Developing countries are systematically underestimating the official size of their sovereign debt, an analysis of World Bank statistics has shown. Over the past 50 years, hidden sovereign debt has reached $1 trillion.

When Zambia and Chad applied for debt restructuring under the new G20 program in 2021, they faced an unexpected problem: it turned out that the countries did not have complete information about how much and to whom they owed. It took the financial advisers of the two states more than six months to collect all the necessary data.   The story illustrates the lack of transparency in sovereign debt: the public debt of low-income countries is difficult to quantify, either because the data is not fully disclosed in official statistics or because some of the debt is hidden due to confidentiality provisions, explained Marcelo Estevan, Global Director of the World Bank Group's Macroeconomic, Trade, and Investment Practice.   In 2021, the World Bank estimated that 40% of low-income countries had not published any data on their debt at all for more than two years. The data of the countries that published and the independent estimates of the multilateral development banks differed, and in some cases the gap in these estimates reached 30% of GDP. And all this against the backdrop of growing debt problems: a 2023 World Bank report notes that one out of every four developing countries is already effectively cut off from international capital markets, and over the past three years, the number of sovereign defaults in these countries has exceeded the previous two decades.   Hidden sovereign debt has been $1 trillion since the 1970s, World Bank economists and co-authors estimated in a recent study. This is more than 12% of the total borrowing for this period of all 146 developing countries included in the sample. Hidden debt is highest in countries with weak institutions, but even emerging economies with relatively strong institutional performance systematically underestimate their debt reporting.  Hidden trends  The calculation of the amount of hidden debt was made possible by analyzing the revision of sovereign debt estimates from various editions of the World Bank's debt statistics. All countries that have obligations to the World Bank must report to it on the amount of their debt, otherwise they risk losing access to financial assistance from this international institution. The World Bank aggregates statistics and checks them for consistency, but the data itself receives only from national authorities. Any revision of data for a particular year means that the country has reported a debt that it has not previously disclosed. It is the data on such revisions that economists are guided by: by comparing statistical releases for different years, they can estimate the amount of previously undisclosed debt for each year and highlight the specific moment at which it still falls into the statements.   The sum of all such debt disclosures gives the total measure of a country's hidden debt. These include government debt, debt of state-owned companies and state guarantees for private obligations. The estimate of $1 trillion is minimal, since not all hidden debt is revealed over time, the authors say.   Hidden debt is not necessarily due to a deliberate violation of debt reporting, the authors clarify. Some countries simply do not have the "capacity" to stably track their own debt (for example, data from state-owned companies); That is why hidden debt is high in countries with weak institutions.   The authors found that hidden debt:  It is extremely common: about 70% of all statistics on debt volumes and about 50% of data on debt flows (new borrowings) are revised at least once after the first publication. On average, as a result of revisions, the volume of debt increases by 1% of GDP, and the annual rate of new borrowing by 0.7% of GDP. Procyclical: hidden debt tends to accumulate during economic booms and is revealed mainly during periods of crisis, often in the context of the need to obtain assistance from international creditors. Thus, the recession caused by the coronavirus pandemic was accompanied by the largest disclosure of debt in 50 years. It corresponds to higher losses of creditors in the event of default and a longer duration of the defaults themselves.  Most often, the problem of undervaluation of hidden debt concerns bilateral loans and commercial loans in non-bond form: these instruments are not traded on the secondary market and are often accompanied by confidentiality agreements. With regard to debt to the World Bank or private bondholders, there are only small adjustments: statistics on tradable bonds are generally freely available, and the World Bank conducts its own monitoring of its loans.   The tendency to detect hidden debt in times of crisis indicates that external monitoring is the driver of this disclosure: in order for a country to receive financing from international institutions or to restructure its debt, it must provide access to its debt statistics to calibrate program parameters, and creditors tend to carefully assess the debt burden of the economy.   During a restructuring, sovereign bondholders fear that hidden debt may dilute the liquidation value of their claims, that is, the amount of funds they can receive in the event of a borrower's bankruptcy. Therefore, the lack of transparency in the country's debt indicators leads to a prolongation of the sovereign debt crisis.   For example, during the restructuring of Zambia's debt in 2020, the fears of private investors significantly slowed down the settlement of its debt. Creditors rejected the government's offer of debt relief (the authorities hoped for a six-month freeze) precisely because of the opacity of Zambia's external debt, especially to China, and the insufficient, in their opinion, guarantees that all creditors would be on an equal footing. Zambia became the first African state to default during the pandemic. A few months later, the country's authorities confirmed that investors' fears were justified by revising the debt statistics - the new president of the country then admitted that "the hole turned out to be larger than expected." In earlier World Bank statistics, Zambia's debt was $12.5 billion, but in later data, it was $3.2 billion more, a difference that corresponds to 14% of the African country's GDP. The restructuring of Zambia's debt has not yet been completed.  Increased debt intolerance  The existence of hidden debt affects the cost of borrowing – to compensate for investors' uncertainty about the real amount of debt, borrowers must pay higher interest rates. Disclosure of debt as a result of monitoring efforts reduces uncertainty, which, accordingly, reduces borrowing costs. However, if the volume of disclosed debt is sufficiently high, then an increase in the total amount of debt does not lead to a reduction, but to an increase in the cost of borrowing, which exceeds the cost reduction from reducing uncertainty. In this way, large hidden debt increases the incentive for the government to default, which needs higher levels of revenue to make this option suboptimal.   Since hidden debt worsens the conditions for attracting borrowed funds for the country, it reduces welfare, the authors write. They estimate that if this debt becomes public information, the economy can sustain higher debt at a lower cost of servicing it – on average, with a welfare gain of 5.5% of consumption growth. This is due to the fact that an economy with complete information has a much higher debt capacity than an economy with hidden debt, the model built by the authors showed; And higher debt in a well-informed economy leads to a much higher level of average consumption. However, again, this gain in the case of hidden debt is possible with a low level of hidden debt. Economists conclude that policies to increase debt transparency have a better chance of success in "good times" – this avoids the negative effects associated with debt disclosure against the backdrop of a crisis.   The phenomenon of hidden debt may explain one of the economic "mysteries": why developing countries have often found themselves in defaults or debt crises at levels of sovereign debt that, by all accounts, should be manageable. In a 2003 paper, Carmen Reinhart and Kenneth Rogoff, then at the IMF, and their colleague Miguel Savastano described the phenomenon as "debt intolerance" and compared it to "lactose intolerance" and linked it to the history of serial defaults that many countries had gone through over the previous two centuries. Reinhart and her co-authors then estimated that "safe" levels of external debt for countries with high "debt instability" could be extremely low – 15-20% of GNP.   However, this "intolerance" may not be due to past defaults – which often do not prevent a country from borrowing more – but to frequent and significant upward revisions of debt levels, according to the authors of the study from the World Bank and their colleagues. The "intolerability" that manifests itself at low levels of reported observed debt may actually be a reaction to high levels of debt, but a latent one. At the same time, the procyclical nature of accumulation and the identification of hidden debt can intensify boom-bust cycles, increasing the amount of resources available to countries in good times and creating additional difficulties in bad times, economists conclude.