined in recent years. But the risk profile of the debt has also improved markedly in terms of susceptibility to financial markets shocks. The 2003 fiscal adjustment unambiguously re-established public sector solvency and it subsequently allowed a tangible improvement in the debt structure. Were a financial shock equivalent to 2002 to occur today, the net debt ratio would fall by 3% of GDP rather than increase by 17% of GDP. It is worth pointing out, however, that, by the standards of the more advanced emerging markets, gross government financing requirements remain significant and the maturity of the domestic debt stock is of relatively short duration.

Net financing requirements (or fiscal deficits) have been on a declining path since 2003. The 2003 fiscal adjustment raised the primary surplus4 to 4.25% of GDP (pre-upward revision of GDP). Primary surpluses averaged 3.5 % of GDP in 2003-07 (following GDP revision). Thanks to sustained primary surpluses, declining inflation and accelerating economic growth, the nominal fiscal deficit fell from more than 5% of GDP in 2003 to a low of 1.3% of GDP in late 2008, before widening again in the wake of the 2009 economic downturn. The deficit will settle at 2-3% of GDP this year, subject to the authorities meeting their announced primary surplus target. The authorities project the nominal deficit to fall to zero within the next few years. This looks unlikely unless a very significant decline in net interest payments occurs. And this looks unlikely given continued low G3 rates and continued high domestic interest rates, the recent Selic rate cut notwithstanding. Declining deficits and the improved government solvency they imply were instrumental in lowering nominal and real interest rates. Although real interest rates remain very high by international standards, they are far below the levels seen during the second half of the 1990s and the first half of the 2000s.

The government is often criticised for frequent changes to the way in which it calculates the primary balance. It is fair to say that the changes thus far do not jeopardize government solvency. That said, the authorities do keep changing the way in which they measure the deficit and this certainly reduces transparency, unnecessarily so. 

First, in 2006 the IMF and the authorities agreed to allow to “abate” from the primary surplus a pre-determined amount for purposes of growth-enhancing public investment. This possibility was introduced

for purposes of IMF conditionality. If the government failed to meet the surplus target, it was allowed to subtract a pre-agreed amount of generally 0.25-0.75% of GDP without being considered in violation of its fiscal targets, provided it managed to implement the pre-agreed investments. Actually, this “abatement” was only used in 2009 and 2010 when the government had trouble meeting its primary surplus due, first, to the economic downturn and, then, to pre-electoral surge in public spending. The 2012 LDO, for instance, allows the government to “deduct” a portion (BRL 40.6 bn or 1% of 2011 GDP) of PAC investments from the primary surplus target.

Second, and more controversially, the government booked the Petrobras capitalisation worth 1% of GDP as revenue. In this asset-liability transaction Petrobras received the rights to explore oil (from the government) and the Treasury (via the BNDES and BSF) received shares and booked them as counting against its primary surplus target. Clearly, this has no impact on the underlying fiscal stance of the public sector and only helps embellish the reported 2010 deficit and thus bring it closer to the (revised) 2.5% of GDP target.

Third, the government has also resorted to using so-called “restos a pagar” (“leftovers to be paid”) to embellish fiscal outcomes. As the primary surplus in Brazil is still calculated on a cash basis, that is, expenditure is only appropriated when paid for, not when it is effectively executed, the primary balance may be raised in any given year by simply postponing the payment of expenditures. This has helped inflate primary surpluses somewhat on several occasions. 

Fourth, the authorities excluded Petrobras, and later Electrobras, from the public sector debt definition. This is legitimate to the extent that, while the government is a majority shareholder, it is not on the hook for its debt. That said, it would be difficult to envision a scenario where Petrobras were to get into financial difficulties and the government would not bail it out.

Fifth, in 2008 the authorities transferred 0.5% of GDP worth of primary surplus to capitalise the newly-created sovereign wealth fund (BSB). Whatever the merits of the sovereign wealth fund, which are somewhat questionable, this arguably reduces the transparency of fiscal policy and public debt. The funds were ultimately invested in Banco do Brasil and Petrobras shares. Last but not least, it is worth noting that in terms of accounting Treasury lending to the BNDES helps embellish fiscal outcomes. The PS primary surplus does not take into account interest paid (on securities issued to provide loans to BNDES) or received (sub-market TJLP received on loans from BNDES), but it does include dividend payments received from the BNDES. This embellishes the primary surplus performance, while adding to the overall balance. 

Overall, the various changes diminish transparency. At this point, none of these changes poses a risk to fiscal sustainability and continued debt reduction. While in some cases good arguments no doubt exist (excl. Petrobras from PS), in others the reasons are less convincing. Above all, these changes are unnecessary. Analysts do their own calculations adjusting their figures for the changes introduced by the Treasury. These changes only serve to reduce transparency. Moreover, if the government is confronted with a sharp economic slowdown (2009), nobody expects it to meet its primary surplus target. On the other hand, if the government is at risk of missing its target during an economic upswing (2010), changing the rules will fool nobody. The government would be better off modifying its fiscal rules by upgrading its target from a headline primary surplus target to a cyclically-adjusted target in the first place.

How will government debt evolve over the next few years? A back-of-the-envelope calculation suggests that the net debt ratio will continue to decline for the foreseeable future. An average real interest rate of 6% and a real GDP growth rate of 4% require a primary balance of only 0.8% of GDP to stabilise the debt ratio at its current level of 40% of GDP. The current primary surplus target is 3.2% of GDP. On current trends, net public sector debt will fall to 30% of GDP by 2017. This assumes an unchanged primary surplus, annual real GDP growth of 4% and average real interest rates of 5%. It also assumes an unchanged real exchange rate. This looks like a slightly optimistic scenario given considerable short-term infrastructure investment needs (World Cup 2014, Olympics 2016) and the political temptation to reduce the primary surplus as debt moves towards the 30% mark. Nonetheless, even under a more adverse scenario, net public debt would continue to decline. The key question is whether the authorities will be able to resist political pressures for greater spending (and a lower primary surplus) as the net debt ratio continues to fall.

Back-of-the-envelope calculations, while simple, overestimate the speed at which the net debt ratio will decline. Applying the average Selic rate to the net debt stock (or even its LCY component) underestimates interest payments. The average weighted interest rate paid on the net debt stock can be derived from the net debt composition. The PS pays an average effective nominal interest rate higher than Selic. Here is why. The PS receives (roughly) TLJP on its BNDES loans, while paying Selic. Leaving aside exchange rate changes, the public sector also faces a negative carry on its FX reserve holdings. On the other hand, 4.7% of GDP worth of debt is interest-free (monetary base). Adding all up results in an effective nominal interest rate about 200 bp higher than Selic, assuming a largely unchanged asset and liability structure and an unchanged real effective exchange rate.

In order to estimate the path of gross domestic government debt, it is necessary to make further assumptions pertaining to public sector asset accumulation, including its effect on real interest rates. Our

projections suggest that net public sector debt will fall below 30% of GDP by 2017. The evolution of gross GG debt will critically depend on the pace of FX reserve accumulation and further Treasury lending to the BNDES. We expect the pace of both FX reserve accumulation and (more so) Treasury lending to the BNDES to slow over the next few years. Assuming a nominal deficit of 2% and nominal GDP growth of 10% and PS asset accumulation worth 4% of GDP, we project gross GG debt (old definition) to decline to 46% of GDP by 2020. Assuming an unchanged fiscal stance, net debt could fall to as low as 25% of GDP by 2020. By comparison, the IMF projects gross GG debt (old definition) to fall from currently 67% of GDP to 57% of GDP by 2016 and net PS debt to fall to 33% of GDP by 2016 (vs our forecast of 31% of GDP).

Why are real interest rates stubbornly high? Debt dynamics will critically depend on real interest rate dynamics. This raises the (perennial) question as to why real interest rates are so high.5 Economic theory (let’s call it ex-post rationalisation) points to national savings that are very low or, which is the same, a national propensity to consume that is very high. If this is so, the solution would consist in a reduction in government consumption (and transfers). A credible medium-term adjustment combined with more immediate fiscal tightening should go a long way in reducing interest rates. Another explanation is to do with the reduced, but still considerable indexation of government debt and its relatively short maturity (for a non-reserve currency country). Practically, pushing out maturities against the backdrop of high real interest rates is financially costly. A fiscal adjustment would seem to be a promising avenue to lower interest rates in the wake of which maturities could be pushed out and debt de-indexed. The problem is perhaps not so much the size of the government deficit, which is relatively modest, but the structure of government spending and the residual risk attached to re-financing.

Comparing Brazil to Turkey is instructive. Brazil’s domestic LCY debt has increased as a share of GDP. It is therefore not entirely surprising that nominal and real interest rates on BRL debt have declined more slowly than the decline in the net debt ratio and the narrowing of the fiscal deficit might suggest. Brazil’s experience differs sharply from Turkey’s. Turkey similarly experienced a major financial and public debt crisis in the early 2000s. Thanks to a tight(er) fiscal policy and higher economic growth, it managed to bring down public debt more aggressively than Brazil. Crucially, Turkey reduced its net public sector debt to 30% of GDP from 70% in 2001, compared to Brazil where the comparable ratio declined from more than 60% of GDP to just below 40% of GDP. More importantly, Turkey’s TRY debt is relatively low at around 30% of GDP compared with more than 50% of GDP in Brazil. Real ex-ante interest rates are currently zero, compared with around 7% in Brazil. Zero real interest rates are probably not consistent with 5% inflation in Turkey, but the equilibrium real interest rate is no doubt significantly lower than in Brazil. A cursory comparison suggests that Brazil should make a greater effort to reduce its net public debt, ideally via a reduction in gross domestic BRL debt. A glance at GG gross interest payments shows that the savings potential stemming from lower interest rates is enormous.

Fiscal reforms would accelerate decline in real rates. While short- to medium-term debt sustainability is not a problem, the government nonetheless ought to implement wide-ranging reforms in order to create fiscal space and to reduce real interest rates more aggressively. It would be desirable to introduce a “structural” primary surplus rule, which could/should – once the indexation of government debt and annual financing requirements have declined further – be upgraded to a nominal balance target. During the first half of the 2000s, the government committed itself to a primary surplus due to the large share of indexed debt, which made interest payments difficult to predict and nominal deficit target difficult to meet due to the unpredictability of interest payments in any given year. It would be desirable to move towards a cyclically-adjusted primary balance target, allowing for tangible short-term deviations from its medium-term target.6 This would also make it unnecessary to resort to accounting changes in an attempt to embellish fiscal outcomes (in the short term). It would also prevent Brazil from being forced to pursue pro-cyclical fiscal policies.

From a more short-term and real interest rate perspective, the government should raise the primary surplus more aggressively by reducing current spending as a share of GDP. Federal primary spending has been rising inexorably over the past decade. In the short term, this may not even require large-scale, big-ticket reforms like social security and pension. However, politically, it is difficult to sell a tighter policy when debt is already declining and solvency concerns are not an issue. The government seems to have finally come round to this view. It remains to be seen if the government is able to meet its recently announced enhanced primary surplus target given the rise in the minimum wage, related fiscal spending and significant pressure to raise public sector wages. The need to raise investment spending won’t help, either, from a short-term adjustment perspective.

Brazil should also implement a medium- and long-term adjustment. Medium- and long-term adjustment would require relatively wide-ranging and politically difficult reforms. For example, both revenue and expenditure are very rigid (“earmarking”). The World Bank7 estimates that only 10% of expenditure and 20% of revenue are “free” (that is, not linked or earmarked). Much of the expenditure is indexed to the minimum wage, introducing a further element of rigidity. Brazil will face medium- and long-term spending pressure arising from social security and pension obligations. The political challenge is to muster the will to put forward reforms and win the large constitutional majorities necessary for their adoption. Sooner or later, these reforms will have to be introduced due to the projected rise in social security payments.

Brazil needs to accelerate the reduction in the public debt ratio, including the stock of domestic debt, in order to reduce interest rates. This will ultimately require both shorter-term fiscal adjustment and longer-term structural reforms. Gross domestic debt of no more than 40% of GDP would be highly desirable. This would roughly translate into a net debt level of 25%, deemed absolutely safe. This is where Brazil is headed in the short- and medium-term. It would nonetheless be highly desirable to accelerate this process. Real interest rate would adjust more quickly to levels seen in economies with similar public debt, investment and savings characteristics (e.g. Mexico), that is, real interest rates of no more than 2-3%, compared with 6-9% today. There is little reason to expect that Brazilian interest rates would be any higher provided it manages to reduce debt, improve the debt structure and in addition implement reasonable medium-term fiscal reforms aimed at both limiting long-term spending commitments and reducing government current spending.8 Brazil has shaken off its “country of the future (and always will be)” image. A further fiscal adjustment resulting in accelerated debt reduction and lower interest rates would at last make Brazil a “normal” country.