5 The External Debt Problem of Kenya

Mohsin Khan, and Simeon Ajayi
Published Date:
May 2000
  • ShareShare
Show Summary Details
N.K. Ng’eno

Mexico’s default on its debt in August 1982 set off a major global debt problem. The problem spread quickly from Latin America and soon reached crisis proportions in sub-Saharan Africa, where it has been more difficult to deal with than in the Latin American case. The high indebtedness of African countries led to the adoption by African leaders of the African Common Position on Africa’s External Debt Crisis (Organization of African Unity, 1987).

The seriousness of the debt problem spawned numerous proposals to deal with it. The first of these was the Baker plan of October 1985, which was followed by the Toronto plan of June 1988 and the Brady initiative of March 1989. In May 1989 the World Bank and the IMF took initiatives to reduce the commercial debts of low-income countries.

The target of the Baker terms was a group of 15 highly indebted countries, whereas the Brady terms were intended for countries highly indebted to commercial creditors.1 The Toronto facility targeted low-income African countries whose official debt could be rescheduled under the Paris Club arrangements, conditional on the implementation of IMF and World Bank structural adjustment programs. The Toronto terms were enhanced by the London terms of December 1991 and the Naples terms of December 1994.

These terms allowed for deeper reduction of the nonconcessional stock of bilateral debt for eligible poor countries. The Naples terms, which allows for relief of up to 67 percent of eligible debt in net present value terms, is accessible to low-income countries after they agree with the donors on debt-service programs that could be sustained over longer periods than is possible under the Toronto or London terms.

The extent of the African debt crisis has been analyzed in various studies, including United Nations (1988), World Bank and United Nations Development Program (1989), Killick and Martin (1989), Krumm (1985), Mistry (1988, 1991), Greene and Khan (1990), Ajayi (1996), and Elbadawi, Ndulu, and Ndung’u (1996). However, individual country case studies are scanty. The few available include Degebe (1992), Ajayi (1991, 1996), Mbire and Atingi (1994), and Osei (1995).

The importance of country studies is that they highlight the differences in the causes of debt and debt management strategies among countries. Subregional studies do not take into account the differences in the structure of the economies of individual countries.2 Debt relief or debt reduction programs, for example, are often linked to economic stabilization and structural adjustment programs. Since differences in economic structures across countries often lead to different policy responses, this would imply different debt strategies. Furthermore, differences in the degree of indebtedness imply differences in ability to service debt and in creditworthiness, which in turn determine debt management strategies and access to new money.

Differences in the sources of debt also imply different debt strategies among countries. For example, countries whose stocks of debt are mainly official would qualify for Toronto- or Naples-type debt relief, whereas those whose debts are predominantly commercial would go for market-based debt reduction strategies such as the Brady, IMF, and World Bank plans.

This chapter analyzes the extent of Kenya’s debt problem, including the size, structure, and source of the stock of the debt. It also examines the causes of the problem and the strategies available to the authorities to manage the problem. Finally, the sustainability of Kenya’s debt is assessed using a deterministic intertemporal model developed by Cohen (1985, 1988).

Debt and the Macroeconomy

This section reviews Kenya’s macroeconomic performance since independence with the aim of highlighting the major events that have contributed to the external debt problem. Kenya’s macroeconomic performance has been examined by Bevan and others (1987), Bevan, Collier, and Gunning (1990), Bevan and Karlstrom (1988), Killick (1984), Killick and Mwega (1990), Ng’eno (1991), Swamy (1994), and Mwega (1995).

The first decade (1964–73) of Kenya’s independence was one of rapid economic growth and macroeconomic stability. The economy grew at an annual average rate of 6.5 percent between 1964 and 1971 (Republic of Kenya, 1989, Table 1.1) and by 5.8 percent in 1972, but growth declined to 5.2 percent in 1973 (Table 1). At the sectoral level, manufacturing, agriculture, and government services grew at 8.2 percent, 4.2 percent, and 9.8 percent, respectively, during 1964–71. The economy also maintained a manageable external balance during the first decade. Although the balance of trade was always in deficit, the current account deficit remained small, and the overall balance and the basic balance were for most years in surplus. Capital inflows were always sufficient to offset the current account deficits.

Table 1.Kenya: Selected Macroeconomic Indicators, 1970–95
(In percent)
Change in real GDP5.
Of which:
Government services7.910.
(In millions of Kenya shillings)
Balance of payments
Balance of trade–50.5–91.4–70.6–51.2–159.9–82.2–77.3–61.3–355.5–299.4–526.6340.5–290.9
Current account17.5–39.9–38.4–57.6–126.1–107.0–64.3–14.7–284.2–0.3–373.3–253.3–159.7
Basic balance13.423.244.758.982.980.538.896.3–87.33.2–125.6–142.2–124.0
Overall balance15.2––31.5–17.635.6112.7–77.670.6–72.2–90.2–108.1
(1990 = 100)
External trade
Export prices43.
Export volumes75.057.086.0100.
Import prices17.
Import volumes119.0140.0121.0117.0133.099.095.0116.0135.0111.0127.099.084.0
Terms of trade84.086.0112.0109.096.090.0112.0117.0108.0106.0109.0103.0100.0
Other indicators
Consumer price index (1986 = 100)
Real exchange rate69.774.175.275.5100.594.695.892.587.199.3114.8121.3109.4
Foreign exchange reserves202.0145.0170.0198.0191.0169.0272.0505.0338.0520.0466.0220.0195.0
Budget deficit (in percent of GDP at current prices)
(In percent)
Sources: Republic of Kenya, Economic Surveys, various issues, and Statistical Abstracts, various issues; and International Monetary Fund, International Financial Statistics, various issues.
Change in real GDP2.
Of which:
Government services4.
(In millions of Kenya shillings)
Balance of payments
Balance of trade–180.6–226.1–270.7–230.7–587.4–696.4–1,067–1,139.9–885.5–951.2–890.8–785.3–1,964.8
Current account–25.5–86.3–79.4–31.1–406.6–403–604.2–595.9–295.4–157.3287.8291.1–1,019.9
Basic balance64.49.6–121.654.2–145.7–124.523–340.2–107.3–418.4418.4–432.6–1,202.2
Overall balance57.432.3–94.273–104.4–67.780–168.9–143.3–4331,284.1264.8–368.9
(1990 = 100)
External trade
Export prices122135117127104111103100106103102109
Export volumes79788193909594100103103121137
Import prices9388918787899310093928276
Import volumes6678728589100105100939095121
Terms of trade94110921038588797182799010196
Other indicators
Consumer price index (1986 = 100)77.184.895.5100108.6122.1138.5160.2191.7244356.3458.9466.1
Real exchange rate104.1100.3102.599.2100.397.4104.4114.3111.7110.4116.742.841.8
Foreign exchange reserves3493773763862222472571849835388539335
Budget deficit (in percent of GDP at current prices)
Sources: Republic of Kenya, Economic Surveys, various issues, and Statistical Abstracts, various issues; and International Monetary Fund, International Financial Statistics, various issues.
Sources: Republic of Kenya, Economic Surveys, various issues, and Statistical Abstracts, various issues; and International Monetary Fund, International Financial Statistics, various issues.

Excessively expansionary policies in 1970 and 1971 (Bevan, Collier, and Gunning, 1990) led to a balance of payments crisis in 1971. But the crisis was swiftly brought under control through credit, price, and exchange controls and by quantitative restrictions on imports. Although these policies were considered short-term measures at the time, they were used together or selectively to deal with macroeconomic imbalances for a long time thereafter.

The achievements made in dealing with the 1971 crisis were immediately reversed by the effects of the oil crisis of 1973–74, which led to a worldwide recession that contributed to Kenya’s poor economic performance and worsened the terms of trade. The terms of trade declined by 24 percent during 1972–75, and the annual growth rate of real GDP declined from 5.8 percent in 1972 to 3.1 percent in 1974 and 1975. The decline in GDP growth was also due to the drought of 1974, which caused a 0.1 percent decline in the real growth of the agricultural sector.

The problems associated with the oil crisis were characteristically dealt with using import compression and credit, price, and exchange controls. But one of the policies used to deal with the problem this time was increased external borrowing, especially from the IMF, the World Bank, and commercial sources. Kenya was the first country to use the IMF Oil Facility in 1975.

The commodity boom of 1976–77 played a major role in halting the balance of payments problems and the economic decline occasioned by the oil crisis. Real GDP grew at an annual average rate of 8 percent in 1977–78. The boom also effectively led to a reduction in the level of protection. The volume of imports increased by 42 percent between 1976 and 1978, but because of strong export performance import liberalization had little impact on the balance of payments. Despite the apparent import liberalization, however, there was no radical change in the trade regime, as the control regime introduced in 1971 remained intact even though trade liberalization was a stated government policy. The government’s commitment to trade liberalization was stated in the 1974–78 Development Plan.

The government’s failure to adjust to the oil shock led to increased private and public consumption, which contributed to a trade deficit of K Sh 355.5 million in 1978, the highest since independence at the time, and a decline in the terms of trade. These events once more triggered the use of credit and import controls, including an import deposit scheme introduced in December 1978. The government also ran down its foreign exchange reserves and borrowed to finance the deficits. Since most of the loans were obtained under harder terms, the cost of servicing the debt began to rise significantly. The debt-service ratio, which stood at about 14 percent in 1975, had almost doubled by 1982.

The overall state of the economy continued to worsen as real GDP growth declined from 8 percent in 1978 to 4 percent in 1980. As the structural weaknesses of the economy became evident, the government began stabilization and structural adjustment programs in 1980.3 The performance of the economy remained poor (Table 1), however, as a result of poor export performance and deteriorating terms of trade, macroeconomic imbalances arising from public sector deficits, and structural rigidities in the economy. The inability of the government to carry through some of the adjustments agreed with the donors compounded the problems. The balance of payments and budget deficits, which began to run out of control in the late 1970s and continued into the 1980s, also began to be financed by increased external debt.

The second half of the 1980s started on a better note. Growth of real GDP averaged over 5 percent for the period. This favorable growth was stimulated by the coffee boom of 1986 and by lower oil prices. Despite reasonable economic growth, the period was characterized by serious balance of payments problems. The volume of exports stagnated after 1985, but the volume and prices of imports rose, causing serious deterioration in the terms of trade, which necessitated large capital inflows.

In 1986 new measures were introduced to deal with the structural weaknesses of the economy. These policies, initially outlined in Sessional Paper No. 1 of 1986 and subsequently in the 1989–93 Development Plan, emphasized sectoral reforms, of which the major programs were financed through World Bank loans. Twenty-two such projects were financed by the World Bank between 1989 and 1996. Balance of payments support was also obtained during the period under various IMF Enhanced Structural Adjustment Facility arrangements started in February 1988.

The policies introduced after 1986 were much deeper and wider ranging than those introduced earlier and were more successfully implemented. Prices were decontrolled in 1988, and interest rates were deregulated in 1991. Trade liberalization moved progressively from 1986, and by the end of 1991 all import and export controls had been abolished except for a small negative list. Tariff reform was also on track, with average import-weighted tariffs declining from 30 percent in 1984/85 to 20 percent in 1991/92 (Swamy, 1994). The effect of the trade reforms was to raise the trade balance to the highest levels from 1987 onward, but healthy capital flows negated any serious balance of payments problems. The shilling was relatively stable between 1985 and 1989 but appreciated slightly in 1990 and 1993. The liberalization of foreign exchange, which began toward the end of 1991, reached a climax in 1993 when exchange controls were removed; the shilling was effectively convertible by June 1994, when Kenya accepted the obligations of Article VIII of the IMF.

It has to be said, however, that despite the apparent success in policy reform, the first half of the 1990s was characterized by the worst economic performance since independence (Table 1). This state of affairs was a result of the lack of credibility of reforms due to inconsistency in implementation, financial mismanagement, the political turmoil leading to multiparty elections in December 1992, and the aid embargo by donors in November 1991. All these contributed to high inflation, low foreign capital inflows, unstable exchange rates, and high balance of payments deficits, which combined to depress the economy.

The Debt Situation

Kenya falls in the category of severely indebted low-income countries. With a ratio of the net present value of debt to exports in excess of 200 percent and a debt-service ratio higher than 25 percent, Kenya was considered to have not only a high liquidity problem but also a large debt overhang in 1991–93 (see World Bank, 1994, p. 40). It will be shown that this does not seem to have worried Kenyan policymakers.

Table 2 shows that turning points in Kenya’s external indebtedness occurred in 1974, 1979, 1985, and 1993.4 Kenya’s external debt problem is largely a problem of the 1980s, however. The total debt rose from less than $0.5 billion in 1971 to $1.3 billion in 1975 and then to $3.4 billion in 1980. It stabilized between $3.2 billion and $3.6 billion during 1980–84 but rose again after 1985, reaching a high of $5.8 billion in 1987. The debt stabilized again at $5.8 billion to $5.9 billion in 1988 and 1989, despite debt write-offs amounting to about $627 million between 1987 and 1990 (Republic of Kenya, 1990). The total debt rose to $7.5 billion in 1991 but declined to $6.9 billion in 1992 as a result of the aid embargo by donors. The end of the embargo in December 1993 led to a rise in debt from $7.1 billion in 1993 to $7.3 billion in 1995.

Table 2.Kenya: External Debt, 1970–95(In millions of U.S. dollars)
Total DebtLong-Term Debt Outstanding and DisbursedPublicly Guaranteed DebtPrivate Nonguaranteed DebtShort-Term DebtUse of IMF Credit
Source: World Bank Debtor Reporting System.
Source: World Bank Debtor Reporting System.

A look at public and publicly guaranteed debt and private nonguaranteed debt, the two components of long-term outstanding debt, indicates that public and publicly guaranteed debt has been a major component, but private nonguaranteed debt was high during periods of crisis. This was true in 1975/76, 1979/80, and 1990/91, implying that as official credit declined, commercial sources became important (Table 2). Short-term debt has been relatively stable since 1978 but peaked in 1980, 1990, and 1993, implying greater reliance on this source during periods of crisis. The use of IMF credit has been stable since 1979/80. Private nonguaranteed debt not only has been low but has been rising relatively slowly over the years as well.

Total public and publicly guaranteed debt has been predominantly official (multilateral and bilateral) and concessional (Table 3). Official debt in 1995 made up 81 percent of total outstanding debt, and 76 percent of this debt was on concessional terms. A decomposition of official debt shows that multilateral debt constituted a major proportion of total outstanding debt. In 1995 this was 46 percent of the total, whereas bilateral debt was 37 percent. The major proportion of the multilateral loans are World Bank (IBRD and IDA) credits, which made up 57 percent and 28 percent of multilateral loans in 1997, respectively, but by 1995 the positions had reversed, with IDA contributing 68 percent of total multilateral loans while IBRD contributed 21 percent. It is also evident that bilateral credits began to rise after 1990 and were almost converging with multilateral credits by 1995.

Table 3.Kenya: Sources of Publicand Publicly GuaranteedDebt, 1970–95(In millions of U.S. dollars)
TotalOfficialMultilateralBilateralPrivateMultilateral ConcessionalBilateral Concessional
Source: World Bank Debtor Reporting System.
Source: World Bank Debtor Reporting System.

The evolution of Kenya’s debt problem can be illustrated further by changes in the volume and pattern of loan disbursements (Table 4). Although the pattern of total disbursements has not been consistent, it is clear that multilateral disbursements rose from 1986 on and peaked in 1989, when they represented 44 percent of total disbursements. Multilateral disbursements declined beginning in 1990 and only recovered in 1993. Disbursements of bilateral loans followed a similar pattern.

Table 4.Kenya: Debt Disbursements, 1970–95(In millions of U.S. dollars)
TotalMultilateralBilateralPrivate CreditorsTotal Private NonguaranteedMultilateral ConcessionalBilateral Concessional
Source: World Debtor Reporting System.
Source: World Debtor Reporting System.

Disbursements from private creditors comprised a small proportion of guaranteed disbursements. Although these disbursements stood at $380 million in 1991, they declined consistently thereafter, reaching $6 million in 1994, their second-lowest ever, but then rose to $130 million in 1995. Similarly, except for a few years, private nonguaranteed debt represented a small percentage of gross disbursements. This emphasizes the point that Kenya’s debt is predominantly official. The sharp decline in disbursements in 1992–94 confirms the downward trend in external debt flows. This is supported by the decline in commitments (Table 5) from $1.1 billion in 1989 to $98 million in 1994. There were no commitments by private creditors in 1993 and 1994, but a commitment of $167 million was made in 1995.

Table 5.Kenya: Debt Commitments, 1970–95(In millions of U.S. dollars)
TotalOfficial CreditorsPrivate Creditors
Source: World Bank Debtor Reporting System.
Source: World Bank Debtor Reporting System.

Total debt-service payments (Table 6) increased from $46 million in 1970 to $359 million in 1980 and then rose steadily, reaching $560 million in 1988. After a decline in 1989, payments rose to $602 million in 1991, then declined in 1992 and 1993 before rising to a peak of $820 million in 1994. It is evident, however, that total debt-service payments for private and private nonguaranteed credit were comparable to debt service for official credit, despite the official debt being larger than private and private nonguaranteed debt. For example, private and private nonguaranteed debt constituted only 18 percent of long-term debt outstanding, which indicates that private credit and private nonguaranteed debt were contracted at high interest rates.

Table 6.Kenya: Debt-Service Payments, 1970–95(In millions of U.S. dollars)
TotalMultilateralBilateralPrivate CreditorsTotal Private NonguaranteedMultilateral ConcessionalBilateral Concessional
Source: World Bank Debtor Reporting System.
Source: World Bank Debtor Reporting System.

This pattern of debt service is evident from data on principal (Table 7) and interest (Table 8) repayments. The negative net transfers on debt from private creditors and on private nonguaranteed debt (Table 9) after 1981 (except for a few years) confirm that debt repayments for nonconcessional debt have been high. The negative transfers on multilateral debt are indicative of the large share of this debt in total debt. Moreover, since this debt does not provide scope for relief, its contribution to Kenya’s future debt burden will be high.

Table 7.Kenya: Repayments of Principal, 1970–95(In millions of U.S. dollars)
TotalMultilateralBilateralPrivate CreditorsTotal Private NonguaranteedMultilateral ConcessionalBilateral Concessional
Source: World Bank Debtor Reporting System.
Source: World Bank Debtor Reporting System.
Table 8.Kenya: Interest Payments, 1970–95(In millions of U.S. dollars)
TotalMultilateralBilateralPrivate CreditorsTotal Private NonguaranteedMultilateral ConcessionalBilateral Concessional
Source: World Bank Debtor Reporting System.
Source: World Bank Debtor Reporting System.
Table 9.Kenya: Debt Net Transfers, 1970–95(In millions of U.S. dollars)
TotalMultilateralBilateralPrivate CreditorsTotal Private NonguaranteedMultilateral ConcessionalBilateral Concessional
Source: World Bank Debtor Reporting System.
Source: World Bank Debtor Reporting System.

Indicators of the debt burden (Table 10) show that the problem began to increase faster after 1980. Most of the indicators peaked in 1987, with a debt-service ratio of 40 percent, a debt-GNP ratio of 75 percent, an interest-export ratio of 17 percent, and a debt-export ratio of 339 percent. The indicators declined between 1988 and 1993, except for the debt-GNP ratio and the debt-export ratio, which had uneven trends. All the indicators were above the World Bank definitions of severe indebtedness (debt-export ratio of over 300 percent, debt-GNP ratio of 80 percent, and debt-service ratio of 25 percent) since 1990, implying that the government has made little headway in solving the debt crisis.

Table 10.Kenya: Debt Indicators, 1971–95(In percent)
Debt-GNP RatioDebt-Service RatioDebt-Export RatioInterest-Export Ratio
Source: World Bank Debtor Reporting System.
Source: World Bank Debtor Reporting System.

The terms of new debt commitments are good indicators of the future debt burden. The average terms for official debt became more favorable (Table 11) in the 1990s. Interest rates came down, while the maturity and grace periods rose and the grant element of the debt increased. These trends are comparable to those of the 1970s. The conditions for private debt, on the other hand, became harder from 1990 on. The increase in the share of public and publicly guaranteed debt from this source suggests a higher future debt burden.

Table 11.Kenya: Average Terms of New Commitments by Official Creditors, 1970–95
Interest (In percent a year)Maturity (In years)Grace Period (In years)Grant Element (In percent)Private interest (In percent)Creditors’ Maturity (In years)Grace Period (In years)Grant Element (In years)
Source: World Bank Debtor Reporting System.
Source: World Bank Debtor Reporting System.

Despite the high indebtedness, Kenya had not rescheduled its debts by the end of 1995, except for the $517 million in arrears restructured in January 1994. The reason behind the avoidance of restructuring debt has been to maintain the country’s credit rating in the financial markets. The government has instead committed itself to a policy of debt reduction through the acquisition of concessional loans, debt write-offs, encouragement of foreign direct investment, and export promotion. Efficient management of the existing stock of debt was also to be enhanced through the use of market-based instruments such as debt-equity swaps, debt buybacks, and discounting in the secondary markets. These policies have not been pursued aggressively, however.

The Causes of Debt

The literature is replete with explanations of the causes of external debt problems, which can be divided into internal and external factors. Mistry (1991), Humphreys and Underwood (1989), and Cline (1995) provide detailed accounts of the factors behind the sub-Saharan African debt crisis. Greene and Khan (1990) and Ajayi (1996) also discuss the major causes of the crisis in the region. The major external factors have been deteriorating terms of trade and rising world interest rates. Other factors include increased protectionism in world markets, irresponsible lending by private and official creditors, poor advice by international financial institutions and official creditors, and reschedulings on punitive terms (Mistry, 1991). Cline (1983) emphasizes the role of external factors in the debts of the least developed countries. On the other hand, domestic factors are seen to arise mainly from economic mismanagement as manifested in unsustainable fiscal deficits, poor economic performance, and the maintenance of unrealistic exchange rates. Ajayi (1991) found that internal factors played a major role in the development of Nigeria’s debt problem. Humphreys and Underwood (1989) list high population growth rates and poverty, commodity concentration, low saving and investment rates, low productivity of investments, and low export performance as the major factors behind the high indebtedness of sub-Saharan Africa.

The review of Kenya’s macroeconomic conditions indicates that a deterioration in the terms of trade has been a major problem and could therefore be a major cause of the country’s high indebtedness. Whereas interest rates could be hypothesized to have had less of an impact, because of Kenya’s low commercial debt, the high interest payments on multilateral debt suggest that a large proportion of these debts were contracted at or close to market rates. Since this debt cannot be rescheduled, it is likely to have a significant impact on indebtedness.

The main internal factors that could be associated with high indebtedness are public sector deficits and exchange rate movements. Public sector deficits have been a major problem since the 1980s. The large fiscal and current account deficits were clearly behind increases in foreign borrowing in 1978–79 and in the second half of the 1980s.

Exchange rate movements can raise external indebtedness in several ways. First, an overvalued currency, by reducing the price of imports, can worsen the balance of payments, thus requiring higher capital inflows to finance imports. An overvalued currency can also raise expectations for devaluation, leading to capital flight. This was the major cause of increased indebtedness, especially in Latin America (Edwards and Larrain, 1989). Currency depreciation can also directly raise the stock of external debt. However, where depreciation leads to increased export earnings, it reduces the debt-export ratio and therefore improves creditworthiness.

This chapter models Kenya’s long-term outstanding debt as a function of the terms of trade, the real effective exchange rate, real foreign interest rates, and budget deficits. The volume of exports is added as a proxy for export performance. The model equation to be estimated takes the following general form:


LDOD=long-term outstanding debt
BD=overall budget deficit
RER=real effective exchange rate
INTR=real foreign interest rate
TOT=terms of trade
EV=export volume.

It is hypothesized that higher budget deficits raise the stock of debt through increased foreign credit required to finance the deficits; hence its coefficient will be positive. If depreciation of the currency disproportionately increases the domestic currency value of exports, it reduces the debt, but it would lead to higher debt if the increase in exports is more than offset by the positive effect on the stock of debt. The coefficient of RER, therefore, cannot be established a priori. Higher world interest rates and a deterioration in the terms of trade would raise the stock of debt. Good export performance is expected to reduce external indebtedness.

Annual data from 1970 to 1995 are used. The real effective exchange rate is derived as the nominal exchange rate indexed by a ratio of the foreign price of tradable goods and the domestic price of nontradables. The foreign price of tradable goods is proxied by the wholesale price index of Kenya’s major trading partners, while the domestic price of nontradables is proxied by the domestic consumer price index. The real world interest rate is determined as the six-month dollar London interbank offered rate (LIBOR) plus a 2 percent markup, deflated by the export price index for developing countries.

The properties of the data were first analyzed by performing unit root tests using weighted symmetric, Dickey-Fuller, and Phillips-Perron tests. The results (Table 12) indicate that all the variables have unit roots. Thus, the hypothesis of stationarity is rejected, implying that the variables can be stationary after first difference.

Table 12.Unit Root Tests
Weighted symmetric–1.591–2.016–1.251–1.463–1.636–1.136
Source: Author’s calculations.Note: Figures in parentheses are the probability values for the null hypothesis.
Source: Author’s calculations.Note: Figures in parentheses are the probability values for the null hypothesis.

Co-integration was tested using the two-stage Engle-Granger procedure, and the result indicates that the stationarity hypothesis could not be rejected when the error correction term was added to the regression. Co-integration was therefore established. The equation was then estimated in first difference with the error correction term.

It should be noted first that the high frequency of the data could not allow for higher-order dynamic specification to avoid the loss of degrees of freedom. A dummy (D78) was introduced to account for the 1978 shock that led to high balance of payments deficits and a deterioration in the terms of trade, which triggered one of the largest increases in long-term debt. Another dummy (D90) was introduced to explain the unusually large increase in capital flows in 1990.

The estimated results (Table 13) indicate that the coefficients of all the independent variables, except the export volume variable, have the expected signs and are significant. The coefficient for export volume has the wrong sign and is significant. All the coefficients, however, are below unity. The positive coefficient of RER indicates that the effect of currency depreciation on the stock of external debt outweighs the positive effect on export performance. As is to be expected for annual data, the coefficient for the error correction term is low, implying slow adjustment.

Table 13.Solved Static Long-Run Equation for Long-Term External Debt
VariableCoefficientStandard Error
Note: The null hypothesis that the coefficients are zero is rejected. Wald test chi2 (8) = 52.18 [0.0000]**.
Note: The null hypothesis that the coefficients are zero is rejected. Wald test chi2 (8) = 52.18 [0.0000]**.

To summarize, the model results indicate that the change in external debt is determined by the real effective exchange rate, the terms of trade, the budget deficit, and foreign interest rates. The impact, however, is moderated by low elasticities. The results indicate that the government can reduce indebtedness by maintaining realistic exchange rates and reducing the budget deficit. Despite the exogeneity of foreign interest rates, the government can minimize the impact of this variable on external debt by limiting access to external capital through commercial markets.

Debt Management

An effective debt management strategy should strive to reduce the current debt burden and hence the growth of future debt. Such a strategy would aim to reduce net transfers and, in the process, improve the country’s creditworthiness. The strategy should also aim to increase capital inflows without necessarily increasing the future debt burden. The success of any debt strategy, however, will ultimately depend on the structure of the existing stock of debt and on sound macroeconomic policies.

Kenya’s debt strategy has concentrated on efforts to increase access to concessional loans, reduce commercial loans and export credit, and request debt forgiveness from creditor governments. The country has not made use of the options available to the heavily indebted poor countries under debt-restructuring schemes such as the Naples terms or the market-based schemes such as debt-equity swaps and buybacks. Even when the country restructured its arrears in 1994, it did not request any write-offs or special concessional interest rates. Indeed, apart from the write-offs of the late 1980s, the country had not obtained any write-offs as of the end of 1996. An insignificant amount of debt was refinanced, and a debt swap on $10.3 million in debt was made in 1994 (World Bank, 1994).

Between 1980 and 1996, 31 sub-Saharan African countries, including South Africa, rescheduled their debts (Republic of Kenya, 1989). It therefore seems odd that Kenya has not rescheduled its debts despite having debt indicators that are higher than the average for sub-Saharan African countries. Moreover, Kenya was considered to have had a debt overhang and severe liquidity problems (World Bank, 1994) from as early as the late 1980s, making the fact that it had not rescheduled by 1996 seem paradoxical.

The truth of the matter is that Kenya has been able to avoid rescheduling its debt because it is not as highly stressed as many other sub-Saharan African countries, including those that rescheduled. Low commercial debt and relatively good performance of the export sector enabled the country to meet its debt obligations. Debt from private sources, which was 24 percent of total public and publicly guaranteed debt in 1991, had been reduced to 12 percent in 1995. The other reason for not rescheduling was the unfavorable political climate toward Kenya, especially on the part of bilateral donors. Since the aid embargo of 1991, Kenya has been viewed by donors to be a reluctant reformer in both economic and political affairs. This clearly played a role in discouraging the government from requesting debt restructuring, to avoid engaging donors on these issues.

Notwithstanding the past reasons for not rescheduling debt, there is no doubt that the economy would benefit from some form of rescheduling. Given that multilateral credits cannot be rescheduled or written off, and that only IDA loans are concessional, the recent shift toward increased IDA loans is a sensible debt strategy. Although private nonguaranteed debt is a small share of total outstanding debt, debt service on it is high. The government should aim to reduce this debt through the IDA debt reduction facility. The importance of using this facility is that it allows for reduction of debt to private creditors on better terms than those offered by Naples terms (Cline, 1996). This would directly improve the country’s creditworthiness and therefore encourage new money on favorable terms and also encourage foreign direct investment. Kenya should also seek debt relief under the Naples terms, given that bilateral debt is rising faster than all other types of debt. This would require the government to be willing and ready to engage the donors in a more constructive dialogue on deeper economic and political reforms.

Sustainability of Kenya’s Debt

Although Kenya has been able to meet all its debt repayments, the current low rate of economic growth and the decline in capital inflows suggest that this position may not be sustainable. The insistence on meeting external debt repayments at the expense of some other necessary expenditures, such as imports, development expenditures, and reduction of domestic debt, may hurt the economy in the short and the medium term. Moreover, if debt overhang is a problem in Kenya, as suggested by the World Bank (1994), then the associated negative impact on investment and private capital inflows will be a drag on economic growth. It is evident, therefore, that unless the problem of indebtedness is tackled through reduction of the growth of the stock of debt and increased access to concessional loans, the sustainability of Kenya’s debt and hence the solvency of the economy may not be tenable in the long run.

The sustainability of a country’s debt can be analyzed in either of two ways. The most common method follows the traditional Harrod-Domar (Harrod, 1948; Domar, 1946) growth model. This method was used by Greene and Khan (1990). The other method, associated with Cohen (1985, 1988), is a simple and deterministic intertemporal model. This model is discussed in Allsopp and Joshi (1985) and van Wijnbergen (1989). The model addresses the issue of debt sustainability more directly than the growth-cum-debt model and for this reason is preferred for our analysis.

Establishing the sustainability of a country’s debt normally involves picking a time horizon and then projecting the growth rate of selected macroeconomic variables for the horizon. The level of indebtedness at the end of the period is compared with some indicators of indebtedness to determine whether the level is sufficient to avoid future debt-service problems. For example, Cohen (1985) divided the future into two sub-periods, 1983–85 and 1996 onward, and projected future growth of exports using past export growth rates and an assumed future rate of interest inclusive of spread plus fees to calculate a solvency index.

This chapter analyzes the sustainability of Kenya’s debt a posteriori. Since Kenya’s debt has been sustainable, the aim is to determine why that has been the case. This is done by calculating Cohen’s solvency index and then deducing from the results why the solvency conditions have been met. This approach is useful in delineating what it would take for the country’s debt to remain sustainable into the foreseeable future.

Following Cohen (1985), the model used in this study is specified as:


d=the debt-export ratio or the debt-GNP ratio
x=real growth of exports
r=real foreign interest rates.

A simple solvency condition compares the rate of growth of the real interest rate (r) and the rate of growth of exports (x). A country is said to be insolvent if (r − x) < 0, so that insolvency can be attributed to rising real interest rates and declining growth of exports. Cohen suggests that this is a simplistic and inadequate method of determining insolvency, because a country is not necessarily insolvent when r > x as long as b satisfies the condition that the present value of future debt tends toward zero in the long term. A country therefore need not repay the principal or the interest due to remain solvent as long as the condition holds.

This model assumes that a debtor country is interested in maintaining the variable d at some constant level in order to be creditworthy. For example, highly indebted countries have tried to maintain their creditworthiness by generating trade surpluses through import compression and/or export promotion. The amount of trade surplus (transfers) necessary to maintain a given level of d is represented by the index b, and this has led Cohen (1985, 1988) to define b as the level of trade surplus or maximum proportion of resources above which a country would rather default than remain solvent. Cohen (1988) established that, on average, debtor countries should not devote more than 15 percent of their exports to service debts. In terms of Equation (2), b is therefore the level of minimum debt to be repaid when the interest rate exceeds a country’s growth of exports.

Two solvency indexes (bx and by) are calculated where d is defined as the debt-export ratio and the debt-GNP ratio, respectively. The averages for the two indexes are 1.0 and 0.3, reflecting a higher debt-export ratio compared with the debt-GNP ratio (Table 14). The trends for the ratios are the same, however. The indexes are positive for most years, reflecting insolvency under the simple solvency criteria. Negative indexes imply r < x, a condition under which the country is perfectly solvent (Cohen, 1985).

Table 14.Kenya: Solvency Indexes, 1971–95
Real Interest Rate (In percent)Export Growth (In percent)bxby
Source: Author’s calculations.
Source: Author’s calculations.

Concentrating on the index based on the debt-export ratio, for comparison with Cohen’s results, the value bx for the periods 1970–80, 1981–90, and 1991–94 was 4.1, 3.0, and -11.5, respectively. These indexes, like the whole-period average, are below Cohen’s critical value of 15.0. The relatively high indexes for 1970–80 and 1981–90 were a result of high interest rates and poor export performance. On the other hand, despite the relatively high interest rates, the negative index during 1991–95 was a result of good export performance.

It can be concluded that, over the period covered by the study, Kenya has neither defaulted nor rescheduled its debt because the solvency index has been well below the critical level. Moreover, moderate interest rates and good export performance have kept the level of the index down. The policy implications of these results are that the government, in its efforts to remain solvent, should carry out policies to improve export performance. The government should also take steps to lower the debt stock and avoid contracting debt that pays high interest rates.


This chapter has established that Kenya’s economic reforms took about two decades to implement and were marked by a stop-and-go pace with a few cases of reversals. The reforms began to be implemented in earnest during the early 1990s, yielding low inflation, stable exchange rates, and declining budget deficits. Economic growth, although rising, continues to be below the expected level, as are foreign direct investment and domestic investment. These results reflect the low credibility of economic reforms arising from high domestic debt, which continues to crowd out the private sector, and a high and rising stock of external debt accompanied by high debt repayments.

Kenya’s stock of debt is mostly official, and although it was earlier dominated by multilateral debt, bilateral debt has been rising and has almost equaled multilateral debt in recent years. Debt-service payments have also been high in recent years.

The structure and volume of Kenya’s debt have important implications for debt management. First, multilateral debt, which consists mainly of IDA loans, although concessional, cannot be written off. Second, although bilateral debt can be written off, the price to pay in terms of conditionality may be high. This is especially true, given Kenya’s poor image abroad. Last, although private debt has been a small proportion of total external debt, debt-service payments on this debt have been almost as high as those on official debt. Moreover, the terms of payment have become increasingly hard, suggesting that debt-service payments will rise in the future.

On the basis of these observations, Kenya’s debt management should concentrate on removing the possible debt overhang by reducing the debt burden through market-based instruments, available restructuring schemes, and appeals for write-offs. The use of all available avenues for debt reduction would lead to maximizing the level of reduction. The most immediate objective, however, would be to reduce commercial debt using the IDA debt reduction facility. This would be a follow-up to the restructuring of arrears in 1994. This step is important to reduce the heavy burden that repayment of private debt imposes on the economy. The government should also make a better case for write-offs by bilateral creditors. This would require greater flexibility on the part of the government toward donors concerned about deeper economic and political reforms. The government should also evaluate whether it is not possible to reduce the burden of debt through the Heavily Indebted Poor Countries debt initiative. This again would require greater commitment by the government to economic reforms and greater understanding and cooperation from creditors.

This chapter establishes that depreciation in terms of the real effective exchange rate, higher budget deficits, and higher foreign interest rates raise external debt, whereas improvement in the terms of trade reduces debt. It is also established that the solvency index for Kenya has been low primarily because of moderate world interest rates and good export performance.

The implication of these results is that maintenance of macroeconomic stability is important in dealing with the debt crisis. It is important to ensure that economic reforms are credible to sustain low inflation and stable exchange rates and to encourage inflows of new money at favorable rates and increased foreign direct investment. Credible reforms would also lower capital flight and encourage domestic investment, leading to more rapid economic growth. Most important, credible economic reforms would facilitate negotiations toward debt restructuring and/or appeals for debt write-offs.

To summarize, the high and rising stock of external debt and debt-service payments mean that the Kenyan government has to take serious steps to resolve the problem. This would involve exploring and using all available channels for debt reduction. The government should also undertake the economic reforms necessary to move the economy into sustainable growth, thus providing a favorable climate for debt reduction in the long run.


The 15 countries were Argentina, Bolivia, Brazil, Chile, Colombia, Côte d’Ivoire, Ecuador, Mexico, Morocco, Nigeria, Peru, the Philippines, Uruguay, Venezuela, and Yugoslavia. The classifications of indebtedness and the number of countries considered to be overburdened with debt have varied over time. In 1996 the World Bank (World Bank, 1996, Vol. 1) listed some 41 countries as heavily indebted poor countries, of which 32 were classified as severely indebted. Twenty-five of the severely indebted countries are in sub-Saharan Africa, and the group includes Kenya.

These points are emphasized by Cohen (1988), van Wijnbergen (1989), and Greene and Khan (1990).

No effort will be made here to examine specifics of the structural adjustment programs, stand-by arrangements, or any other agreement between the Kenyan government and donors or creditors on such issues as conditionalities or their success or failure. Some of these issues have been examined in Hecox (1988), Mosley (1986), Godfrey (1987), Bevan, Collier, and Gunning (1990), Swamy (1994), and Mwega (1995).

There are two sources of debt data: the World Debt Tables of the World Bank and CS-DRMS of the Debt Management Division of the Kenyan Ministry of Finance. The data from the two sources are more or less comparable. However, since CS-DRMS has five shorter series, World Bank data were used in this study. The data were obtained on diskette and used with the aid of the STARS system prepared by the International Economics Division of the World Bank.

    Other Resources Citing This Publication