Chapter

9 Capital Flight from Tanzania

Editor(s):
Mohsin Khan, and Simeon Ajayi
Published Date:
May 2000
Share
  • ShareShare
Show Summary Details
Author(s)
Timothy S. Nyoni

This chapter focuses on the measurement and determination of capital flight from Tanzania. Capital flight has been regarded as a major reason for the mounting foreign debt problems and inhibition of development efforts in developing countries (see, for example, Cuddington, 1986). In real terms, external debt in Tanzania increased from $677.4 million in 1971 to $3,863.6 million in 1993, growing at an average annual rate of 8.2 percent (Table 1).

Table 1.Tanzania: Selected Indicators of Financial Flows and Economic Performance, 1971–93
Total External Debt (In millions of 1997 U.S. dollars)Per Capita GDP at 1977 Prices (1977=100)1Current Account Deficit (In percent of GDP)Net Private Transfers (In millions of 1997 U.S. dollars)2Short-Term Capital Inflows (In millions of 1997 U.S. dollars)1
1971677.497.58.1266.8
1972978.8101.04.7–51.7265.7
19731,092.8101.16.6–33.4284.7
1974891.499.614.5–16.7135.3
19751,048.6101.010.018.2228.5
19761,330.5102.11.114.367.5
19771,449.5100.02.119.4111.0
19781,589.298.711.624.8188.3
19791,581.498.98.827.1113.0
19801,582.698.611.419.5139.8
19811,615.195.57.323.0149.4
19821,917.993.48.726.5127.2
19832,132.388.85.519.6189.4
19842,343.789.27.463.67.8
19852,630.190.86.6256.5–50.7
19862,980.190.08.1235.36.6
19873,333.091.916.1319.639.1
19883,333.892.311.6230.120.9
19893,190.593.410.3186.413.0
19903,313.894.812.5178.068.5
19913,623.797.511.8461.760.7
19923,687.498.212.2554.648.8
19933,863.699.412.9562.641.2
Sources: Tanzania, Bureau of Statistics, National Accounts of Tanzania, various years; International Monetary Fund (1996); and World Bank (1987, 1994).

Deflated by the import price index.

Nominal transfers deflated by the export price index.

Sources: Tanzania, Bureau of Statistics, National Accounts of Tanzania, various years; International Monetary Fund (1996); and World Bank (1987, 1994).

Deflated by the import price index.

Nominal transfers deflated by the export price index.

Despite positive real net private transfers and short-term capital inflows during most of the years between 1971 and 1993, Tanzania’s per capita GDP from the late 1970s to the early 1990s fluctuated at levels lower than those registered during the first half of the 1970s; the current account deficit worsened as well (Table 1). The current account was in persistent deficit throughout 1971–93. As a percentage of GDP, the current account deficit increased from an average of 6.7 percent during 1971–77 to 16 percent during 1987–93 (Table 1). These observations suggest that much of the capital inflow to Tanzania was probably not used for domestic investment or for reducing the current account deficit and external debt problems. If so, there is reason to suspect capital flight.

Conceptual differences in defining capital flight have led to concerns about the need for policy intervention to stem it and about the interaction between economic policies in the developing countries and those of industrial countries. Capital flight also raises questions about the role of international financial institutions, which lend to developing countries and act as depositories for capital exports from these countries.

There are conflicting views about the concept of capital flight. Some observers regard it as a symptom of a sick society characterized by the breakdown of social cohesion, reduction in growth potential, erosion of the tax base, failure to recover from the debt problem, and a redistribution of wealth from poorer to richer social groups (Lessard and Williamson, 1987). Others consider the use of the term “capital flight” unnecessarily pejorative. Why, for example, label capital movements from developing countries “flight,” while those from the United Kingdom and other industrial countries are termed “foreign investments” (Pastor, 1990; Lessard and Williamson, 1987)? Part of the explanation for the use of the term “capital flight” is the notion that capital should be flowing from capital-abundant (industrial) countries to capital-scarce (developing) countries. In economic theory, so-called capital flight from developing countries results from the natural and economically rational behavior of wealthy residents of these debtor countries to diversify their portfolios to protect themselves against the riskiness of any particular investment.

Any study on capital flight must first sort out what capital flight really is. Is capital flight something bad that requires policy intervention to stop or reverse it? Is it not true that capital flight would deter policymakers from making “wrong” political and economic decisions? Such arguments make it difficult to define the concept of capital flight. This is especially true because capital flight has had damaging consequences for the economies concerned despite any reflows that some countries have experienced. On the one hand, the damaging effects of capital flight may make rational foreign lenders hesitant to increase credits to the debtor countries. On the other hand, the behavior of individual agents to diversify their portfolios by keeping some of their wealth in foreign assets is postulated as rational by economic theory and accepted as normal in industrial and some developing countries.

The main objective of this chapter is to measure the magnitude and determine the causes of capital flight from Tanzania. After empirically determining the causes of capital flight, we shall discuss the policy implications of containing or reversing capital flight.

This introductory section is followed by a statement of the hypothesis, along with a discussion of the definition and measurement of capital flight. Next, political and macroeconomic developments in Tanzania are discussed, and an empirical analysis of capital flight is provided. The chapter concludes with a summary of the findings and implications for policy.

Capital Flight: Definition, Measurement, and Consequences

The main hypothesis of this chapter is that capital flight is a positive function of the growth differential between the GDPs of the United Kingdom and Tanzania, the parallel market premium, domestic inflation, the financial incentive for capital flight, capital availability, political and external shocks, and lagged capital flight. This hypothesis will be tested using the capital flight model specified below.

Definition

The term “capital flight” connotes illegal movement of capital from one country to another. This connotation implies that there may be “normal” or “legal” and “abnormal” or “illegal” flows (see, for example, Lessard and Williamson, 1987). Normal capital flows are those that are sanctioned by the government. Abnormal capital flows are illegal capital movements that take place without the government’s approval. The question of the legality of capital flows, however, implies that the country concerned imposes exchange or capital controls. This then complicates the matter, since so-called capital flight is known to have taken place even in developing countries, such as Argentina, Mexico, and Venezuela, that had no capital controls (Lessard and Williamson, 1987). Thus, it is difficult to come out with an operational definition and measurement of normal and abnormal capital flows.

The concept of capital flight refers to “capital that runs away” or “flees” abnormal risks at home regardless of whether or not the flight is legal (Lessard and Williamson, 1987, p. 2). Measuring capital flight thus defined requires an attempt to measure normal capital outflows and deduct these from total outflows. Dooley (1986) attempted this. Although Dooley’s concept of abnormal outflows as those propelled by the desire to escape the control of the domestic authorities corresponds to the concept of capital fleeing abnormal risks at home, his method of measuring normal capital flows is controversial and unacceptable (Lessard and Williamson, 1987).

The emphasis in this chapter is on all resident capital outflows rather than “normal” outflows. This is primarily because there is no sensible way of separating normal from abnormal flows. Furthermore, the consequences of capital outflows to the national economy are the same whether the outflows represent normal portfolio diversification or abnormal flight. If normal capital outflows are small, the resulting economic cost from the resident capital outflows will be modest. However, the cost will escalate when normal capital outflows are reinforced by money running away from abnormal risks at home.

The justification for using the emotive term “capital flight” is in the extent of the damage that resident capital outflows inflict on the home economy. If the damage is small, one may be content with using the term “resident capital outflows”; otherwise, “capital flight” is the preferred alternative.

Measurement

It is difficult to measure capital flight for two main reasons. One is that the concept itself is ambiguous, which makes it obvious that different definitions of capital flight may yield different estimates of the problem. The other reason is that even if there may be a consensus that capital flight means “money that runs away,” such capital is in most cases not reported to the domestic authorities or compilers of balance of payments statistics. This makes it difficult (if not impossible) to deduct capital that flees abnormal risks at home from total capital outflows. For these reasons, therefore, we shall not attempt to distinguish normal from abnormal capital outflows. We shall instead concentrate on measuring all resident capital outflows.

Table 2 presents the various items of the balance of payments and external debt for estimating capital flight. There are three main approaches to the measurement of capital flight: the balance of payments accounts approach, the residual approach, and the bank deposits approach. We discuss each of them in turn.

Table 2.External Debt and Balance of Payments Items for Measuring Components of Capital Flight
A=Changes in external debt
B=Current account balance
C=Net foreign direct investment
D=Net private short-term capital flows
E=Portfolio investment
F=Change in foreign assets of the banking system
G=Net errors and omissions
H=Changes in reserves

The Balance of Payments Accounts Approach

In the pioneering studies of capital flight, the phenomenon was measured using the balance of payments accounts (see, for example, Cuddington, 1986). The objective in this approach is “to isolate short-term capital movements that might reasonably be considered capital flight” (Cuddington, 1986, p. 3). In this approach, changes in short-term foreign assets, often called “hot money,” are considered speculative and an indication of recorded capital flows (Cuddington, 1986). The net errors and omissions item of the balance of payments is considered equivalent to unrecorded capital flows. According to Cuddington (1986), private short-term capital movements are either imprecisely reported or not reported at all, especially in countries that impose capital controls. The failure to precisely record short-term private capital flows shows up in the net errors and omissions entry of the balance of payments.

In the balance of payments approach, capital flight is measured as the negative of the sum of recorded short-term capital flows and unrecorded net flows or net errors and omissions.

That is,

where KFB is capital flight measured by the balance of payments approach, D is net private short-term capital flows, and G is net errors and omissions (Table 2). Positive values of this capital flight (KFB) measure represent capital outflows, whereas negative values are capital reflows or reverse capital flight.

The balance of payments approach has several problems. In this approach, the short-term capital flows capture recorded capital flows, and the errors and omissions are supposed to capture unrecorded capital flows. However, not everything in the net errors and omissions entry is capital flows. Errors and omissions include not only unreported or smuggled exports and imports (which may be related to capital flight) but also items that are not related to capital flight, such as true measurement and recording errors and errors because of lagged registration. Hence, the balance of payments approach fails to adequately capture unrecorded capital flows.

Another problem with this approach is that the measure ignores capital that flees to finance long-term investments such as equity and real estate abroad. The approach also fails to distinguish clearly between short-term and long-term investments, especially when the two are close substitutes. Long-term bonds, for example, can be substituted easily for short-term investments without significant loss of liquidity (Eggerstedt and others, 1995).

The Residual Approach

The residual approach was developed by the World Bank (1985) and Erbe (1985) and was modified further by Morgan Guaranty Trust Company (1986). This approach arose out of the belief that the balance of payments accounts approach was not sufficient to estimate resident capital outflows.

In the World Bank (1985) and Erbe (1985) versions of the residual approach, capital flight is calculated as the difference between sources and uses of capital inflows. The sources of capital inflows are increases in external debt and foreign direct investment. These capital inflows are used to finance either current account deficits or increases in official reserves. Inflows that finance neither current account deficits nor increases in reserves constitute capital flight.

Capital flight in the World Bank (1985) and Erbe (1985) versions of the residual approach (KFW) is measured as

where A is changes in external debt, C is net foreign direct investment, B is the current account balance, and H is changes in reserves (Table 2). Positive values for KFW represent capital flight, whereas negative values are capital reflows.

Morgan Guaranty (1986) adjusted the World Bank (1985) and Erbe (1985) measure for changes in foreign assets held by domestic agents other than the banking system. According to Morgan Guaranty (1986), capital flight is measured as

where KFM is the measure of capital flight, F is changes in foreign assets of the domestic banking system, and A, B, C, and H are as defined above. As with the other residual measure of capital flight, positive values of KFM are capital flight, whereas negative ones are reverse capital flight.

The residual approach to the measurement of resident capital outflow faces the problem of inadequacies in the debt statistics. The statistics may not always be correctly recorded, especially if the private sector also borrows from foreign creditors. When some debts are denominated in currencies other than the U.S. dollar, changes in exchange rates may increase (or decrease) the debts even if no new borrowing has taken place (Zedillo, 1987). The changes in debts denominated in currencies other than the U.S. dollar should thus be adjusted for exchange rate variations.1

The Bank Deposits Approach

A third approach to the measurement of resident capital outflows involves measuring the increase in recorded foreign bank deposits of a country’s residents. This is controversial because, even if there are statistical sources that distinguish between private and official holdings, it cannot be argued convincingly that all private funds held abroad are recorded by the relevant authorities. The difficulty in compiling bank deposits data is reinforced by the fact that some of the funds are deposited in banks that do not report to the IMF or other relevant authorities, whereas others may be held in nonbank foreign assets. Another difficulty arises from the desire of the depositors to hide their nationality in order to minimize any potential risk they may perceive. Therefore, foreign bank deposits owned by a country’s residents are likely to underestimate resident capital outflow.

The Magnitude of Capital Flight from Tanzania: Unadjusted Measures

The magnitude of capital flight from Tanzania unadjusted for trade misinvoicing in 1977 is shown in Table 3. The corresponding measures in current prices are shown in Table 4. Using estimates from the balance of payments approach, it can be seen that there were only two episodes of capital flight in 1971–93: during 1985–86, when capital flight totaled $102.24 million in 1977 prices, and during 1988, when it totaled $5.08 million in constant prices.

Table 3.Tanzania: Real Measures of Capital Flight Unadjusted for Trade Misinvoicing, 1971–93(In millions of 1977 U.S. dollars)
Balance of Payments ApproachWorld Bank Residual ApproachMorgan Guaranty Residual Approach1Bank Deposits Approach2
1971–233.18–113.95–72.17
1972–305.30174.70154.97
1973–241.68113.5488.76
1974–166.63–172.08–157.98
1975–198.11–6.29–37.61
1976–60.85227.25221.61
1977–180.10356.20375.70
1978–197.77–274.0–330.14
1979–131.13–60.7153.32
1980–110.18–130.73–130.73
1981–209.74–176.57–176.57
1982–169.55–164.90–164.90
1983–152.39-49.27-49.2721.09
1984–98.31–72.91–72.910
198580.47–15.17–15.1721.57
198621.77190.71190.7121.34
1987–100.89242.52242.5252.24
19885.08–38.39–38.3912.55
1989–24.50–272.31–272.3124.14
1990–191.17267.21267.2166.80
1991–50.27–58.04–58.045.71
1992–74.96–31.27–31.27–33.69
1993–31.67–173.33–173.3322.46
Totals
1972–73–546.98288.24243.73
1976–77–240.95583.45597.31
1983–91–585.17163.07163.07191.76
1986–87–79.12433.23433.2353.58
1971–93–2,821.04–237.87–186.00214.25
Sources: Author’s calculations based on data from World Bank, World Tables, various issues, and International Monetary Fund, International Financial Statistics, various issues.

The two residual approaches coincide after 1979 because the banking system’s foreign assets are zero.

Data for calculating the bank deposits approach during 1971–82 were not available.

Sources: Author’s calculations based on data from World Bank, World Tables, various issues, and International Monetary Fund, International Financial Statistics, various issues.

The two residual approaches coincide after 1979 because the banking system’s foreign assets are zero.

Data for calculating the bank deposits approach during 1971–82 were not available.

Table 4.Tanzania: Nominal Measures of Capital Flight Unadjusted for Trade Misinvoicing, 1971–93(In millions of current U.S. dollars)
Balance of Payments ApproachWorld Bank Residual ApproachMorgan Guaranty Residual Approach1Bank Deposits Approach2
1971–88.2–43.1–27.3
1972–125.371.763.6
1973–147.369.254.1
1974–140.6–145.2–133.3
1975–179.6–5.7–34.1
1976–56.1209.5204.3
1977–180.1356.2375.7
1978–222.6–308.5–371.6
1979–174.1–80.670.8
1980–172.1–204.2–204.2
1981–321.2–270.4–270.4
1982–251.6–244.7–244.7..
1983–216.8–70.1–70.130
1984–138.2–102.5–102.50
1985111.9–21.1–21.130
198630.6268.1268.130
1987–154.5371.4371.480
19888.1–61.2–61.220
1989–40.6–451.2–451.240
1990–343.4480.0480.0120
1991–88.0–101.6–101.610
1992–133.5–55.7–55.7–60
1993–56.4–308.7–308.740
Sources: Author’s calculations based on data from International Monetary Fund (1996) and World Bank, World Tables, various issues.

The two residual measures coincide after 1979 because banking system foreign assets are zero.

Data for calculating capital flight by the bank deposits approach during 1971–82 were not available.

Sources: Author’s calculations based on data from International Monetary Fund (1996) and World Bank, World Tables, various issues.

The two residual measures coincide after 1979 because banking system foreign assets are zero.

Data for calculating capital flight by the bank deposits approach during 1971–82 were not available.

In 1971–93, in general, Tanzania experienced reverse capital flight totaling $2.82 billion in real terms (Table 3). The balance of payments measure—which is narrow by definition—understates the magnitude of capital flight and yields lower estimates than the broader residual approach.

The residual approach indicates higher values and more episodes of capital flight than the balance of payments measure (Table 3). The World Bank (1985) and Morgan Guaranty (1986) residual measures yield more or less the same magnitudes and episodes of capital flight. According to the residual measures, capital flight during 1972–73 ranged between $243 million and $288 million in constant prices. Another episode of capital flight occurred in 1976–77, when between $583 million and $597 million fled the country. During 1986–87 capital flight totaled $433 million, and in 1990 it totaled $267 million. During 1971–93 Tanzania experienced capital reflows of between $186 million and $238 million in real terms according to the Morgan Guaranty (1986) and World Bank (1985) residual approaches, respectively.

The bank deposits approach indicates that during 1983–91 a total of $192 million (in 1977 prices) fled the country. This approach, as discussed above, gives a lower estimate of capital flight relative to the residual approach. Hence, the residual approach gives higher estimates than either the balance of payments approach or the bank deposits approach.

The Morgan Guaranty (1986) residual approach captures more episodes of capital flight than does the other broad residual approach. As will be explained below, the capital flight episodes indicated by the residual approach reflect the major political and external shocks the country experienced. The balance of payments and bank deposits approaches are too narrow to be credible and are conceptually inferior to the residual approaches. We thus prefer the Morgan Guaranty measure of capital flight to the other approaches discussed in this chapter.

Trade Misinvoicing and Capital Flight

The three approaches to measuring capital flight have been criticized on several grounds, and none can be considered satisfactory. One weakness is that they ignore trade misinvoicing. In countries where currencies are overvalued and foreign exchange and trade controls are prevalent, invoices for imports or exports are often faked and the registration of imports is avoided. In such cases, balance of payments data are recorded imprecisely and do not fully reflect actual trade and capital flows. If smuggling and underinvoicing of imports conceal a larger current account deficit, capital flight (measured by the residual approach) will be overstated unless it is adjusted for trade misinvoicing.

Capital flight can be effected by overinvoicing imports and under-invoicing exports (Eggerstedt and others, 1995; Ajayi, 1992). By under-invoicing exports, exporters surrender only a part of their export receipts to the domestic authorities and may use the remaining receipts to build up their foreign exchange holdings, some of which may be shifted to foreign countries. When importers overinvoice imports, they obtain foreign exchange in excess of their import costs. The importers can then use this excess foreign exchange to add to their foreign assets in the domestic economy and abroad. Since such operations are not recorded in the balance of payments statistics, omissions of trade misinvoicing will yield an inaccurate measure of capital flows in the residual and balance of payments approaches. Some of the trade misinvoicing, however, may be detectable in foreign bank deposits if those involved in trade faking and smuggling deposit the money in (reporting) foreign banks.

Depending on the nature of trade policies in the domestic economy, trade misinvoicing may occur for motives other than financing capital flight. Thus, rather than imports being overinvoiced to transfer funds abroad, imports may be underinvoiced to lower the cost of customs duties and value-based quantitative restrictions (Eggerstedt and others, 1995). If there are export incentives, it will be more rational for exporters to overinvoice their exports to take advantage of the export subsidies and preferred access to subsidized credits. The net effect of underinvoicing imports and overinvoicing exports will be reverse capital flight.

Data on smuggling and unreported trade flows are hard to come by, especially because such activities are not reported to the domestic authorities. To track down trade misinvoicing and smuggling, one has to compare trade statistics of the trading partners. Since the imports of any one country are the exports of another country, the ratio of the free-on-board values of the imports and exports—called the valuation ratio—should be unity (Ajayi, 1996). If the valuation ratio is not unity, it is an indication of trade misinvoicing. However, there are other reasons—apart from capital flight motives—for the recorded values of the exports and imports not to match. These reasons include differences in the recording systems, improper identification of the origin and destination of goods, differences in valuation methods, and differences in time spans covered (Eggerstedt and others, 1995; Ajayi, 1996).

Trade misinvoicing is calculated as

and

where XMIS and MMIS are, respectively, export and import misinvoicing; kT and kw are, respectively, the c.i.f./f.o.b. correction factor for Tanzania and its world trading partners; XD is Tanzania’s exports to world trading partners as reported by Tanzania; MW is imports from Tanzania as reported by Tanzania’s world trading partners; MD is imports from world trading partners as reported by Tanzania; and XW is exports to Tanzania as reported by the country’s world trading partners.

When export misinvoicing (XMIS) is negative, it reflects export underinvoicing and thus capital is flowing out of the country. Conversely, a positive XMIS denotes export overinvoicing, which implies reverse capital flight. When import misinvoicing (MMIS) is positive, it reflects import overinvoicing. A negative MMIS represents import underinvoicing. Import overinvoicing connotes capital outflows, whereas underinvoicing of imports implies capital inflows. To adjust our measures of capital flight, we take the sum of import overinvoicing and the negative of export underinvoicing, since capital flight appears with positive values in our measures.

Trade misinvoicings in constant and current prices are presented in Tables 5 and 6, respectively. The general case during 1971–93 appears to be that of export and import underinvoicing as reflected by negative values in either case. Export misinvoicing as a percentage of imports from Tanzania as reported by the country’s trading partners averaged –9.33 percent during 1971–93. (See International Monetary Fund, Direction of Trade Statistics Yearbook, for information on Tanzania’s trading partners.)

Table 5.Tanzania: Trade Misinvoicing in Real Terms, 1971–93(In millions of 1977 U.S. dollars)
Export Misinvoicing1Import Misinvoicing2Misinvoicing Adjustment3
1971114.93–14.42–129.35
1972197.38–153.94–351.33
1973–18.22–172.04–153.81
1974–44.85–136.40–91.55
1975–28.3339.5767.89
1976–45.16–115.54–70.38
1977–33.37–106.53–73.16
1978–96.09–68.7727.32
1979–117.34–105.3212.02
1980–95.34–154.85–59.51
19815.15–82.22–87.37
1982–2.24–10.29–8.05
1983–47.88–39.128.75
1984–51.20–72.65–21.45
1985–97.69–47.6050.09
1986–64.87–97.97–33.10
1987–97.71–141.49–43.78
1988–120.95–118.852.10
1989–193.57–131.9561.62
1990–20.00–58.21–38.21
1991–15.45–65.62–50.16
1992–26.75–76.91–50.16
1993–17.71–70.81–53.09
1971–93–917.26–2,001.92–1,084.67
Sources: Ajayi (1996) for the misinvoicing concept and formula; and International Monetary Fund, Direction of Trade Statistics Yearbook, various issues, for trade flows.

Export misinvoicing (XMIS) was calculated as XMIS = XDMW/kw, where XD is exports to world trading partners as recorded by Tanzania, MW is imports from Tanzania as recorded by world trading partners, and kw is the world c.i.f./f.o.b. correction factor. The world export price index for Tanzania was used as the deflator.

Import misinvoicing (MMIS) was calculated as MMIS = MD/kT – XW where MD is imports from world trading partners as reported by Tanzania, XW is exports to Tanzania as reported by world trading partners, and kT is Tanzania’s c.i.f./f.o.b. correction factor. The world import price index for Tanzania was used as the deflator.

Sources: Ajayi (1996) for the misinvoicing concept and formula; and International Monetary Fund, Direction of Trade Statistics Yearbook, various issues, for trade flows.

Export misinvoicing (XMIS) was calculated as XMIS = XDMW/kw, where XD is exports to world trading partners as recorded by Tanzania, MW is imports from Tanzania as recorded by world trading partners, and kw is the world c.i.f./f.o.b. correction factor. The world export price index for Tanzania was used as the deflator.

Import misinvoicing (MMIS) was calculated as MMIS = MD/kT – XW where MD is imports from world trading partners as reported by Tanzania, XW is exports to Tanzania as reported by world trading partners, and kT is Tanzania’s c.i.f./f.o.b. correction factor. The world import price index for Tanzania was used as the deflator.

Table 6.Tanzania: Trade Misinvoicing in Nominal Terms, 1971–931(In millions of current U.S. dollars, unless otherwise noted)
Export MisinvoicingImport MisinvoicingMisinvoicing Adjustment2Export Misinvoicing as Percentage of Trading Partners’ Imports to TanzaniaImport Misinvoicing as Percentage of Trading Partners’ Exports to Tanzania
197126.41–5.45–31.8610.38–2.09
197247.35–63.18–110.5317.65–18.36
1973–8.90–104.85–95.96–2.48–23.92
1974–36.13–115.09–78.96–7.73–15.55
1975–20.0535.8755.92–4.806.01
1976–34.71–106.51–71.81–6.53–17.43
1977–33.37–106.53–73.16–5.67–14.80
1978–90.26–77.4012.86–14.84–7.23
1979–118.83–139.83–21.01–17.44–13.63
1980–116.28–241.88–125.60–18.20–19.34
19815.59–125.91–131.500.96–12.52
1982–2.32–15.27–12.95–0.50–1.66
1983–47.88–55.66–7.78–11.09–7.78
1984–51.85–102.14–50.29–11.57–12.71
1985–94.15–66.1827.97–23.93–7.20
1986–70.44–137.73–67.29–15.93–17.21
1987–82.25–216.67–134.42–20.96–23.86
1988–113.62–189.47–75.85–27.62–21.26
1989–181.84–218.63–36.80–38.96–24.65
1990–18.30–104.56–86.26–4.11–9.64
1991–13.95–114.86–100.91–3.42–10.47
1992–22.20–136.97–114.77–4.83–10.17
1993–14.70–126.11–111.41–2.97–10.26
Average
1971–93–9.33–12.86
Sources: Ajayi (1996) for the misinvoicing concept and formula; and International Monetary Fund, Direction of Trade Statistics Yearbook, various issues, for trade flows.

Export and import misinvoicing are defined in Table 5.

Misinvoicing adjustment is the sum of import misinvoicing and the negative of export misinvoicing. When this sum is positive (negative), it reflects capital flight (reflows).

Sources: Ajayi (1996) for the misinvoicing concept and formula; and International Monetary Fund, Direction of Trade Statistics Yearbook, various issues, for trade flows.

Export and import misinvoicing are defined in Table 5.

Misinvoicing adjustment is the sum of import misinvoicing and the negative of export misinvoicing. When this sum is positive (negative), it reflects capital flight (reflows).

Import misinvoicing as a percentage of exports to Tanzania as reported by the country’s trading partners averaged –12.86 percent during 1971–93 (Table 6). Expressed in 1977 prices, export underinvoicing during 1971–93 totaled $917 million; import underinvoicing in the same period amounted to $2 billion (Table 5). During 1971–93, therefore, there was net reverse capital flight of about $1.1 billion.

The fact that exports and imports in Tanzania are underinvoiced suggests that export misinvoicing in the country is done mainly for capital flight motives, whereas import misinvoicing is done for reasons other than financing capital flight. Using an econometric analysis, Mpango (1996) found that in Tanzania imports are underinvoiced mainly to evade import duty.

Since there is evidence of trade faking in Tanzania, it is appropriate to adjust the capital flight measures for trade misinvoicing. The adjusted capital flight measures in constant and current prices are shown in Tables 7 and 8, respectively.

Table 7.Tanzania: Capital Flight Measures in Real Terms Adjusted for Trade Misinvoicing, 1971–93(In millions of 1977 U.S. dollars)
Balance of Payments ApproachWorld Bank Residual ApproachMorgan Guaranty Residual Approach1Bank Deposits Approach2
1971–362.53–243.29–201.52
1972–656.63–176.63–196.36
1973–395.49–40.28–65.05
1974–258.18–263.63–249.52
1975–130.2161.6030.28
1976–131.23156.87151.23
1977–253.26283.04302.54
1978–170.45–246.76–302.82
1979–119.10–48.6865.35
1980–169.69–190.24–190.24
1981–297.11–263.94–263.94
1982–177.60–172.95–172.95
1983–143.64–40.52–40.5229.84
1984–119.76–94.36–94.36–21.45
1985130.5734.9234.9271.67
1986–11.33157.61157.61–11.76
1987–144.66198.75198.758.46
19887.18–36.29–36.2914.65
198937.12–210.69–210.6985.76
1990–229.38229.00229.0028.59
1991–100.44–108.21–108.21–44.45
1992–125.12–81.43–81.43–83.85
1993–84.76–226.42–226.42–30.63
Totals
1975–77–514.70501.52484.05
1985–87–25.43391.28391.2868.37
1971–93–3,905.69–1,322.52–1,270.6546.84
Source: Author’s calculations.

The two residual approach measures coincide after 1979 because the banking system’s foreign assets are zero.

Data for calculating capital flight by the bank deposits approach during 1971–82 were not available.

Source: Author’s calculations.

The two residual approach measures coincide after 1979 because the banking system’s foreign assets are zero.

Data for calculating capital flight by the bank deposits approach during 1971–82 were not available.

Table 8.Tanzania: Capital Flight Measures in Nominal Terms Adjusted for Trade Misinvoicing, 1971–93(In millions of current U.S. dollars)
Balance of Payments ApproachWorld Bank Residual ApproachMorgan Guaranty Residual ApproachBank Deposits Approach
1971–120.06–74.96–59.16–31.86
1972–235.83–38.83–46.93–110.53
1973–243.26–26.76–41.86–95.96
1974–219.56–224.16–212.26–78.97
1975–123.6850.2221.8255.92
1976–127.91137.69132.49–71.81
1977–253.26283.04302.54–73.16
1978–209.74–295.64–358.7412.86
1979–195.11–101.6149.79–21.00
1980–297.70–329.80–329.80–125.60
1981–452.70–401.90–401.90–131.50
1982–264.55–257.65–257.65–12.95
1983–224.58–77.88–77.8822.22
1984–188.49–152.79–152.79–50.29
1985139.876.876.8757.97
1986–36.69200.81200.81–37.29
1987–288.92236.98236.98–54.42
1988–67.75–137.05–137.05–55.85
1989–77.40–488.00–488.003.20
1990–429.66393.74393.7433.74
1991–188.91–202.51–202.51–90.91
1992–248.27–170.47–170.47–174.77
1993–167.81–420.11–420.11–71.41
Sources: Author’s calculations.
Sources: Author’s calculations.

Capital Flight Measures Adjusted for Trade Misinvoicing

After adjusting for trade misinvoicing, the capital flight measures are significantly changed, and in some cases (especially in the bank deposits approach) capital flows in the reverse direction relative to the unadjusted measures (Tables 3 and 7).

According to the balance of payments approach, capital flight occurred only in 1985 and during 1988–89, when $174.87 million in real terms fled the country (Table 7). The general case for the rest of the years was one of capital reflows. During 1971–93, reflows totaled $3.9 billion. According to the bank deposits approach, there were several episodes of capital flight, and net capital flight during 1971–93 totaled $46.84 million. However, as discussed above, the balance of payments and bank deposits approaches are narrow measures and show only the minimum levels of capital flight.

The adjusted residual approaches indicate higher levels of capital flight than under the balance of payments and bank deposits approaches. Three episodes of capital flight are jointly reported by the two residual approaches, in 1975–77, 1985–87, and 1990. According to the residual approach, between $484.05 million and $501.52 million in real terms fled Tanzania during 1975–77, $391.28 million during 1985–87, and $229.0 million in 1990 (Table 7). The Morgan Guaranty (1996) approach reports an additional episode, that of 1979, when $65.35 million fled the country.

One can contemplate many other modifications to the measures of capital flows than trade misinvoicing and others discussed above. Such modifications would include adding interest income earned on bank deposits held abroad and income from tourism. All these aspects are taken into consideration when capital flight is defined as resident capital outflow (Lessard and Williamson, 1987).

A better picture of the magnitude of capital flight may be seen from some capital flight ratios. In analyzing capital flight ratios in Tanzania, we use our preferred Morgan Guaranty residual measure and express it as a percentage of total external debt, GDP, and merchandise exports and imports. These ratios are presented in Table 9.

Table 9.Tanzania: Capital Flight Ratios, 1971–931
Capital Flight as a Percentage of
Total external debtMerchandise exportsMerchandise importsGDP
1971–18.11–22.58–17.13–4.30
1972–11.59–14.84–13.04–3.00
1973–7.52–11.51–9.56–2.24
1974–26.83–53.17–32.14–9.46
19752.185.853.260.85
197611.1327.0223.844.55
197720.8756.1846.788.69
1978–19.85–75.37–36.15–8.60
19792.379.125.181.13
1980–13.32–56.60–30.28–6.42
1981–15.33–65.56–37.87–6.78
1982–8.84–62.40–27.06–3.86
1983–2.48–20.32–10.99–1.27
1984–4.51–38.34–20.10–2.74
19850.182.090.790.11
19864.6859.7821.994.43
19874.6182.3123.697.45
1988–2.53–35.46–13.27–3.71
1989–9.12–117.56–45.60–11.97
19906.4496.5533.1910.13
1991–3.14–55.91–15.76–4.75
1992–2.54–42.54–12.98–4.49
1993–6.13–90.93–32.32–12.12
Averages
1975–7711.3929.6924.634.69
1985–873.1648.0615.494.00
1971–93–4.32–18.44–8.50–2.10
Sources: Author’s calculations.

Capital flight refers to the Morgan Guaranty (1986) measure adjusted for trade misinvoicing. The ratios were of nominal variables in U.S. dollars.

Sources: Author’s calculations.

Capital flight refers to the Morgan Guaranty (1986) measure adjusted for trade misinvoicing. The ratios were of nominal variables in U.S. dollars.

Capital flight as a percentage of total external debt was most severe during 1975–77, when it averaged 11.39 percent. In 1979 it was less than 3 percent, and in 1985–87 the average was 3.16 percent. The ratio was 6.44 percent in 1990. As a percentage of merchandise exports, capital flight averaged between 29.69 percent and 48.06 percent during 1975–77 and 1985–87, respectively. In 1990, capital flight reached the remarkable level of 96.55 percent of the country’s merchandise exports. For imports, capital flight averaged 24.63 percent during 1975–77 and 15.49 percent during 1985–87; in 1990, capital flight was equal to 33.19 percent of the country’s imports (Table 9). Capital flight as a percentage of GDP ranged from 0.11 percent in 1985 to 10.13 percent in 1990. These figures indicate that, despite the reflows that occur, when capital flees the country it may reach critical levels and may have undesirable consequences for the domestic economy.

Consequences of Capital Flight

Whether or not a country imposes capital controls, the immediate consequence of capital flight is to reduce foreign exchange reserves, which may, in turn, require increased external borrowing to finance development expenditures. When both foreign exchange reserves and external borrowing capacity have been exhausted, the country will be forced to initiate balance of payments adjustment through either devaluation or equivalent expenditure switching policy, or by expenditure reduction. Devaluation will result in the reduction of domestic saving required for financing domestic investment and hence reduce future growth potential. The reduction in expenditure or demand also reduces output in the current and in future periods.

In countries that impose capital controls, capital flight is associated with a rising parallel market premium as the demand for foreign exchange increases owing to the need to finance capital flight. The premium will also result in the drain of some foreign resources from official reserves into the parallel market. In the extreme case, the diversion of reserves into the parallel market may exhaust the country’s foreign exchange reserves, forcing both the government and the private sector to increase their foreign borrowing.

Capital flight also has some consequences for income distribution, especially if the country follows the “origin” principle of taxation, since there may be little foreign investment that can be taxed. Even if Tanzania followed the “residence” principle of taxation, it would be difficult to tax flight capital since such capital is normally not reported to the tax authorities.

It is not easy to reach any strong conclusions on the consequences of capital flight without an empirical analysis, which, however, is beyond the scope of this chapter.

Political and Macroeconomic Developments in Tanzania During 1967–93

Episodes of capital flight from Tanzania reflect political and external shocks, which in turn have an impact on the performance of the macroeconomy. The shocks referred to here include the 1967 Arusha Declaration on African socialism and the consequent nationalization of private banks, manufacturing, estate farms, and wholesale and retail trade. Another political shock was the 1971 Buildings Acquisition Act, which nationalized private buildings and real estate. Other shocks included the 1973 oil price hike, the breakup of the East African Community in 1977, the 1978 Kagera War, the second oil price shock in 1979, the crackdown on “economic saboteurs” in 1983, the presidential elections of 1985 and 1990, and the implementation of economic recovery programs in 1986 and 1989.

Tanzania’s economic performance was better during the late 1960s to the mid-1970s than in the late 1970s and early 1980s. During 1967–75, for example, per capita GDP in real terms grew at an average annual rate of 1.0 percent; during 1976–85 it declined at an average annual rate of 1.3 percent (Table 10).2

Table 10.Tanzania: Per Capita GDP, International Reserves, Inflation, and Parallel Market Premiums, 1967–93
Per Capita GDP (In 1997 Tanzania shillings)International Reserves1 (In millions of 1997 U.S. dollars)Inflation2 (In percent)Parallel Market Premium3 (In percent)
19671,235257.1913.0421.85
19681,266341.0015.3819.05
19691,261338.2916.6721.85
19701,277260.002.8641.46
19711,287262.405.5662.46
19721,333498.547.89112.89
19731,334296.189.76106.55
19741,31562.3220.0089.34
19751,33392.3825.93179.51
19761,348146.136.62161.34
19771,320281.8011.72159.35
19781,303106.3511.7369.91
19791,30668.0413.8145.99
19801,30216.6430.20156.10
19811,26017.3125.37233.33
19821,2334.6329.27251.29
19831,17219.4026.96255.48
19841,17826.5636.12273.45
19851,19916.6033.33280.65
19861,18856.2732.40392.35
19871,21337.7829.98172.33
19881,21982.7131.20112.51
19891,23357.7025.8283.78
19901,252210.7119.7149.90
19911,287225.9622.2959.02
19921,296394.3522.0736.34
19931,312244.9525.3211.04
Sources: Computed using data from World Bank (1995); Tanzania, Bureau of Statistics; Bank of Tanzania; International Montary Fund (1996); and Kaufmann and O’Connell (1992).

World price of Tanzanian exports was used as the deflator.

Changes in the average domestic consumer price index.

Parallel market premium (PREM) defined as 100*(EUEO)/EO, where EU and EO are, respectively, the unofficial and the official nominal exchange rates between the Tanzania shilling and the U. S. dollar.

Sources: Computed using data from World Bank (1995); Tanzania, Bureau of Statistics; Bank of Tanzania; International Montary Fund (1996); and Kaufmann and O’Connell (1992).

World price of Tanzanian exports was used as the deflator.

Changes in the average domestic consumer price index.

Parallel market premium (PREM) defined as 100*(EUEO)/EO, where EU and EO are, respectively, the unofficial and the official nominal exchange rates between the Tanzania shilling and the U. S. dollar.

International reserves in constant prices declined at a relatively low annual rate of 12.0 percent during 1967–76, compared with the drastic annual decline of 21.5 percent during 1976–85 (Table 10).

During 1986–93 growth was restored and the balance of payments crisis was ameliorated. Real per capita GDP during 1986–93 grew at an average annual rate of 1.4 percent, while international reserves in constant prices increased at the impressive rate of 23.4 percent a year.

Domestic inflation seems to have been fairly constant throughout 1967–93, when it averaged 20.4 percent. During our capital flight sample period of 1971–93, inflation averaged 21.9 percent (Table 10).

The performance of the economy has implications for capital flight. Poor economic growth is an indication of low profitability of domestic investment; capital will thus tend to flee the country. A balance of payments crisis is a symptom of possible devaluation as one of the measures to correct the balance of payments deficit. Since a devaluation reduces the value of domestic financial assets, private agents will protect the real value of their assets by acquiring foreign assets, some of which will find their way out of the devaluing country.

The performance of the Tanzanian economy and the political aspirations of the country decided the nature and level of government intervention in the economy. Government intervention, in turn, had varying impacts on resident capital flows.

Before the 1967 Arusha Declaration on African socialism and self-reliance, Tanzania had a minimum level of government intervention in economic affairs. This noninterventionist policy was supported by conservative fiscal and monetary policies (Kaufmann and O’Connell, 1997). During this time, there was less incentive for capital flight, especially during the operation of the East African Currency Board, when there were no capital controls within the sterling area. The desire for capital flight from Tanzania was intensified following the dismantling of the East African Currency Board in 1965 and the promulgation of the Arusha Declaration of 1967 (O’Connell, 1990). The dismantling of the currency board was followed by the prohibition against moving capital out of East Africa (that is, outside Kenya, Tanzania, and Uganda).

Following the Arusha Declaration, the government of Tanzania nationalized all the “commanding heights of the economy.” The major private import-export firms were nationalized and replaced by the State Trading Corporation. The banking sector was also nationalized, and the National Bank of Commerce took over the activities of trade finance. In line with the Arusha Declaration, the Building and Acquisition Act was passed in 1971, under which most buildings and real estate belonging to the private sector were expropriated. The act increased expropriation risks and the incentive for capital flight.

Government control of the economy—especially capital and foreign exchange controls—was evident in the parallel market premium. The controls drove a wedge between the parallel and official market exchange rates, giving rise to a premium, the behavior of which reflected the changing nature of the controls (Ndulu and others, 1995). The trade and foreign exchange controls remained relatively low during 1967–71, as reflected in low parallel market premiums, which averaged 33.33 percent in that period. With the heightening of the nationalization policies after 1971, coupled with balance of payments crises, trade and foreign exchange controls were intensified; this gave rise to a booming parallel market in foreign exchange in which the premium increased from 62.46 percent in 1971 to a peak of 392.35 percent in 1986, growing at an average annual rate of 13.03 percent. With the more liberal policies beginning in 1986, the controls were eased and the parallel market premium declined to 11.04 percent in 1993, falling at an average annual rate of 60.04 percent between 1986 and 1993.

In response to the mini–balance of payments crisis in 1974, the government instituted and then intensified capital and foreign exchange controls (Kaufmann and O’Connell, 1992 and 1997). Despite the strict and severe trade and exchange controls instituted in the early 1970s, the balance of payments showed no significant improvement and the economic decline continued. The country was then forced to implement some structural adjustments. Beginning in mid-1984, the government instituted the own-fund import license scheme, under which individuals with access to their own (unofficial) foreign exchange sources were allowed to obtain import licenses without being asked about the source of their funds. Licenses issued under the own-funds scheme accounted for about 40 percent of total import licenses issued in 1988. The share of own-funded imports in total imports is estimated to be larger than that of the licenses, probably exceeding 50 percent of total imports (Kaufmann and O’Connell, 1992).

The capital flight that occurred during 1985–87 could be partly explained by the need to finance own-funded imports. The value of such imports is therefore the reverse of capital flight. The sources of funds for own-funded imports are the decumulation of foreign exchange and illegal trade. Individuals with their own foreign exchange (abroad) may sell it in the parallel foreign exchange market or use it directly to purchase own-funded imports. Those who do not hold foreign exchange may receive grants or borrow from relatives and associates abroad. Illegal trade includes, among other things, smuggling of minerals, elephant tusks, rhinoceros horns, and export crops (mainly coffee), as shown in Table 11. However, not all foreign exchange generated through illegal trade is used to finance own-funded imports. Some of it may be used to increase private foreign assets abroad and thus constitutes capital flight.

Table 11.Tanzania: Some Reported Cases of Corruption, 1987–94
Amount Involved
In millions of U.S. dollarsIn percent of exportsDescription
19870.180.06Smuggled elephant tusks ($0.07 million), smuggled coffee ($0.11 million)
19880.310.08Reported traveler’s checks stolen from the National Bank of Commerce
19898.862.13Smuggled elephant tusks
19901.450.36Money lost primarily through fake contracts at the Tanzania Breweries Limited ($0.61 million) and smuggled garments ($0.67 million)
19910.160.04Gold and gemstones seized at the Dar es Salaam International Airport
199259.2214.78Money lent to government institutions that has not been accounted for
199343.5026.85Payments for services not rendered at the Ministry of Home Affairs
19943.200.63Open general license and debt conversion program funds illegally paid to a private businessman ($3.19 million) and illegal drugs ($0.01 million)

Economic reforms were further enhanced with the introduction and implementation of economic recovery programs in 1986 and 1989. The main item of the reforms included devaluation of the local currency. The recovery programs were supported by massive inflows of foreign aid that increased the availability of capital (Nyoni, 1995). These capital inflows and the devaluations of 1986–87 provided a great incentive for capital flight during these two years and in 1990.

Another important political development was the crackdown on “economic saboteurs” in 1983. Under this policy, many private entrepreneurs were jailed or were forced to go into hiding with their capital (Bagachwa and Maliyamkono, 1990). The “lucky” ones managed to acquire foreign assets and shift them to foreign countries.

The combination of uncertainties following the breakup of the East African Community in 1977, together with increased capital availability and looser exchange controls following the 1976 coffee boom, led to a massive exodus of capital during 1975–77, totaling $484.05 million in real terms using the adjusted Morgan Guaranty approach (Table 7). The risks associated with the 1978–79 Kagera war and the 1979 oil price hike seem to explain the capital flight of $65.35 million that occurred during 1979.

The uncertainties associated with the 1985 and 1990 presidential elections and the increased availability of capital following the implementation of the 1986 and 1989 economic recovery programs seem to explain the capital flight that occurred during 1985–87 and in 1990, when flight capital in 1977 prices totaled $391.28 million and $229.00 million, respectively.

Another plausible explanation for the capital flight that occurred in the last half of the 1980s and early 1990s in Tanzania is corruption. Corruption in the country is well documented in Msambichaka and others (1994), Chachage (1994), and Mbatia (1994). The reported amounts of money involved are sometimes quite substantial, reaching up to 27 percent of the country’s exports (Table 11). Money obtained through corruption is most likely be deposited in foreign banks to avoid detection and prosecution at home.

From the discussion of the measurement of capital flight and trade misinvoicing, one may say that the conduits of capital flight in Tanzania include foreign exchange bureaus, illegal trade, corruption, and trade misinvoicing. Foreign exchange can also be carried physically across the borders through legal or illegal means. Money obtained through illegal trade or corruption can only safely flee the country through illegal means. Funds purchased through foreign exchange bureaus or other official sources are allowed by law to be taken out of the country for financing current account transactions. Capital can then flee the country through overinvoiced imports that are financed by foreign exchange purchased in either the official or the parallel market. Capital can also flee the country through underinvoiced exports. The foreign exchange obtained from trade misinvoicing may be used by those involved to increase foreign asset holdings by domestic residents.

Empirical Analysis of Capital Flight in Tanzania

Causes of Resident Capital Outflow

The discussion in this chapter of the causes of capital flight is based mainly on Lessard and Williamson (1987) and Pastor (1990). Other contributors to the subject include Cuddington (1986), Dornbusch (1985), and Khan and Ul Haque (1985). In discussing the causes of resident capital outflows one must also look at the circumstances in both source and haven countries and identify factors that would distinguish “normal” from “abnormal” flows.

In general, “resident capital outflow results from a difference in perceived, risk-adjusted returns in source and haven countries” (Lessard and Williamson, 1987, p. 215). The risk (and the returns) may be economic or financial. Economic or social returns are the total returns to a given society derived from a particular investment and adjusted for price distortions and externalities. Financial or private returns are the total returns that accrue to private investors and are calculated at market prices, net of both explicit and implicit taxes.

Capital flight occurs when the overall investment climate is negative for domestic investors. “Overall investment climate” refers to factors that influence the financial attractiveness of source country assets relative to the world standard. These factors include domestic inflation, fiscal deficits, the real interest rate differential between the domestic economy and other economies, expected devaluation, and the differential in expected rates of profitability between the domestic economy and other economies.

Increasing domestic inflation reduces the real value of the domestic currency. To protect the real value of their domestic currency holdings, private agents may substitute foreign assets for these holdings. Some of these foreign assets may be siphoned out of the country, thus constituting capital flight. Increasing fiscal deficits create inflationary expectations and increased tax liabilities to the private sector, since the deficits will have to be financed by either printing money or selling government bonds (Ajayi, 1995). To avoid increased tax liabilities and protect the real value of their assets, domestic investors will reduce their domestic assets and increase foreign assets, some of which will be shipped abroad.

When there is financial repression or there are ceilings on interest rates, real interest rates will be negative, and this will fuel capital flight. Capital flight will also occur if real interest rates on savings in the domestic economy are lower than those obtainable in foreign countries.

Other factors explaining capital flight include devaluation and the relative profitability of the economy. Depreciation of the local currency will reduce the value of domestic assets. When the currency is considered overvalued, a devaluation is expected. The expectation of devaluation will, in turn, propel the exodus of capital from the domestic economy. Capital flight is also caused by low anticipated profitability in the domestic economy relative to that in other countries. Thus, if growth in the real sector in Tanzania is lower than real growth in, say, the United Kingdom, it is a signal that profitability in Tanzania is low; capital will flee to the United Kingdom, where the expected profitability is higher.

Another explanation for capital flight is political instability and the potential for confiscation (Khan and Ul Haque, 1985; Lessard and Williamson, 1987). With high expropriation risks, capital flight will occur even if the relative profitability of capital is higher in the domestic economy than abroad.

Capital availability is a potential cause of capital flight (see, for example, Pastor, 1990; Cuddington, 1986). Increased capital availability may be due to increased foreign aid inflows, a boom in export earnings, or disbursement of public and publicly guaranteed loans. In countries where both the private and the public sector can borrow abroad and private investors may purchase government debts, private investors can shift their own capital abroad and acquire publicly guaranteed external loans at home. When domestic investment is no longer profitable, so that the loans become unserviceable, the private sector may avoid payments on the debt obligations by either “declaring bankruptcy or pressuring the local government to bail out private borrowers through the assumption of their liabilities” (Pastor, 1990, p. 8). When this happens, there will be increased external debt simultaneous with increased capital flight as capital becomes more available. The government thus ends up with the debts, and the elites with the dollars (Duwendag, 1989).

As Lessard and Williamson (1987) point out, it may not always be possible, to take into account all the factors that explain international capital flows. Some economists have decided to take only a subset of the factors by focusing on the overall investment climate (see, for example, Cuddington, 1986; Conesa, 1987; Dornbusch, 1985). The emphasis in this doctrine is on overvaluation in terms of the real exchange rate, on fiscal deficits, and on inflation. Others emphasize discriminatory tax treatment of residents versus nonresidents, financial repression, capital denominated in different currencies, and asymmetric investment guarantees (Dooley, 1986; Khan and Ul Haque, 1985; Eaton, 1987; Ize and Ortiz, 1987).

Earlier Studies

Most studies on capital flight have been done for Latin America. Cuddington (1986) and Conesa (1987), for example, emphasized the overall investment climate, while Dooley (1986) focused on the discriminatory treatment of resident versus nonresident capital. Studies on capital flight from Africa include those by Ajayi (1995) for Nigeria, Awung (1996) for Cameroon, Baruani (1995) for Tanzania, Ng’eno (1994) for Kenya, and Olopoenia (1995) for Uganda.

Cuddington (1986) estimated the economic determinants of resident capital outflow for four major flight countries: Argentina, Mexico, Uruguay, and Venezuela. In his portfolio-adjustment model, the factors that explained resident capital outflow were the expected inflation rate, the foreign interest rate, the domestic interest rate, and disbursement of public and publicly guaranteed loans. Depending on the perceived risks and returns, domestic investors will allocate their wealth among domestic financial assets, domestic inflation hedges such as land, and foreign financial assets. An increase in the foreign financial assets owned by domestic residents constitutes resident capital outflow.

The empirical findings by Cuddington (1986) differed from country to country. For Mexico, the author found that capital flight was highly correlated with currency overvaluation, disbursements of public debt, and lagged capital flight. In Argentina and Uruguay, only the lagged real effective exchange rate and the lagged error of the model were correlated with capital flight. In Venezuela, the only statistically significant determinants were overvaluation and foreign interest rates.

The model results of Conesa (1987) are similar to those of Cuddington (1986), except that Conesa had 16 annual observations whereas Cuddington had only 9, and Conesa had (lack of) growth as an additional explanatory factor. The other difference between the two studies is that, unlike Cuddington (1986), Conesa (1987) did not attempt to estimate overvaluation in terms of the real effective exchange rate, but used the level of government external borrowing and reserves as proxies for overvaluation. Like Cuddington (1986), Conesa’s (1987) results also differed from country to country.

In a study of seven developing countries—Argentina, Brazil, Chile, Mexico, Peru, the Philippines, and Venezuela—Dooley (1988) found that capital flight was significantly determined by domestic inflation, financial repression, and a measure of the country risk premium.

The study by Ajayi (1995) provided a link between capital flight and external debt in Nigeria. He concluded that most capital flight from Nigeria was recorded in the balance of payments and external debt statistics and that capital flow is explained not only by economic factors but also by political instability.

Baruani (1995) did an economic analysis of capital flight from Tanzania. The author measured capital flight using both Cuddington’s (1986) balance of payments approach and the World Bank (1985) residual approach. This chapter complements the Baruani (1995) study with adjustments for trade misinvoicing and the Morgan Guaranty (1986) residual approach and the bank deposits approach.

Econometric studies on Africa seem to suggest that capital flight results mainly from macroeconomic mismanagement, especially domestic inflation. Olopoenia (1995), for example, found that inflation was the only significant variable (with a positive coefficient of 0.176) propelling capital flight from Uganda. The Ng’eno (1994) study for Kenya found that the coefficient for inflation was positive 0.9, whereas that for lagged capital flight was positive 0.7.

The R2 in Ng’eno’s study was 0.98, which contrasts sharply with a similar study by Olopoenia (1995) for Uganda in which the R2 was less than 0.30. In an earlier study on economic analysis of capital flight, Ajayi (1992) found that capital flight from Nigeria was triggered by the real interest rate differential, growth of the domestic economy, changes in the nominal exchange rate, the fiscal deficit, and the foreign interest rate adjusted for changes in the nominal exchange rate. The models in Ajayi (1992) had low adjusted R2 ranging from 0.32 to 0.49. The low explanatory power in Ajayi (1992) and Olopoenia (1995) raises concern about the difficulty of estimating capital flight in African countries, which arises mainly from short sample periods, low data frequency, and poor quality of the data.

The Data

In carrying out econometric analysis of capital flight, we used annual data for 1971–93. However, after transforming the data into first differences, the estimation sample was reduced to 20 years, covering only 1973–92. The brevity of the sample period was dictated by the availability of data, most of which are compiled on an annual basis.

The main data sources for the empirical analysis were the IMF’s International Financial Statistics and Direction of Trade Statistics, the World Bank’s World Tables and African Economic Indicators, and Kaufmann and O’Connell (1992). The data on political and external shocks were obtained from the author’s own observations and from Bagachwa and Maliyamkono (1990). Data on domestic interest rates, inflation, and GDP were obtained from the Tanzanian Bureau of Statistics (1995a, 1995b) and various issues of the Bank of Tanzania’s Economic and Operations Report.

The Model and Model Results

Following Cuddington (1986), Lessard and Williamson (1987), and Pastor (1990), we specify an implicit capital flight model as follows:

where t is a time subscript, KF is capital flight in 1977 prices (measured by the Morgan Guaranty approach adjusted for trade misinvoicing), KFt-1 is lagged capital flight, GUTD is the difference between the U.K. and Tanzanian real GDP growth rates (serving as a proxy for the relative profitability of investment in the domestic real sector), PREM is the parallel market premium (as a proxy for foreign exchange controls and the degree of real exchange rate overvaluation), DINF is domestic inflation, FINC is the financial incentive for capital flight as defined in Table 12, KAV is a measure of capital availability proxied by net long-term capital flows expressed in 1977 prices, and POS and EXS are, respectively, dummy variables for political and external shocks. Capital flight is an increasing function of all the explanatory variables specified above. The regression variables are further explained in Table 12 and data for these variables are given in Table 13.

Table 12.Definitions of Variables in the Regression Model
Expected signVariableDefinition
KFtCapital flight in 1977 prices measured using the Morgan Guaranty approach (deflated by import price).
+KFt-1Lagged capital flight.
+FINCtFinancial incentive for capital flight. It is measured as FINC = (iUK – (iTZ – Δe)) where iUK is the U.K. deposit interest rate; iTZ is the Tanzanian deposit interest rate; and Δe is devaluation or the change in the Tanzania shilling-pound sterling nominal exchange rate.
+EXStDummy variable for external shock. EXS = 1 for year with shock, 0 if otherwise.
+POStDummy variable for political shock. POS = 1 for year with shock, 0 if otherwise.
+PREMtPremium on the U.S. dollar in the parallel market. Measured as 100*(EUEO)/EO, where EU and EO are the unofficial or parallel market and the official Tanzania shilling–U.S. dollar nominal exchange rates, respectively.
+DINFtDomestic inflation in Tanzania. Measured as the change in the average consumer price index.
+KAVtAvailability of capital. Measured as net long-term capital flows deflated by the world price of imports.
+GUTDtGrowth differential. Measured as GUTD = gUK – gTZ where gUK and gTZ are, respectively, growth rates of U.K. and Tanzanian real GDP.
Table 13.Data for Selected Variables Used in the Regressions1
GUTDPREMEXSPOSFINCDINFKAV
1971–2.1962.46013.695.56364.04
1972–3.22112.89002.447.91264.12
19734.30106.5510–3.659.76228.88
1974–4.2089.08002.1020.00147.78
1975–6.41179.89002.8325.92227.67
1976–3.83161.3400–5.086.62111.08
19771.96159.35100.0311.7296.60
19781.3383.99102.5111.73120.83
19790.3845.991019.4913.81169.76
1980–5.15156.100015.5330.10185.08
1981–0.79233.3300–4.7525.38192.83
19821.13251.2900–4.3429.16232.56
19836.05255.480121.5526.96167.50
1984–1.06273.450020.3836.12–32.93
19851.13280.65018.8933.3314.81
19861.02392.3511223.7532.40857.01
1987–0.27172.330093.1029.99144.44
19880.77112.510032.4831.20163.72
1989–54.3283.781119.5025.82172.97
19904.0749.90017 9019.71215.89
1991–7.6659.02001.3522.29130.37
1992–4.0636.34003.0422.0791.07
1993–1.9111.040020.0123.46180.74
Sources: International Monetary Fund (1996); World Bank (1995) and World Tables, various issues; Kaufmann and O’Connell (1992); Bagachwa and Maliyamkono (1990).

The variables are as defined in Table 12.

Sources: International Monetary Fund (1996); World Bank (1995) and World Tables, various issues; Kaufmann and O’Connell (1992); Bagachwa and Maliyamkono (1990).

The variables are as defined in Table 12.

The variables were first tested (results not shown) for stationarity using the augmented Dickey-Fuller test (Hendry, 1989). The only variables found to be stationary were the contemporaneous and lagged capital flight, the U.K.-Tanzania real growth differential, and the two dummy variables for external and political shocks.

The regression analysis was carried out using PC-GIVE version 8 software developed by Hendry (1989). Attempts to run a cointegration model yielded very poor results (not reported here) with no explanatory variable significant at the conventional significance levels. The main problem could be that of the short sample period and low-frequency annual data. To avoid spurious regression, we ran the capital flight model by using the variables in first differences.

The empirical model is explicitly specified as

where αi (i = 1, …, 8) are coefficients; KFt, KFt-1, GUTDt, PREMt, DINFt, FINCt, KAVt, EXSt, and POSt are as defined above; μt is the error term; and Δ is the difference operator for the indicated variables.

The regression results are presented in Table 14. The model has a goodness of fit of 0.63. We also attempted to run the model without capital availability or lagged capital flight. Omitting either of the two variables reduced the R2 by more than 50 percent. This strengthens the F-test for the R2 that the explanatory variables are jointly significant and explains 63 percent of the variations in capital flight.

Table 14.Tanzania: Determinants of Capital Flight
VariableCoefficientStandard Errort-Value
Constant–43.26372.892–0.594
ΔKFt-1–0.5550.202–2.744
ΔGUTDt7.0562.8842.446
ΔPREMt1.3061.1571.129
ΔDINFt–3.9827.348–0.542
ΔFINCt2.5311.9621.290
ΔKAVt–0.8830.503–1.755
POSt144.280127.3301.133
EXSt51.109139.7000.366
R2 = 0.63; F(8,11) = 2.31; RSS = 445636.77; Adjusted R2 = 0.36; period = 1973–92
Source: Author’s calculations using ordinary least-squares estimation.
Source: Author’s calculations using ordinary least-squares estimation.

At the level of individual variables, the only statistically significant variables in explaining capital flight from Tanzania are the U.K.-Tanzania real GDP growth differential. The growth differential variable bears the expected positive sign. Lagged capital flight, however, bears an inverse sign, suggesting that lagged capital flight in Tanzania is associated with reverse capital flight.

As mentioned above, since the model passes the F-test for goodness of fit, all the variables in the model are jointly significant. Capital flight is then jointly determined by the U.K.-Tanzania real GDP growth differential, domestic inflation, political and external shocks, lagged capital flight, the parallel market premium, financial incentives for capital flight, and capital availability. Looking at the coefficients of the variables, it can be seen that political and external shocks in Tanzania have the overriding impact of propelling capital flight from the country.

Conclusion

Summary of the Findings

This chapter has analyzed the magnitude and determinants of capital flight from Tanzania. Capital flight was measured by the balance of payments approach, the World Bank (1985) and Morgan Guaranty (1986) versions of the residual approach, and the bank deposits approach. We argued that the measures should be adjusted for trade misinvoicing to get more accurate estimates of capital flight. We found that the Morgan Guaranty measure was more robust than the other two measures. The adjusted Morgan Guaranty measure indicates that capital flight episodes in Tanzania occurred during 1975–77, 1979, 1985–87, and 1990. The total amount of capital that fled the country during those years was $1.2 billion in 1977 prices.

The analysis of the causes of capital flight was carried out using both historical and empirical approaches. From the historical analysis, we found that trade misinvoicing and smuggling were important mechanisms of effecting capital flight. For this reason, the capital flight measures were adjusted for trade misinvoicing.

Corruption was also found to be an important vehicle for effecting capital flight from Tanzania. Cases of corruption were identified to include stealing of foreign exchange from the state-owned National Bank of Commerce, foreign exchange lost through fake contracts in public enterprises and government departments, and siphoning out of aid monies.

Political instability increased fears of expropriation risks and damage to private property. Such risks, in turn, increased the desire for capital flight. Instances of political instability were cited to include the nationalization policy following the 1967 Arusha Declaration on African socialism, the 1971 Buildings Acquisition Act, the Kagera war of 1978–79, the crackdown on “economic saboteurs” in 1983, and the presidential elections of 1985 and 1990.

From the empirical analysis, we observed that the only statistically significant factor explaining capital flight from Tanzania was the U.K.-Tanzania real GDP growth differential. Lagged capital flight was also statistically significant but was associated with reverse capital flight. It was also observed that capital flight was jointly determined by the U.K.-Tanzania real growth differential, lagged capital flight, external and political shocks, the parallel market premium, domestic inflation, capital availability, and financial incentives for capital flight.

Policy Implications

We can draw several policy implications from the historical and empirical analyses. To control capital flight, it is important to restore public confidence in the government and the sustainability of macroeconomic policy. This, in turn, requires a realistic exchange rate, positive but moderate real interest rates, robust economic growth, and an honest government.

The econometric analysis of capital flight showed that the exodus of capital from Tanzania was strongly determined by the U.K.-Tanzania real growth differential. This underscores the importance of a pro-growth bias of official policy in the country. It is important to avoid excessive austerity in monetary and fiscal policies so as not to choke off production. Pro-growth policies include market liberalization and export promotion measures.

To control capital flight and recapture flight capital requires getting the macroeconomic fundamentals right; that is, policymakers should avoid the habit of allowing the local currency to be overvalued and thus minimize the necessity of devaluation. This may be achieved by allowing the currency to float or by setting it at an appropriate level and making sure that inflation is controlled so that the currency is not overvalued in real terms. This also warns against expansionary fiscal and monetary policies, as these generate expectations of inflation and the desire for capital flight. For this policy to be effective, it is important to maintain public confidence that the policy will not be reversed under any circumstances.

Restoration of public confidence in the sustainability of macroeconomic policy also requires setting domestic interest rates above international rates. However, care must be taken not to set the rates too high, since too-high interest rates will discourage domestic productive investment and trigger capital flight.

Even if the macroeconomic fundamentals are set right, a country may continue to experience repeated bouts of capital flight if corruption is rampant. Some of the income illicitly gained from bribes on public sector contracts or from import licenses is likely to be shifted abroad to avoid detection and prosecution in the home country. Policy change is thus necessary to reduce such illicit incomes as well as incentives for corruption and capital flight. The establishment of a uniform tariff rate structure, lower tax rates, and an end to the rationing of underpriced foreign exchange and credits will greatly reduce incentives not only for corruption but also for trade misinvoicing and capital flight.

From both the historical and the econometric analyses, we found that political shocks are important in propelling capital flight. Hence, policymakers should strive to maintain peace and calm in the country and avoid policies that are hostile to the private sector and property. Such hostile policies include the crackdown on “economic saboteurs” and the expropriation of private property.

References

I would like to thank Professor S. Ibi Ajayi for pointing out the need to adjust the change in external debt for exchange rate variations. However, since the sources used for the external debt statistics report the data in U.S. dollars, there is less need to adjust the external debt for exchange rate variations.

The averages and growth rates referred to in this section were computed from Table 10.

    Other Resources Citing This Publication