Abstract

This workshop provides background information and discusses possible techniques for forecasting balance of payments developments in Kenya during the period of a hypothetical financial program.1 Balance of payments forecasting is particularly important in developing countries with large foreign sectors2 and potentially unstable export receipts. Further, forecasting the possible need for balance of payments assistance and the impact of proposed adjustment policies represents a crucial element of financial programs linked to the use of the International Monetary Fund's resources.

This workshop provides background information and discusses possible techniques for forecasting balance of payments developments in Kenya during the period of a hypothetical financial program.1 Balance of payments forecasting is particularly important in developing countries with large foreign sectors2 and potentially unstable export receipts. Further, forecasting the possible need for balance of payments assistance and the impact of proposed adjustment policies represents a crucial element of financial programs linked to the use of the International Monetary Fund's resources.

One of the more difficult parts of an overall macroeconomic forecast is the balance of payments. By definition, the external sector involves interrelationships with the rest of the world and therefore necessitates assumptions about developments in other countries. Items entering the balance of payments of one country must be recorded in the balance of payments of another.3 Even apart from the need to consider developments in other countries, the diversity of balance of payments components complicates analysis and considerable differences may be expected in the types of behavorial relationships used to explain items in the trade, services, or capital accounts. Further, the development of stable behavioral relationships for specific items may be particularly difficult in countries with some form of import or foreign exchange controls.

Reflecting these difficulties, the workshop provides material for an eclectic approach to balance of payments forecasting. The first section contains information on the structure of Kenya's external sector and recent economic developments affecting balance of payments. The second section presents a general analysis of factors that might influence developments in the external sector. The final section gives some guidelines for forecasting balance of payments developments.

THE EXTERNAL SECTOR IN KENYA

Analysis of the structure of the foreign sector, knowledge of the major policy measures significantly affecting it, and data on recent balance of payments developments are a prerequisite of any forecasting exercise.

Structure of the Foreign Sector

Imports disaggregated according to economic category are presented in Table 1. About 85 per cent of Kenya's 1976 imports were intermediate and capital goods, with industrial inputs and fuels each representing about a third of this total. Consumer goods imports were relatively unimportant, accounting for only 14 per cent of the 1976 imports. Western Europe supplies about 45 per cent of imports, with the United Kingdom and to a lesser extent the Federal Republic of Germany as the major sources (Table 2). Reflecting the increased price of oil, the share of Middle Eastern countries in total imports has risen sharply.

TABLE 1.

Kenya: Total Imports by Economic Category, 1973–77

article image
Source: Kenya, Economic Survey, 1977 and 1978.

Imports are influenced by both economic forces and the system of trade and exchange controls. All nongovernment imports require a license by the Central Bank of Kenya, but this approval is granted automatically for imports on open general license. In addition, Kenya maintains four import schedules. Goods under Schedules I and II require specific import licenses to be issued by the Director of Trade and Supplies. Foreign exchange approval is granted freely for goods under Schedule I and, on the basis of quotas reflecting past imports, for goods under Schedule II. Imports under Schedule III are generally prohibited; for goods on Schedule IV no specific import license is required, but foreign exchange is allocated on a quota basis. Goods listed under Schedules I, II, and III are items produced locally, and the lists were introduced to protect local producers. Items under Schedule IV are not produced locally, and the foreign exchange allocation was introduced for balance of payments reasons. Schedules II, III, and IV cover approximately 25 per cent of the 1973 imports. About 11 percentage points of this total reflect an expansion in coverage introduced in June 1974 and a further 4 percentage points arose from changes in June 1975.4 No major changes were made in the system of exchange control in 1976. In addition to changes in the list of restricted items, the effective tightness of controls can be influenced by administrative decisions on the allocation of quotas for goods under Schedules II and IV.5

TABLE 2.

Kenya: Direction of Trade, 1973–77

(As percentage of total)

article image
Source: Kenya, Economic Survey, 1977 and 1976.

Including Australia.

Components may net add to totals because of rounding.

Exports are dominated by primary products, of which the most important are coffee and tea (Table 3). Although these two commodities account for over 45 per cent of non-oil exports, Kenya has only a small share of total world exports.6 Kenya refines imported crude oil for export to neighboring countries and for sale as fuel for aircraft and ships. Oil products accounted for about 20 per cent of 1976 exports. In 1976 the United Kingdom was the market for only 10 per cent of Kenya's exports, which was a little less than the share of the Federal Republic of Germany. African countries, and particularly the East African Community (EAC),7 represented important export markets. In part reflecting oil exports, Kenya had a substantial trade surplus with the EAC countries.

Kenya has in recent years benefited from a substantial surplus in the service account, and in 1976 receipts represented 59 per cent of exports while payments represented 33 per cent of imports. Major receipts arise from port services and tourism, while investment income represents the most important debit item. The balance of payments is also affected by substantial long-term capital inflows (Table 4).

TABLE 3.

Kenya: Composition of Exports, 1973–77

article image
Sources: Kenya, Economic Survey, 1978 and Statistical Abstract, 1977.

Non-oil and oil exports do not total 100 per cent because of the exclusion of re-exports.

Includes petroleum by-products.

Includes aircraft and ships' stores.

Consists of imported goods re-exported in the same form to countries outside the East African area.

On October 27, 1975, Kenya adopted a new central rate of KSh 9-66=SDR 1, a depreciation of 12.4 per cent from the exchange rate prevailing at that time. The Kenya shilling was then pegged to the SDR rather than to the U.S. dollar, as it had been previously.

TABLE 4.

Kenya: Balance of Payments Summary, 1973–77

(In millions of Kenya shillings)

article image
Sources: For Items A and B (except commercial banks), the data for 1976 and 1977 are taken from Kenya, Economic Survey, 1978; the data for 1974 and 1975 from Kenya, Economic Survey, 1977; and the data for 1973 from Kenya, Economic Survey, 1976. The data for commercial banks (In Item B) and financing (Item E) are taken from International Monetary Fund, International Financial Statistics, July 1978, in order to ensure consistency with the monetary accounts for purposes of financial programming (see Workshop 9: Financial Programming). The adjustment for the resulting discrepancy is contained in the residual item, errors and omissions (Item C).

Data have been adjusted to include capital grants in government transfers and exclude them from capital movements.

Includes valuation changes.

Recent Balance of Payments Developments

There was a sharp deterioration in the balance of payments situation in 1974 as a result of an approximate tripling of the trade deficit. Imports grew by about 68 per cent, with the value of oil imports rising to 3-5 times the 1973 level and non-oil imports increasing by nearly 50 per cent. To an important extent, these increases reflected higher import prices, but there was also a substantial rise in the import volume of intermediate goods for stockbuilding purposes (Charts 1 and 2). Exports grew by more than 30 per cent, with much of the rise occurring in oil exports. The deterioration in the trade balance was partially offset by a substantial rise in the surplus in services. This largely resulted from the increased value of sales of fuel for ships and airplanes (Chart 3). There was also a substantial inflow of capital; net private nonbank borrowing in 1974 increased by 42 per cent, in part reflecting trade credits for the higher imports and in part due to the effect of the restrictions introduced in June on foreign company borrowing in the Kenyan market (Chart 4). The overall balance of payments position changed from a surplus of KSh 187 million in 1973 to a deficit of KSh 544 million in 1974, or about 3 per cent of the nominal GDP in 1974.

In 1975, the trade deficit declined by 19 per cent to KSh 2,566 million, owing to a reduction in imports of about 25 per cent in volume and 6 per cent in value. In part this decline reflected an adjustment from the high level of inventory accumulation in 1974, but there may also have been some administrative tightening in the system of import controls. The value of exports grew by only 2 per cent, with a rise in oil exports more than offsetting the decline in non-oil exports. This latter development reflected the impact of drought on coffee exports, a sharp decline in the volume and price of sisal exports, and a reduction in exports to Uganda. The surplus on services remained at about the same level as 1974. There was a decline in net capital inflow, with an increase in net public sector borrowing failing to offset a fall in net private foreign borrowing. The overall deficit in 1975 declined to KSh 376 million, or about 1.8 per cent of nominal GDP.

Chart 1.
Chart 1.

Kenya: Import Ratios, 1970–771

(As percentage of GDP)

1 Imports and gross domestic product at factor cost, measured at constant 1972 prices.Note: Derived from the data in the Appendix.
CHART 2.
CHART 2.

Kenya: Import Prices, 1970–77

(1969=100)

Note: Derived from the data in the Appendix.
CHART 3.
CHART 3.

Kenya: Services and Unrequited Transfers, 1970–77

(In millions of Kenya shillings)

Note: Derived from the data in the Appendix.
CHART 4.
CHART 4.

Kenya: Nonmonetary Capital Flows, 1973–77

(In millions of Kenya shillings)

Note: Derived from the data in the Appendix.

There was a sharp turnaround in the balance of payments situation in 1976, and a substantial overall surplus emerged.8 In large part this improvement was due to a reduction in the trade deficit as a result of a 130 per cent rise in the export price of coffee.9 The 31 per cent rise in the value of exports also reflected higher tea prices and the impact of improved weather conditions on coffee and tea production. Imports grew in nominal terms by only 11 per cent, while in real terms they continued to decline. There was a further rise in the surplus on services, with higher receipts from tourism and transportation offsetting a rise in investment income outflows; the “current account” deficit was more than halved. An increase in net capital inflow allowed an overall surplus of KSh 776 million, equivalent to about 3 per cent of nominal GDP.

METHODOLOGY FOR BALANCE OF PAYMENTS FORECASTING 10

For a financial program, forecasting the balance of payments may start from the definitional relationship between changes in net foreign assets (NFA) and the other items in the balance of payments:
ΔNFAt=XtMt+INVt+CMt(1)

where

  • ANFA = change in net foreign assets

  • X = export receipts

  • M = import payments

  • INV = net invisibles, i.e., services and unrequited transfers

  • CM = net capital inflow

The exact definitions adopted and the level of aggregation for each variable should depend on the availability of data and the development of stable behavioral relationships. For exports and imports a separate determination of unit price and volume may be desired. Although the currency used to obtain the initial forecasts may vary, an eventual aim should be to produce an estimate of NFA in domestic currency for inclusion in the monetary survey.

Imports

A simple equation for forecasting the volume of imports could take the following form:
(MPM)t=a0+a1Yta2(PMPD)t(2)

where

  • M = value of imports

  • PM = unit value of imports

  • Y = domestic real income

  • PD = domestic price level.

This equation is based on the hypothesis that real import demand depends on domestic income and on the ratio of import to domestic prices as an indicator of possible substitution between domestic and imported goods.

Estimation of the equation in this form involves several important assumptions. In order to predict imports by using only a demand equation, it is necessary to employ the so-called small-country assumption that supply elasticity is infinite or sufficiently large to enable the effect of changing demand on price to be ignored. To the extent that this assumption is not warranted, explicit consideration needs to be given to the supply equation.11 A corollary of the small-country assumption is that import prices need to be determined from an analysis of world markets and developments in particular trading partners. For a country such as Kenya, the small-country approach would seem reasonable. The specification of equation (2) also assumes that actual imports adjust to the desired level within the observation period. Full adjustment may not, however, occur within one period because of the costs involved in short-run changes in the level of imports or because of limitations imposed by contractual arrangements. The following partial adjustment specification may then be more appropriate:
(MPM)t(MPM)t1=δ((MPM)t*(MPM)t1)(3)
(MPM)t*=b0+b1Ytb2(PMPD)t(4)
where
(MPM)*=desiredimportvolume
and
0<δ<1.
Substitution of equation (4) in equation (3) and rearrangement yields:
(MPM)t=δb0+δb1Ytδb2(PMPD)t+(1δ)(MPM)t1(5)

where the coefficients δb1 and δb2, represent the first period effect of real income and relative prices, respectively.12

A further question concerns the appropriate degree of disaggregation for imports. Recent international developments suggest that a disaggregation into oil imports and non-oil imports may be warranted. Such a disaggregation could be further pursued by breaking down non-oil imports according to the major commodity flows. Particular commodity flows could then be related to specific rather than general indicators—for example, import of capital goods as related to domestic investment. Even if imports are not disaggregated, it might be necessary to make allowance for differences in the marginal propensity to import among different demand components. Such an analysis is likely to be especially important when significant changes are expected in the structure of demand. Particular attention may need to be paid to possible changes in imports demanded for stock-building.

An alternative specification of the import equation assumes that there is excess demand for imports at current incomes and prices and that the effective constraint on imports is the availability of credit. It might be argued further that this constraint operates only on the private sector.13 An equation of the following form could be tested:
(MtMGt)=c0+c1ΔDCPt(6)

where

  • MG = value of government imports

  • DCP = domestic credit to the private sector.

Exports

For a small primary product exporting country such as Kenya, the small-country assumption is that exports are essentially determined by supply and that changes in exportable output do not affect world market prices. Forecasting of exports should, in these circumstances, be closely linked with the projections of the appropriate items of domestic production. Exports would still be affected by world market prices to the extent that there are variations in output or in the relationship between domestic consumption and exports. The relevant price determining production incentives is the price received by the producer. In Kenya, nearly all the coffee export price has accrued to the producer. In principle, the relative movement of export and domestic prices could influence export supply. The proportion of the main export crops consumed domestically in Kenya is, however, very limited.14 In the long run, coffee prices would affect domestic production and exports through their impact on coffee plantings. Some short-run price responsiveness may also occur through the intensity of cultivation.15 For many primary producing countries, the central problem is to obtain a forecast of the world market price for their major export commodities.16

In cases where it may be appropriate to forecast exports using a demand equation, the following type of specification may be used: 17

(XPX)t=d0+d1VYWd2(PXPW)t(7)

where

  • X = value of exports

  • PX = unit value of exports

  • VYW = real world income

  • PW = world price level.

Many of the problems of estimation encountered in the use of the above equation are similar to those mentioned earlier in the case of the import equation. Particular aggregation problems exist, however, in developing adequate indicators of world prices and incomes.18

Other Items

Determination of services, unrequited transfers, and capital flows would usually require more eclectic methods, and the scope for econometric analysis is probably limited.19 Some items of the service account, such as transportation and merchandise insurance, may be related in a fairly stable manner to trade flows. Other items, such as travel and investment income, may be forecast by using sample survey and other detailed information provided by the relevant government department. Balance of payments forecasts of unrequited transfers of the government should utilize, and be consistent with, budgetary estimates of foreign grants.

Capital flows may, particularly in a country with a well-developed financial market, reflect interest rate differentials and expected changes in exchange rates.20 In many developing countries, changes in the system of exchange control are also important. Direct investment may be influenced by the provision of tax incentives or other incentives, and forecasts of this component should take into account development plans and policy. Similarly, information on official capital flows may be obtained from development plans and the budget. Sample survey and banking data may be sources for forecasts of private capital movements. Provisions of trade credit may be related to trade flows.

GUIDELINES FOR FORECASTING BALANCE OF PAYMENTS IN KENYA

This section provides some empirical estimates for the equations discussed in the preceding section. In addition, some indicators of developments in the program period and other background information are presented. The purpose is to stimulate discussion of the issues and exercises pertaining to balance of payments forecasting and to provide a starting point for development of the workshop participants' own estimates during the financial programming exercise.21

Trade Account

Several formulations of the import demand equation were estimated. 22 The actual and predicted values for some of the equations are presented in Table 5.

MR=2164.256(3.602)+0.30463.607GDP2076.4712.805(PMPGDP)(8)R¯2=0.501DW=2.152196576
MR=2471.379(8.816)+0.2495(6.264)GDP1886.755(5.534)(PMPGDP)+860.072(5.9)DMI(9)R¯2=0.895DW=2.651196576
LnMR=0.5779(0.478)+0.8093(6.253)LnGDP0.6011(5.134)Ln(PMPGDP)+0.2231(5.116)DMI(9')R¯2=0.913DW=2.395196576
MRL=3202.041(10.617)+0.263(5.903)GDP3178.603(5.625)(PLPGDP)+798.489(5.698)DMI(10)R¯2=0.917DW=2.082196576
MR3=104.337(3.018)+0.021(7.259)GDP+52.461(2.931)DMI(11)R¯2=0.897DW=2.22196576

where

  • MR = total imports at constant 1972 prices on a customs basis (in millions of Kenya shillings)

  • GDP = GDP at factor cost at constant 1972 prices (in millions of Kenya shillings)

  • PM = price index (unit value) for total imports (1972 = 100)

  • PGDP = GDP deflator (1972 = 100)

  • DM1 = dummy for inventory accumulation (1971 and 1974 = 1)

  • MRL = non-oil imports at constant 1972 prices on a customs basis (in millions of Kenya shillings)

  • PL = prices of non-oil imports (1972 = 100)

  • MR3 = mineral fuel imports at constant 1972 prices on a customs basis (in millions of Kenya shillings).

In addition, the disaggregated import equations presented in Table 6 were estimated.23

TABLE 5.

Kenya: Actual and Predicted Values for Selected Equations, 1965–761

(In millions of Kenya shillings)

article image
Source: Appendix.

Definition of variables is as in the text. Series under the variable name indicate the actual value, while PR refers to the predicted value obtained from the equation. The number in parentheses indicates the relevant equation in the text.

TABLE 6.

Kenya: Estimated Import Equations1

article image
Source: Appendix.

From equationLnMR=a+bLnGDP+cLnPM+dDM1whereMR = imports for the respective categories at constant 1972 pricesGDP = gross domestic product at factor cost in 1972 pricesPM = import price deflator (unit value) for the appropriate importcategoryDM1 = dummy variable, taking the values 1 in 1971 and 1974 and 0 inall other years. (This variable is excluded for some importcategories.)

Figures in parentheses show the percentage share in the value of total imports In 1976. ** and * show that the coefficients are significant at the 99 per cent and 95 per cent levels, respectively. The equations were estimated for the period 1965–76.

The credit constraint hypothesis was tested: 24

M=1583.792(2.881)+4.575(6.51)ΔNDC(12)R¯2=0.84DW=2.45196776

where

  • M = imports, c.i.f., on a balance of payments basis (in millions of Kenya shillings)

  • NDC = total domestic credit of the banking system (in millions of Kenya shillings).

In the preceding section, under Exports, it was suggested that forecasts for Kenya's major export commodities should be closely linked to projections of domestic production.25 Data on production and exports for some of the major export crops are presented in Table 7. Coffee exports are higher than production in a number of years, mainly because of changes in stocks. Indicators of possible price and market developments for the major export crops are presented in Table 8.

Other Items

As indicated in the preceding section, under Other Items, there is usually only limited scope in a financial programming exercise for a rigorous behavioral analysis of items outside the trade account. Rather, items in the services and capital accounts would usually be determined in an ad hoc manner, with considerable reliance on the forecasts of the relevant government departments. The use of such methods should, however, be based on a careful analysis of the structure of services and capital flows, any trends in the behavior of the major items, and available information on the factors affecting particular components.26

TABLE 7.

Kenya: Agricultural Production and Exports, 1973–76

(In thousands of metric tons)

article image
Source: Kenya, Statistical Abstract, 1977.

Value of gross marketed production divided by total output at current prices.

TABLE 8.

Kenya: Semiannual Indicators of Developments in Program Period, 1976–781

article image
Sources: Commodities Division of the International Monetary Fund's Research Department and Organization for Economic Cooperation and Development, Economic Outlook, December 1977.

Actual data for 1976 and 1977 and forecasts for 1978.

Seasonally adjusted at annual rates. Export prices represent average values In terms of U.S. dollars.

The major credit items in the service account are freight and insurance, other transportation, and travel—accounting, respectively, for 15 per cent, 41 per cent, and 23 per cent of service receipts. Freight and insurance receipts can be related through a simple regression to exports, f.o.b.:
SCI=59.335(3.497)+0.0756X(15.803)(13)R¯2=0.969DW=1.426196876

where

  • SC1 = freight and insurance credit (in millions of Kenya shillings)

  • X = exports, f.o.b. (in millions of Kenya shillings).

In 1976, two thirds of Kenya's receipts from other transportation were from port disbursements, and the remainder was from passenger fares. More than half of the port services were for fuel and ships' stores, with a further 30 per cent representing port dues. Receipts from port services are to a large extent dependent on trade flows. Kenya's receipts from such services may, however, be generated by carriers involved in imports, exports, or trade with third countries. In view of the importance of fuel and ships' stores, the forecast for other transportation should take into account the projections for import prices of fuel.

Receipts from travel largely result from tourist expenditure. Such receipts can be related to the number of tourists and their average length of stay.27 A relationship of this type, however, only shifts the problem to forecasting the number of tourists and does not deal with the average expenditure of visitors. Nearly two thirds of the visitors to Kenya in 1976 came from Europe and North America. Projections of receipts from travel should therefore consider the cyclical position in those countries and possible changes in air fares.

Investment income represented 46 per cent and other transportation accounted for 21 per cent of total payments for services.28 The former category covers retained profits, other direct investment income, and other investment income. The shares of these components in 1976 were 18 per cent, 44 per cent, and 38 per cent, respectively. In principle, investment income payments should be related to the actual rate of return for each type of investment and the value of asset holdings.29 The information necessary for such a sophisticated approach is not, however, available for Kenya or for most developing countries. Insofar as foreign direct investment takes place largely in the export sector, some relationship may be expected between direct investment flows and exports.30 Allowance should also be made for the definitional relationship between investment income and direct investment arising from the convention of treating retained profits as distributed and reinvested.31 Payments for other transportation would be affected by the same type of factors as those that were discussed for receipts from this source. For payments, however, passenger fares are relatively important, accounting in 1976 for 46 per cent of other transportation debits, with the remainder represented by port services.

Government receipts are the major item in unrequited transfers. The forecast for such receipts should be consistent with the projection for grants as shown in the government budget.

Increases in long-term liabilities represent the major nonmonetary capital flows. In 1976, the Central Government, other public sector, and the private nonbank sector accounted for 50 per cent, 13 per cent, and 37 per cent, respectively, of the increase in long-term liabilities. Projections for the public sector components should be consistent with budgetary projections and policy. A substantial backlog of undisbursed loans exists, and the size of the inflow in a particular period should be linked to the expected timing of major capital projects. In 1976, about 60 per cent of the private sector inflows were direct investment. Projections for such inflows should reflect government policy toward direct investment as well as any changes in the availability of domestic credit. Forecasts for short-term capital flows should allow for changes in trade credit to finance imports and for the availability of domestic credit.

EXERCISES AND ISSUES FOR DISCUSSION

  1. 1. The first section (on the External Sector) presents information on Kenya's balance of payments structure and developments for the period 1973–76. On the basis of this information, what do you see as the major problems involved in forecasting balance of payments developments for 1977?

  2. 2. Compare and contrast the various import equations, utilizing the information on actual and predicted values presented in Table 5. In particular:

    • (a) Compare equations (8) and (9). Why do you think the addition of the dummy variable improves the fit of the equation? Consider alternative ways of allowing for the effects of inventory accumulation.

    • (b) Should imports be disaggregated according to oil imports and non-oil imports?

    • (c) To what extent do the results appear consistent with the credit constraint hypothesis?

    • (d) Do the data used adequately approximate the variables suggested by the theoretical specification of the import models?

    • (e) The specification of the partial adjustment equation (equation 5) imposes the same lag structure on the income and the price variable. Do you consider this appropriate?

    • (f) What alternative specifications of the aggregate import equations might you want to consider?

  3. 3. Assume that for 1977 import prices increase (in terms of Kenya shillings) by 7.7 per cent, the GDP deflator rises (in terms of Kenya shillings) by 19.5 per cent, and real GDP increases by 7.34 per cent. (These were the 1977 actual increases.) On the basis of equation (9):

    • (a) Forecast the percentage change in the volume and the value (in terms of Kenya shillings) of imports in 1977. Comment on whether you find the forecast plausible.

    • (b) How might your forecast be affected by a 5 per cent currency devaluation in 1977?

    • (c) Compute the income and price elasticities of imports for 1977, using 1976 as a base, and compare them with the corresponding elasticities in equation (9′). Comment on the plausibility of these price and income elasticities.

    • (d) Comment on how you might allow for an expected liberalization of import controls in 1977.

    • (e) How might your forecast for 1977 depend upon what you assume about inventory accumulation for that year?

  4. 4. Comment on the disaggregated import equations presented in Table 6. Do you consider the price and income elasticities plausible? How would you improve the specification of the individual commodity equations? Do you consider this degree of disaggregation warranted?

  5. 5. Discuss possible methods of forecasting exports in Kenya. In particular:

    • (a) To what extent would it be reasonable to forecast exports using only a demand equation?

    • (b) How would you attempt to estimate exports, using the small-country assumption that demand is infinitely elastic? In this context what use would you make of the information in Tables 7 and 8? Would the price of exports still have an important role?

  6. 6. Utilize the information in Charts 3 and 4 and in the Appendix, Tables B and C, to analyze the pattern of invisibles and capital flows. What methods might you use for forecasting these items?

APPENDIX

Kenya: Selected Variables Influencing Balance of Payments, 1965–77

article image
article image

Kenya: Data on Selected Variables Influencing Balance of Payments, 1965–771

article image
article image
Sources: Data for imports, exports, and gross domestic product are taken from Kenya, Economic Survey, 1978 and Statistical Abstract, 1977 and also from earlier issues of these publications. Monetary and exchange rate data are from International Monetary Fund, International Financial Statistics, July 1978. World income data are from Organization for Economic Cooperation and Development (OECD), National Accounts, 1976 and Main Economic Indicators, May 1973.

For definitions of variables and units of measure, see above table on Kenya: Selected Variables Influencing Balance of Payments, 1965–77.

TABLE A.

Kenya: Quarterly Changes in Balance of Payments Components, 1976–77

article image
Source: International Monetary Fund, International Financial Statistics, July 1978.

Growth over the same quarter of the previous year. Value and prices based on data in Kenya shillings.

Excludes exports to and imports from Tanzania and Uganda. Prior to 1976, imports also include transfers from Tanzania and Uganda of foreign goods.

Change from previous quarter in millions of Kenya shillings. Data are taken from the monetary survey.

TABLE B.

Kenya: Services and Transfers, 1973–77

(In millions of Kenya shillings)

article image
Source: Same as Table 4.

Consistent with imports on a c.i.f. basis.

Includes capital grants.

TABLE C.

Kenya: Nonmonetary Capital Flows, 1973–77

(In millions of Kenya shillings)

article image
Source: Same as Table 4.

Excludes capital grants.

1

The present workshop utilizes data up to the end of 1976 and presents forecasting exercises for 1977. Data are also provided, however, for the actual outcome for 1977, in order to allow comparison with the results of the forecasting exercise and to aid in the completion of the hypothetical financial program for 1978 presented in Workshop 9.

2

Exports and imports (national accounts basis) averaged 33 per cent of gross domestic product in Kenya during the period 1971–76.

3

There are some exceptions to this principle, such as the purchase of gold from the private sector by domestic monetary institutions (see Workshop 3: Balance of Payments Statistics). The interrelationships between external sectors have led to the development of consistent models for forecasting the balance of payments for a group of countries. See, for example, Michael C Deppler and Duncan M. Ripley, “The World Trade Model: Merchandise Trade,” Staff Papers, Vol. 25 (March 1978), pp, 147–206.

4

In June 1974 the Central Bank of Kenya also restricted foreign-owned firms, except in certain sectors, from borrowing domestically more than 20 per cent of their local capital requirements.

5

For a description of exchange restrictions on other items in Kenya's balance of payments, see International Monetary Fund, Twenty-Eighth Annual Report on Exchange Restrictions (Washington, 1977), pp. 277–80.

6

Kenya accounts for about 2 per cent of world coffee exports and about 6 per cent of world tea exports. The latter share makes Kenya the world's third largest exporter of tea.

7

Composed of Tanzania, Uganda, and Kenya.

8

Quarterly data for 1976 and 1977 on the trade component of the balance of payments and on changes in net foreign assets are provided in Appendix Table A.

9

In July 1975 a severe frost destroyed a substantial part of Brazil's coffee crop. Brazil during 1974–75 accounted for 31 per cent of exportable coffee production. See Horst J. Struckmeyer, “Coffee Prices and Central America.” Finance and Development, Vol. 14 (September 1977), pp. 28–31,

10

A rigorous survey of the technical problems involved in forecasting imports and exports is presented by George McKenzie, “Imports and Exports,” in David F. Heathfield (ed.), Topics in Applied Macroeconomics (London: Macmillan, 1976), pp. 144–63.

11

See Mohsin S. Khan, “Import and Export Demand in Developing Countries;” Staff Papers, Vol. 21 (November 1974), pp. 678–93.

12

The long-run effects are b1 and b2.

13

In principle, allowance should also be made for trade credit from abroad. Accurate data on such credit are, however, often difficult to obtain (see Major Sources of Data in Workshop 3: Balance of Payments Statistics). A further hypothesis that may be applicable in some developing countries is that imports are constrained by the availability of foreign exchange reserves or the purchasing power in terms of imports of export receipts. See William L. Hemphill, “The Effect of Foreign Exchange Receipts on Imports of Less Developed Countries;” Staff Papers, Vol. 21 (November 1974), pp. 637–77.

14

Domestic consumption of coffee is less than 3 per cent of total production.

15

Coffee plantings have a gestation period of about 5 years, mature in 7–10 years, and have an economic life of about 40 years. Newer hybrids mature, however, somewhat sooner. Heavy picking of coffee beans in one year may adversely affect output in the following year. See Shamsher Singh, and others, Coffee, Tea and Cocoa; Market Prospects and Development Lending, World Bank Staff Occasional Paper No. 22 (Baltimore: Johns Hopkins University Press, 1977).

16

Such forecasts might be obtained from agencies specializing in the commodity concerned or from various international organizations. For an approach based on a time series technique, see Ke-Young Chu, “Short-Run Forecasting of Commodity Prices: An Application of Autoregressive Moving Average Models,” Staff Papers, Vol. 25 (March 1978), pp. 90–111.

17

See Khan (cited in footnote 11), pp. 683–84. Particularly for short-run analysis, the indicator of world demand might be weighted according to market shares. In some cases, industrial production might be a more appropriate variable than income. Consideration might also be given to the cyclical position in the domestic economy and the prospective markets. Publications such as the Organization for Economic Cooperation and Development, Economic Outlook, and United Kingdom, National Institute of Economic and Social Research, National Institute Economic Review, provide forecasts of short-run developments in industrial countries.

18

See, for example, Jacques R. Artus and Susana G Sosa, “Relative Price Effects on Export Performance: The Case of Nonelectrical Machinery,” Staff Papers, Vol. 25 (March 1978), pp. 25–47.

19

See, however, Marian R Bond, “A Model of International Investment Income Flows,” Staff Papers, Vol. 24 (July 1977), pp. 344–79, and “The World Trade Model: Invisibles,” Staff Papers, Vol. 26 (June 1979), pp. 257–333.

20

Where a forward market exists, the covered interest rate differential may be important. This differential represents the difference between the foreign and domestic interest race, plus the forward premium or discount on the foreign currency.

21

It should be emphasized that, given the limited number of observations and the complexity of the factors affecting the various balance of payments components, considerable care and judgment are needed in interpreting the regression equations. Where commodity patterns are changing, the measured price indices may be sensitive to the weighting system used. Import and export unit values in Kenya are calculated using a “Fisher Ideal Index.” See Kenya, Statistical Digest (June 1977).

22

Import data based on customs information have a slightly different coverage than data from the balance of payments. To use the customs data for balance of payments forecasting, the projected percentage change may be applied to the last observation of imports calculated from balance of payments sources. The sources of the data and a full listing of the definitions of the variables appear in the Appendix. Considerable splicing of series and some approximations were necessary to obtain the time series for the various import price indices. The data and the derived constant price series should therefore be interpreted cautiously.

23

The adjustment model presented in equation (5) was rejected because of the negative and insignificant coefficient on the lagged endogenous variable.

24

Data are not available for imports net of central government imports. Consequently, NDC rather than DCP was used as the explanatory variable. Empirically, a lower level of explanation was found when using the latter variable.

25

An analysis of factors affecting coffee and tea supply and world market prices is presented in Shamsher Singh, and others (cited in footnote 15).

26

Reference should be made to Charts 3 and 4 and to the Appendix, Tables B and C, for an indication of the trend developments in the services and capital accounts.

27
For example:
SC3=27.723(0.264)+0.1384(5.36)TOR¯2=0.776DW=0.7196876
where
SC3=travel credit (in millions of Kenya shillings)TO=total stay of departing visitors (in thousands of days).
28

Imports are treated on a c.i.f. basis.

29

See Bond (cited in footnote 19), p. 348.

30
For example:
SDI=81.42(1.192)+0.2046(10.615)XR¯2=0.933DW=2.434196876
where
SDI=investment income debit (in millions of Kenya shillings)X=exports, f.o.b. (in millions of Kenya shillings).
31

See Workshop 3: Balance of Payments Statistics.