A recent report by a working party of Robert Triffin International documented the major financial stability risks created by the post-Global Crisis surge in dollar liquidity (RTI 2019). The worry then (in December 2019) was the same as that of Acharya et al. (2015) – that an eventual rise in dollar interest rates would hurt non-US companies which had borrowed dollars. “The risk of an unexpected and unplanned reversal of abundant global liquidity,” it warned, “hangs over the world economy.” The dollar bonds issued by emerging market economy (EME) corporates seemed most at risk.
Covid-19 upended this scenario. At first, the worries seemed vindicated. Companies faced a drop in dollar earnings, and there was a flight from EME financial assets. But then the Fed eased monetary policy quickly. The ‘world’ long-term rate – which the RTI report took as a price indicator of global liquidity – hit new lows.1
Figure 1 shows a decomposition of the ‘world’ long-term interest rate into the long-run expectation of the short-term interest rate and the term premium.2 It is the drop in the long-run expectation of future short-term rates (currently at about 1.5%) which has driven the decline in ten-year yields in recent months. Markets have received the message that major central banks will keep interest rates low for very long. But markets have not yet received the message about fiscal policy. Despite the large rise in government debt/GDP ratios everywhere, the term premium remains near record lows. Perhaps investors believe that governments will lean on central banks to keep long-term rates down no matter how bad the fiscal position.
Figure 1 Decomposition of the ten-year world yield
By mid-2020, money had begun to flow back into EME bonds and equities. Given even lower long-term rates, many EME corporations issued more dollar debt. As Wall Street sets out its ideas for 2021, the Financial Times noted on 25 November that EME paper is on top of the list. Does this mean that the earlier worries about the impact of an eventual reversal of global liquidity on EME companies were exaggerated? The answer is ‘no’: corporate solvency risks have increased and many firms are vulnerable to tightening of dollar lending conditions.
The dangers of non-US dollar debts
The RTI report documented several risk-prone characteristics of recent international borrowing which could be highly destabilising. The dollar debt of EME companies was only the most striking example.
The first reason to worry is the scale of the global expansion in international foreign currency debt of non-banks.3 Following the Landau Report (BIS 2011), the BIS developed this measure of global liquidity, which was first presented to the G20 in April 2013. After a sharp contraction after the onset of the global crisis, there was a strong rise in foreign currency debt relative to world GDP, indicating increased leverage (Figure 2). This has continued since the pandemic and may have accelerated. By mid-2020, US dollar credit to non-banks outside the US exceeded $12 trillion, more than 14% of world GDP – up from less than 10% of world GDP in 2007.
The second reason is that most of the increase in foreign currency credit has come from bond issuance rather than from banks. Low long-term rates (and a negative term premium) have encouraged EME companies to borrow long term in international capital markets rather than from banks, a trend reinforced by the post-Global Crisis determination of regulators to force banks to de-risk. The many diverse channels of capital market intermediation create systemic risks which have always been difficult to manage. Risks have become opaquer, have migrated to less regulated entities, and have often been hedged in ways that could magnify contagion.
Figure 2 International foreign currency debt of non-banks (as a percentage of global GDP)
Source: BIS Global Liquidity Indicators
Such risk migration has made EME bonds especially vulnerable to liquidity shocks. Asset managers use bond funds to acquire what they imagine to be a diversified portfolio of EME bonds. Such funds, often offering a daily price even when the underlying assets are illiquid, create an illusion of liquidity. Investors (especially when leveraged) may be more prone to herd, making markets even more volatile.
The third concern is the domination of dollar-denominated credits, which is more extreme than before the global crisis. Because the share of debt denominated in dollars is much larger than the share of trade with the US (or dollar-based economies), a rise in the dollar against third currencies such as the euro adds to the burden of foreign debt without necessarily improving international competitiveness. Devaluations may therefore be contractionary: increased local currency payments on dollar debt reduce disposable income and may lead companies to cut investment.
The fourth worry is the fragility of foreign exchange hedging markets. Many EME companies will hedge their long-term dollar debts with short-term instruments (e.g. three-month swaps), counting on being able to renew them easily. But in periods of financial stress, when the demand for dollars usually rises, the terms of such hedges tend to turn against firms which are short dollars. Sudden changes in pricing, or even the disappearance of foreign exchange hedging instruments for some currencies, can destabilise indebted companies.
Larger currency exposures have led to greater dependence on foreign exchange derivatives. These markets have grown rapidly but liquidity strains have become more frequent.4 One indication of such strains is the widening spreads (the cross-currency basis) in foreign exchange swaps markets. A good proxy for global foreign exchange hedging pressures is the IMF’s calculation of the median of 22 currencies (Barajas et al. 2020). Before 2007, this median was close to zero (indicating that covered interest parity was generally satisfied). Since the global crisis, however, the dollar cross-currency bases of many currencies have become more volatile. Figure 3 shows that the median widened in the March 2020 Covid-19 panic to 55 basis points (70 basis points for emerging market currencies). A dollar liquidity crisis was averted only by the rapid activation of the Fed’s dollar swap lines with other central banks. Barajas et al. (2020) estimate that such action narrowed the cross-currency basis for the currencies of swap line countries, but not for those of non-swap line countries.
Figure 3 Marginal dollar funding costs for non-US banks (three-month cross-currency basis)
Currency mismatches of EME corporates
The financial stability threat from increased foreign currency borrowing by EME corporates depends on the size of their currency mismatches. It is not enough to look only at aggregate international bond debt. Account also needs to be taken of: (a) foreign currency assets which have also risen strongly, (b) any offsetting reductions in other forms of foreign currency liabilities (such as bank loans), and (c) foreign currency earnings (exports).
In the absence of corporate sector data, Chui et al. (2016) calculated currency mismatches for the non-official sector (in which companies have a large weight). Figure 4 shows, for a sample of major medium-sized EMEs, aggregate net foreign currency assets – that is, foreign currency assets minus all foreign currency liabilities – as a percentage of exports. The dotted line shows that the country aggregate is still positive, not negative as it was in the late-1990s. Hence a currency depreciation still improves the local currency value of country’s external balance sheet, and so reinforces the stabilising effects of currency depreciation on the current account.
Figure 4 Currency mismatches in medium-sized EMEs
However, this ratio has fallen over the past decade, so aggregate balance sheet protection is smaller than before the global crisis. Vulnerability is much greater because of the large net debts in foreign currencies of the non-official sector (continuous line in Figure 4), which reached around 40% of annual exports by end-2017. The scale of dollar-denominated debts, the component which gets most attention, varies substantially across geographic areas (Figure 5). China’s dollar debts rose to 20% of exports by end-2019 (6% at end-2008). The ratio was much higher in South East Asia (70%, up from 24%) and higher still in Latin America (106%, up from 70%).
Figure 5 Dollar-denominated debt as a percentage of exports
There are three reasons why large net foreign currency debts of companies could have systemic consequences:
- Destabilising foreign exchange market dynamics. Chui et al. (2016) provide evidence that many firms without dollar revenues had borrowed dollars without fully hedging their foreign exchange exposures. Indeed, many firms had invested the proceeds of the dollars they had borrowed in higher-yielding assets denominated in their own local currency (Bruno and Shin 2020). Companies with unhedged dollar debts tend to buy dollars (directly or by purchasing hedges) whenever the local currency comes under pressure in foreign exchange markets. This can set off a destabilising market dynamic: a drop in the exchange rate makes their dollar debts even harder to service, inducing further dollar purchases and thus a still weaker currency.
- Magnification of corporate fragility from leverage. Higher leverage has increased the financial fragility of EME companies. Alfaro et al. (2019) show that such fragility is magnified by foreign currency debts, especially for firms in non-tradable industries. This can have macroeconomic consequences as firms with dollar debts cut business investment (Avdjiev et al. 2019).
- Damage to local banks. If companies find it harder to borrow dollars abroad, they may react in ways that transmit the shock to local banks. They may activate under-priced credit lines and squeeze out other borrowers. They might reverse the carry-trade borrowing dollars and cut their wholesale local bank deposits. Banks may be hit just when a downturn has already impaired their loan book.
A new challenge for monetary policy
When domestic companies have large dollar debts, the exchange rate constraint on monetary expansion tightens. At the onset of the pandemic, many EMEs faced a classic quandary: how to ease monetary policy to counter recession without triggering a large currency depreciation which might cripple companies with dollar debts. Cutting the policy rate beyond a certain point in such circumstances might be contractionary – akin to the reversal rate of Brunnermeier and Koby (2019) but arising from the damage done to corporate balance sheets.
The strategy followed by some central banks (e.g. Indonesia, the Philippines, South Africa) was to keep up the policy rate in order to protect the currency but at the same time to buy government bonds. New legal charters were introduced to allow some Latin American central banks to buy public and private securities in secondary markets. Many central banks used quantitative easing (QE) without reaching the zero lower bound on interest rates. This change in monetary policy implementation has favoured more government borrowing (Forni 2020).
The remarkable development of local financial markets, often with a deeper local investor base, has given EME central banks new possibilities for balance sheet policies. Foreign investor demand for government bonds denominated in local currency was stimulated by a decade of low yields on advanced economy government bonds.
The drawback is that the greater importance of EME local currency bonds in the portfolios of foreign investors has also increased the foreign exchange market/bond market interactions, magnifying the domestic consequences of external financial shocks. Foreign investors learnt from the 2013 taper tantrum that an EME currency crisis often goes hand-in-hand with a bond market crisis (Carstens and Shin 2019). Foreign investors without foreign exchange hedges are thus doubly exposed.
The larger stock of government bonds and other financial assets that trade in open markets nevertheless means that the central bank asset purchases can be more ambitious. In addition, QE can be supported by official measures (such as regulatory relaxation, offering investors hedges which put a floor under future bond prices, and so on) to encourage local banks and other domestic investors to buy government bonds that foreigners were selling.
Measures to remove the tail risk of a bond market collapse help to stabilise the domestic financial system because bonds serve as a safe asset for banks and as reliable collateral for borrowing. They also re-assure foreign investors, and so support the exchange rate. This challenges the orthodox view that QE tends to weaken the exchange rate. The sharp rises in EME bond spreads in March 2020 have been decisively reversed, exchange rates have risen, and many corporates have issued more dollar bonds.
New policy frameworks in three areas are required to address the financial risks of much higher EME corporate debt in dollars. None of these will be easy.
First, repeated episodes of bond and derivatives market turbulence point to serious gaps in the international regulatory framework covering bond markets. The strong warning by the chairman of the Financial Stability Board (FSB) about the underlying fragility of the more diverse and interconnected non-bank sector globally suggests that regulatory action is finally high on the international policy agenda (Quarles 2020). There seems to be strong political support from a US Treasury under Yellen.
Second, macroprudential policies in borrowing countries - focused almost entirely on banks - have often failed to cover the risks from excessive leverage as well as currency and maturity mismatches created by non-bank financial institutions. In 2017, the ECB Vice-President warned that new financial crises would be inevitable unless macroprudential policies covered capital markets more effectively (Constâncio 2017). EME central banks should develop new macroprudential tools to discourage the risky borrowing strategies of non-financial companies and to contain any collateral damage to domestic banks.
Finally, EME central banks face more complex monetary policy choices. On the one hand, their ability to fight recessions by cutting their policy rate is constrained by the substantial dollar debts of their companies. On the other hand, larger and deeper domestic financial markets have made QE a more effective policy tool both to stabilise the domestic banking system and to support aggregate demand. How well QE works depends on the fiscal position. It will also depend on confidence that a credible and independent central bank will keep inflation low and avert a flight to dollars or euros. Too often in EMEs, alas, government attempts to hobble the central bank have made crises worse.
Acharya, V, S Cecchetti, J de Gregorio, S Kalemli-Ozcan, P Lane and U Panizza (2015), “Emerging economy corporate debt: the threat to financial stability”, VoxEU.org, 5 October.
Alfaro, L, G Asis, A Chari and U Panizza (2019), “Corporate debt, firm size and financial fragility in emerging markets”, NBER Working Paper 25459, January.
Avdjiev, S, V Bruno, C Koch and H S Shin (2019), “The dollar exchange rate as a global risk factor: evidence from investors”, IMF Economic Review, 5 February.
Bank for International Settlements, BIS (2011), “Global Liquidity – concept, measurement and policy implications”, CGFS Papers 45.
Barajas, A, A Deghi, S Fendoglu and Y Xu (2020), “Strains in offshore US dollar funding during the COVID-19 crisis”, MCM Analytical Notes 20/01, IMF.
Brunnermeier, M and Y Koby (2019), “The reversal interest rate”, IMES Discussion Paper Series 19-E-06, Institute for Monetary and Economic Studies, Bank of Japan
Bruno, V and H S Shin (2020), “Currency depreciations and emerging market corporate distress”, Management Science 66(5).
Carstens, A and H S Shin (2019), “Emerging markets aren’t out of the woods yet”, Foreign Affairs.
Center for Global Development and Inter-American Development Bank, CGD-IDB (2020), “Sound banks for healthy economies”, A report of a CGD-IDB working party chaired by A Powell and L Rojas-Suarez, October.
Chui, M, E Kuruk and P Turner (2016), “A new dimension to currency mismatches in the emerging markets: non-financial companies”, BIS Working Papers 550
Constâncio, V (2017), “The evolving risk landscape in the euro area”, Speech at Banco do Portugal conference on Financial Stability, 17 October.
Forni, L (2020), “Private and public debt interlinkages in bad times”, Journal of International Money and Finance 109.
Hördahl, P, J Sobrun and P Turner (2016), “Low long-term interest rates as a global phenomenon”, BIS Working Papers, no 574, August.
Robert Triffin International, RTI (2019), “Managing global liquidity as a global public good”, RTI Paper no 11.
Quarles, R K (2020), “The FSB’s roadmap for addressing non-bank financial institution vulnerabilities”, Speech of the Vice Chairman for Supervision of the Federal Reserve Board of Governors, 20 October.
1 The simplest price indicator of global liquidity in the past was dollar LIBOR because short-term international bank lending in dollars was a large component of capital movements which reacted quickly to changes in global financial conditions.
2 This is derived from a macro-financial model and is based on a principal components analysis of movements in dollar, euro, and sterling government bond markets (updating Hördahl et al. 2016).
3 International foreign currency debt or credit is the sum of bank loans plus outstanding international bonds.
4 According to the BIS’s OTC derivative statistics, the notional outstanding for foreign exchange contracts had reached almost $94 trillion by mid-2020, with $26 trillion in currency swaps.