Issue #2: November/December 2019

Chinese overseas lending, Sri Lanka

First off, apologies for missing the edition which should have come out in late October/early November. Family issues and travel prevented me from putting one out. I’ve also held the release of this issue back a few days to comply with the UCU strike. The next issue should be out late December/early January and hopefully we can keep to a regular monthly schedule from then on.

Right now my own research is mainly focused on official Chinese loans to other governments, which are the main mechanism used to finance Belt and Road projects around the world (this sometimes gets confused, with the loans being called ‘aid’, or ‘investments’). This month I thought I’d share a few bits and pieces of what I’ve learned so far- first on the topic of Chinese loans in general, and then on how this ties in to recent developments in Sri Lanka, the case that has been perhaps the most controversial of all when it comes to China’s lending practices.


The idea of ‘debt trap diplomacy’ is at this point pretty much the standard in much English-language media as a way to frame Chinese lending. This is the charge that China is acting like a global loan shark- knowingly lending unsustainable amounts, waiting till countries can’t pay and then using the debt as leverage to gain political loyalty and/or the surrender of key assets. ‘Debt trap diplomacy’ was only coined in 2017, by Brahma Chellaney of the Centre for Policy Research in New Delhi (described here as a ‘polemicist’ by Chas Freeman, who was the US interpreter when Nixon met Mao in 1972).  Since then, talk of debt traps has become something of a meme, appearing in headline after headline about Chinese official lending (here’s a small sample from the FT, Al Jazeera, the Washington Post, Quartz, Deutsche Welle, and the Straits Times). It’s also become a mainstay of the Trump administration’s increasingly critical rhetoric on China, with Mike Pompeo, John Bolton and Mike Pence all repeating the debt trap accusation in the past year.

In the last few months, a new and slightly different charge has also dogged Chinese lending- that a large proportion of it is hidden, to the extent that ‘…outside of the government in Beijing, nobody knows much about where th[e] money has been invested and what conditions and risks are attached…’, as a recent Spiegel article put it. There have been worries about BRI’s transparency from the start. But fears have been amplified with the publication this Summer of a high-profile paper by economists Sebastian Horn, Carmen Reinhart and Cristoph Trebesch. They claim to have uncovered US$200 billion of previously unknown debts to China, owed by governments of emerging markets and developing economies (EMDEs- essentially countries of the global south, plus a few others in places like Eastern Europe).

The idea of hidden debts brings back bad memories of the late 1970s and early 80s, when a boom in private bank lending to governments, especially those in Sub-Saharan Africa and Latin America, turned into a major debt crisis. Back then, when countries started struggling to repay and began calling in the IMF, it was discovered that several owed more than had been publicly revealed. Today, the outlook for EMDEs does appear uncomfortably similar to the late 70s in some respects, with overall debts reaching comparable levels. Though it does not seem likely to do so any time soon, if the US raises interest rates at some stage, the cost of borrowing for EMDEs will rise, making it much harder for countries to manage their debt burdens. This is what happened from 79-82, with catastrophic results.


But how worried should the world be about Chinese lending in particular? In first place, the debt trap idea can be more or less safely filed away as US propaganda at this point, even if it’s not hard to see why it has gained such prominence. The vast majority of EMDE loan deals with China come through one of the two main state-owned ‘policy’ banks- China Development Bank (CDB) and China Export-Import Bank (CHEXIM)- and are effectively treated as official government to government lending from the Chinese state. Assessing exactly how much is being lent and what for is tricky, because neither bank provides full details of its overseas loans (though then again, neither do many other G20 governments’ official lending institutions). One recent estimate puts CDB and CHEXIM’s combined overseas loan portfolio at $671 billion, equivalent to about one and a half times that of the World Bank’s total.

Lack of solid information has allowed space for all kinds of speculation as to the nature and purpose of Chinese loans. Compounding this, they often seem to go to places and projects that don’t make much commercial sense, at least in the view of some experts and traditional players in global finance. If some of these loans don’t look a good bet economically for the policy banks, then it’s not hard to make the logical leap to assuming there must be some other (usually nefarious) purpose involved- and that’s exactly what many analysts and journalists have done, which has then been seized upon by others for the sake of geopolitical point scoring.

It’s absolutely true that some Chinese-financed projects have been, at the very least, economically questionable, from railways in Venezuela and Kenya to motorways in Montenegro. Nevertheless, evidence that the policy banks are actively trying to turn bad loans into strategic investments, by swapping debt relief for control of assets in borrowing countries, remains thin on the ground. There are many rumours circulating about loan deals with secret clauses or, for example, the unconfirmed claim that Tajikistan gave up a slice of territory to China in exchange for debt relief in 2011. But the single concrete case of debt trap diplomacy invariably cited in the coverage is the Hambantota Port in Southern Sri Lanka- and on closer examination even that turns out to be mostly hype (more on which below).

There are a number of factors which, in combination, provide a more plausible explanation for the pattern of Chinese lending than the debt trap claim. It’s definitely true that, unlike private creditors, direct profit from loan deals themselves is only one part of what CDB and CHEXIM are trying to accomplish through their loan activities abroad. As Yiping Huang puts it,

“There are three types of [Chinese] external finance: commercial finance which is purely based on market principles and is conducted in pursuit of profit maximization; policy finance which is conducted for national strategies, not for profit; and development finance which is conducted for national strategies but which is commercially sustainable.”

Almost all the money lent to EMDEs comes under the banner of development finance. What this means is that while CDB and CHEXIM are expected to be financially self-sustaining over the long run, maximising returns is often not the sole concern behind every loan deal. As for the ‘national strategies’ Huang mentions, these can be both economic and political. On the economic side, CDB especially tends to plan not in terms of single loans, but packages of related deals. The idea is to boost the recipient country’s economic growth and tax revenue overall, so that borrowers will be able to make payments even if an individual project or two loses money.

In some instances the goal is to create new markets practically from scratch, which then might be profitable destinations for Chinese trade and investment. More directly, loan deals usually stipulate Chinese firms as the main contractors tasked with building whatever the loan is for (whether a road, power plant, stadium or airport). In this way, loans are a means to push and subsidise the global expansion of China’s domestic companies. Beyond this, they support the export of capital goods like railways, nuclear plants and renewable energy, where China is trying to establish itself as a global leader- and a setter of standards in these industries.   

Additionally, there is undoubtedly a political or strategic element involved to many loans, with a willingness to extend credit more freely and for a wider range of purposes to strategically important states like Pakistan, as well as to China’s oil suppliers such as Venezuela (at least until the recent collapse). Decisions to grant loans are often simply driven by requests from governments (which may in turn be the result of lobbying from Chinese companies angling for contracts)- and the degree to which a particular project is prioritised by a borrowing government often (though not always) seems to trump economic feasibility (or social and environmental concerns) when it comes to Chinese willingness to lend. 

Because there is a tendency in some quarters to view BRI as some grand Machiavellian scheme, western observers especially often seem to implicitly rule out the possibility that overly risky loan deals might just be the result of bad decisions. But there’s plenty of evidence of this happening. One way Chinese lenders have tried to balance risk in some cases is via commodity-backed loan deals, whereby the proceeds from a proportion of the borrowing country’s commodity exports are pledged to pay down the debt. In places like Venezuela, Ecuador and Angola, the idea is that no matter what financial stress these countries are facing, oil production will always be sufficient to meet loan payments- so tying loans to oil sales is a way of trying to guarantee repayment. But as the Venezuelan economy has collapsed in recent years under the pressures of a commodity bust, economic mismanagement and US sanctions, so too has the capacity of state petroleum firm PDVSA to keep the oil flowing. Contrary to the debt trap thesis, some scholars argue that China’s huge exposure in Venezuela (more than $60bn in loans) has actually resulted in a creditor trap- that is, far from debt being a lever for Chinese influence in Venezuela, China has found itself having to keep lending defensively in order to keep the Venezuelan government afloat in the hopes of salvaging repayments on the original loans. There are smaller scale examples from elsewhere, too. In Ukraine, for example, a $1.5 billion grain-backed CHEXIM loan to the State Food and Grain Corporation of Ukraine ran into problems when grain shipments meant to fund repayments amounted to only a fraction of the required amount. Exact details are hard to come by, but it does not seem as though Chinese authorities were able to force Ukraine to make full repayments.

Last, it’s important to remember that the decision-making around which projects get loan financing is quite fragmented across the various arms of the Chinese state- and some of those bodies (including firms lobbying for contracts) have an interest in pushing for deals, even if they present risky prospects. There are some indications that, probably in response to some of the failures, there’s been an effort to tighten up debt sustainability concerns recently.


The key to the ‘hidden loans’ charge is understanding what the word ‘hidden’ means in this context. China is not a member of some of the major organisations which monitor international finance (such as the OECD and the Paris Club), so the loans it makes to other governments via the policy banks do not get systematically recorded (and neither the Chinese government nor the banks release comprehensive information either). Various teams of researchers have attempted to shed light on this area, using a combination of media reports, official documents and on the ground research to estimate bilateral Chinese loans (and sometimes aid) in Africa, Latin America and the Pacific, as well as globally. Working from these sources (plus a few others)  Sebastian Horn, Carmen Reinhart and Cristoph Trebesch have compiled their own estimate of how much 106 EMDEs owe China. They then compare this total with the combined figures reported by the same countries to the World Bank and find a huge gap, of $200bn in hidden debts to China which are not officially declared.

Frankly, I’m sceptical of these claims. First, some of the media reports on the paper have suggested that the authors uncovered previously unknown debts. But since almost all their data is complied from publicly available sources, it’s not as if any of it was some closely guarded secret only known to a select few inside the walls of Zhongnanhai, as you might think from the coverage.  

I’m really at the beginning stages of research in this area, but so far I’ve taken a close look at the numbers for four Central/Eastern European states (Montenegro, Belarus, North Macedonia and Ukraine) and thought I’d give a preview of my findings here (I’ll be putting out a paper on this in future). Essentially, I think that Horn, Reinhart and Trebesch have overestimated debts to China in each case. Now, obviously, even assuming I’m correct here, this does not necessarily mean the same holds for all the other countries they examined. But I raise it because my experience of looking at these four states in detail highlights some of the common problems encountered by pretty much any attempt to figure out how much countries owe to China- problems that tend to push such attempts towards overestimation.

Because there is no central repository of Chinese overseas loans, most of the efforts to compile estimates rely to varying degrees on press reports. The problem with this is that quite a high proportion of deals announced in the press don’t actually turn into loans. Sometimes early-stage talks between a government and Chinese authorities about a loan are reported as though they were done deals. In other cases, talks result in a framework agreement with a high price tag attached (as I discussed in the October newsletter in relation to Iran). These documents are essentially wish lists of possible projects for which funds could potentially be made available- but quite often are mistaken for concrete deals. In other instances loan deals are cancelled, scaled down or simply misreported. Last, even when loans have been agreed, the money is usually handed over in instalments, meaning that a country’s debt will increase gradually over time, rather than the full amount of the loan being added the moment a deal is agreed (as is often assumed).

Horn, Reinhart and Trebesch suggest that countries might be underreporting their Chinese loans to the World Bank because of inaccurate accounting- whether this is due to lack of statistical capacity (especially in poorer countries) or a deliberate attempt to hide debts. But in the four cases I looked at, it was possible, with a bit of digging, to find quite a lot of detail on amounts owed to China- and broadly this matched with figures reported to the World Bank. One of the most useful sources of data here have been bond prospectuses. When governments want to issue bonds for sale to international investors they are obligated to compile a prospectus containing usually a very detailed rundown on their economic situation- including a breakdown of their debts. While there’s no guarantee that the information is always correct, both the country concerned and the banks which help them prepare the report can be liable if the prospectus is inaccurate, so there is a big incentive for them to get it right. For example, here below is a screenshot from a prospectus issued by the government of Belarus- hardly a country known for openness and transparency.

Comparing this kind of information against the various existing estimates of Belarus’ Chinese debts, as well as media reports, provides insight into where some of the likely inaccuracies are coming from. For instance, the ‘Great Stone’ referred to in the list above is the name of a Chinese-financed industrial park on the outskirts of Minsk. Early reports on this project claimed that Belarus would be borrowing $3bn for its development- which, on the face of it, seems like an awful lot for an industrial estate. In fact, the $3bn figure probably relates to part of a $15bn framework agreement of the kind described above- an amount earmarked to fund future loans for possible ventures within the park, rather than the amount actually borrowed at that point. Other media reports concur with the $170m figure for the park itself, seemingly extended with another $110m loan after this bond prospectus was published. However, attempts to estimate Belarus’ Chinese debts (including Horn, Reinhart and Trebesch) have apparently taken the $3bn figure at face value, leading to an inflated total. 

In other cases across the four countries I looked at, I’ve been able to find debt details in government documents, audit reports of public enterprises, and, in one instance, even CHEXIM letters detailing the full schedule of payments for a major loan to Montenegro.  All this leads me to believe that the Chinese policy banks themselves do not generally treat their overseas loans (or their repayment terms) as matters of any great secrecy. The fact that media reports often overestimate rather than underestimate loan amounts adds to the sense that there is no systematic attempt to hide or minimise loans going on here. There may be some exceptions, but generally it appears that the onus is on receiving governments themselves to publish as many or as few details of their Chinese loans as they see fit. If this is the case, it would provide an opening for civil society groups concerned about transparency, in that pressuring their own governments to release figures will often be a more realistic task than demanding the same from Beijing.

Relying on governments to be transparent in every case of course presents its own problems. It is more likely that the kind of discrepancies Horn, Reinhart and Trebesch point to might occur in low income states where statistical capacity is less developed and where it may be easier for various parts of the state to conceal how much they have borrowed, and from whom. In Zambia, for instance, there is confusion over the country’s debt situation, including the possibility that some Chinese loans have been contracted by the Presidency without going through the required process of parliamentary approval. But it is important to remember that the one major confirmed case of hidden loans in recent years- in Mozambique- had nothing to do with China. In 2013-14 three quasi-state-owned firms liked to the Mozambican security services borrowed $2bn via Credit Suisse, Russia’s state-owned VTB Bank and via the European bond market. In theory, the money was supposed to fund a new tuna fishing fleet and security infrastructure to protect the country’s coastline. But more than half of the funds were kept secret from everyone but a small circle within the government and security forces, until coming to light in 2016 via the Wall Street Journal. Mozambique had to endure the withdrawal of support from most lenders and donors as a consequence, causing economic crisis. What happened to most of the money is still not entirely clear, but several figures from both the banks and the Mozambican government have been indicted by the US Justice Department.

Credit Suisse Headquarters in Zurich. Source: Thomas Wolf via Wikimedia Commons

So, if in the near future we enter an early-80s style debt crisis and it turns out some countries have larger debts than previously thought, it seems at least as likely that any secret loans will have come from commercial sources as from Chinese policy banks. This points to a larger issue with the way government debt (especially in poorer countries) is currently being discussed. There is no doubt that debts to China are significant in many countries- and that some of these funds have been both lent and borrowed in reckless, perhaps corrupt ways. But the fascination with Chinese loans in particular across media and policy circles has tended to obscure the much greater threat to economic stability presented by debts to private investors. According to the World Bank, in 2018 external debts to bondholders among low- and middle-income countries stood at $1.39 trillion. That’s more than double the amount of government-to-government loans (including debts to China, but also to all other states), even if Horn, Reinhart and Trebesch’s $200 billion in supposedly missing Chinese debts is included.

Because interest rates in the global north have been so low in recent years, investors have had to look further afield to find investments with higher returns. One consequence is that a much larger range of states- even low-income countries- have found it relatively easy to find investors willing to lend them money. Just as with Chinese loans, this can be a very good thing in terms of providing much-needed finance for development. But, as the Mozambique example shows, we shouldn’t assume that borrowing from these sources is necessarily more transparent, nor that it is subject to any particularly stringent tests in terms of whether the funds are being lent and borrowed sensibly or with the public good in mind. Commercial lending also tends to have higher interest rates- and unlike loans, bonds have to be paid back in one go when they become due.


On 18 November Gotabaya Rajapaksa was inaugurated as the Sri Lanka’s new President. Rajapaksa was defence minister during the Sri Lankan civil war and campaigned as a hardliner on security, arousing concern among both Tamil and Muslim minorities. But a good deal of international coverage has also been framed around the notion that Rajapaksa’s victory will mean a return of closer ties between Sri Lanka and China.

Gotabaya’s brother Mahinda Rajapaksa (appointed Prime Minister in the new government) previously served as President between 2005 and 2015, borrowing perhaps $5bn from China following the end of the civil war, when traditional lenders were less keen to provide funds. Mahinda’s tenure also saw a $1.4bn investment from China Communications Construction Company in building Port City, a land reclamation project on which an ambitious new financial district is slated to be constructed- and which has been the subject of worries over environmental damage.

Land reclamation at Colombo Port City. Source: Rehman Abubakr via Wikimedia Commons

Of all Mahinda Rajapaksa’s Chinese deals, the most infamous is the series of loans (totalling around $2bn, according to Aid Data) taken on to build the Hambantota Port in the country’s south. Later, in 2018 and with Sri Lanka facing a severe debt crisis, President Maithripala Sirisena handed operational control of the facility to China Merchants’ Port, receiving $1.1bn in return. For this reason Hambantota is often pointed to as the key (some would say sole) example of the ‘debt trap diplomacy’ discussed above. Details vary, but the basic outline as usually reported is that the port was never economically viable, but that kickbacks provided an incentive for Sri Lankan officials to borrow money from China to have it developed. Years later under Sirisena, with the Hambantota facility finished but nearly empty, and the country struggling to make loan repayments, the Chinese government forced Sri Lanka to hand over the port in exchange for debt relief. As to why a Chinese firm would want an apparently useless white elephant of a port, this is where many reports offer groundless speculation that Hambantota might eventually be used as a Chinese naval base. 

Ships docked at Hambantota Port. Source: Deneth17 via Wikimedia Commons

While all this makes for a good story, the truth seems to be more mundane. For starters, it’s not clear that Hambantota always looked like a nonsensical venture, with a range of feasibility studies going back to 2001 reaching mixed conclusions on commercial prospects. And while the port still isn’t profitable today under Chinese management, business does seem to be picking up. Either way, the most charitable interpretation of the project is that it was envisaged as part of reconstruction efforts in an area devastated by the 2004 Tsunami. A less generous version would be that the port (along with a surrounding industrial zone, plus a Chinese-built motorway and an airport, now empty) were pet projects for President Mahinda Rajapaksa in his home district (the port now bears his name).

It’s absolutely true that by 2016 Sri Lanka was mired in debt, with the government struggling to meet repayments. Chinese loans were an important part of this burden (around 10%), but hardly the overwhelming factor one might think based on some of the media coverage. Hambantota-related loan repayments only amounted to about 5% of the total. In this context, and with Hambantota Port losing money, Sirisena’s administration did a deal in 2017 by which CMPort would effectively get a 70% share on a 99 year lease for the facility, in exchange for $1.12bn. The deal is usually reported as a straight swap- control of the port for writing off the debts associated with it. But this is not accurate, since most of the money went towards paying off (much larger) debts to private investors, while the Hambantota loans remain on the books to this day. Deborah Brautigam instead calls the agreement a privatisation, making the comparison with the Greek port of Piraeus (now majority owned by China’s COSCO). But the Hambantota deal was not quite a straight privatisation, either- while protests around the site led to a reduction of CMPort’s share from 80 to 70 percent, the company successfully demanded control of a much-expanded industrial district, which, assuming the land is fully developed, will displace hundreds of local people in the process. 

One spin off from the debt trap idea has been a tendency to interpret Sri Lankan politics as partly being divided by attitudes to China, with the return of the Rajapaksa family to power signalling a likely turn back towards the PRC. But while Sino-Indian rivalry no doubt plays a role, Sri Lankan politics has been far more complicated in recent years for a simplistic framing in these terms to make much sense. 

In practice, the Sirisena administration did not mark a break with the PRC. On gaining power in 2015, Sirisena initially suspended work on most Chinese projects in the country, including the Port City development, before allowing them to restart after only a few months (and with some minor changes to terms). Sri Lanka also continued to borrow from China, including around $1bn earlier this year for a new motorway. There was also another $1bn loan in 2018, with this one earmarked to help repay other loans. That is significant, because such deals are rare, with China usually only lending to fund specific projects rather than to ease government finances, except in strategically important countries (Venezuela and Pakistan being other examples).

All this seems to fit something of an emerging pattern internationally. In Argentina Mauricio Macri came to power in 2015 with the intention of cancelling Chinese-financed hydropower projects in Patagonia, but could only negotiate a reduction in their scale. In Pakistan, Imran Khan’s presidency began with attempts to agree an overhaul of the China-Pakistan Economic Corridor, which quickly turned into an extension of existing agreements, with a few nods to Khan’s priorities in areas such as agriculture. What Sri Lanka, Argentina and Pakistan all shared during these episodes- and in contrast to the likes of Malaysia which was arguably more successful in its attempts to scale down BRI projects- was high levels of indebtedness. Importantly, debts to China are significant, but not dominant, in all three countries (the IMF has Chinese loans as making up 25% of Pakistani debt, for example). But with projects already underway and China presenting a relatively easy option for new funding, governments under the strain of heavy debt burdens have found it more advantageous to keep Beijing onside.

Interestingly, and seeming to confound the predicted turn back towards China, a few commentators on the recent Sri Lankan election have noted that the manifesto of Gotabaya Rajapaksa’s party, Sri Lanka Podujana Peramuna, included a pledge to renegotiate the terms on the Hambantota Port deal. It seems pretty unlikely that the Chinese side would offer anything more than symbolic concessions here, but again this may be more evidence of a pattern. In some countries Chinese influence is now a salient political issue- and it makes electoral sense for politicians of all stripes to at least gesture towards standing their ground, or pledging to conduct some sort of audit into the deals signed with China under previous governments. That such politicians often do not then make many substantive changes to the country’s relationship with China after coming to power suggests that, while the ‘debt trap’ story might be demonstrably false, there is no doubt that Chinese loans can be a source of influence in some countries- and one which lasts beyond changes of government.