The surge in Tunisia foreign debt: causes and possible ways out

Abstract

Over the last decade (2010–2020), Tunisia foreign debt has experienced a phenomenal jump. It has increased at a faster pace than domestic debt. It has doubled without any significant positive effect on investments or economic growth. This study aims to identify the major causes of this abnormal increase in Tunisia foreign debt. Since the model variables are not all stationary at level, the paper applies the autoregressive distributed lag technique to quantitative quarterly economic data covering the 2009–2020 period. Findings show that the economic growth, the exchange rate, the current deficit, the coverage ratio, and the lagged foreign debt itself are the major causes leading to the phenomenal increase of Tunisia foreign debt. Based on these findings, the paper suggests boosting foreign currency-generating activities, containing current public expenditures, stabilizing the exchange rate, and making structural economic adjustments as possible escape roots.

Supplementary Information

The online version contains supplementary material available at 10.1007/s43546-023-00443-2.

Introduction

Tunisia foreign debt, which is basically public, has phenomenally increased over the post-revolution period (2010–20). Foreign currency-denominated debt records an overall growth rate of 250% and an annual average growth rate of 14%. This growth is significantly higher than the total debt growth, which records an overall growth rate of 148% and an average rate of 10%. The foreign public debt records the highest overall growth, about 285% (these statistics are calculated from Table 4 in the Online appendix). External debt distress is indeed an international fact and not specific to Tunisia. Indeed, the World Bank (WB) (2020) states in its international debt statistics report that prior to the onset of the COVID-19 pandemic, rising public debt levels and heightened debt vulnerabilities were already a cause of concern. However, even if we consider this fact, the Tunisian external debt remains exceptionally high and an issue of great concern. Compared to the average of all low- and middle-income countries, Tunisian external debt statistics are very high and stand as an outlier. For instance, from 2009 to 2019, the Tunisian external debt to gross national product ratio jumped from 55 to 101%, compared to an average increase from 22 to 26%. Meanwhile, the Tunisian external debt to exports ratio leaped from 113 to 191%, compared to an average increase from 85 to 107%. This proves that even if Tunisia external debt trend is in line with the international one, the huge increase in external debt is much more related to national economic and political factors and local events than to international economic conditions. Over the same period, the Tunisian Dinar (TND) exchange rate exhibited a steady downside movement against the United States Dollar (USD) as well as against the Euro (EUR). At the end of 2019, the TND recorded an overall depreciation rate of 95% against the USD and 63% against the EUR. Accordingly, the exchange rate is suspected to be one of the main causes of the phenomenal increase in foreign debt: the currency exchange effect explains about 40% of this increase.

Therefore, the foreign debt problem can be considered one of the most urgent issues in Tunisia, considering a continuous weakening of economic and financial fundamentals, a steady downgrade of the sovereign rating, and an increasingly worrying country risk. The challenge for the Tunisian authorities today is to curb this phenomenal increase in foreign debt. Therefore, we can claim that the main contribution of this paper lies essentially in the vitality of the issue it deals with in the studied country (Tunisia). It aims to counter the surge of Tunisia foreign debt by identifying its principal economic and financial causes and proposing some possible ways out.

The paper aims to identify the major causes of the steady increase in Tunisia foreign debt during the last decade. For that, the paper adopts a quantitative analysis. The first step in the analysis process is an extensive literature review intending to identify the factors that can affect the evolution of Tunisia foreign debt. The second step is data collection. Data consists of annual and quarterly time series of Tunisia foreign debt and other factors that can affect it, published essentially by the Central Bank of Tunisia (CBT). The annual data give preliminary ideas about external debt patterns and their predictors. Afterward, an econometric analysis is applied to quarterly time series to detect the main factors that have led to the explosive growth of the Tunisian foreign debt and to find possible ways out. The quarterly dataset has two advantages: Check for short-term relationships and get long time series. Given that the time series are not all integrated of order zero (not stationary at level), the study applies autoregressive distributed lag (ARDL) modeling.

Accordingly, the first section of the study is devoted to the foreign debt phenomenon by studying its causes, effects, and solutions. The second section provides an overview of the evolution of Tunisia foreign debt over the post-revolution decade. It presents some preliminary statistics about Tunisia foreign debt and related variables that might have influenced it. The third section consists of an empirical study of the factors probably leading to the explosive growth of Tunisia's foreign debt. The fourth section discusses the results and proposes some recommendations. The last section presents some concluding remarks.

The foreign debt constraint

The foreign debt constraint, named by Hausmann and Panizza (2003)‘Original Sin’, means the inability to borrow in domestic currency for the long term and even for the short run in the international market. It is a common problem even for well-developed countries. It takes a lot of time to change and has a seriously negative impact on the economic and financial stability of the countries affected. In fact, over the last decades (since the financial crisis of the late 1990s), the ability of emerging economies to borrow outside in their local currencies has significantly improved (Engel and Park 2022; Bertaut et al. 2022). Du and Schreger (2016) state that local currency debt has become the primary form of financing for many emerging economies. However, this trend did not last long, and global investors have started moving away again from local currency debt in recent years (Bertaut et al. 2022).

The causes of the foreign debt constraint

There is extensive literature on foreign debt constraint determinants. Even though Artus (2003) classifies them into macroeconomic and microeconomic causes, this paper adopts an exogenous–endogenous classification to separate controllable causes from uncontrollable ones.

The principal exogenous cause is the size of the country to which we can add a lender-side cause named by Carstens and Shin (2019) ‘original sin redux’. Hausmann and Panizza (2003) find that among seven explanatory theories, the absolute size of the economy is the only variable robustly correlated with the original sin, a result recently confirmed by Engel and Park (2022). It is more correlated than the level of development, institutional quality, monetary credibility, or fiscal solvency. Hausmann and Rigobon (2003) explain this size effect by the international transaction costs: large countries have an advantage in shedding the original sin because the large size of their economies and currency issue make their currency attractive as a component of the world portfolio. By studying the case of European countries in the nineteenth century, Flandreau and Sussman (2002) attribute the size effect to the large presence of European countries in international trade.1 With this large presence, they were able to avoid the original sin quite independently of the quality of their institutions because there emerged spot and future currency markets in their currencies, arising out of the demand by traders to hedge their exposures. Whatever the reason, size is an important determinant, but it cannot be the only one. Switzerland is an obvious counterexample. As for the original sin redux, borrowing from global investors in domestic currency transfers exchange risk from borrowers to lenders: the appreciation of the lenders’ national currencies undermines the value of their loans (or investments). To mitigate this risk, lenders reduce their transfers denominated in the local currency of the host countries (Bertaut et al. 2022).

In addition to the exogenous causes, the literature on foreign debt constraint determinants presents a wide range of endogenous causes. For multiple reasons, the underdevelopment of the domestic financial markets, namely reflected in market incompleteness and information asymmetry, presses borrowers to borrow in foreign currency. In an incomplete market, the limited inter-temporal choices prevent the use of domestic currency to borrow for the long term. To avoid maturity mismatch, borrowers borrow in foreign currency (Eichengreen and Hausmann 1999). In a non-efficient market marked by high information asymmetry, foreign debt may be used either as a tool to reduce agency costs and increase efforts or as a signal. Entrepreneurs borrow abroad in foreign currencies as a signal that they have a good quality (Artus 2003) or as an incentive to increase efforts. Governments of emerging markets use foreign debt as an incentive to more discipline because it makes the depreciation of their currencies more costly and reduces the temptation to inflate their debts (Artus 2003; Jeanne 2003).

The exchange rate policy can also increase the dependence on foreign debts. Given that emerging countries are inherently unstable, a floating regime choice produces high exchange rate volatility, which increases the interest rate differential. High domestic interest rates attract hot money and increase the likelihood of the original sin (McKinnon and Pill1 1999). The controversy is that a peg may not be a good solution because fixed exchange rates create moral hazard and lead to excessive foreign currency borrowing (Burnside et al. 2001). Moreover, since a fixed exchange rate is generally achieved at the expense of a stable interest rate, risk-averse borrowers and lenders would then prefer to use foreign currency debt denomination even in the absence of the moral hazard problem (Chamon and Hausmann 2002). Lee (2022) finds empirically that emerging market sovereigns borrow even more in foreign currency when exchange rate volatility is higher, precisely when it is riskier for them to do so. Increased exchange risk drives investors to seek refuge in world currencies. Furthermore, the causality direction is ambiguous, and there is a possible reverse causality between foreign debt and exchange rate stability. Hausmann et al. (2001) and Calvo and Reinhart (2002) show that emerging market countries do not tolerate exchange rate flexibility because of the ‘‘fear of floating’’ caused by the original sin problem.

The Monetary policy credibility may also be another cause of foreign currency financing. Engel and Park (2022) show that a country with weak monetary policy discipline (with a low cost of inflation) wants to borrow more in local currency, but it is obliged to borrow mainly in foreign currency due to enforcement constraints. In the same vein, Ottonello and Perez (2019) attribute the gradual dissipation of the original sin, to a great extent, to the prolonged expansion and stabilization of inflation.

The lack of law enforcement and the weakness of the institutions that address commitment problems are other endogenous causes of foreign debt constraint. Eichengreen and Hausmann (1999) argue that foreign creditors are unlikely to lend to a sovereign in its currency if it can manipulate the exchange rate and, therefore, the real value of the external debt. Myers and Majluf (1984), Fan et al. (2012), and Demirguc-Kunt and Maksimovic (1999) find that short-term debts are likely to dominate in highly corrupted economies because they are safer than long-term debts (the domestic side of the original sin). In addition, as stressed by Poirson(2001), Kimakova (2008), and Russell (2011), countries characterized by political instability and a low degree of accountability use the exchange peg as a commitment mechanism and a sign of credibility. Overall, corruption and weak law enforcement make it difficult to borrow for the long term in domestic as well as in foreign capital markets in domestic currency. So, governments are constrained to borrow from outside in foreign currencies to meet their needs in foreign currencies or to avoid the maturity mismatch problem. Nevertheless, while the exchange rate peg can resolve the foreign currency reserves problem, at least provisionally, it increases the hot money problem, fails to solve the maturity mismatch, and creates a currency mismatch.

Except for the economy size and the lender-side effects, the other causes, such as the economic and financial infrastructure, monetary and exchange policies, law enforcement and institutional robustness, are endogenous. Some of them are structural problems and their resolution takes a while. Therefore, it is unexpected that Tunisia, a small and underdeveloped country, can quickly mitigate all these causes and completely unload the foreign debt burden. The study of these root causes of the foreign debt problem has in fact a less ambitious goal. That is to identify the leading factors of the steady increase of the Tunisian foreign debt because a high level of foreign debt is harmful and can have serious adverse effects.

The adverse effects of foreign debt

Foreign currency indebtedness has destabilizing effects: it causes financial fragility (Eichengreen and Hausmann 1999) and domestic economic instability (Eichengreen et al. 2002). Under an original sin situation, incomes and capital flows become more volatile (Eichengreen et al. 2002). Limiting the recourse to foreign-currency-denominated debt cannot be a solution because it undermines economic growth. Countries in this situation face a Hobson’s take-that-or-take-none choice. In addition to the capital flows volatility, financial fragility is unavoidable because foreign debt produces a currency mismatch and sometimes a maturity mismatch (Eichengreen and Hausmann 1999). Countries with external liabilities denominated in foreign currencies cannot hedge, and their firms cannot match the maturity and currency structures of their assets and liabilities.

Foreign currency debt also has a negative wealth effect. The real depreciation of the exchange rate increases the external debt service (and consequently), creating a balance sheet effect when residents’ assets and cash-flows are denominated in domestic currency (Bleany and Ozkan 2011). For firms, the increase in foreign debt reduces profits and net present values, lowers firm values, and makes banks more reluctant to lend, which can stunt economic growth. This adverse balance sheet effect can even outweigh the positive effect of depreciation on competitiveness and net exports (Jonas 2003), especially when the fraction of foreign currency debt is high and the responsiveness of exports to depreciation is low (Céspedes et al. 2003). Furthermore, Jonas (2003) and Lee (2022) show that the high exchange rate volatility can affect even the cost of borrowing in domestic currency via the interest rate channel. Given the trouble created by the exchange rate volatility, monetary authorities will actively use the interest rate to limit exchange rate fluctuations, and foreign lenders in local currency will request high risk premium (Lee 2022). The borrowing costs for floating interest rates and new debts aggravate the balance sheet effect. Consequently, emerging economies suffering from the original sin problem face a major dilemma: On the one hand, a fixed exchange regime in open international capital markets is an accident waiting to happen because of the high-interest rate variability and the maturity mismatch problem. On the other hand, a floating exchange regime is counterproductive since banks and firms will suffer the currency mismatch problem (Eichengreen and Hausmann 1999; Chamon and Hausmann 2002).

Excessive external borrowing can equally cause the overinvestment problem when domestic savings are already very high (Jonas 2003). In this case, foreign debts will finance inefficient projects, create bubbles, and foster corruption. When domestic savings are too low, foreign debts can bridge the gap. But the output and the economic growth incur the risk of excessive fluctuation, especially in transition periods or political instability (due to the foreign inflows instability).

These over-mentioned adverse effects of foreign debts may be translated into a higher sovereign risk, a downgrading of the country rate (Baksay et al. 2012), and an increased risk premium (Artus 2003). Further, they limit the choices for the decision-makers and make it hard to react rapidly to shocks. By studying the original sin in the Great Depression, Bordo (2020) finds that the higher the share of foreign currency debt to total debt was, the longer nations waited to devalue.

Finally, because of the possible interaction between foreign debt and its adverse effects, the latter can turn into causes. For instance, the balance sheet effect of foreign debt can undermine economic performance and make the national authorities and local firms forced to contract new foreign debts to service former foreign debts.

Solutions to the foreign debt constraint

The solutions to the foreign debt constraint are cause-oriented and effect-oriented. The first category aims to eradicate the endogenous causes of the original sin, such as the underdevelopment of the financial markets, weak law enforcement, and commitment problems. The other intends to reduce the destabilizing effects of the original sin.

As cases of the first category, Jonas (2003) recommends the correction of the financial market deficiencies that produced the vulnerable borrowing structure. Hausmann and Panizza (2003) suggest the development of long-term domestic markets while stressing that it may be a necessary but not sufficient condition to eliminate the international original sin. In this respect, Bertaut et al. (2022) and Lee (2022) show that over the last decades, emerging economies have started borrowing in local currency thanks to the deepening of their financial markets. The reinforcement of the monetary policy credibility can also be an efficient solution. Ottonello and Perez (2019) show that inflation stabilization, linked to prolonged expansion, gradually dissipates the dependence on the foreign debt problem. Ogrokhina and Rodriguez (2019) find that inflation targeting and high financial integration (high capital account openness) reduce the reliance on foreign debt. Furthermore, Krugman et al. (1992), in their report on the Philippines crisis, concluded that while the high internal and external debt remained important, the major challenge was to change the economic structure toward exportable. In other words, the philosophy of this solution is to economize foreign currencies use by boosting activities that would earn foreign currencies and contract the import-based activities. The goal is not to remove the original sin, but to evade it with a foreign currency auto-satisfaction strategy. Finally, given the exceptional importance of the size (Hausmann and Panizza 2003), the outright solution that one first thinks about is the one that addresses it. Flandreau and Sussman (2002) suggest, in light of their historical study of the origins of original sin, that size and liquidity are sufficient to escape original sin. However, it is not easy to change radically the size, except if a country joins an existing union or creates a new one with others. For that, we have supposed that the size is an exogenous cause.

The second category includes exchange regime-based solutions. For instance, Jonas (2003) advances dollarization because it fixes the currency mismatch problem and reduces the maturity mismatch since it makes long-term borrowing easier. However, it has shortcomings and cannot be generalized to all countries. Goldstein (2002) proposes a currency regime called ‘managed float plus’. This special regime includes inflation targeting to provide a credible nominal anchor for the economy and measures to reduce currency mismatches. Note that exchange rate management deserves special care when the foreign debt is associated with inflationary pressure and fiscal imbalances. In this situation, Faini and Gressani (1998) state that the anti-inflation solution that includes conservative exchange rate management overcomes the pro-growth solution. The conservatively managed exchange rate contributes to containing inflation and rationalizing public expenses. Devereux and Wu (2022) show that foreign exchange reserves can mitigate original sin redux by reducing risks for foreign investors.

Another form of the second category solutions is the pro-growth or the export-boosting solution. Faini and Gressani (1998) argue that a weak current account position represents the biggest constraint to a sustained recovery. Therefore, an export-led recovery strategy based on structural reforms and a more aggressive exchange regime may be the solution. As a matter of fact, Eichengreen and Sachs (1985) and Campa (1990) argue that economic recovery after the Great Depression depended crucially upon devaluation (a form of aggressive exchange regime). Nevertheless, it should be borne in mind that the effectiveness of this solution depends on the nature of exports and the foreign currency-dominated debt level. For that reason, with a low exchange rate pass-through and high foreign debt, the negative balance sheet effect may outweigh the positive competitiveness effect. Hargreaves and Watson (2011) state that in New Zealand, characterized by a low level of foreign currency debt and elastic exports, an exchange rate depreciation would help to adjust New Zealand’s trade balance at a relatively rapid pace.

To summarize, it is unexpected that countries suffering from the financial problem of foreign debt can remove it once and for all because it has diverse causes, the main of which are exogenous. For most countries, original sin should be considered a benign tumor. Since they cannot rapidly remove it, they should cope with it while taking measures to keep it under control and palliate its adverse effects. In the short term, Faini and Gressani (1998) propose debt workouts. Debt workouts reduce the budgetary burden of the debt service, which reduces the sensitivity of fiscal balances to changes in the exchange rate, even in the short run, and makes a combination of anti-inflation and pro-growth policies possible and more effective. Sound monetary and exchange policies, high foreign reserves, low current account deficit, and high financial integration can also contribute, at least, to gradually dissipating the problem. As for the long run, first, note that the over-mentioned solutions are not mutually exclusive. Countries suffering from the original sin should simultaneously address as much as possible the causes of the original sin (underdevelopment of financial market, weak law enforcement, commitment problems) as well as its disrupting effects (mismatching problems, balance sheet effect, fiscal imbalances). Second, the exchange rate regime seems to be one of these main factors. It deserves special care.

Overview of Tunisia foreign debt

The evolution of Tunisia foreign currency debt

The foreign-currency-denominated debt of Tunisia records an overall growth rate of 250% and an annual average growth rate of 14% (see Table 4 (the top) in the Online appendix). This growth is significantly higher than the total debt growth, which records an overall rate of 148% and an average rate of 10%. The foreign public debt records the highest overall growth rate, about 285%. Though when standardized by the GDP the overall growth rate of the foreign debt falls to 96%, it still remains highly more important than the growth of the total debt (38%) and less than the growth of the foreign public debt (114%). This debt is basically denominated in EUR (55% on average) and in USD (22%) with reversed trends: downward trend of the EUR weight (61.3–48%) and upward trend of the USD weight (14.3–28%).

The Tunisian foreign debt financing constraint is not new. Like any other small developing country, for a long time, Tunisia has suffered from this problem. The concern is not the existence of foreign debt but its dramatic trend, as shown in the previous paragraph. The following subsection presents a preliminary study of the probable causes of this surge.

Preliminary notes on the causes of the external debt increase

Table 4 exhibits three principal causes of this phenomenal foreign debt increase: the exchange rate depreciation (the exchange effect) and the degradation of the revenues of tourism and phosphates and mining activities.

Over 2010–2019, the TND exchange rate exhibited a steady downside movement against the USD and the EUR. At the end of 2019, the TND recorded an overall depreciation rate of 95% against the USD and 63% against the EUR. This sizeable depreciation of the TND weighs heavily on the foreign debt evolution: the exchange effect explains about 40% of the increase in foreign debt.

The exports of phosphates and mining denominated in USD have recorded a free fall (-62%). This declining trend cannot be attributed to a significant fall in the world demand for phosphates and mining products. Inversely, Heckenmüller et al. (2014) have predicted a 3% average positive growth rate, which is not more optimistic than the growth rate calculated by the Food and Agriculture Organization (FAO) (2017) for Africa between 2015 and 2020. This downward trend instead has a supply-side cause: continuous sit-ins and strikes in the Tunisian mining basin have paralyzed phosphates extraction and heavily disturbed its value-creation chain.

Like the phosphates and mines activities, the tourism revenues have recorded a downward trend, but they have a seesawing curve and a less acute trend (an over-the-period decrease of about 22%). The decline in tourism revenues is due to terrorist attacks and economic and social instability. Note that in 2020–2021, the tourism sector faced a new crisis triggered by the COVID-19 pandemic.

By summing the normal revenues of phosphates and mining exports and those of tourism and deducing the effective revenues of the two activities, we find a revenue gap that could have covered totally or partially the need for new foreign debts (See the middle of Table 4). Precisely, between 2010 and 2015, the total revenue gap is less than the new foreign debt and covers it only partially, but since 2016, it has completely offset it.

Prospects of Tunisia foreign currency debt

In addition to the unresolved problem of the phosphates sector and the world’s inability to eradicate the COVID-19 pandemic which have damaged the two main sources of foreign currencies, the bottom of Table 4 shows three other indicators helpful to assess the perspective of Tunisia foreign-denominated debt: savings, investment, and rating.

On the one hand, the savings rate has recorded a very high overall decline (-57%), twice the overall decline of the investment rate (− 28%). That means that the national investment becomes more and more foreign-debt-dependent. On the other hand, Tunisia’s rating by Fitch Rating has recorded a continuous downgrading, which makes new foreign debts costly and scarce.

Finally, it is worth noting that external debt distress is an international fact and not specific to Tunisia. The WB (2020) states in its 2021 statistics report on international debt that prior to the onset of the COVID-19 pandemic, rising public debt level and heightened debt vulnerabilities were already a cause of concern. However, even if we consider this fact, Tunisia external debt remains exceptionally high and an issue of great concern. Compared to the average of all low- and middle-income countries, the Tunisian external debt statistics are extremely high and form an outlier. For instance, from 2009 to 2019, the external debt (ED) to gross national product ratio increased from 55 to 101%, compared to an average increase from 22 to 26%; the ED to exports ratio increased from 113 to 191%, compared to an average increase from 85 to 107%; and the foreign reserves to ED ratio decreased from 48 to 20%, compared to an average decrease from 125 to 72% (WB 2020). That clearly shows that even if the Tunisian ED trend is in line with the international trend, it is already higher than average and is recording very high growth rates. Therefore, the deterioration of ED indicators is much more related to national economic and political factors and local events than international economic conditions.

Empirical modeling

Data and Methodology

The empirical study is based on collected data from the Central Bank of Tunisia, (CBT (2010–2019) annual reports, the Tunisian Foreign debt annual reports and website statistics), the International Monetary Fund, (IMF website statistics), the WB (International Debt Statistics reports and website statistics), and the Tunisian Ministry of finance (Public Debt brochures). The collected data are quarterly and cover the 04/2009–04/2020 period.

The objective of the econometric modeling is to detect the main factors that have led to the explosive growth of Tunisian foreign debt, which is the dependent variable. The dependent variable is proxied by the "Foreign debt to gross national income" ratio and is an I(1) variable (See Table 1). The independent variables, derived from the literature review, are constrained by data availability. They are exchange rate, economic growth, coverage rate, current deficit, net foreign reserves, foreign debt services, savings, and revenue gap (See Table 1).

Table 1.

Variables definition and integration levels

Variable (in log transformation)	I(0)	I(1)	Model
Foreign debt to GNI (LFD)		*	(a) and (b)
GNI growth (LEG)		*	(a)
Foreign debt service to revenue (LFS)		*	(b)
Exchange rate (LEX)		*	(a)
Savings to GNI (LSA)		*	(b)
Net foreign reserves to GNI (LFR)		*	(b)
Current deficit to GNI (LCD)	*		(b)
Coverage rate (LCR)	*		(a)
Revenue gap (LRG)	*		(a)
Foreign direct investment to GNI (LFI)	*		(b)

A high exchange rate depreciation produces a foreign debt increase because it has a direct adverse effect on the foreign debt, the exchange effect, and an indirect adverse effect, the balance sheet effect. In addition, emerging economies borrow more when exchange rate volatility is high (Lee 2022) or if the sovereign can manipulate the exchange rate (Eichengreen and Hausmann 1999). Equally, net foreign reserves have direct and indirect effects on foreign debts. High foreign reserves reduce the need for new foreign debts to repay previous debts or to finance necessary imports (the direct effect), as they allow to stabilize the exchange rate and reduce risks for foreign investors (Devereux and Wu 2022) (the indirect effect). Significant economic growth has a positive impact on foreign debt. Ottonello and Perez (2019) attribute the gradual dissipation of the original sin partly to a prolonged economic expansion. Hausmann and Panizza (2003) and Engel and Park (2022) find that the size of the economy is the most correlated variable to the original sin problem. Thus, sustained economic growth might gradually relieve the foreign debt constraint. FDI is an alternative foreign financing that is more sustainable than foreign debt. High FDI reflects large currency reserves and a financial integration level able to reduce the reliance on foreign debt (Ogrokhina and Rodriguez 2019). High savings combined with high foreign debts are a sign of an overinvestment problem (Jonas 2003), and a deterioration of savings associated with excessive foreign debt reflect a heavy dependence on foreign financing. An increase in foreign debt services might reflect the increase in the risk premium (Artus 2003) and, in recession periods and weak economic performance, might also translate the need for new foreign debt contracts to repay maturing debts. A deteriorating coverage rate implies a growing foreign currency deficit and a heavy reliance on foreign debts. It also indicates that the positive expenditure-switching effect (Towbin and Weber 2011) does not outweigh the adverse exchange and balance sheet effects. In the same vein, current deficit degradation has the same expected effect on foreign debts. The current account deficit is broader than the trade deficit and serves to estimate the possible association between other factors (essentially remittances from the diaspora) and foreign debt growth. Lastly, the revenue gap describes the shortfall of revenues in the tourism and phosphates sectors (foreign currency-generating activities) stemming from post-revolution social and security instabilities. High revenue gaps create foreign currency shortages and push sovereigns to contract more foreign debts. Evidently, the literature review advances other variables that are not less important, like monetary credibility, financial market completeness, law enforcement, and corruption. Unfortunately, they cannot be modeled because of data availability and frequency constraints.

The revenue gap values are calculated by deducing the actual revenues of tourism and phosphates and mines from the normal revenues of these two activities (the revenues calculated under the assumption that Tunisian growth rates are equivalent to the world growth rates). This time series and the time series of savings published by the CBT are annual. They are converted to quarterly data by frequency conversion techniques, namely the Chow-Lin technique. The independent variables are not all stationary at level. Some of them are stationary at level, I (0), and others are stationary at first-difference, I(1) (See Table 1). For this reason, we use the autoregressive distributed lag (ARDL).

Determinants of the explosion of Tunisia foreign debt

To choose the appropriate model, we start with the unit root test, Table 1. Note that: (1) when the variable has zero or negative values, such as the economic growth, we take the logarithm of one plus the variable instead of the log of the variable. (2) For the savings and revenue gap, there are only annual data transformed into quarterly time series, and in the regression, we have tried not to include them in the same model. (3) The quarterly GNI growth rate is calculated based on the quarterly current sold deficit and the quarterly current sold to GNI ratio published by the CBT.

Some regressors are stationary at level, I(0), and others are stationary at first-difference, I(1). To model appropriately these time series and extract both long-run and short-run associations, we will use the Autoregressive Distributed Lag (ARDL)/ Bounds testing methodology. The model has only one dependent variable, the log of foreign debt to gross national income (GNI) adjusted to seasonality. The choice of exogenous variables has two constraints: (1) the limited number of observations prevents the use of all the regressors and their lags in the same model. For that reason, we will make two regressions, indicated by (a) and (b) in the last column of Table 1. (2) The variance inflation factor (VIF) test shows that some variables have a relatively high risk of multicollinearity. So, they should be managed in such a way as to avoid the multicollinearity problem and to consider the limited size of the data.

The basic form of the ARDL model (a) is as follows:

$${\text{LFD}}_t={\alpha }_0+\sum _{i=1}^p{\beta }_i{\text{LFD}}_{t-i}+\sum _{i=0}^{q1}{\gamma }_i{\text{LEX}}_{t-i}+\sum _{i=0}^{q3}{\rho }_i{\text{LCR}}_{t-i}+\sum _{i=0}^{q4}{\tau }_i{\text{LRG}}_{t-i}+\sum _{i=0}^{q2}{\delta }_i{\text{LEG}}_{t-i}.$$ 1a

The ARDL bounds test model (or the unrestricted error correction model) (a) is:

$$\begin{array}{cc}\hfill \mathrm{\Delta }{\text{LFD}}_t& =\alpha +\sum _{i=1}^p{\beta }_i\mathrm{\Delta }{\text{LFD}}_{t-i}+\sum _{i=0}^{q1}{\gamma }_i\mathrm{\Delta }{\text{LEX}}_{t-i}+\sum _{i=0}^{q3}{\rho }_i\mathrm{\Delta }{\text{LCR}}_{t-i}+\sum _{i=0}^{q4}{\tau }_i\mathrm{\Delta }{\text{LRG}}_{t-i}+\sum _{i=0}^{q2}{\delta }_i\mathrm{\Delta }{\text{LEG}}_{t-i}+{\theta }_0{\text{LFD}}_{t-1}+{\theta }_1{\text{LEX}}_{t-1}+{\theta }_2{\text{LCR}}_{t-1}\hfill \\ \hfill & +{\theta }_3{\text{LRG}}_{t-1}+{\theta }_4{\text{LEG}}_{t-1}.\hfill \\ \hfill \end{array}$$ 2a

The first part of the model (the sums of differences) represents the short-run component, and the last part (the levels) represents the long-run component. As for model (b), we change only the regressors as defined in Table 1. The lags are not necessarily the same. Each variable may have its optimal lag. Fortunately, Eviews allows choosing automatically and at once the optimal lags of both the dependent and independent variables based on the information criteria. With a maximum lag equal to four, the SC (Schwarz information criterion) and the AIC (Akaike information criterion) give the same optimal ARDL model, ARDL (4,0,1,0,3).

Table 2 exhibits the regression output of model (a) including the short-run and the long-run coefficients, the bounds test, and the residual diagnostics. The model (a) bounds test, the middle of Table 2, shows a long-run joint association between the dependent variable and some independent variables because the F-statistic (8.2) is superior to the I(1) critical value even at the 1% level. These variables are the exchange rate (LEX), economic growth (LEG), coverage rate (LCR), and revenue gap (LGG). The error correction form confirms this association because the error correction term (CointEq(− 1)) is negative and statistically significant. The long-term component of the output provides more details. Except for the revenue gap, all the independent variables have statistically significant coefficients. The exchange rate has a statistically significant positive coefficient, which means that the TND depreciation causes the increase of foreign debt in the long run. Economic growth has a statistically significant negative coefficient. In the long run, poor performance has a negative impact on foreign debt. The coverage ratio also has a statistically significant negative coefficient. The less exports are cover imports, more foreign debt is needed. The short-term component of the output, the top of Table 2, shows that the foreign debt increase may be attributable to the poor economic performance and the precedent level of the foreign debt itself.

Table 2.

Model (a) regression output: short-term and long-term association tests and goodness-of-fit tests

Model (a): ARDL (4,0,1,0,3), Endogenous: LFD, Exogenous: LEX, LCR, LRG, LEG
Variable	Coefficient	Std. Error	t-Statistic	Prob
Short-term component
C	– 0.522569	0.101586	– 5.144086	0.0000
D(LFD((− 1))	– 0.200705	0.111909	– 1.793467	0.0837
D(LFD((− 2))	– 0.350483	0.099290	– 3.529903	0.0015
D(LFD((− 3))	– 0.462224	0.124109	– 3.724347	0.0009
D(LCR)	– 0.041281	0.120172	– 0.343514	0.7338
D(LEG)	– 0.626860	0.145224	– 4.316510	0.0002
D(LEG((− 1))	1.255014	0.318054	3.945912	0.0005
D(LEG((− 2))	0.684615	0.176565	3.877406	0.0006
Long term component
LFD((− 1)*	– 0.420	0.102	– 4.095	0.0003
LEX**	0.391	0.095	4.087	0.0003
LCR((− 1)	– 0.442	0.131	– 3.353	0.0023
LRG**	– 0.001	0.004	– 0.382	0.7048
LEG(− 1)	– 2.433	0.529	– 4.594	0.0001

F-bounds test	Null hypothesis: no levels relationship
Test statistic	Value	Signif.		I(0)	I(1)
		Asymptotic: n = 1000
F-statistic	8.2	10%		2.45	3.52
K	4	5%		2.86	4.01
		2.5%		3.25	4.49
		1%		3.74	5.06
Actual sample size	41	Finite sample: n = 45
		10%		2.638	3.772
		5%		3.178	4.45
		1%		4.394	5.914
Alternative long-term association test: Error correction form
Variable	Coefficient	Std. Error		t-Statistic	Prob
CointEq(-1)	− 0.420	0.061		− 6.84	0.000
Goodness of fit tests
R-squared: 0.793			F-Statistic: 15.356 (0.000)
Serial correlation LM test			Normality test
F-statistic	Prob. F(2,26)		Jarque–Bera		Prob
0.553620	0.5815		1.786		0.409

*p-value incompatible with t-Bounds distribution

**Variable interpreted as Z = Z(− 1) + D(Z)

The bounds test of model (b), the middle of Table 3, does not confirm the existence of a long-run joint association between the regressors and the Tunisian foreign debt. The F-statistic (2.693) lies in between the I(0) and the I(1) critical values at the 5% level. At the 1% and 2.5% levels, the value is inferior to the I(0) critical value. The more detailed long-term component supports this result: among five independent variables, only two have statistically significant coefficients [foreign direct investment (LFI) and current deficit (LCD)]. The two variables have a negative long-term association with foreign debt. Nevertheless, the error correction term has a negative sign, and it is statistically significant.

Table 3.

Model (b) regression output: short-term and long-term association tests and goodness-of-fit tests

Model (b): ARDL (3,0,0,3,3,2), Endeg.: LFD, Exog.: LFS, LSA, LFR, LFI, LCD
Variable	Coefficient	Std. Error	t-Statistic	Prob
Short-term component
C	– 1.920	0.562	– 3.415	0.002
D(LFD(– 1))	0.175	0.174	1.008	0.322
D(LFD(– 2))	– 0.408	0.186	– 2.189	0.038
D(LFR)	0.109	0.085	1.279	0.212
D(LFR(– 1))	0.130	0.116	1.121	0.272
D(LFR(– 2))	0.238	0.083	2.869	0.008
D(LFI)	– 0.047	0.031	– 1.499	0.146
D(LFI(– 1))	0.119	0.048	2.458	0.021
D(LFI(– 2))	0.050	0.024	2.056	0.050
D(LCD)	– 0.029	0.017	– 1.672	0.106
D(LCD(– 1))	0.029	0.014	2.070	0.048
Long-term component
LFD(– 1)*	– 0.222	0.109	– 2.023	0.053
LFS**	0.038	0.027	1.420	0.167
LSA**	– 0.028	0.068	– 0.409	0.685
LFR(– 1)	– 0.268	0.137	– 1.951	0.062
LFI(– 1)	– 0.206	0.077	– 2.654	0.013
LCD(– 1)	– 0.077	0.035	– 2.168	0.039

F-bounds test	Null hypothesis: no levels relationship
Test Statistic	Value	Signif.		I(0)	I(1)
		Asymptotic: n = 1000
F-statistic	2.693	10%		2.26	3.35
K	5	5%		2.62	3.79
		2.5%		2.96	4.18
		1%		3.41	4.68
Actual sample size	42	Finite sample: n = 45
		10%		2.458	3.647
		5%		2.922	4.268
		1%		4.03	5.598
Alternative long-term association test: Error correction form
Variable	Coefficient	Std. Error		t-statistic	Prob
CointEq(− 1)	− 0.222	0.05		− 4.4	0.000
Goodness of fit tests
R-Squared: 0.658			F-statistic: 5.264 (0.000)
Serial correlation LM Test			Normality test
F-statistic	Prob. F(2,26)		Jarque–Bera		Prob
0.27	0.76		0.156		0.924

*p-value incompatible with t-Bounds distribution

**Variable interpreted as Z = Z(− 1) + D(Z)

In the short run, for certain variables, the lagged values have significant coefficients, and the current has not (i.e., the current deficit). The Wald test shows that the current deficit and foreign direct investment have a short-run association with foreign debt.

The F-statistic (15.356 for model (a) and 5.264 for model (b)) and the R-squared (0.793 for model (a) and 0.658 for model (b)) show that the two models fit sufficiently the data. The residual diagnostics show that residuals are normally distributed and not serially correlated for the two models. The CUSUM curves lie between the boundaries (red lines in Figs. 1 and 2), which means that the two models are stable.

Fig. 1.

CUSUM stability text of model (a)

Fig. 2.

CUSUM stability text of model (b)

Results and discussion and possible ways out

Results and discussion

The economic growth rate, the coverage rate, the foreign direct investment, the current deficit, and the exchange rate have a long-run association with foreign debt. Economic growth has a statistically significant negative coefficient. The long-run negative impact of the poor performance can have several different causes. First, a very low or negative economic growth rate, while the foreign debt growth rate remains constant or even increasing, makes the external debt burden heavier. Second, under economic crises and poor economic records, sticky public expenses make local authorities more dependent on external financing. Third, there is possible reversal causality between foreign debt and economic performance. The balance sheet effect, resulting from exchange rate depreciation combined with high foreign debt, has a negative effect on investments and economic growth. This result confirms Ottonello and Perez (2019) findings that prolonged expansion gradually dissipates the original sin and Kim et al. (2021) result that the debt shock adversely affects future economic output. The coverage ratio (the ratio of exports to imports) has negative long-run association with the external debt level. The coverage ratio deterioration makes it difficult for the surplus of the service (mainly tourism) and factor income (mainly labor income) accounts to cover the trade deficit. One evident cause (not the only) of the ratio deterioration is tied to the phosphates and mining activities. Continuous strikes in this sector have highly dampened its exports. Without a radical political solution to this problem, the Tunisian external debt will probably have the same trend, all else being equal. The revenue gap is the difference between the normal revenues of the tourism and phosphates and mining sectors (that could have been realized with a growth rate equal to the world average growth rate) and their actual revenues recorded under the special conditions of Tunisia. I have expected that the revenue gap is the main cause of the huge increase in foreign debts. Astonishingly, the output does not confirm the long-run association. The revenue gap is the only variable in model (a) that has not a statistically significant coefficient. However, even with this result, it is still difficult to preclude the negative impact of the revenue erosion of tourism and phosphates (two main foreign currency sources) on foreign debt. Remember that the revenue gap time series is initially annual and converted into quarterly. This can bias the results.

Foreign direct investment has a negative long-run association with foreign debt, which is the expected result. High foreign direct investment inflows can be a good substitute for external debts. For instance, they increase foreign currency reserves and reduce the need to finance basic imports by foreign debt. This fits with Rodriguez (1988) findings that foreign direct investments reduce the effects of indebtedness. On the contrary, the coefficient of the current account deficit has a negative sign, which is unexpected. This result can be partially explained by labor transfers: they are not hampered by the local social and political instability and they are positively affected by the TND depreciation. The coefficient of the net foreign reserves has a negative sign, but it is statistically insignificant (prob = 0.062). The sign supports Devereux and Wu (2022) who find that large holdings of foreign reserves are associated with less foreign debts.

The long-run effect of the exchange rate on foreign debt may be interpreted as a sign of a weak expenditure-switching effect because imports (energy and basic commodity goods) are inelastic and exports are constrained by production capacity and not price competitiveness (e.g., phosphates and olive oil). This weak positive effect is counterbalanced by the negative balance sheet effect of the exchange rate (e.g., increase in foreign debt service). It may also be attributed to the high depreciation of the TND. Lee (2022) finds that high sovereign external debt is linked to high exchange rate volatility.

In the short run, the output shows that the poor economic performance, the foreign direct investments (FDI), and the lagged level of foreign debt cause the foreign debt increase. This may mean that the excessive recourse to foreign debts creates a “foreign debt trap”. The heavy reliance on easy solutions such as foreign debt financing instead of making the necessary structural adjustments has locked the country in a vicious circle of more and more dependence on foreign debts. The longer the structural adjustments delay, the higher the cost to break the circle and get out of the foreign debt trap. The short-run negative relationship between FDI and foreign debts can be interpreted as a sign of the misuse of foreign debt. Foreign debt funds are used to finance consumption or refinance maturing debts and not investment. The absence of a short-run association between foreign debt and the exchange rate can be attributed to the active intervention of the CBT in the exchange market to stabilize the exchange rate despite its claims of a floating exchange regime (Bouabidi 2020).

Even if the Tunisian savings have recorded a drastic decrease over the period, as shown in Table 4 (from 21% in 2010 to 9% in 2019 and 4% in 2020), the econometric regression does not confirm the existence of a causality relationship between it and foreign debt neither in the short nor in long run (Table 3). This may be attributed to the conversion of annual data to quarterly data. Anyway, the drastic decrease in savings raises some remarks. (1) This continued decrease cannot be attributed to international conditions. For instance, the World Bank statistics show that the world average savings rate hit a new record in 2018 (27.005%) and experienced a slight decrease in 2019 (26.794%). Hence, the downward trend of Tunisian savings has national causes. (2) A drastic decrease in savings accompanied by a phenomenal increase in foreign debts is a sign of a national investment dependency on foreign financing. (3) A drastic decrease in savings over a period marked by real salary growth (at least till 2018), low economic growth, high foreign debt increase, and national investment decrease, is confusing and a real concern. This can confirm the use of foreign debts to refinance foreign debts (debt rollover) and consumption, as it can be a sign of the booming of the informal sector and the black market at the expense of the formal economy.

Recommendations

How to get out of the foreign debt trap and stop its phenomenal increase? This study does not claim that it has the remedy, but the theoretical background and the output results inspire to suggest some possible escape routes.

The literature review and the paper findings indicate that the dramatic increase of foreign debt is a complex multicausal problem. The Tunisian authorities must take emergency measures and make structural reforms to contain it.

First and urgently, the Tunisian authorities must improve currency-generating engines. They must especially have enough determination to curb problems in the primary foreign-currency-generating sectors. They must end the unrest in the phosphates and energy sectors, improve their governance, and support the tourism sector to overcome the COVID-19 crisis.

Second, the Tunisian authorities are urged to restrain foreign currencies outflows. Faini and Gressani (1998) propose debt workouts. Debt workouts reduce debt services payments, which reduces the sensitivity of fiscal balances to changes in the exchange rate, even in the short run, and makes a combination of anti-inflation and pro-growth policies possible and more effective. They may also temporarily control imports. Under these exceptional circumstances and foreign reserves shortage, the authorities can temporarily stop unnecessary imports and clamp down on informal imports and smuggling. The findings show that foreign debt is negatively associated with the coverage ratio and these measures help to improve the latter and reduce the need for foreign debt financing.

Three, since several years ago, the Tunisian sovereign rating has been recording a steady downgrade (see the last line of Table 4), but very recently the pace has accelerated. On March 18, 2022, Fitch downgraded Tunisia to ‘CCC’ and on September 30, 2022, Moody's placed Tunisia's Caa1 long-term foreign currency on review for downgrade (Moody’s 2022). These developments have made access to the international capital market highly expensive, if not impossible, and may cause foreign currency shortage and foreign debt repayment challenges. Tunisia must conduct some priority economic policy reforms needed for an IMF lending facility (the Extended Fund Facility) to ensure minimum foreign currency funding, rebuild international investor trust, and restore external stability. The main reforms are: (1) turning around loss-making public enterprises; (2) improving tax equity by bringing the informal sector into the tax net and ensuring stronger contributions from liberal professions; (3) reducing the growth of the civil service wage bill; and (4) gradually phasing out energy subsidies (IMF 2022a, b). These reforms help to contain current public expenditures and reduce the need for new debts.

As structural adjustments, the Tunisian authorities can learn from the Krugman et al. (1992) approach to overcome the Philippines crisis. Krugman et al. (1992) conclude that while the high internal and external debt remained important, the major challenge was to change the economic structure toward more exportability. The Tunisian authorities should boost exporting activities, reduce the national deficit in energy and commodities (by boosting renewable energy and agriculture sectors), and contract import-based activities. Additionally, Tunisian authorities must develop a capital market deepening policy. As Bertaut et al. (2022) and Lee (2022) noted, financial market deepening is necessary to borrow globally in local currency. These reforms help to contain the foreign debt problem by boosting exports, reducing imports, and allowing to borrow outside in local currency.

The study shows that the high exchange rate depreciation has increased the foreign debt burden. This negative effect on foreign debt and its service is not offset by a significant export recovery or sustained economic growth. Therefore, aggressive depreciation fails to achieve the intended objective. Alternatively, the Tunisian authorities should think about an alternative strategy, an anti-inflation and stable exchange rate strategy. Ottonello and Perez (2019) and Ogrokhina and Rodriguez (2019) show that inflation stabilization (or inflation target) reduces the reliance on foreign debt. Harrigan (2006) findings show that floating has a substantial inflation cost for low-income countries without any significant improvement in export competitiveness or the balance of trade. This relies on a number of factors (e.g., high price elasticity of imports and exports, low foreign debts, mature financial system) which often do not hold in these countries. Otherwise, the Tunisian authorities must be aware of misalignments that may compromise the credibility of a more stable exchange rate and produce currency crises (e.g., high inflation, high budget deficit, persistent current account deficit). As Husain et al. (2005) note, “no exchange-rate system can ultimately act as a substitute for sound macroeconomic policies.” In fact, the CBT has already started monetary policy tightening to curb the accelerating inflation partly caused by the Ukraine war (IMF 2022a).

Concerning the adjustments of the exchange rate and the capital flows management, Rodrik’s advocation (2011) may be helpful. Rodrik (2011) argues that the expansion of financial globalization and the free movement of capital make crises more frequent. Hence, a stable exchange rate and an independent monetary policy may outweigh a flexible exchange rate and free movement of capital. Given that emerging markets are inherently more unstable and more fragile, and the original sin problem causes financial fragility (Eichengreen and Hausmann1999) and domestic economic instability (Eichengreen et al.2002), Tunisia may temporarily use the neglected side of the Mundell-Fleming impossible triangle, which is the control of the capital flows movement. Tunisia may negotiate with its foreign investors to reinvest their incomes in Tunisia instead of repatriation. It can also negotiate with lenders to convert debt securities to stocks or other equity ownership forms. Nevertheless, Ogrokhina and Rodriguez (2019) find that high financial integration is a cofactor to reduce reliance on foreign debt.

Conclusion

The phenomenal increase in Tunisian foreign debt is a matter of growing concern. The paper fits in this context. It is an attempt to identify the principal causes of this worrying steady increase and propose some possible escape routes. Given that the different variables do not have the same level of integration, the paper applied ARDL modeling.

The findings show that the economic growth rate, the coverage rate, the foreign direct investment, the current deficit, and the exchange rate have a long-run association with foreign debt. In the short run, they show that poor economic performance, foreign direct investments (FDI), and the lagged level of foreign debt cause the foreign debt increase. Unexpectedly, the results do not support the existence of a causality relationship between the Tunisian foreign debt and the revenue gap variable nor with savings.

The paper recommendations can be categorized into four escape routes that are not conflicting, but each one of them deals with the problem from a special angle. The first consists of boosting the coverage ratio by boosting foreign currency-generating activities (stopping the chaos in the phosphates sector and helping the tourism sector to recover) and restraining foreign currency outflows (rationalizing imports and negotiating a debt workout). The second consists of limiting the need for foreign debts by containing current public expenditures (broadening the tax base, reducing the wage bill, and rationalizing subsidies). The third aim is to mitigate the negative impact of the exchange rate depreciation by stabilizing the exchange rate and inflation (inflation target) and keeping controlling capital movement. Note that this solution depends on the two previous ones because an exchange peg needs a minimum level of foreign currency reserves (at least four months of imports). The fourth consists of making economic (an export-oriented economic structure) and financial (deepening the financial market) structural adjustments.

It should be noted that this study has two main limitations. The first is related to the database. Even though it covers the last decade in which Tunisian foreign debt highly increased, it is limiting because it is difficult to find quarterly time series for all the variables. Second, the foreign debt needs, availability, and costs are constrained by the sovereign risk, which is a function of political, social, economic, and financial factors. So, the foreign debt explosion is a complex phenomenon and should not be viewed only from an economic perspective.

Finally, while trying to diagnose the economic causes of the high increase in Tunisia foreign debt, I am deeply convinced of two facts. First, even if the economic and financial factors are important, political instability is of crucial importance. Second, because of its economic and financial destabilizing effects, foreign debt may be a sin, but it is not the original sin. The original sin is “dictatorship” as Al-Kawakibi (1902) stated long before: “despotism is the root of all evil (corruption)”. Tunisia emergent democracy is actually under the mercy of the mid-2021 coup, the specter of dictatorship is back, and the foreign debt problem may be devastating.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Not applicable.

Funding

No financial support.

Data availability statement

The data are available on the Mendeley repository under the following specifications: Bouabidi (2022). Quarterly data of the Tunisian foreign debt and some other indicators (savings, exchange, current deficit...).

Declarations

Conflict of interest

The author confirms that there are no conflicts of interest to disclose, neither financial nor non-financial.

Ethical approval

This article does not contain any studies with human participants performed by any of the authors.

Informed consent

Not applicable.