Policy, Research, andExternal Affairs WORKING PAPERS Country Operations Eastern Africa Department Africa Regional Office The WorldBank September 1991 WPS767 Framework for Macroeconomic Analysis (Applied to Kenya) Colin A. Bruce and David Ndii Models of the RMSM-X genre can - while preserving their logical structure - incorporate behavioral equations and pro- vide useful insights into policy actions thiat would correct internal and external macroeconomic imbalances. The Policy, Research, and Extemal Affairs Complex distobutes PRE Working Papers todissemuate thefindingsof work in progress and to encourage thecxchange of ideas ar-ong Bank staff and all others interested in deselopmnent issues. These papers carry the names of the authors, reflect only their views, and should beusedand cited accordingly. The findings,unterpretauons, and conclusions are the authors' own. They should notbeattributed to theWorld Bank, Its Board of Directors,its managcment, or anyof its member countries. Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized
110
Embed
Framework for Macroeconomic Analysis - World Bank · Behavioral Relationships, Accoundng Rules and Flow-of-Funds Consistency 2.3 Macroeconomic accounting rules and flow of funds consistency
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Policy, Research, and External Affairs
WORKING PAPERS
Country Operations
Eastern Africa DepartmentAfrica Regional Office
The World BankSeptember 1991
WPS 767
Frameworkfor Macroeconomic
Analysis
(Applied to Kenya)
Colin A. Bruceand
David Ndii
Models of the RMSM-X genre can - while preserving theirlogical structure - incorporate behavioral equations and pro-vide useful insights into policy actions thiat would correctinternal and external macroeconomic imbalances.
The Policy, Research, and Extemal Affairs Complex distobutes PRE Working Papers to dissemuate the findings of work in progress andto encourage the cxchange of ideas ar-ong Bank staff and all others interested in deselopmnent issues. These papers carry the names ofthe authors, reflect only their views, and should be used and cited accordingly. The findings, unterpretauons, and conclusions are theauthors' own. They should not be attributed to the World Bank, Its Board of Directors, its managcment, or any of its member countries.
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Pub
lic D
iscl
osur
e A
utho
rized
Policy, Research, and External Affairs
Country Operations
WPS 767
This paper- a product of the Country Opcrations Division, Eastern Africa Department, Africa RegionalOffice - is part of a larger PRE effort, led by the Macroeconomic Adjustment and Growtlh Division,Country Economics Department, to develop a macroeconomic projection model that will improve andextend the Bank's RMSM model. Copies are available free from the World Bank, 1818 H Street NW,Washingtor DC 20433. Please contact Margaret Lynch, room 110-271, extension 34046 (10(4 pages,including figures and tables).
Bruce anl Ndii develop a macroeconomic model projections. For ease of exposition, these gapsfor Kenya that projects detailed national, fiscal, are assumed to be independent.monetary, and private sector accounts, and thebalance of payments. With no changes in its Next, Bruce and Ndii illustrate the model'slogical structure, the model also calculates the usefulness in calculating the policy adjustmentsmagnitude of policy adjustments that would - exchange rate, interest rate, and monetizationeliminate external and internal imbalances. - that would eliminate intemal and extemal
macroeconomic imbalances. The policy adjust-Their approach emphasizes transparcticy and ments are applied independently and jointly.
user-friendliness, which they achieve in two Predictably, the exchange rate depreciates toways. First, historical data - the basis for eliminate the external financing gap. Similarly,projections - are entered in worksheets that are the real domestic loan interest rate rises, crowdssimilar (in format, coverage, and units of ac- out private investment, and releases financing forcount) to standard tables produced by the data the residual PSBR. The model also calculatesgenerating agencies. Second, key behavioral the level to which domestic inflation would haveassumptions and targets are entered in a clearly to rise to finance the residual PSBR throughi andesignated worksheet. This worksheet contains inflation tax.details such as commodity price projections andthe LIBOR rate that would nornally be provided When activated jointly with exchange rateby the Bank's International Trade Division. This depreciation, a smaller increase in the interestworksheet also accommodates coefficients for rate eliminates the residual PSBR. This isthe import and export price and income elastici- because the exchange rate adjustment improvesties by major commodity groups; the interest and the net foreign assets position and increases theincome elasticities of private investment; and the scope for creating domestic credit. Understand-impact of inflation, income growth, and changes ably, therefore, the increase in interest rate isin interest rates on the demand for currency, even smaller whCIl action in the exchange rate isdemand deposits, and time deposits. The targets combined with partial monetization of the fiscalinclude domestic inflation and sectoral growth deficit.rates.
As an important by-product. these simula-Initially, they use the f'ramework to project tions also illustrate the impact of exchange rate
detailed national and sectoral accounts, including and interest rate adjustments on the fiscal deficit.the financeable fiscal deficit. But they arcespecially interested in the external financing Subsequent versions of the model willgap and the residual public sector borrowing endogeniize growth andi extend coverage to therequirement (PSBR) that emerge in these base consolidated public sector.
The PRE Working Paper Serics disseminates thc findings of work under way in the Bank's l'olicy, Research, and ExtemalAffairsCornplcx. An objectiveofthcseries is to get these finidingsout quickly. even if presentations are less than fully polishedThe findings, interpretations, and conclusions in these papers do not necessarily represent offic ial Bank policy.
Introduction .10The Economics of the Exchange Rate Solu.on .10The Econouics of the Interest Rate Solution .13The Economics of Monetizing the Fiscal Defcit .15The Economics of Joint Policy Action. . 16Summary .18
IV. POLICYSIMLATIONS .18
The Base Projections ... . ........................................ 18Exchange Rate Solution ......................................... 22The Interest Rate Solution ..................................... 26Monetization ....... 29loint Exchange Rate and Interest Rate Policy Action.ii*v***@*********** 30Joint Exchange Rate and Interest Rate Policy Action and Monetzati.on .34Summary of Simulation Results. .............. ....... ... 34
V. LIMITATIONS AND FUTURE EXTENSIONS .......... 37
ANNEXES
Annex I: Descripdon of Model . ............................. 40Annex 11: Fiscal Deficits and Macroeconomic Consistency .................... 49Annex III: Import Elasdcities in Kenya: 1968 and Beyond 56Annex IV: Key Assumptions and Targets Input Worksheets and Output Tables 66Annex V: Time Bound Formulae ..................... I 100
Figure One: External Gap and Residual Public Sector BorrowingRequirement .......... ......................... 25
Figure Two: Exchange Ra Simulations, Real Loan nterest RateSimuaion, flation Rate Simulaions ........................ 33
We gtezly wknowledge the ugemeand helpful suggestio of John Holem,Pew Miovic and coleague i AF2CO. Katheen B. Jordan provided valuable editonalausitanc whil plyabha Kongsamut wiligly shared her experence with JAVELIN.
Introduction
1.1 This paper presents a transparent, user-friendly framework for analyzing macroeconomic
policy issues in Kenya. It adheres to flow-of-funds consistency, covers the fiscal, monetry,
national and private sector accounts, and balance of payments, and is an analytical tool for
grappling with the economic relationships and institutional realities which ultimately dictate the
design, implementation and effectiveness of macroeconomic policies. The model is programmed
in JAVELIN.
1.2 A recent report concluded that Kenya's "... recovery ... is somewhat fagile as it has
relied on foreign savings, leaving the economy vulnerable to external shocks, and the remaining
domestic macroeconomic imbalances, unless corrected, threaten to undermine recent economic
and stabilization gains." (World Bank, 1990). The model presented in this paper is usefil to the
policy analyst who may wish to project the broad magnitude of these imbalances under different
fiure growth assumptions. With these results in hand, the policy analyst may choose to revise
growth targets in an ad hoc, piecemeal manner until acceptable" imbalances are obtained and
then focus on mobilizing inflows to fill the resource gaps. Alternatively, the policy analyst may
wish to determine what levels of imbalances are sustainable, financeable and consistent with other
development goals. Here judgment about institutional/political feasibility is indispensable. But
so too is a map of inviolable accounting macro-relationships, their evolution under different
behavioral assumptions, and the technical limitations they impose on policy selection and
effectiveness. The model developed in ftis paper is such a map which could also demonstrate
-2-
the usefulness of instruments such as the interest rate, exchange rate and monetary policy in
managing macroeconomic imbalances and steering the economy doser to its development goals.
1.3 The paper is organized as follows. Chapter Two sketches the analytical firamework and
discusses the econemics of its structure. Chapter Three contains the policy solutions (the
technical details are presented in Annex I). Chapter Three illustrates three policy options-
exchange rate adjusunent, interest rate adjustment and monetization-and highlights their
implications when used individually and jointly. Chapter Five identifies the model's limitations
and scope for expansion.
II. Framework for Macroeconomic Policy Analysis
Ovrview
2.1 The model comprises .ix data input worksheets, debt modules for existing and new
external debt, and output tables (Diag:am One). There is also one central worksheet which
contains key behavioral assumptions and macroeconomic targets. The model covers the national
accounts as well as the accounts of the central bank, the monetary authority, the commercial
banks, the central government, the external sector (trade, balance of payments and external debt)
and the non-financial private sector. 11 The projection period is 1990-96.
2.2 The data inpit worksheets contain annual historical data for 1987-89. In general the
format, units of account and coverage of these worksheets are similar to standard tables produced
by the data-generating agencies of the GovernAent of Kenya. The exceptions are the worksheets
.L1/. TM accounts of monewy auhority consolidate the accounts of the central bank with the accounts arising frommoncay funiions undertaken by other insitutions or tho cntral govemment.
3 -
for central government finances and exteumal debt which eome from the IM and the World Bank
respectively. Ihese historical data must be consistent within and across sectors, and conform to
standard macroeconomic accounting rules. Both of these requirements obtain in the flow of finds
famework. Accordingly, the model reconciles historical data for flow of fumds consistency and
then uses them as the basis for projections.
Behavioral Relationships, Accoundng Rules and Flow-of-Funds Consistency
2.3 Macroeconomic accounting rules and flow of funds consistency set limits within which
economic behavior takes place in the model. The accounting rules and flow of funds consistency
requirements are the same as those set out in Holsen (1989), Khadr et al. (1989) and Easterly et
al. (1990). The behavioral relationships are the projection rules for variables. Some originate
in econometric analysis that has yielded coefficients linking the variable of interest to its
determinants. Others are constant ratios of other vaiables with which they had, or are exp -- ted
to have, a stable long run relationship. Yet others are projected on the basis of their relationship
to variables whiclt are exogenous to the model.
2.4 To simplify the exposition of the model, this section only specifies and discusses the main
behavioral relationships in the monetary sector, the central government accounts, the external
sector and the national accounts. A more comprehensive notational presentation of the behavioral
relationships, with their flow-of-fimds accounting identities, is provided in Annex I.
The Monetar Sector: the demand for financial assets
2.5 The model identifies demand functions for three financial assets-currency, demand
deposits and time deposits (quasi-money). The demand for currency is a positive fniction of real
income growth and a negative function of the real deposit interest rate and the rate of inflation.
The demand for demand deposits and time deposits is a positive fnction of real income growth,
a negative function of the rate of inflation, and a positive function of the real interest rate on
deposits. In essence, the demand for currency, demand deposits and time deposits increases with
national income but is eroded by inflation. Demand and time deposits, unlike currency, earn
interest and therefore move in the same direction as interest rates. The coefficients entering the
framework are based on econometric estimates discussed in Annex H. The model ensures
consistency between financial stocks and flows by expressing flows as changes in the respective
stocks over the specified period. Where appropriate, stocks are derived by aggregating flows.
Government revenue
2.6 In the model, government revenue falls into two main categories, recurrent revenue and
capital revenue. Recurrent revenue is subdivided into direct taxes, indirect taxes and other
recurrent items. Direct taxes throughout the projection period are assumed to maintain the ratio
of direct taxes to GDP (at factor cost) estimated for the base period, 1987-89. Other recurrent
items obey a similar projection rule. Indirect taxes are disaggregated into sales taxes on domestic
manufactures and imports, duties on imports, and other taxes and licenses. 21 In turn, each of
these revenue categories is projected with respect to the activities which generate them. Sales
taxes on imports (in local currency) are assumed to maintain the ratio of sales taxes to total
imports (in local currency) esimated for the base period, 1987-89. The same relationship is
assumed for sales taxes on exports with respect to exports and for duties on imports with respect
to imports. Other indirect taxes are projected in a similar way but are linked instead to the pool
of tax revenues (net of indirect taxes).
Z/ Duties on exports have been discontinued by the Government of Kenya.
-6-:
Governm t penditure
2.7 Govenment consumpdon expenditr has two main cowponents. The first is wages and
salaries which are determined by an exogenous nominal growth target. The second component
is expenditure on goods and services. Its relationship to government consumpdon expenditue
is assumed to be the same in the future as in 19874B9. Subsidies, trnsfers and other current
expenditures are projected in a similar mner but with respect to GDP. Interest on foreign debt
is calculated by the debt module on the basis of the debt stock. Interest on domestic debt is
calculated in a similar manner.
2.8 External inflows are an important source of funding for and determinant of government
capital expenditure. Often these inflows finance imported investment goods. Accordingly, the
model projects government capital expenditure as a fixod ratio of government imports, based on
the ratio in 1987-89. In turn, govermment imports are linked to long-term capital inflows (as
at 198749) which are computed in the debt module.
The balance of payments
2.9 Exports are disaggregated into commodity groups-coffee, tea, horticulture, petroleum,
maize, manufacatring, re-exports, other goods and non-factor services. The import groups
idendfied are food, petroleum, other consumption goods, primary intermediate goods,
manufacturing intermediate goods, capital goods, leases, government imports, and non-factor
services. With the exception of leases of aircrafts and government imports, these traded goods
are projected solely on the basis of their estimated elasticities with respect to GDP and to their
prices in local currency (Annex o).
2.10 Thrughout the projecon peiod net crret fer, nct direct rbr?ign investment,
offici gram and short-term captal are deemined by their respective ratios to GDP estiated
for 1987-89. Long-term public and publicly guamteed inflows, Don-guaranteed private medium
and long-tem inflows together with their amorzation and interest obligadons are calculated in
the debt module. Short-term public and private debt are assumed to be a fixed ratio of total long-
term inflvuw. Interest on short-term extenal debt and on reserves is determined by applying
the exogenous foreign interest rate to their stocks. Reserve accumlation is derived from the
overall balance of payments or stated as an explicit target.
National accounts
2.11 Several items for the national accounts are determined in the sectoral accounts described
in paras. 2.6 to 2.10. They include indirect taxes, subsidies, the net current transfers in the
balance of payments, Government consumption and investment. Domestic savings is simply the
difference between total investment and the external resource gap. LikbwiLe, national savings
is the difference between total investment and the current account balance. Private savings is the
residual after public savings-defined as the current balances in the fiscal accounts-is subtracted
from national savings.
2.12 GDP at ftctor cost is broken down into value added in agriculture, mining,
imanufacturing, other industry and services. Growth in each of these sectors is projected
exogenously. GDP at market prices is then determined in conjunction with net direct taxes which
are calculated in the fiscal account.
2.13 Public investment and consumption are projected in the public sector accounts. Private
Iinestment is a negative function of the change in the real domestic loan interest rate and a
-8-
positive function of GDP growth. Private consumption is the residual item for balancing GDP
at market prices with gross domestic expenditure
Miscellaneous
2.14 Tle projected population growth rate, dollar prices of Imports and exports including the
manufactuing uiit index, and LIBCOR are external to the model and are obtained from sources
within the World Bank. The commodity prices from these sources serve as inputs into the
exl,ort, import, investment, consumption and GDP deflators. All deflators are weighted by the
shares of commodities whose prices enter their calculation.
Flow of Funds Consistency
2.15 Economic flows generated in one sector may be used within that sector or other sectors
of the economy. This is the fundamental organizing principle of the flow of funds consistency
framework as presented in Holsen (19S9), Khadr et al. (1989) and Easterly et al. (1990). It
classifies each variable as a source of funds or a use of funds. The sources and uses may be
current or capital. Consistency prevails when current sources of funds equal current u;es of
funds within eack. sector and in the overall economy. The same condition holds for capita
sources and uses of finds.
2.16 The flow of funds identities for the nonetary sector, central government finances, the
balance of payments and the national accounts are presented in Annex I. They include the stocks
and flows of the central bank, monetary authority and commercial banks, as well as the non-
financial private sector. The projectigns derived from the behavioral relationships satisfy these
identities at the sectoral and aggregate level. Historical data, however, required several
adjustments to achieve consistency. An adjustment was made for consistency among the capital
-9-
fiows and stocks implied by the debt data, the balance of payments, the fiscal accoun and the
monetary accounts. Similar calculations were done after balance of payments data in the national
accounts were compared with the figures in the balance of paymen. The rule of thvmb was to
identify those adjustments which made the data in each input worksheet internally consistent,
reconciled stocks with flows and achieved intersectoral consistency in relation to the national
accounts.
2.17 Historical net foreign assets of the central bank and monetary authority were also adjusted
for consistency with their capital sourcm. This adjuement became the balancing item in the
capital sources of the balance of payments. An adjustment was also needed to make the capital
uses of government revenue equal to the capital sources. This adjustment balanced the private
sector's sources and uses of funds. Finally, the total historical sources of the capital funds of
the central government were changed by an amount which became the balancing item for the
historical uses of capital funds by the private sector. After reconciliation for flow-of-funds
consistency, the historical data are used by the model as the basis for projections and policy
simulations.
- 10-
HI. POLICY SOLUTIONS
Introduction
3.1 This chapter discusses the economics of using exchange rate policy, interest rate policy
and monetization of the fiscal deficit, independently and jointly, to redress macroeconomic
imbalances. These policy instruments have been selected because of their increasing role in
managing the Kenyan economy. Prior to 1982, Kenya had a fixed exchange rate regime.
Increasingly since 1985, the exchange rate adjustments have been used to ensure Kenya's external
competitiveness, support the import liberalization program, and strengthen the overall
macroeconomic policy environment. The regulation of interest rates in Kenya dates back to
1974 but movement toward market-determined interest rates is expected to be completed by June
1991. Other instruments for managing monetary aggregates are also being strengthened.
The Economics of the Exchange Rate Solution
3.2 The exchange rate policy simulation computes the exchange rate at which the specified
external financing gap would be closed at a given GDP level. ai The solution depends largely
on the negative relationship between import demand and import prices, and the positive
relationship between export supply and export prices. A devaluation of the exchange rate would,
other things being equal, raise the domestic currency prices of traded goods, increase the supply
of exports, decrease the demand for imports and narrow the resource gap. Because the exports
and imports are each subdivided into major commodity groups-each with its own price and
income elasticity-the model can account in a realistic way for differences in response lags. For
/ 'MTho ternal fancing gap is the additional capital inflows-over and above those already accounted *br-whichare required to finance the current account deficit in the balance of payments and to meet the reserveaccumulaton target.
- I1 -
example, in the absence of large stocks, the supply response of coffee exports to coffee price
increases is likely to be slower than manufactured items. Imported food, with its many local
substitutes, is likely to be more sensitive to price changes than imported crude oil which has no
close local substitutes. Similarly, to the extent that government imports are financed by donors,
their demand is expected to be fairly inelastic with respect to prices.
3.3 All other variables in the model which are linked to exports, imports and the exchange
rate are affected by this solution. When the exchange rate is devalued, governnent revenue from
imports falls if there is a reduction in the local currency value of imports. An exchange rate
devaluation would also increase the domestic currency expenditure on foreign debt service and
externally financed activities. In addition, it would increase the domestic currency receipts from
external transfers, loans and grants. The impact of the exchange rate depreciation on the fiscal
deficit therefore would be ambiguous.
3.4 To maintain macroeconomic consistency, domestic investment and/or savings must adjust
when an exchange rate devaluation results in the importation of fewer goods and the export of
more goods. Consider the simple macroeconomic identity:
+ ?-
GDP(fc) + Ti (net) = Cp + Cg + Ip + Ig + X -M (1)
where GDP(fc) is gross domestic product at factor cost, Ti (net) is net indirect taxes, Cp and Cg
are private and public consumption respectively, Ip and Ig are private and public investment
respectively, X is exports and M is imports-all in domestic currency units. Equation (1) equates
the total current sources of fimds with the total current uses of funds in the national accounts. The
12 -
signs above the variables indicaze their possible response ('+' is positive,'-' is negative) to an
exchange rate devaluation. Equation (1) shows that variables on left and right hand sides of
equation (1) may have to adjust to maintain equality.
3.5 GDP at factor cost (measured in domestic prices) is exogenous and therefore does not
change with exchange rate movements in the model. In contrast, indirctames could rise or fall.
Public consumption comprises government consumption and wage targets, and therefore is not
affected in the model by exchange rate changes (paras. 2.7, 2.12). Private investment is
deternined by the real domestic loan interest rate and GDP at factor cost. Neither is affected by
exchange rate movements in the model. By definition, the increase in the resource balance (X-M)
would be greater than any increase in net indirect taxes Ti) (for a given tax rate). Accordingly,
Cp and/or Ig must fall to maintain equality in equaion (1). Ig could only fall on economic
grounds if government imports, including inputs into investment, fall in response to the exchange
rate devaluation. Ultimately, if government investment does not contract sufficiently, the
residual variable-private consumption-must fall. In essence, private domestic savings must rise
to finance the unchanged level of private investment and a largely unchanged or slightly reduced
level of government investment, becuse of the reduction in foreign savings (X-M).
3.6 The exchange rate solution affects monetary aggregates via its impact on the net foreign
exchange assets and the demand for financial assets. Equation (2):
represents equilibrium in the monetary sector, where MI is currency (negative function of the
real interest rate) and demand deposits (positive function of the real interest rate), QM is time
deposits (quasi-money), rd is the real domestic loan interest rate, X is the rate of domestic
inflation, NOL is the flow of net other liabilities of the monetary system, NFA is the net foreign
assets, Lcg is the flow of net credit from the central bank to the government, Ldg is the flow of
net lending from the commercial banks to the government and Lmp is the flow of net credit to
the private sector. Equation (2) equates the total capital sources of funds with the total'capital
uses of funds in the monetary survey; that is, equality of the liabilities and assets of the banking
system as a whole. While it is obvious that NFA could be affected by an exchange rate
devaluation, it is also conceivable that only variables on the left hand side of the equation would
change. MI and QM are negatively affected by inflation which could be induced by the exchange
rate policy but they are driven in different directions by real interest rates. The important point
is that equation (2) has economically meaningful behavioral flexibility to achieve overaU balance
after an exchange rate adjustment.
The Economics of the Interest Rate Solution
3.7 The interest rate simulation computes the interest rate at which the residual public sector
borrowing requirement would be met by borrowing from domestic commercial banks, given the
level of GDP and the private sector's demand for credit from these banks. _4/ Because of data
limitations, the 'public sector" is limited to the central government in the model.
3.8 The recent Country Economic Memorandum ascribes three main objectives to interest
rate policy in Kenya. The first is to maintain positive real interest rates in order to encourage
41 The residual public sector borrowing requirement is defined in the model u the additional domestic financing--net of the anticipated current and capital revenues, and domestic and foreign financing-which would covertotal government expenditure.
- 14 -
savings and thereby contribute to the mobilization of financial resources. The second objectve
is to move gradually to market determined interest rates which in turn would encourage a more
efficient allocation of credit across and within sectors. The third goal is to encourage competition
between commercial banks and near-bank financial institutions by reducing the artificial
differentials between their lending rates (World Bank, 1990). Towards this end, restrictions on
commercial bank fees and charges were lifted as of April 1, 1990, providing these institutions
with greater flexibility in determining the cost of borrowing and lending. The Government plans
to adopt a fully liberalized interest rate structure by June 1991.
3.9 In recognition of this policy direction, the model uses the real interest rate to reallocate
domestic credit between the private sector and the public sector, given the pool of credit in the
economy. This means that net credit to the private sector must fall when net credit to the public
sector increases, other things being equal (equation (2)). Accordingly, the private sector's
demand for credit is negatively related to real interest rates, reflecting the negative relationship
between the real interest rate and private investment.
3.10 Research in several countries suggests that the transition from negative to positive
interest rates, rather than the levels themselves, has a more significant impact on the level of
domestic savings. Once the rate becomes positive, further increases appear not to affect savings.
The reasons may be two-fold. First, target savers-persons wishing to reach a certain level of
savings-would not need to increase their savings because of rising interest income from a given
level of savings. This income effect may actually lower savings at higher interest rates. Second,
and in contrast to the first, higher real interest rates raise the opportuniy cost of consumption-
savings and its interest income. Savers may therefore switch from consumption to savings when
real interest rates rise. Because of this indeterminate impact of interest rates and savings in
- 15 -
theory, and indeed in practice, no direct link is forged In the model between domestic savings
and interest rates. Increases in the government's demand for credit would raise the real domestic
loan interest rate and cause private investment (and d.rivatively, private sector credit demand)
to contract. However, aggregate consumpdon need not fall or savings rise if the additional credit
to the govermment is spent on consumption.
3.11 Increases in the real domestic interest rates in the model affect tke demand for currency
negatively, but affect the demand for demand deposits and quasi-money positively (para. 2.5).
In equation (2) therefore, both right and left hand side variables may change. Accordingly, the
policy solution for the residual public sector borrowing requirement of the government which is
fimanced by domestic borrowing would be higher real domestic interest rates, subject to the
indirect changes induced in the assets (currency, deposits, etc) of the monetary system. The
interest rate solution provided by the model also covers any incremental financing which may be
required if higher real, and derivatively nominal, interest rates increase the residual public sector
borrowing requirement. Such increases in the residual public sector requirement are possible
because the debt service obligations of the government would rise as interest rates rise.
The Economics of Monetizing the Fiscal Defcit
3.12 Instead of borrowing domestically to finance the public sector borrowing requirement,
the Government could print money. However, 'excessive' monetary expansion creates inflation.
Loosely speaking, monetary expansion is excessive if it exceeds what individuals would
ordinarily hold at given levels of real income. Annex m shows how the willingness of
individuals to hold mon'v helps to determine the *financeable" fiscal deficit. One of its
conclusions is that the Government mighg be able to monetize a larger fiscal deficit if it is also
willing to tolerate a higher level of domestic inflation.
- 16 *
3.13 Ihe model computes the rate of inflation, measured as the change in the GDP deflator,
which would be required to maintain macroeconomic consistency when given fiscal defichs (or
portions thereof) are monetized at the given level of GDP. Higher inflation raises nominal values
of the domestic interest rates, the exchange rate and GDP growth (at factor cost) but leaves their
real values unchanged. In consequence, real private investment remains unchanged. In contrast,
the demand for currency, demand deposits and quasi-deposits all fail as inflation rises. To
maintain macroeconomic consistency at the fixed level of GDP, the reduction in the demand for
the financial assets-currency, demand deposits and quasi-deposits-must be offset by an increase
in the demand for government debt; that is, the residual public sector borrowing requirement.
In practice, the portfolio reallocation to government debt would be induced by monetary
expansion which raises inflation, reduces real money balances, and yields the government an
inflation tax which eliminates the public sector borrowing requirement.
The Economics of Joint Policy Action
3.14 This solution calculates the exchange rate, interest rate and level of inflation which, if
applied together, would eliminate the residual public sector borrowing requirement and close the
external financing gap. Put differendy, it shows how the same outcome may be obtained by
various combinations of policies. The underlying economies is the same as in the application of
individual policy instruments. But the magnitude of the required adjustment may differ because
of the indirect effects of each policy instrument on the policy target to which it is not directly
assigned. For example, exchange rate depreciation reduces the external financing gap but may
simultaneously worsen govenament revenue (from indirect taxes). It may also worsen the residual
public sector borrowing requirement if the reduction in recurrent revenue is less than the
reduction in govermnent investment expenditure. At the same time however, higher real interest
- 17-
PolicyTarget External Residual Primary
Financing Public Sectr AdjustingPolicy Gap Borrowing VariablesSolution Requirement
Exchange Rate Exports &
X Imports
Interest Rate X PrivateInvestment
Deficit MoneyMonedzation Private Sector
x ~~~~CreditX Net ForeignAssets
Joint PolicyAction: ExportsExchange Rate Imports& Interest X X PrivateRate Investment
Joint PolicyAction:Exchange Rate, All of theInterest Rate X X AboveandMonetization
Poligy Matrix
rates squeeze private investment and thereby reduce the private sector demand for credit while
the improvement in the extemal account increases the scope for domestic credit creation.
- 18 -
Summary
3.1S Ita policy solutions discussed in paras. 3.2 to 3.14 are distlled in the policy matrix
above for easy reference. In brief, the exchange rate solution depends on export expansion and
Import contraction to meet the economy's surplus demand for foreign exchange. The interest rate
solution crowds out private Investment and releases domestic credit to finance the residual public
sector borrowing requirement. Finally, the fiscal deficit could be financed through monetary
expansion which raises inflation and the resulting inflation tax. Our next task is to illustrate these
solutions and thereby demonstrate how the analytical framework presented in this paper could be
helpful in designing macroeconomic policy packages for Kenya.
IV. POLICY SIMULATIONS
4.1 This chapter is illustrative, not predictive or prescriptive. The first and only order of
business is to demonstrate that the tool-the analytical framework sketched in Chapter E-can
calculate policy adjustments which would be required to achieve particular policy targets, given
selected behavioral assumptions and inviolable macroeconomic accounting rules. S/ However,
the leap from illustration to prescription would merely involve agreement on base year data and
behavioral coefficients, most of which are conveniently assembled in one worksheet (Annex IV).
Base Projections
4.2 Table 1 shows key indicators of the Kenyan economy as they would evolve under the
projection rules and assumptions discussed in Chapter 2 and base year data in the input
worksheets in Annex IV. In these base projections, domestic interest rates are based on
I/ . To better demonstmte how the interest rate solution of the modd works, an adjustment factor was used toInflate the residual public sector borrowing requirement (albeit without violating flow of funds consistency).
v 19 -
exogenous fbrecasts for LIBOR, while the exchange rate is adjusted to reflect the differential
between the future inflation rates which are assumed for Kenya and the rest of the world.
Accordingly, the exchange rate (K£1US$) moves from 1.202 in 1991 to 1.581 in 1993 and 1.856
in 1996. The real domestic loan interest rate is 11.2 percent in 1991, 10.5 percent in 1993 and
10.3 percent in 1996. The corresponding domestic inflation target is 9.5 percent, 6.0 percent and
S.0 percent. The complete set of input and output tables of the model fori e base projections,
including the flow of funds consistency check, is at Annex IV.
4.3 In this base scenario, the current account deficit falls from 8.4 percent of GDP
(excluding grants) in 1991 to 8.3 percent of GDP (excluding grants) in 1996. Exports (goods
and nonfactor services) and imports (goods and nonfactor services) are 21.1 percent of GDP and
28.9 percent of GDP respectively in 1991 compared with 23.7 percent of GDP and 31.4 percent
of GDP in 1996. 6/ In dollar terms, the external financing gap moves from US$19.8 million
in 1991 to US$159.5 million by the end of the period (Table I). Meanwhile, government revenue
and expenditure reach 26.2 percent of GDP (excluding grants) and 31.4 percent of GDP
respectively in 1996, from 19.8 percent (excluding grants) and 32.2 percent of GDP in 1991.
The residual public sector borrowing requirement is 27.8 million Kenya pounds in 1991, 97.8
million Kenya pounds in 1993 and 450.7 million Kenya pounds in 1996.
f lThroughout, exports and imports refer to goods and nonflactor services.
interest rates, adopted an appropriate exchange rate policy, and generally managed exenal
finance have begun to remove structural constaints and to foster an eninm ent which Is more
conducive to efficient resource use.
4.13 Ihe simulations discussed in paras. 4.24.12 suggest that exchange rate, interest rate and
fiscal policy have an important role to play in managing future macroeconomic imbalances in
Kenya. A 5 percent devaluation of the real exchange rate appears to reduce the current deficit
by about one percent of GDP. However, it also appears to have a negative impact on the fiscal
deficit Likewise, in a constrained credit envivonment, every one percentage point (of GDP)
increase in domestic borrowing to finance the iscal deficit appears to increase the real domestic
loan rate by around 0.5 percentage points. However, the required increase in the real domestic
interest rate falls as domestic inflation rises.
4.14 In summary, the simulations suggest the need for joint exchange rate and interest rate
flexibility in managing the macroeconomic imbalances in Kenya. More important perhaps,
aggressive exchange rate action is necessary if the economy is o reduce its dependence on
foreign savings. On the fiscal front, greater discipline is warranted since it would reduce the
public sector borrowing requirement, lower the equilibrium real interest rate and accommodate
a higher level of investment in the private secmr.
V. Limitations and Future Extensions
5.1 The model is only as useful as the accuracy of the response functions. The coefficients
used in the simulations reported in Chapter IV are based on our econometric analysis using
Kenyan data (import elasticities, the demand for financial assets), the econometric studies of other
-38 -
researchers on Kenya and comparable economies, and judgment based on casual empiricism
(export elasticities, private investment elasticities). But the most sophisticated quantiative
techniques are still guty of projecdng the future on the basis of past behavior. nhis approach
is particularly uncomfortable because current adjustment policies in Kenya are changing,.however
slowly, the macroeconomic Incentives and the scope for responding to them. Estimates of import
eldasticities deal with these changing circumstances explicitly. As observation points become
available, this would need to be done especially for exports and private investment.
S.2 The interest rate and exchange rate adjustments required for solving the model are
cumulative. That is, the adjustment required in any given year is based on the interest rate and
exchange rate level in the previous year and not to some fixed point in the past. In contrast, the
gaps are dealt with annually, not cumulatively. As a result, the adjustments in the policy
variables could be large fxom year to year if the gaps exhibit large annual fluctuations. For
example, to finance a large public sector borrowing requirement in a given year, the model may
produce a large increase in the real domestic loan interest rate. Because interest rate adjustments
are cumulative, the real interest rate may actually fall or even become negadve in the next year
If the public sector borrowing requirement is small. The economic logic of this solution is that
the previous year's high real interest rate must fall if private sector investment is to absorb the
additional domestic credit which is no longer required to finance the public sector. Similar
results obtain for the exchange rate when there are large fluctuations in the external gap. Such
drastic adjustments in policy variables from year to year are not likely to occur in practice. Still,
this is not a problem for the model to deal with as much as a signal about how to frame the
policy questions and interpret the policy solutions. When the gaps fluctuate widely from year to
year, it may be appropriate to ask the model to solve for the average annual gap over the period.
Another approach would be to treat the solution for the largest annual gap as an indication of the
upper limit of the policy adjustnents needed during the projection period.
-39 -
5.3 Ihe model treats the residual public sector borrowing requirement and the external
fundlg gap as mutually exclusive and independen. This simplifies the exposition of the model.
However, the residual public sector borrowing requirement may be a component of the external
financing requirement. Under this scenario, closure of the exteral gap would amount to
elimination of the public sector borrowing requirement. Reality is somewhere between the two
extremes, one in which the extenal gap includes the residual public sector borrowing requirement
and the other in which they are completely independent. To capture this reality, a simple
specification would be required to indicate the relationship between the public sector borrowing
requirement and the external financing gap. The model could then solve for the interest rate
and/or the exchange rate which closes the net gap.
5.4 Recent economic work on Kenya recognizes the need for data on the consolidated public
sector but this information is only just being gathered and coUated by the Central Government
In time, this data should be used to extend the coverage of the public sector accounts in the
model. Parastatal imports also need to be disaggregated from private imports and a more reliable
series on government imports obtained. As progress is made in public sector investment
programming, the recurrent expenditure implications of capital outlays should also be captured
explicitly.
5.5 With relatively few modifications, the model could endogenize economic growth (at
factor cost). At this stage however, it may be more rewarding to concentrate on refining the
response functions and to widening the coverage of the public sector finances. Econometric
work should also be done on the income and price elasticities of exports, the interest elasticity
of private investment and the behavior of private consumpdon.
-40-
ANNEX I
DESCRIPTION OF MODEL
Overview
1. The model comprises six data input worksheets, a debt module for pipeline and new externaldebt, and output tables. There is also one central worksheet wh contains key behavioralassumptions and macroeconomic targets.
7782. The data input worksheets are similar in fornat, units of account and coverage to standardtables produced by the Government. These input worksheets are:
(a) Original Data: CB (GOK)(b) Original Data: BOP (GOK)(c) Original Data: Fiscal Data (FY) (GOK & IM)(d) Original Data: DRS (World Bank)(e) Original Data: National Accounts (GOK)(t) Original Data: Trade (GOK)
3. The debt module comprises two files which are separate from the central model, the first fileis analogous to the Input Debt file for the LOTUS-based RMSM. Historical data are entered inmillions of US dollars. Medium and long-term debt is subdivided into multilateral and bilateraldonors and financial markets, including commercial banks and suppliers credits. The standarddefinitions of the World Bank's debtor reporting system are used. The second file is analogous tothe Debt file of the LOTUS-based RMSM. Projected commitment of loans together with their terms-interest rate, maturity, grace period and disbursement profie-and grants are entered individually orby donor ty,e. The module calculates annual disbursements and debt service of ihe projectedcommitments and aggregates them by donor types in a manner which is consistent with the pipelinedata. It also aggregates the projected debt variables with the historical pipeline data. A buildingblock is then used to import this combined pipeline and projected debt data into the central model.
4. There are eleven output tables covering:
(a) National Accounts (percentage of GDP)(b) National Accounts in Constant Prices(c) Private Sector Account(d) Balance of Payments(e) Fiscal Accounts(f) Financeable Fiscal Deficit(g) Monetary Survey(h) Monetary Authority Accounts(i) Commercial Bank Accounts(j) External Debt(k) Flow of Funds Consistency Check
Most variables in the output tables, apart from growth rates, are expressed as percentages of GDPat market prices but could be easily expressed in levels. These levels would generally be denominated
- 41 -
in million of Kenya pounds, with the exception of balance of payments data which are in million ofUS dollars. The model produces a flow of funds consistency output table. The sectoral aggregatesin these tables allow easy inspection of the consistency between the capitaland current sources and uses of funds in the central government, the balance of payments, themonetary sector-disaggregated into the monetary authority and commerical banks-the nationalaccounts, and the prvate sector. The key flow-of-funds identities and the projection rules for theimportant variables are summarized in Table 1.
Inputting Historical Data
5. For convenience, the input worksheets closely resemble the format and presentation of databy the key data-generating government agencies. Information on banking comes from the CentralBank's Quarterly Economic Review and is expressed in millions of shillings as in the original source.Assets and liabilities are disaggregated and entered separately for the central bank, the monetaryauthority, commercial banks and the monetry survey. The model checks for intemal consistencyacross these historical accounts and adjusts other liabilities" when there are imbalances in individualaccounts. A more difficult problem arises when the inconsistencies are found between the monetarysurvey and its components, the accounts of the monetary authority and the commercial banks. Therule of thumb used is to treat the monetary survey as accurate and adjust "other liabilities" of themonetary authority if the source of the discrepancy is not obvious.
6. The source of the balance of payments data, in millions of Kenya shillings, is the Balance ofPayments (BOP) Unit within the Ministry of Finance. The format and sequence of data items areidentical to the standard table generated by this government agency. For ease of comparison withother data sources, the worksheet routinely produces a table in millions of US dollars which isidentical to the original input table in all other respects.
7. Data on stocks of medium and long-term and short-term external debt come from the DRSsystem of the World Bank and accords with that system's nomenclature and definitions. DRS is alsothe source of historical data on debt disbursements, IMF purchases, and interest and amortizationpayments. Information on grants comes from the BOP Unit of the Ministry of Finance.
8. The fiscal data input worksheet is a composite of standard tables produced by the IMF andis designated in millions of kenya pounds in fiscal years (July to June). Only data on centralgovernment finances are currently available. Sources of revenue are identified in a way which allowseasy transcription of data from the standard data sources and allows revenue projections to respondto policy simulations. For similar reasons, recurrent expenditure is subdivided into wages andsalaries, other goods and services, interest payments on foreign and domestic debt, etc. Fiscal datais converted from fiscal to calendar year, on the recommendation of the Budget Department of theMinistry of Finance, by using a factor of .351.65. For example, government expenditure in 1990would comprise 35 percent of the expenditure in FY1989/90 and 65 percent of the expenditure in FY1990/91.
9. The national accounts data cover gross national product by expenditure and are entered inmillions of current Kenya pounds. The source of this information is the Central Bureau of Staistcsof the Government of Kenya. A memo item, showing the breakdown of GDP at factor cost bysector-agriculture, mining, manufacturing, other industry, services-is also compiled from this sourceand serves as the base year data for projecting GDP growth at factor cost. This breakdown replicatesthe categories from the national accounts section of the LOTUS-based RMSM, to which manyanalysts have become accustomed.
TABLE I OENAVIU. -eJATIUSUIP FUol Of mm-DS IDENITY
Denvd for Selected Ffafact Iawssets
3
Nd,. I C~16.s1j).C1 t ea).C(g*fc).(t+ aa.(r.)).Cl * @-.(i))1 TOTNOCapS a goL * Ni * G
k-1 TOTNonetpa a Lcg t Ldg t Lp .NFA
where: where:
as - stock of currency TOTHonCepS - swietary survey: total capital
a - stock of demand deposits sources of fundsa, - stock of quasi money (time deposits) TOTNmeapU monetary survey: total capital
uses of fiAsigdpfc - real GDP (at factor cost) growth Nol - flow of other liabilities (net)
"I flow of currency and demandrd - growth rate of real daomstic deposit rate depositsI - domestic inflation ON - flow of qas-mey (tim deposits)k - 1987* 198B, 1989 Lcg - flo of net credit from the central
bank to the govermmnta - coefficients Lds flow of net lendino frm cinwrcula
banks to the govrmntLap - flow of net credit to the private
setorUFA - flow of net foreign assts
Central nnnt fInuEAs Cntral Gow_ement Flnce
levi
3
I Id lOTGoiCapS - Sg Lce g dg + Notg Lfgk-I + Lsf*
Td _ _. GDPfc, TOGovCapU - 19 Lgp
3 TOTGovCurS l Td + Othrm-Sihg
L TOTGoVCurt - Cg + Tp * lNgp * Ilgf * Sg.
GDPfc , llsgfk-i
ujhere: ibere:
Td - direct taxes TOTGOvCa46 - Central gwneruent: totalsources of capital funds
TOTGovCapU - Central goveranent: totaluses of capital ftwds
TA81E I CodT D wuvEUAVIAL EELATIINSUP FLOW OF MNS IDENTITY
Central roernot finences cont'd Central CoveIvaet Finnces cont'd
Sg - current balence3 Xotg flow of other lisbilitiesI Ti , (net) of monetary sector to
central goverimentLpg flow of net lending by private
Tic Sector, sector to central governmontLfg - floa of medi and long term external
borrowing (not) (NLT)3 Lsff taflow of external short-termY. Setor ^ borrowing (net)k-i l gross Investnt
Lgp - flow of net lending by centralgovenMent to private sector.
sAere:
n (t - 5) - correspond to the five min categories of OthrR 1 other governnent revemuesindirect taxes: sales taxes on domestic Subsg susidiesanufactures, sales taxes on Iqports, datles Co central governmnt consumption
on imports, dutles on exports, and other Tgp - goverrnent transfersindirect taxes Nga.Ngp - interest on dmoestic borrowingSector - refers to activity shich generates the taxes Hof - Interest on NLT foreign debtlsgf Interest on short-term forefgn de't
3
E wtk-i
CApr, _ _ . GDPfc,3
I GDPfc 4",k-i
where:
CAPr - capital reverne
I/ Includescapia ve.
TAILE I CfTfD ENAIm AL REIATIOU8UIP
Central CovenwMt Ffnwices cont$d
Recurrent expaiitsure:
* V,. wugr
k-IGds, * . U,
3
k-1
where:
V waes aN salwieswar wages and salaries: nominal growth rate targetGd9S - goods nd servicesk - 197, 198. 1t89
Int 9 DCD6g#. r
I OthREX..k-I
OthiEX, u _ _ _ _ __ _ _ DPIc,3
I SPfc 5 ,k-I
where:
Into - Interest on debt outstanding and disbursed (foreignor domestic depending on specification of k)DoDg - Governrent debt outstanding and disbursedothREX, - Other recurrent empenditure
k - 187, 1988, 1989
TIJ I D I£T1AVWAI Ilt I &C OF fUWS IDEIITIIY
Central Ginnut FInenes ant'd
capitat EDpwituwe
3
Ik-I
leg, 3 _ _ _ _ _ INPg,3
I IIPO..k-I
where:
lcg - capital expenditure of the central governmtIlPg - goverment lwortsk 1 1987. 1988, 1989
3ialncef PaI Ut of Pemnt
9
Expt I geq ., tl + 4 gdpfc,).t1 + a.pd.) TOTOPCpS a NFA + Sfk-i TOTOPCaIU a Lfg + Lsfg + Lfc + Ifd +
Lfp + DFI + Lsfp + Kne
lapt w I (~l ti + , gcltc).(l + a-.p4) TOTBOPCurS a Mfg + Vf.Vpf I aplsfpI+m Ine + Prof + Yrgf
TOTBoPCurU -RC * Tfp .I p + Sf + IFA
where: where:
exp - 9 categories of exports: coffee, tea, horticulture, Sf - current account balancepetroleus, mile, mmafacturlng, re-exports, other goods, Lfg t net overn entnon-factor services capital '(NIt)
1E - 8 categories of iprts: food, petroleun, other cnsumption Lfe - foreign borrowing by thegoods, primary Intermediate goods, mafacturfng Intermediate central beatn 03goods, capital goods, non-factor services. Lfd . foreign borrorn by the
ccmerclat bssks (net)Lfp - foreign borrecng by the
private non-finwnclet sector(net)
DFI - not direct foreign3 investmentI DFI0 Lsfp - short-term capital
inflows to privet sectorOFI, (Lfe*Lfp), a Kel net capital not elsere
ircluded
I (Lfgttfpl. Ifg - Interest on LfgNs llUpf - interest an foreian borrwing
by central bank ancecrclat bait
Isfp - Interest on LsfpEsgf - Interest an short-tem foreign
debt11Wi intert on Knolhere: Prof reittances of profits
nd dividends
k - 1987. 1988. 1989 C rsrce balatUtp - trfer to the private sectorSf - current ecount balUFA - Interest an NFA
NMIin
K=Sgi4ifc, a gifc, I gdf .(1 . kgr) TOTIApS a l *Ip
kIl TOTUCFaI a so + Sp; Sf
GDPx% a 0Pfc , Ti - Subg TOTIUCuarS * C + l + Cp. Ip - oTOTUACurU a GDPfc Ti -abg
OODg, Lsfp, Lsfg, Prof and Tfp are projected in a similar way.
.JTA [IfO FLW O uLAT IlP FLOWF RDS gm tpeII
Redl *M cont'd al Ac
Ip, Ip,1-C 1 a$.gtcfl.(1 e e,.(r 1 ) TOTUACpS - ational acccnts:capital sxurce of funds
TOTCI - natinal accosuts: capitalues of fus
TOTNAurS nat onal accounts: currentsources of ftun
TOTNACuIrU national acosuts: awrrentuses of ud
where: where:kgr - refers to growth rate of real value added In the k sector: 19 - total pltlic Investment agriculture, mining, manufacturing. other industries, Ip - total private investmentservices. Sp - private sector savingsgtWfe - real growth of iDPfc Cp private censu ptfonGDPsp 0P at market prices
mm:
1. Cp. private consumptfon, is the balancing Item In the national accounts.
2. ProJections of export and isport prices. including the maeufacturing Index, and LIBOR are taken from the Plwning Assuspti4um for KajorMacroeconomic Indicators, international Economic Analysis and Prospests Division, International Economics Departent, irid Bank.
4/ includes increases in stocks.
- 48 -
10. Trade data come from the BOP Unit of the Ministry of Finance. Because this data comesfrom the same source as the balance of payments figures, they are expected to be consistent. Thetrade table provides information on major items in the merchandise account, including export earningsof coffee, tea, petroleum, horticulture and manufactured goods as well as imports offood, petroleum, primary intermediate goods, manufacturing intermediate g&. Is and capital goods.These items serve as the basis for projecting the trade account.
11. Apart from ensuring that columns add up correctly, the model checks for consistency betweenstocks and flows within given worksheets, and identifies inconsistencies aross data input sheets. Forexample, it checks for consistency between the capital flows implied by the DRS debt data, thebalance of payments, the fiscal accounts and the monetary accounts, and calculates the adjustmentswhich are required for consistency between capital and current sources and uses of funds. Similarcalculations are done after trade data in the nadonal accounts are compared with the figures in thebalance of payments. Because the national accounts are viewed as the 'umbrella,' the rule of thumbis to identify those adjustments which make the input tables internally consistent, reconcile stockswith flows and achieve intersectoral consistency in relation to the national accounts. The modelspecifies these adjustments explicitly.
Updating Historical Data
12. The model has been designed so that in general, historical data entered in the data inputworksheets automatically override projected data. The task of updating the model is simply one ofentering the most recent year's data in the appropriate data entry worksheet. For example, thenational accounts projected by the model for 1990 could be replaced by the actuals for 1990when they are entered in the national accounts data entry worksheet. If similar data is not availablefor monetary aggregates or other sectors, the model would continue to apply the projection rules foryears (including 1990) for which historical data are available. However, the projections would nowbe consistent with national accounts information for 1990.
13. Exceptions to this simple rule for updating the model are projected variables which usehistorical input data for explicidy specified years Mn are programmed to apply from a given startingdate. Many of these formulae and their calculated variables are time-bound to avoid simultaneitywhich JAVELIN is unable to handle or to anchor the base year of an index. The time-boundformulae are presented in Annex V.
Programming Information
14. The model is programmed in JAVELIN, a spreadsheet and database management softwarepackage. The program solves the model recursively, subject to the convergence criteria which theuser of the model could change easily. The solutions reported in this paper satisfy the criterion offalling within -5 and 5 units of the target.
15. A 'macro' has been written for each of the solutions discussed in Chapters m and IV:exchange rate solution, interest rate solution, monetization, exchange rate and interest rate solution,and the exchange rate, interest rate and monetization solution. Each of these macros operatessequendally from 1990 through 1996. For flexibility, additional macros have been written whichactivate each of the foregoing solutions one year at a time.
- 49 -
ANNEXII
FISCAL DEFICITS AND MACROECONOMIC CONSISTENCY
A. Introduction
1. Fiscal deficits could be financed through external borrowing, domestic borrowing or/andmonetization, that is, creation of base money. Each of these optiom as important macroeconomicconsequences. The most frequently discussed of these consequences in applied macroeconomicanalysis is the crowding out of private investment which could occur if the budget is financed throughdomestic borrowing. Often reference is also made to the potential inflationary effect of monetizingthe deficit. By this is usually meant the potential impact of increasing the money supply on thedemand for goods and services. However, at a given level of output, an increase in the moneysupply would only raise inflation if it exceeds what individuals would normally hold, other thingsbeing equal. Similarly, at a given level of inflation, a money supply increase would only be"excessive' if it exceeds what individuals would normally hold as output increases, other things beingequal. Accordingly, the level of inflation and the demand for financial assets limit the extent towhich the fiscal deficit can be financed through monetization. One corollary is that there may bea unique "financeable fiscal" deficit which is consistent with a macroeconomic policy package whichsets targets for inflation, external borrowing, domestic borrowing and output growth.
2. This Annex explores this issue in the context of Kenya. It uses a simple framework whichlinks the financing of fiscal deficit to domestic borrowing, the interest rate, the inflation rate, outputgrowth rate and the real exchange rate. This approach draws on the methodology developed byAnand and van Wijnbergen (1989) and Catsambas and Pigato (1989) but includes disaggregatedbehavioral equations for money demand.
3. The' Annex begins by reviewing definitions of the budget and then selects the onerationAconcept as the one most suited for determining the financeable deficit. Next, a budget constraint isspecified and manipulated to obtain an accoundng idendty which includes money creation, inflationand borrowing. Tbe specification also captures the role of the inflation tax and seignorage in deficitfinancing. Linkages are then made to selected policy targets, the behavior of money demand and theconditions for consistency between fiscal imbalances and selected macroeconomic objectives.
4. After sketching and discussing the integrated framework, the section assembles data fromChapter 3 to evaluate the consistency of fiscal performance with other macroeconomic parameters.
B. The Budget Constraint and Macroeconomic Targets
5. The operational, primary and cash deficits are three measures of fiscal imbalance. Theoerational deficit is expressed in real terms and excludes the inflation component of interestpayments by the government. The rationale for using real values is the view that interest paymentsby the government may contain an inflation premium which is not regarded as income by therecipients but as a restoration of inflation-induced capital losses. In consequence, the inflationcomponent of interest receipts would not be spent, at least not fully, and therefore would not (fully)translate into a claim on resources. The primary budget goes a step further and focuses on the non-interest component of the budgetary accounts. Hence, this concept is useful for analyzing the impactof curren fiscal policy on aggregate demand, since the interest paid on debt is only the result of ps
so.
fiscal policies. £ Finally, there is the cash or observed deficit which is pacticularly useful whentracing the movement of financial resources between the central government and the rest of theer.onomy. 'The view is now widely held that the ogeradjonalg deficit best reflects the abso.pdion ofreal resources by a government, and thus is the most appropriate measure of fiscal plicy within thefamework of stabilization and adjustment." #6. The analysis which follows Is anchored in the govemnment budget identity whichacknowledges that the sum of the real non-interest defcit (d), including the real inerest due ondomestic debt (ib) and foreign debt ([be), is equal to the sum of financing from all soum es.Financing could take the form of real external debt (b'), domestic debt (b) or central bank credit tothe government (dc). Formally,
d 4' ib + ib-e 3 b + b"e + dc, (1)
where i is the real domestic intarest rate, i^ is the real foreign interest rate, e is the real exchangerate, and " ' indicates the change in the specified variable.
7. Eq. (1) would come closer to reality if it were to include deficit financing which may not bereflected in the account of the central government. It is conceivable, for istance, that some foreigndebts of the central government could be handled by the central bank without being adequatelyrecorded in the budget. To cover such cases, consider a simplified balance sheet of a central bank.The assets therein are real credit to the central goverment, dc, and real net foreign assets (valuedin local currency), nfae. The liabilities consist of currency held by the public, Cu, and commercialbank reserves, RR, so that:
M/P = (Cu + RR)IP o
where P is the domsstic price level and M/P is the real value of the central bank's nominal basemoney or the central bank's net liability to the p,ivate sector. The balancing item is assumed to benw, the real net worth of the central bank, where:
nw M/P = dc, + ae .. (3)
Eq. (3) indicates that the net worth and base money are used, in this illustration, to extend credit tothe central government and to obtain foreign assets. Intuitively,:
nw ifwe ... (4)
that is, that the real value of interest earnings on foreign assets Increases the central bank's real netwoeth.
S. Eq. I can now be rewritten to reflect the inc!usion of the central bank in the budget constraint:
I/ Tis claim makes th simplisic aumpion tha no idtrs payments oa deb conttd during the ourent period all duewithin dha patiod.
I/ Catamba at i. (t989). p. 4.
d + ib + i(b - ab)e = b + (b- -nfale + dc,+ nfe -nw
The &eft hand side of eq. S consolidates the defbit of the central govenmuent and the central bank bynettng out the net worth of the latter. On the right hand side, the change in net worth-equivalentin this context to the change in net foreign assets-is introduced (and then subtracted to mainainequality between the right and left sides) to forge a link between the deficit and the money supply,MIP, where:
MIP = dc, + nfae - .. (
Eq. (6) is the dynamic form of eq. 3 and permits eq. S to be rewritten as:
d + ib + i7(b* - nfa)e = b + (b- -nfa)e + m ... (7)
In eq. (7), *m" excludes private sector deposits in the banking system which would be offset in aninflationary situation by the erosion of liabilities outstanding to the private sector. Otherwise, "m"would not represent a U revenue increase-the inflation tax-which partly finances the deficit. 1'A logical extension of this argument is the conclusion that the ggnratonal definition of the budget,the one which excludes the inflationary component of interest payments, is the one most appropriatefor this analysis (para. 2).
9. In eq. (7) a link is establisned between the deficits of the cenal government and the centralbank, and the money supply. The next step is to find a way to reflect capital losses which arise fromreal exchange rate changes and which are a part of the cost of servicing the foreign debt. This isdone by decomposing the change in real net foreign debt (valued, in local currency) into the impactof the real exchange rate movement on the sto^k and flow of debt; that is,
[(b-nfa)e] - (b-nfa)e + e'(b - nfa)e ... (8)
If The appropriate deriniclon of money for this exercise should coincide with the central bankc's aet liabides; tha is,ba money (currency plus commecial bank reevs with the central bank) should exsclude credits which the centnabank may have extended to the ptivae sectot and to commercial benka. This is becas net revenue via the inMadtiontax should exclude revenue derived feom the inflationaty erosion of the priva setom's deposit in the banking system.which Is offset by the inflationary erosion of loans ousanding to the private sactor. What remains is the central bankd non.wnterestbwaring liabilities to the privat sector. For this eaon it would bo neccuay to adjust da reportedas reserve money in the Interntional Finance Sistics to exclude any loans by the canta bank to the nongovernmentsetot.
2 2-I
where "'" is the percentage change of the variable. Similarly, the change in the real value of thenominal base money may be decomposed into:
MIP m + P'm ... (9
The first variable on the right hand side of eq. (9) represents ana egL-t change in the realmoney balances which economic agents hold at a Ma rate of ifation. Conceivably, as nationalincome grows, agents may wish to hold higher real money balances. 1 y be a source of deficitfinancing in that it would allow the government to increase the real money stock (in line with thisincrease holding of real money balances) without fueling inflation. The second term on the right handside of eq. (9) represents the inflation tax or the amount by which the nominal money balances haveto be increased to keep the real value of the money stock constant.
10. Eqs. (8W) may be substituted in eq. (7) to produce:
d *- ib + (i! + e')(b- n)e = b + (be -nfa')e + [(b- nfa)e]
+ m + P'm .. (10)
Eq. (10) restates the budget constraint as the requirement that the operational fiscal deficit, includingreal interest payments and the real value of the profits of the ceural bank, must equal the changesin the real value of domestic and foreign debt (net of the real value of the profits of the central bank),and revenue from seignorage and the inflation tax.
C. The Budget Deficit and Macroeconomic Targets: From Design to Application
11. The practical question for the policy analyst is to determine the level of fiscal imbalance whichis consistent with the growth target, the inflation target, the real exchange rate and interest rate, andthe limits on domestic and external borrowing. The first step towards an answer is to derive theactual annual deficits of the central government in accordance with the left hand side of eq. (10).The second step is to calculate the financeable deficit using the target value for P' and the ratio ofdebt of GDP a la the right hand side of eq. (10). However, although eq. (10) is useful, it ignoresthe relationship between exchange rate changes and inflation and vice versa. It also ignores the factthat inflation may have an independent negative impact on the size of the monetary base from whichthe inflation tax is derived and omits the influence of factors such as interest rates on differentfinancial instruments. e'
12. The econometric results in Tables I and 2 are an input into the calculation of the demand forbase money and thereby, an input into the calculation of the seignorage, the inflation tax and thefinanceable deficit. Tae signs and magnitudes of these results are theoretically plausible andeconomically sound. The relationship between real income growth and real money balances ispositive. Inflation appears to have an understandable negative impact on the growth rate of real
11 Finnncial rogulaions-reserve requirenents. cash ratios, a-e-would also play a pan in determining the currency held bycommercial banks and nonbanic uLancial instiions, and thereby, in determining the finaAceable deficit. Thea factors didnot appar to be sazaiaically significant and cconomically meaningful in the esimated money demand functions reported inTable 1. This is probably understandable since flnaeal seatr ratoms were sed only recently In Keya. However.subsequene reserch will analyze the impact of finaneal forms on the demand tor money.
T he. sam'.e contain a a ' . peroed: .o . :..: p .: ' ' .: | -i: .:T:. h :,igue e estatti Mrefers+ variously: to* thea eavalue. O crrency es a tame deosts
by the. Nairobi consumer price index. Al varables are exese&.exq d ir .natul logaFithms. The. dt-mmfor the currency equ;tioni. ^sone for119.81-84 and 1-98849(adjustingyears) and. zero,otherwiseThe dumnmyfor the demand deposits.eqt~ation is one f' h cffeboo er (arnd three er
. .~~~~~~~~~~~~~~~frtec .:e .- ,c yers a. .: . .,:
thereafterdandzero otherwise T-he dummy for the timedeposits:equat. t-is.one for8 8 an
zera otheswise.
./ nterest rate on. deposits wit .bui.l.ing. societies.
money demand while the demand for demand deposits and time deposits and their respectiveinterest rates is understandably positive. Currency does not eam interest and is therefore negativelycorrelated with the interest rate on time deposits which are an alternative asset. Table 2 shows themagnitude of these relationship in the long-nm. These coefficients are used tO simulate the behaviorof base money as inflation and real growth change, to esdmate the inflation tax and seignorage andthereby, to determine the fiscal imbalanceswhich are consistent with the stabilization and.t.adjustment objectives.
13. The results in Table 3 are obtained fromthe financeable deficit shell of the model usingthe methodology and coefficients in pra. 6-12.The operational and financeable deficits arecalculated and the sources of fiacing-seignorage, the inflation tax, domesticborrowing and foreign borrowing - are shown.By disaggregating the sources offinancing, the Table is also able to provide a pardatl it ...equUibrium perspective on what the consistent j0e ̀:t:>financeable deficit would be under each scenario{f that source of Jinancing does not materialize Ol97 ;. .001
14. Generally, the availability of foreign " ' _'''';____>:
financing is the largest single determinant of thefinanceable fiscal deficit when exchange rate and , , . .; Oe.4trora .Ta , .
the interest rate policy are used and in the basicprojections. The inflation tax is also a . .... ..
significant contributor especially when the fiscaldeficit is monetized.
D. Limitations of the Framework
15. Whereas this framework is useful in calculating the level of fiscal imbalance which is consistentwith given macroeconomic targets, it does not deal with the desirability of one form of deficitfinancing versus another. 2 It also neglects the implications of financial sector reform for moneydemand.
2/ 'Me model is aiso inappropriate for dealing with the microeconomics of various revenue instruments.
- 55 -
Table 3Tbe Fiancable Fisal Deicit in Selected Scenarios
A deficit carries a positive sign. A negative deficit is a surplus.
-56 -
ANNEX m
IMPORT ELASTIC IS IN KENYA: 196849 AND BEYOND
1. In Kenya import restrictions were used pervasively during 1968-89 to close the gap betweenthe notional demand for imports and the available foreign exchange, as weil as to protect domesdcproduction. This Annex explores how Import elasticides are likely to behave after import restrictionsare lifted. The broad conclusion Is that traditional esdmates witi1gnort import restrictionsunderstate the otrue" import and price elasticities of demand. Comparative stadc analysis suggestsdtt with import liberalization in Kenya, the income elasticities of demand for imports of goods andnonafctor sevices would rise to about 1.00 but would be dominated by an aggregate price elasticityof approximately -1.45.
2. Section One reviews the traditional and modified traditional approaches to import demandspecification, idendfies their weaknesses and suggests modified, disaggregated import demandfunctions for Kenya that take into account import restrictions and the influence of sector specificactivity. Section Two reports and analyzes the econometric results and enumerates their implicationsfor import behavior after trade liberalization. Finally, Section Three identifies areas for additionalempirical work on import elasticities in Kenya.
&WnnAnoroaches to the Sgeciflcatdon of m=t Elasticities
3. The traditional specification of import demand is:
In M bo + b, In Y, + b 2 In (P/PJ + u, (1)
where M is real imports, Y is real income, P. is import prices, Pd is domestic prices and u is theerror term with the properties u, - NID (0,oa:). As the equation is specified in logarithms, theparameters b, (usually bt >0) and b2 (usually b2 <0) are the real income and price elasticities ofimports, respectively. Tbis traditional approach has three major operational weaknesses. First, ittends to focus on aggregate imports but for developing countries in pardcular, different categoriesof imported items may behave very differently. Second, the traditional approach tends to emphasizeaggregate real income even when imports and prices are treated at a disaggregated level. i' Itis therefore unable to capture movements in imports, aggregate or disaggregate, which may beassociated with changes in the sectoral origin of national income. Third, eq. (1) does not take intoaccount trade restrictions which may suppress actual demand below its notional level.
4. To overcome these shortcomings of the traditional approach, this paper relates broad categoriesof imports to aggregate income as well as valued added in specific sectors. The methodology of Chuet al. (1983) and more recently Faini et al. (1988) and Moran (1989) is also followed by introducing
indicators of foreign exchange availability into eq. (1). The new specification is:
j2i Kha (1975) discusses the problem of using aggegage imports in the traditional demand equaion but ignores thepractice of using aggrega income. See M. S. Khan, The Strtucture and Behavior of Imaport of Venaeuela.Rovi n ics anStatistics. vol. 57. pp. 221424.
-57-
lnM4 bo + bt In Ys, + b2 In (P,P&J - b, ln F,+ b, ln R.L + bs ln M, + 4u (2)
whre Ys is valued added in the named sector, F is the sum of export earnins and not capitalinflows, R is the end-year reserve posidon and MN,, Is the laged values of the dependent variable.All variables are expressed in millions of real Kenya pounds exceptn prices which are In indicesderived therefrom.
S. Eq. 2 is a mixed blessing. k diaggrgates the relationship between categories of Impors andvalue added by sector and includes foreign exchange variables which, at least In the Kenanl case,were likely to have influenced import demand. However, it does not allow the stuctural parameters-the real income and price elasticities of demand-to be reovered. IV Moreover, as with eq. (1),prices are assumed to be exogenous. w
6. Faini (1988) and Moran (1989) suggest a system of simultaneous equations which includes theforeign exchange constraint but which also allows unconditional elasticities to be recovered. Theequations are as follows:
InM - ao + a, 1 nYs + 4 2 In [1P.JPJ + a3 InM .+ut (3)
InM.A bo + b, InF + + In Pt +b,lnIMn t+ v, (4)
where M, is the supply of foreign exchange; other variables are as defined in eq. (2).
IV If eq. (1) were the correct import demand fucdon. b, and b would be inctpru resecively as the real income end
price cagltjides. If eq. (2) were the cocrec specitlladon, bt and b2 would reresent the influence ot Income andprice, sM the influence ot the other stauistically gnliticant variable.
W The a.i.f. pricm ot impocu may be exogenous but policy-induced masures such as txe. tAriffs and non-ariff
baniers may affect the price ai is aculy paid. Similarly. demcsdc prices may be influened to considerablemeasure by governmantierlvetons rsch as price controls. Thus, PJP may not be exovocus.
Section EIEconometric Estimation and Results
A. Econometric Results at a Glance
7. Eqs. (1) and (2) were estimated by the method of ordinary least squares (OLS); eqs. (3) and(4) used two-stage least squares (2SLS). The estmates which appear to be economically meaningfuland sutistically significant are reported in Tables 1-4. This secdon focuses on the range of valueswhich the various specifications and esdmation techniques yield for the elasticities of given variables.(Box 1).
8. The influence of foreignexchange availability (1.27 for -Box 1foreign exchange and 0.34 for *. *e6n lOIErastlcitfs
lagged reserves) appears todominate the income (0.55) and .g s- - P<;ep
the price (-0.35) elasticity of ._:_:::_____i;___-demand for imports at theaggregate level. When relative i 055. .0.25.1 lJ7fflet(GP4FW 0.75~~W b/ 4.46,4. bl O34W14prices are endogenized and -1.60 @1
foreign exchange availability is .03501 0.3t-1 ..taken into account, the "true" ... : ;. Q?9 i -. 8W -income of demand rises to about . .O1: .1.00 and the price elasticity levels ; ,;o02ot.o. ;7agout at approximately -1.45. > .Foreign exchange availability is , tste, XY.s1
also dominant in the demand for0 38 .1.4~I02OtX eprimary intermediate goods XM'; o -1.3461 .
(1.40) and 'other consumergoods' (0.85). This result is 79 t t.64@1 1
Qdw ~~~~3.79b1 15.consist,.nt with a policy stancewhich favored the importation of o -0.22(01W-2.73 el 0.85 a
producer items and carefully COe. 0.58 el -370 bI
rationed imports of consumptiongoods. a/ModIfiedtadJ,trzfonaI'OLSI G - GOP
IV MgdtedtIat (2SLS) f. fo,.exc.9. Foreign exchange c/.t: 6 SI r-Resis.availability appears to equal the odl ot;* stX --..6p GNFS'influence of income (1.35 . aggudvauab/.compared to 1.28) in the demandfor capital goods but is upstagedby income in the equation for fuel .and lubricants (0. 19 compared to1.02). The probable explanationof the latter is that some fuel and lubricants are refined and re-exported, and thereby earn (insteadof being constrained by) foreign exchange. The fact that constrained income elasticities and foreignexchange appear to dominate the price elasticities is a particularly interesting finding because theseapproaches treat prices as exogenous. The intuitive conclusion suggested by this finding is that whenprices were exogenous, the controls over imports may have focused on foreign exchange allocation.
59 e
The converse is probably true. Indeed, once foreign exchange constraint are incorporated and pricesare endogenized, the income elasticity rises and dominates the price elasticity of demand for capitalgoods and manufacturing. However, the price elasticity takes pride of place in the demand for otherconsumer goods.
10. Comparison of statistically significant OLS and 2SLS estimates fiuther suggests that as therelative prices of imports rise with trade liberalization, import growth would be dampened, ceterisparibus. The share of food and other consumer goods in imports would also fall. Liberalizationwould cause the income elasticities of demand for import of goods and nonfactor services in Kenyato rise but cause the price elasticity ofdemand to fall marginally; the incomeelasticity of demand for food would fallbut its price elasticity would rise; the prb¢ g i miii tanEStc.income and price elasticities of demandfor fuel would fall; the income elasticity ; ;; ,s -of demand for primary intermediate goods . * .
would fall but the price elasticity wouldremain virtually unchanged; and theincome and price elasticities of demand : E . i P ;for manufactured, capital and 'other m: --consumer goods' would rise (Text Box J.|,. ;- v h Pe
2). The dominant influences suggest that ras the economy grows and trade isliberalized, the shares of capital goods, Pd . ' Le - a;manufactured items, fuel and 'other ;-consumer goods' are likely to increase, _A A Ihusoi
ceteris paribus. Furthermore, if the :relative prices of imports rise with trade caita A A ince.
liberalization, it is likely to dampen aimport growth, ceteris paribus. The od ncs A .I.A
share of food and other consumer goods Gos
in total imports would fall.
A*.... .- .. . - ... :.
B. Econometric Estimation andResults in Detail
11. Turning now to the estimates ingreater detail, this section finds that theaggregate elasticities which are derived from disaggregated import groups differ from those whichare based on aggregate data. However, only disaggregation of imnports and prices was meaningful;sectoral value added was often not a statistically significant replacement for national income or GDP.
60-
12. The coefficients reported in Tables 14 have convendonal signs in that the 'income' elasticitiesare positive and the 'price' elasticities are negative. One exception occurs in Table 2 for food andbeverages: imports are positively correlated with value added in the service sector but nagativelycorrelated with value added in agriculture. Moreover, the influence of the foreign exchangeavailability appears to be weak. These findings may be explained by the fact that: (1) a significantshare of imported food and beverages is consumed in the foreign-exchange-earning tourism sector;and (2) Kenya produces a large share of its food supply. However, the positive sign of the dummyvariable suggests that imported food supplemented domestic production during major droughts. Theoaly other unusual sign is for the "income' elasticity for other consumer goods, also in Table 2. Theexplanation for this anomaly may lie in the apparent strong statistical significance of the foreignexchange constraint, which is consistent with the fact that this category of 'discretionary' imports wasparticularly suscepdble to Import controls.
13. Overall, Table 2 appears to betuer explain the behavior of imports in Kenya during 1987-89than Table 1 (which excludes indicators of foreign exchange availability). In fact, the equation forprimary intermediate imports seems to have been exceptional among Table 1 in that it did flDundergo at least two structural shifts, shifts which were probably induced by changes in foreignexchange availability.
14. Table 3 highlights the fact that aggregate income elasticities from disaggregated data werehigher than those based on aggregated data (0.96 versus 0.78 in the tradidonal approach; 0.69 versus0.55 in the modified traditional approach). I The differences are more pronounced for price (-1.60 compared with (-1.14). Given the size of the differences in income elasticities and the fact thatincome is variously GDP and national income, the results from aggregate and disaggregated areassumed not to be operationally significant. This interpretation is consistent with the fact that theuse of resources at the disaggregat2d level must fall within the resource envelope set by aggregateincome. Ordinarily, prices at the aggregate level are related to prices at the disaggregated level. Itis also true that the domestic price level is a common denominator in all relative prices and thatdisaggregate import prices are a component of the aggregate import price level. However, the importprices used are border prices and thus may not be the final prices to the consumer. The distortionof these price relationships at the disaggregated level may then become amplified at the aggregatelevel.
15. Table 3 shows that with the exception or income elasdcity based on the traditional approach,there appears to be no appreciable difference between weighted and unweighted average elasticitieswhich are based on disaggregated data. What is very significant operationally however, are thedifferences between various categories of imports. For instances, the income elasticity of demandfor capital goods and intermediate primary goods exceeds 1.00 in both versions of the model and ishigher than the "income' elasticities of demand for other consumer goods. This may well have beenconsistent with foreign exchange rationing which deliberately favored producer items over consumergoods.
16. Table 4 indicates that disaggregated elasticities which have been estimated using the traditionalapproach differ, at times significandy, from those which have been obtained from the non-traditionalspecification. This is not true for capital goods and food for which the "structural" coefficients risemarkedly when the impact of foreign exchange availability is incorporated. In general, the estimatesreported in Tables 14 support the broad conclusion of Faini (1988) that "a less restricted traderegime is associated with higher responses of imports to economic incentives' (p. 21).
at1 Htiweovr. the reults rtomn weighted and unweighwcd aggregation arc quite similar.
- 61-*
Table IUmpirical Results: TraditIonal Appvoach
(OLS Estimates)
Importsof goods Food Fuels Primaryand non- and and inter- Manu- Othetfactor beve- lubr- mediate factured Capital cons.services rages cants goods goods goods goods
The dummy (ai ls one for the oil shock years and zero otherwise; (bJ is oine for post-1973 oiol shock) periodand zero otherwise; (cJ Is one for the adjustment perlod beginning In 1987 and zero otherwise; (dii one forthe adJustment period in the early and late 1980s and zero otherwlse; (el is one for the droughts in the 1970sand 1980s and zero otherwise; (f Is one for the period after the second oil shock and zeo otherwise; and (giIs one for the perlod years following the oil price Increases up 40 the commencement of sector adjustmentoperations, and zero otherwise.
imports of goods Food Fuels Pdmaryand non- and and Inter- Manu- Otherfactor beve- uhdb- meato e fwctwd Capital cons.sewices rages cants goods goods goods goods
Au equations used a constant term, foreign exchange and reseres as lnstnUments. In addition (81 used a tfmetrend and the dummy (bh; (Ci used a time trend also along with dummIes 0b) and (el; (DI used the dummy tei;(EJ used the dummy fci; and Inf used the dummies (cJ and Ifl. See Table I for descriptfon of dummyvaritiAes.
-65 -
In keeping with the fact that import controls in Kenya dudng 196849 focused more on incomevariables - particularly foreign exchange - than on prices, the income elasdicity of demand estimatedIs found in this paper to rse markedly with liberalization while price responsiveness appears to faimarginally.
Secdon IVUnfinished Buslness
17. Capital goods prices and income elasticides in all Tables are very large, possiblyreflecting the bulky nature of several items in this category. At a lar stge it would be useful todisaggregate this category of goods and to separate government imporm from private imports.
18. Foreign exchange receipts, including export eanings, are assumed to be exogenous inthis paper. This is not likely to be the case in practice, because exports volume may be determinedin part by the availability of imported inputs. In tur, the capacity to import may be affected byexport earnings. The precise linkages between dtese variables could be the subject of futre research.
-66-o
ANNEX IV
KEY ASSUIPTtONS AND TARGETS Base-- --- -----.--- Projections
Quasi-momny 16246.8 18387.6 20987.6Other Its (net) 1/ 7160.0 8017.4 8287.0
TOTAL LIABILITIES 45394.5 49039.0 54583.4
1/ Entries in the ccoumts of the monetaryauthority ond the commericat banks havebeen added for Internal consistency.other Item net Ctlabilities) have ben adjustedaccordingly.
If Is etrsted from the resourco balance In he Doerfved UP9 Tab*2/ Ixeludes diret foreign fnvestmet3/ Assued to be 7X percent of the totat shwt-tem4/ 19T error and owl"fum h. bee adJusted for consistencS/ Total discrepany betwnen tapitt tnflos
re: thie SP and capitat inflow re: DOSCaim. sian indicates that shortfatll Itn O dota vsgP data)
-77 -
DATA INPUT WARKSHEIT................................ _
external Debt..... ....... 0
(USS Millions)
1987 1988 1989
Medfun and long-term 4777.0 4711.0 4633.0
Public and publicly guranted 4281.0 4084.0 4001.0
Central goverrssnt n.a n.a fi.n
Parastatals n.e n.a n.4
Central bank nMa n.a n.e
Financiel sector n.s n.s n.a
Private sector guaranteed n.e n.a n.a
Private sector non-guaranteed 496.0 627.0 632.0
Use of IMF credift 401.3 455.3 415.3
Short-term 608.4 564.4 641.4
TOTAL OEBT Cincl. ItF & short-term) 5786.7 5730.7 5689.7
DISUURS!MENTS
Disbursements CMLT) 531.0 527.0 491.0
Publfc and publicty guranteed 441.0 331.0 471.0
official credftors 300.0 271.0 384.0
Multilateral 153.0 187.0 335.0
Concessionat 98.0 151.0 259.0
IDA 7;i.0 136.0 227.0Nonconcessiona 56.0 36.0 76.0
1/ From Goverrhnt of Kenya (see CEM, 1990. StatisfcaL Appendix)and IMF Staff Report for the Nidterm Review of the SecondAnnual Arrangewnt Under the Enhanced Structural AdjustmentFacility. Nov. 14, 1990.' 1987 refers to FY 86/87.
2/ IncLudes IMf adjustment to cash basis and uwillocated
3/ f tlects dinmstif borrowing *nd cretin inflows which are projectedunder thG senwrfo. No additfonas ltfits on financing laposedbut the mechanics of doing so fn the model are straightforward.Exeludes grants.
4/ As uwns that aLl PPG foreign borrowing acrues to the central government.
PRF.IO USe(mp) + 4 m)1)Z (1 9 901)z a.2 .3- XFYFA TA XI .2.3;VVA L UE (FlA TA XI .2.3/((PE VMO US (Imp) +Imp) /2), ,(7 990])
z fy.2.4=FF TAX .2.4dPRWOUS(jp)+Fp)12* VA LUE(FYATA X 2.4/((PREO US(Ex) +Ekp)/2),l199J)
z fiyfa=w.2Z5.=BEGN(f9JFYFATMML2.5,3), 0
- 103 -
,1 I~~~~~~~RFRENC:ES
Anand, R., and van Wijnbergen, S. (1989), Inflation and Financing of Government Expenditure:An Introductory Analysis with an Applicadon to Turkey. The Wgrld Ban Economic
vi January 1989, Vol. 3, No. 1, pp. 17-39.
Buiter, W. (1985), A Guide to Public Sector Debts and Deficits, Econgmic Plicy. November 1985,Vol. 1, No. 1, pp. 14-79. Camsambas, T. and Pigato, M. (1989), The Consistency ofGovernment Deficits with Macroeconomic Adjustment An Application to Kenya and Ghana.World Bank PPR Working Paper, No. 287.
Catsambas, T., and Pigato, M. (1989), The Consistency of Government Deficits withMacroeconomic Adjustment: An Application to Kenya and Ghana, The World Bank, PPRWorking Papers, No. 287, p. 3,
Chhlbber, A., Cottani, T., Firuzabadi, R. and Walton, M. (1989), Inflation, Exchange Ratesand Fiscal Policy in Zimbabwe, mimeo, World Bank.
Chu, K., Hwa, E.C. and Krishnamurty, K. (1983), "Economic Instability and Adjustment ofImports, Capital Flows, and External Reserves: A Short-Run Dynamic Model," in D). Bigmanand T. Taya (Eds.), Exh e ant d de Sabil Es dRimedia.
Easterly, W., Rwa, E.C., Kongsamut, P., and Zizek, J., "Modeling the MacroeconomicRequirement of Policy Reforms, PPR Working Paper, May 1990.
Faini, R., Pritchett, L. and Clavijo, F. (1988), Import Demand in Developing Countries. WorldBank PPR Working Papers, November.
Ielpman, E., Razin, A., and Sadka, E. (Eds.) (1988), Egonomic, Effet of the GoveMent BAugw.Cambridge, Massachusetts: MIT Press.
Rolsen, J. (1989), "KENYA.MOD: A Simplified Example of RMSM-X," mimeo, May1989.
Kbadr, A. and Schmidt-Hebbel, K. (1989), "A Framework for Macroeconomic Consistency forZimbabwe," PPR Working Paper 310, December 1989.
Khan, M.S. (1975), "Cyclical and Secular Income Elasticides of Demand for Imports," Rlview oEconomics and Statistics, Vol. 57, pp. 351-361.
-104-
Rhan, M.S., (1975), "MTh Structre and Behavior of Imports of Venezuela," RBIwofEoomica and Stitistdis Vol. 57, pp. 221-224.
McKenzie, G.A. (1989), Are All Summary Indicators of the Stance of Fiscal Policy Misleading?f IMF= Sp ege. Vol. 36, No. 4, pp. 743-'70.
Moran, C. (1989), "Imports Under a Foreign Exchange Constrlut," Wodd fank EcoMicReview Vol 3, No. 2, pp. 279-295.
van Wijnbergen, S. (1988), Inflation, Balance of Payments Crium, ad Public Sector Deficits.In Helpman, E., Razin, A. and Sadka, E. (FAs.) (1988), Economig Effectz of the
mmmeg &g. Cambridge, Massachusetts: Mfl Press, pp. 287-309.
World Bank, (1990), Kenya. Stabilizadon and Adjustment: Toward Accelerated Growth. Report No.9047-KE. Washington, D.C.
PRE Working Pa r .eries
ContactJJUE Author for paer
WPS742 The Cost of the District Hospital: A. J. MillsA Case Study from Malawi
WPS743 Antidumping Enforcement in the Angelika Eymann August 1991 N. ArtisEuropean Community Ludaer Schuknecht 37947
WPS744 Stainless Stee! in Sweden: Anti- Gunnar Fors August 1991 N. Artisdumping Attacks Good International 37947Citizenship
WPS745 The Meaning of "Unfair in U.S. J. Michael Finger August 1991 N. ArtisImport Policy 37947
WPS746 The Impact of Regulation on Dimitri Vittas August 1991 W. PitayatonakarnFinancial Intermediation 37666
WPS747 Credit Policies in Japan and Korea: Dimitri Vittas August 1991 '; PitayatonakarnA Review of the Literature 3,666
WPS7A8 European Trade Patterns After the Oleh Havrylyshyn August 1991 N. CastilloTransition Lant Pritchett 37947
WPS749 Hedging Commodity Price Risks in Stiin Claessens August 1991 S. LipscombPapua New Guinea Jonathan Colenian 33718
WPS750 Reforming and Privatizing Poland's Esra Bennathan August 1991 B. GregoryRoad Freight Industry Jeffrey Gutman 33744
Louis Thompson
WPS751 A Consumption-Based Direct Tax for Charles E. McLure, Jr. August 1991 CECSECountries in Transition from Socialism 37188
WPS752 Inflation and Stabilization in Roberto de Rezende August 1991 L. LyYugoslavia Rocna 37352
WPS753 The CMEA System of Trade and Martin Schrenk August 1991 CECSEPayments: The Legacy and the 37188Aftermath of Its Termination
WPS754 Esti .-.ing Returns to Scale with James R. Tybout August 1991 S. FallonLarge imperfect Panels M. Daniel Westbrook 37947
WPS755 Hedging Crude Oil Imports in Stijn Claessens August 1991 D. GustafsonDeveloping Countries Panos Varangis 33714
WPS756 Taxes Versus Quotas: The Case Arvind Panagariya August 1991 S. Fallonof Cocoa Exports Maurice Schiff 37947
WPS757 Managing the Transition: Enhancing Eric Rice August 1991 P. Infantethe Efficiency of Eastern European 37642Governments
PRE WorkqEPaper Series
ContactJIR Author Dae for paper
WPS758 Is There Excess Co-Movement of Theodosios B. Palaskas August 1991 D. GustafsonPrimary Commodity Prices? A Panos N. Varangis 33714Co-integration Test
WPS759 The Profamilia Family Planning Jesus Amadeo August 1991 0. NadoraProgram, Columbia: An Economic Dov Chernichovsky 31091Perspective Gabriel Ojeda
WPS760 How Conflicting Definit'ons of Alexander J. Yeats September 1991 J. Jacobson"Manufactures" Distort Output and 33710Trade Statistics
W' 1761 Uncertainty and the Discrepancy Gerhard Pohl September 1991 P. Leebetween Rate-of-Return Estimates Dubravko Mihaljek 81950at Project Appraisal and ProjectCompletion
WPS762 Debt, Debt Relief, and Growth: Caniel Cohen September 1991 S. King-WatsonA Bargaining Approach Thierry Verdier 31047
WPS763 A Valuation Formila for LDC Debt Daniel Cohen September 1991 S. King-Watson31047
WPS764 African Financing Needs in the 1990s Jorge CulagovskiVictor GaborMaria Cristina GermanyCharles P. Humphreys
WPS765 Withholding Taxes and International Harry Huizinga Ceptember 1991 S. King-WatsonBank Credit Terms 31047
WPS766 Economic Crisis, Structural Francois Diop September 1991 0. NadoraAdjustment, and Health in Africa Kenneth Hill 31091
Ismail Sirageldin
WP3767 Framework for Macroeconomic Colin A. Bruce September 1991 M. LynciAnalysis (Applied to Kenya) David Ndii 34046