China’s overseas lending and the looming developing country debt crisis
While China’s dominant role in global trade is well-understood (e.g. Autor et al. 2016, Acemoglu et al. 2014), its growing influence in international finance has remained relatively obscure, mostly due to lack of transparency and data.
The chronic problem of lack of transparency has assumed a more urgent role since the COVID-19 pandemic has all but paralyzed the global economy. Many developing countries now face deep financial strain and have approached the IMF for emergency lending in record numbers. In mid-April, the G20 agreed to a temporary debt moratorium on bilateral (government to government) debt, calling on private investors to follow suit. However, there is considerable uncertainty about the magnitude, timing, and modalities of China’s intended debt relief, as China remains reluctant to join the Paris Club of official creditors.
Figure 1 illustrates China’s overseas lending boom using official balance of payments data. Outstanding debt claims from direct loans and trade advances alone have risen from almost zero in 2000 to more than 1.5% of global GDP in 2019. This surge in lending has financed many projects in infrastructure, mining, and energy, but we know little beyond those aggregate numbers.
Figure 1 China’s overseas lending boom
Sources: Horn et al. (2019), PBoC BoP Statistics and IMF.
China’s opaque lending
China’s lending is opaque for several reasons:
The Chinese government has not published a break-down of its overseas lending activity.
China is not a member of the Paris Club or the OECD’s Creditor Reporting System that share data on bilateral lending and trade credit flows.
Commercial data providers and credit rating agencies focus on private bank lending and bond offerings.
Many Chinese loans go to state-owned enterprises and special purpose vehicles whose debts are often not properly reported, thus complicating transparency. According to the IMF, fewer than one in ten low-income countries report debts of public corporations that are outside the general government. As a result, the debtor countries may themselves have an incomplete picture of how much they have borrowed from China and under which conditions.
In Horn et al. (2019), we address these data shortcomings and construct a database of Chinese loans with corresponding debt service and debt stock estimates by country and year. Our ‘consensus dataset’ is granular and was gathered from an extensive list of sources, including international treaties, debt contracts, policy reports, as well as the work of academics such as the AidData team at William and Mary (Dreher et al. 2017). We collated details on 5,000 loans and grants extended by the Chinese government and state-owned creditor agencies since 1949, to more than 150 countries worldwide, with total commitment amounts of $520 billion.
The paper and database have contributed to the ongoing debate on the true scope and nature of Chinese lending. A recent contribution by the IMF/World Bank (2020), for instance, suggests that we may be over-estimating debts owed to China. Yet, as detailed in the Appendix to this note, we demonstrate that the opposite is likely to be true; our debt estimates should be regarded as a lower bound.
Three features of Chinese lending to developing countries
First, almost all of China’s overseas lending is official, in the sense that it is extended by the Chinese government and various state-owned entities. This makes China by far the world’s largest official creditor, with outstanding claims in 2017 surpassing the loan books of the IMF, World Bank and of all other 22 Paris Club governments combined.
Second, the data show that China tends to lend at commercial terms, meaning that the average interest rates on China’s claims is close to those in private capital markets. Other official entities, such as the World Bank, typically lend at concessional or zero interest rates, and longer maturities. In addition, many Chinese loans are backed by collateral, meaning that debt repayments are secured by revenues, such as those coming from commodity exports.
Third, debt to China has been accumulating at a very rapid pace in some countries. For the 50 most exposed countries, we estimate that debt owed to China has increased from less than 1% of debtor country GDP in 2005 to more than 15% in 2017. A dozen of these countries now owe debt of at least 20% of their nominal GDP to China. Figure 2 illustrates the geographic distribution.
Figure 2 External debt to China: End-2017 (excludes portfolio asset holdings and short-term trade credit)
Source: Horn et al. (2019).
From hidden loans to hidden debt restructuring
Another of our key findings is that China has extended many more loans to developing countries than the official debt statistics suggest. We compare our database to an extract of the World Bank’s Debtor Reporting System, which forms the basis of the most widely used debt databases published by the World Bank and the IMF. Specifically, we identify Chinese commitments to developing country public sector recipients that have not been reported to the World Bank.
We find that about 50% of China’s lending to developing countries goes unreported, meaning that these loans and the corresponding debt stocks do not appear in the ‘gold standard’ data sources. As of 2016, the unreported lending from China has grown to around $200 billion. While unreported loans are low for the median country, they create sizeable discrepancies to official debt statistics in at least two dozen developing countries.
The ‘hidden debts’ to China can distort the views of the official and private sector in several ways. Official surveillance work is hampered when parts of a country’s debt are not known. As noted earlier, in the context of emergency debt relief efforts, assessing repayment burdens and financial risks requires detailed knowledge on all outstanding debt instruments.
The private sector will likely misprice debt contracts if it fails to grasp the true scope of debts that a government owes. This problem is aggravated by the fact that many Chinese official loans have collateral clauses, so that China may be treated preferentially in case of repayment problems. The upshot is that private creditors may be underestimating the risk of default on their claims–including during the unfolding pandemic crisis.
Apart from the debt database, our paper also documents the substantial number of credit events/debt restructuring of Chinese loans, which have not appeared in the reports of international credit rating agencies. Since 2011, two dozen developing countries have already restructured their debt to China (see Figure 3). While the documentation is also scanty and opaque, we can conclude that China’s approach to debt restructuring has been dominated by debt maturity extensions with outright debt cancellations being limited to the (very few) zero interest loans.
Figure 3 Restructurings and debt write offs: Chinese overseas loans versus private external debt
Sources: Authors’ calculations and Cruces and Trebesch (2013), Dreher et al. (2017), Hurley et al. (2018), Kratz et al. (2019) and Meyer et al. (2019).
Echoes from an earlier boom-bust: The 1970s-1980s cycle
China’s overseas lending boom shares many characteristics with some earlier lending booms, most notably with the 1970s episode in which Western banks lent record amounts to developing countries. Like China’s counterparts of the past 20 years, the borrowers of the 1970s were developing and resource-rich countries, many of which had not been able to access international capital markets for decades. Then and now, the purpose of lending were investments in infrastructure, resource extracting industries, or financing other potentially growth-enhancing projects. Moreover, the bulk of the lending of the 1970s was in US Dollars, maturities were rather short, and interest rates implied risk premia, very much in line with the terms of most of China’s lending. Moreover, much of the bank syndicated lending was not picked up by official statistics and transparency was a major issue.
The boom ended in a wave of sovereign defaults in the early 1980s, after commodity prices began their collapse in late 1979, and new lending dried up. Most of these defaults were resolved only after protracted negotiation, resulting in a ‘lost decade’ for dozens of countries. The wave of defaults of the 1930s followed a similar grim boom-bust pattern.
China’s overseas lending has started to slow recently, in tandem with their domestic economic slowdown and the decline in global commodity prices. Compared to the onset of the Global Crisis during 2008-2009, for example, initial conditions and debt and growth fundamentals were weaker for China and emerging markets by end 2019, even before the emergence of the pandemic.
The COVID-19 crisis has fuelled a synchronous global recession, a crash in commodity prices (alongside a historic collapse in oil prices), and a reversal of capital flows to developing countries. These reversals have unfolded at a speed and on a scale that recalls the antecedents of the very worst earlier debt crises.
The historic contraction in China’s GDP in the first quarter of 2020 is in line with the ‘sudden stop’ and capital outflows recorded for developing countries in March and April. As international lending (including Chinese) continues to dry up, many governments will face severe problems in rolling over their existing debts and finding alternative sources of financing. The number of sovereign credit rating downgrades have already skyrocketed in 2020 with new ones posted on almost a daily basis. New defaults and restructurings usually follow. Dealing with the precarious debt situation in so many developing countries simultaneously is at the forefront of global policy initiatives.
Figure 4 China is the largest official creditor to developing countries
Note: This figure shows aggregate public debt to different official creditors for all 122 developing and emerging market countries contained in the World Bank International Debt Statistics (excluding China). Debt to China is estimated by Horn, Reinhart and Trebesch (2019). Debt to all 22 Paris Club governments is taken from the Paris Club website (available since 2008). Debt to the IMF and the World Bank Group (IBRD plus IDA) is from the World Bank’s International Debt Statistics.
To date, China has restructured its debt in a piecemeal approach (case by case, involving only the debtor government) while the Paris Club governments, the World Bank, the IMF, and private investors have taken part in coordinated and comparatively well-documented debt relief initiatives. However, today, any effort to provide meaningful debt relief to the most vulnerable countries must encompass the debts owed to China, as Figure 4 highlights. Furthermore, there is urgent need for transparency in the present circumstances, as governments and private creditors demand assurance that the debt relief is channeled to confront the COVID-19 pandemic and not to repay pre-existing debts to China.
Autor, D, D Dorn and G Hanson (2016), “The China Shock: Learning from Labor Market Adjustment to Large Changes in Trade”, Annual Review of Economics, 8, 205-240.
Atkins, L, D Brautigam, Y Chen and J Hwang (2017), “Challenges of and opportunities from the commodity price slump” CARI Economic Bulletin #1. China Africa Research Initiative, Johns Hopkins University School of Advanced International Studies, Washington DC: CARI.
Bandiera, L and V Tsiropoulos (2019), “A Framework to Assess Debt Sustainability and Fiscal Risks under the Belt and Road Initiative”, World Bank Policy Research Working Paper 8891.
Cruces, J J and C Trebesch (2013), “Sovereign Defaults: The Price of Haircuts”, American Economic Journal: Macroeconomics 5(3): 85–117.
Dreher, A, A Fuchs, B Parks, A M Strange and M J Tierney (2017), “Aid, China, and Growth: Evidence from a New Global Development Finance Dataset”, AidData Working Paper 46.
Gallagher, K P and M Myers (2019), “China-Latin America Finance Database”, Washington DC: Inter-American Dialogue.
Horn, S, C M Reinhart and C Trebesch (2019), “China’s Overseas Lending”, NBER Working Paper No. 26050.
Hurley, J, S Morris and G Portelance (2018), “Examining the Debt Sustainability Implications of the Belt and Road Initiative from a Policy Perspective”, CGD Policy Paper 121.
International Monetary Fund and World Bank (2020), “The Evolution of Public Debt Vulnerabilities in Lower Income Economies”, Policy Paper No. 20/003. Washington DC: IMF and World Bank.
Kratz, A, A Feng and L Wright (2019), “New Data on the ‘Debt Trap’ Question”, Rhodium Group.
Meyer, J, C M Reinhart and C Trebesch (2019), “Sovereign Bonds since Waterloo”, NBER Working Paper No. 25543.
Appendix: Why our estimates of debts to China are conservative (i.e. a lower bound)
Our paper has contributed to the debate on the true scope of Chinese lending. A recent report by the IMF and World Bank (2020), for example, uses data from 14 low income countries and suggests that our debt stock estimates may be too large and “that any hidden debt may be smaller” than what we report. The main point raised in the report is that our estimates are largely based on loan commitments rather than disbursements, which could lead to an upward bias.
However, despite using loan commitments, our numbers are conservative in comparison to official numbers and estimates by other researchers, including from a recent World Bank survey.
Since reliable disbursement data is not available for most countries, our data, like those of others, are indeed based on (adjusted) commitments. When disbursements fall short of commitments, this can lead to an overestimation of outstanding debts.
To tackle the committed-but-not-disbursed problem, we dropped transactions such as cancelled projects, projects that are still in an initial pledge phase, and large framework agreements for which we do not have detailed project and loan specifications. This excluded a significant list of projects. In addition, we have conducted benchmarking exercises, to check whether we are over-estimating the scope of Chinese lending and outstanding debt. Since our data is granular, we can compare our estimates both in aggregate to balance of payments data, but also at the level of individual Chinese creditor banks.
Figure A1 summarizes these benchmarking exercises, which are also discussed in the latest draft. The main take away is that our estimates are substantially below comparison figures and can thus be considered a lower bound of the magnitude of Chinese overseas lending.
Figure A1 Our conservative (lower bound) estimates viz official figures and others
First, our estimated debt stocks are lower than the few available comparable official numbers. As of end-2017, the aggregate PBoC data on China’s International Investment Position showed outstanding overseas lending claims of $637 billion. This is more than 50% higher than our aggregated external debt stock estimate for end-2017, which is close to $400 billion. Another example is given by the overseas loan claims of China Development Bank, which released an aggregate figure in its 2015 report. In that year CDB reported outstanding overseas loans of $210 billion, of which we only capture $127 billion (by adding up all debts to this creditor). Our tally on commitments is also below those reported by state-owned but commercial creditor banks such as Bank of China and ICBC, which have become large overseas creditors in recent years, but for which very little systematic data collection exists.
Second, our database yields a lower volume of lending than those of other researchers and policy institutions, who focus on individual regions. For example, Figure 1 shows that our aggregate commitment numbers of Chinese loans to Africa and Latin America are lower than the aggregates reported by the widely used scholarly databases by SAIS CARI (Africa database by Atkins et al. 2017) and the Inter-American Dialogue (Latin American database by Gallagher and Myers 2017), respectively.
Third, our loan data yields lower estimates than those presented in a World Bank survey of Belt and Road countries (BRI) as reported by Bandiera and Tsiropoulos (2019). They estimate a total of $300 billion in public and publicly guaranteed Chinese loan finance to 50 BRI countries (between 2013 and 2018), while we identify only $105 billion for those same countries and a slightly shorter time span (2013 to 2017).
We conclude from these comparisons that our estimates are not too high, as suggested in the IMF/World Bank report, but if anything, too low. Despite our best efforts to merge data from multiple sources, we still miss substantial amounts of Chinese overseas lending.