A Simulation Model for Financial Programming

This paper describes a simulation model that can serve as a basis for a developing country growth-oriented adjustment program. The model has been designed to provide explicit links between fiscal, monetary and exchange rate policies and major macroeconomic variables. While the model is applied to and solved for the case of Turkey, its simplicity and flexibility make it sufficiently general to be applicable to a wide range of countries. The model integrates demand-determined output with a supply side that responds to policies which affect investment and it allows the relative shares of domestic and foreign factors of production to be determined by their relative prices. The model is solved using Lotus 1-2-3, software that is familiar to Fund economists and which allows the user to quickly evaluate alternative assumptions and policies.

Abstract

This paper describes a simulation model that can serve as a basis for a developing country growth-oriented adjustment program. The model has been designed to provide explicit links between fiscal, monetary and exchange rate policies and major macroeconomic variables. While the model is applied to and solved for the case of Turkey, its simplicity and flexibility make it sufficiently general to be applicable to a wide range of countries. The model integrates demand-determined output with a supply side that responds to policies which affect investment and it allows the relative shares of domestic and foreign factors of production to be determined by their relative prices. The model is solved using Lotus 1-2-3, software that is familiar to Fund economists and which allows the user to quickly evaluate alternative assumptions and policies.

I. Introduction

Fund-supported adjustment programs are designed with the goals of promoting balance of payments viability, price stability and economic growth. 1/ In particular, the promotion of economic growth has been emphasized in recent Fund programs and there is currently a strong interest in developing a general framework for growth-oriented adjustment. Khan, Montiel and Haque have developed a model in which elements of the monetary theory of the balance of payments are integrated with an open economy growth model. 2/ An important consideration in formulating their model was to design a framework that linked major macroeconomic policy instruments with growth, inflation and balance of payments targets; moreover, they explicitly adopted a simple structure so that the model would not be difficult to adapt to a wide variety of developing economies. 3/ In this approach limitations on the complexity of the equations were imposed in order to make the model amenable to a reduced form solution. In order to serve as an input in the design of economic programs, however, a model must be comprehensive enough to adequately represent the structure of an economy, although a more comprehensive model may be obtained only by sacrificing an analytical solution. In this case, numerical methods can be used to solve the model.

For operational work involving multi-year programming it would be desirable, if not essential, to have a small, computable model in which explicit behavioral relationships between key macroeconomic variables and policy instruments are specified within a consistent accounting framework. 4/ In the absence of an explicitly specified model it is extremely difficult to design a sufficiently disaggregated, medium-term economic program which is consistent with behavioral relationships. Moreover, a key element in the design of an economic program is the ability to alter assumptions, relationships, and policies to provide a range of alternatives, even if the only purpose of the alternatives is to provide a broader basis of discussion of proposed policy measures. An explicit fully integrated model is necessary for this type of exercise.

The model presented in this paper, while applied to a particular country, Turkey, is sufficiently general to be applicable to a wide range of developing economies. The model is designed to provide a framework for medium-term analysis with explicit links between fiscal, monetary, and exchange rate policies and key macroeconomic variables. Moreover, keeping in mind the requirements of potential users of the model, it is written and solved for use on a microcomputer with particular attention being given to its flexibility—both in terms of its adaptibility to other countries and in terms of being able to use it to evaluate the effects of alternative economic policies.

The model has several broad distinguishing features: it integrates demand-determined output with a supply side that responds to policies which affect investment; it allows the relative shares of domestic and foreign factors in the production process to be determined by their relative prices; adjustment and growth are linked to fiscal, monetary and external policies; it allows for a flexible external debt policy; and finally, it has been solved for the case of Turkey using LOTUS 1-2-3, software that is familiar to Fund economists and which allows the user to quickly evaluate alternative assumptions and policies as well as behavioral relationships.

The remainder of the paper is as follows: in Section II a general overview of the structure of the model is described followed by a detailed description of the equations. Section III presents the estimation results for the behavioral equations. Section IV and Appendix I present the results of some simulation exercises to demonstrate how the model works and the concluding section summarizes the highlights of the model. A user’s guide to the computer program is included as Appendix II.

II. The Model

1. General overview

The economy produces a single (composite) good—the home good—that is either exported, consumed at home, or invested. All imports are assumed to be intermediate, that is, on average there is some domestic value added to all imports before final sale. Output is produced by the use of domestic factors of production and intermediate imports. Domestic value added or GDP is equal to output less the intermediate imports. The relative amounts of each factor employed—the factor shares—are a function of their relative prices. The level of output produced is determined by final demand for output. If output is less than or at capacity, so that there is no demand pressure on prices, the price of the home good is determined by the cost (prices) of the two factors of production, production taxes and import duties. If output is above capacity, prices rise above cost to reflect the pressure of excess aggregate demand. The model, therefore, allows for both demand-pull and cost-push generated inflation.

Final demand is composed of two segments. An exogenous part which comprises government consumption, government investment, and changes in stocks, and an endogenous part which includes private consumption, private investment and exports. Real private consumption is a function of real disposable income. Real private investment is a function of the real value of the flow of domestic credit extended to the private sector. Exports are a function of a domestic scale variable, real gross domestic product, and relative prices. Real domestic value-added is generated by multiplying gross output by the (endogenously determined, relative price sensitive) share of output contributed by domestic factors of production—the value-added coefficient. With an exogenous “value-added price” index that, in principle, accounts for all wage settlements, payments to land, capital, etc., real value-added is converted into nominal terms. Of total nominal value-added in the economy, a certain proportion goes directly to the private sector as income; the remainder accrues to the government. 5/ Value-added accruing to the private sector, adjusted for “export profits,” along with factor income and transfers from abroad, and government transfers determines gross private income. 6/ After deducting direct taxes, private disposable income is generated. Real private consumption is then computed as a function of real disposable income.

The fiscal deficit is computed as the difference between government expenditure and revenue. Government expenditure is comprised of exogenously specified consumption and investment expenditures, interest payments on outstanding official external debt, transfers to the private sector, and endogenously determined interest payments on debt contracted after the base year. Government revenue is comprised of endogenously determined direct and indirect tax revenue, transfers from abroad, and the share of total value-added in the economy that is attributed to the government. By deducting from the fiscal deficit an exogenously specified proportion of domestic nonbank financing, the flow of domestic credit extended to the government by the banking system, and known external financing (scheduled disbursements less scheduled amortization payments), an external financing gap is computed. The external debt service module which accepts detailed assumptions on loan types, maturity and grace periods, and interest rate paths, then computes the service required on this financing gap and updates government expenditure for interest payments on the newly acquired debt.

The balance on the external current account is computed as the resource balance (i.e., the difference between exports and imports of goods and non-factor services), plus net factor income and transfers to the private sector from abroad and transfers to the government from abroad, less interest payments on official external debt. 7/ The capital account of the balance of payments consists of a set of exogenous entries for the public and private sectors, an endogenously determined amount of external financing of the government deficit, and private capital flows in response to disequilibrium in the money market. The reserve movement is simply the sum of the current and capital accounts. The reserve movement will therefore reflect developments in the current account, the extent of external financing of the government deficit, and the extent of disequilibrium in the market for domestic money.

The demand for real domestic money balances is a function of real GDP and the rate of inflation. It is posited that there is imperfect capital mobility. If, for example, money supply exceeds money demand there will be a gradual private capital outflow, as domestic residents attempt to reduce excess holdings of domestic money balances in order to balance their portfolios.

2. Detailed description

This section provides a detailed description of the model equations. A complete list of the definitions of the variables is contained in Appendix II.

The real sector: output and expenditure

The material balance identity relates real output of the home good, Xt, to the real demand for it, where real demand consists of real private consumption, Ct; real private investment, It; real government consumption and investment; Gtc and GtI; export volume, Etv; and the real change in stocks, ΔSTKSt:

Xt=Ct+It+Gtc+GtI+Etv+ΔSTKSt(1)

Real private consumption is determined as a function of a permanent or trend component of real disposable income, YDRPt, and a transitory or deviation from trend component, YRDTt:

Ct=C¯+k1YDRPt+k2YDRTt(2)

Real private investment (in log terms) depends on (the log of) one-and two-period lags of real private investment and (the log of) real flows of credit to the private sector from the banking system, ΔDCRt/Pt:

Ln(It)=I¯+k3Ln(It2)+k4Ln(It3)+k5Ln(ΔDCRPt/Pt)(3)

Export volume (in log terms) is determined as a function of (the log of) real gross domestic product, GDPt/PGDPt, and the (log of) price of exports in domestic currency terms, etPt*E, determined as the product of the exchange rate, et, and the foreign price of exports, Pt*E, relative to the price of the home good, Pt.

Ln(Etv)=E¯+k6Ln(GDPt/PGDPt)+k7Ln(etPt*E/Pt)(4)

Import volume, Mtv, is determined by multiplying the contribution of imports to a unit of output, mt, by the level of output. The contribution of (the log of) imports to output depends on the (log of) price of imports in domestic currency terms, which is determined as the product of the exchange rate, the foreign currency price of imports, P*M, and one plus the import duty rate, tm, relative to the price of the home good:

Ln(mt)=M¯+k8Ln(etPt*M(1+tM)/Pt)(5)

Gross domestic product at market prices is equal, by definition, to the nominal value of gross output or the nominal value of the sum of the demand components less the value of imports:

GDPt=Pt(Ct+It+Gtc+GtI+ΔSTKSt)+etPt*EEtve.Pt*MMtv(6)

To obtain gross domestic product at factor cost, GDPFCt, excise taxes net of subsidies, calculated as the excise tax rate, b, multiplied by nominal gross output, and import duties, are deducted from gross domestic product at market prices:

GDPFCt=GDPtbPtXttMetPt*MMtv(7)

Gross private sector income is largely determined by the proportion of real value added attributable to domestic factors of production, vt, which is jointly determined with the import coefficient, and the exogenously specified price of value-added, Ptv. Gross private sector income is equal to the nominal value-added accruing to the domestic economy, vt.Ptv.Xt, less that part of value-added accruing to government-owned factors, n.Ptv.vt.Xt, plus any difference in the domestic currency value of exports and exports valued at the home good price, (etPt*EPt)Etv, plus the domestic currency value of net factor income and foreign transfers to the private sector, et.Rt*, and government domestic transfers to the private sector, Rt:

Yt=(1n)PtvvtXt+(etPt*EPt)Etv+etRt*+Rt(8)

Disposable income, YDt, is generated by deducting direct taxes, d.Yt, from gross private sector income, where d is the direct income tax coefficient.

YDt=(1d)Yt(9)

Real disposable income, YDRt, is equal to disposable income deflated by the price of the home good:

YDRt=YDtPt(10)

The permanent component of real disposable income is obtained by regressing real disposable income on a constant, YDRP, and a time trend, t. The residual is transitory real disposable income:

YDRPt=YDRP¯+k9t(11)
YDRTt=YDRYDRPt(12)

Capacity output, Xt, (in log terms) is a function of a time trend and (the log of) real investment:

Ln(X¯t)=X¯+k10t+k11Ln(It+GtI)(13)

The real sector: prices

The price of the home good, Pt, depends on cost factors and demand relative to capacity output. Cost components include the cost of value added and the cost of imported goods, each of which depends on the relative shares of domestic and imported factors in the production process which, in turn, depend on relative factor prices. 8/ To the extent that demand for output exceeds capacity output, the price of the home goods will reflect an additional excess profit factor. Once the price of the home good is determined, nominal GDP can be calculated and the GDP deflator may be computed as the ratio of nominal to real GDP.

More specifically, the price of the home good, Pt is determined as a weighted average of the cost of the domestic factor of production, Ptv, and the domestic currency cost of imports, adjusted for excise taxes, times a retained profit markup factor, wt. The weights represent the shares of output produced by the domestic and imported factors, respectively.

Pt=1(1b)[vtPtv+mt(1+tM)etPt*M](1+wt)(14)

The retained profit factor is related to the gap between actual and capacity output, where the size of the markup is equal to a proportion, st, of the percentage deviation of output above capacity in the previous period. For simplicity, it is assumed that the additional revenues generated by the sale of output at higher prices are retained by the producing firms:

wt=max(o,st[Xt1X¯t1X¯t1])(15)

Since the factor shares of output adjusted for taxes must sum to one, this relationship may be used to solve for the share of the domestic factor of production in total output:

1=[vt+mt(1+tM)]1(1b)(16)

Finally, the GDP deflator, PGDPt, is defined as a weighted average of the price of the home good, and the domestic currency price of exports and imports:

PGDPt=GDPt[Ct+It+Gt+Gt+ΔSTKSt+EtvMtv](17)

The government sector

The government sector is defined as general government, including central and state and local governments and extrabudgetary funds. Total government revenue GOVREVt is equal to the sum of: the share of value added by the domestic economy contributed by the government, n, multiplied by total nominal value added by the domestic economy; excise tax revenue, revenue from import tariffs; direct income tax revenue, and transfers to the government from abroad.

GOVREVt=nPtvvtXt+bPtXt+tMetPt*MMtv+dYt+etTt(18)

Government expenditure, GOVEXPt consists of the sum of consumption and investment expenditures, transfers to the private sector and interest payments to the rest of the world, which are divided into an exogenous component, interest payments in foreign currency terms on debt outstanding in the base year, D0t, and two endogenous components, both in foreign currency terms—interest payments on external debt contracted by the government in the current year, D1t, and interest payments on external debt contracted by the government since the base year but excluding the current year, D2t:

GOVEXPt=Pt(Gtc+GtI)+Rt+et(D0t+D1t+D2t)(19)

In order to satisfy the government budget constraint it is necessary to first compute the gap between expenditures excluding interest payments due on the inflows required to close the gap and revenues and other net financing flows, which include domestic bank financing, ΔDCRGt, domestic nonbank financing, DOMNONBANKt, scheduled disbursements of foreign loans, OFFINFLt, and amortization payments on previously incurred debt—those resulting from debt incurred up to and including the base year, AMORTt, and those resulting from debt incurred after the base year but prior to the current year, ENCAPt. The net financing gap is therefore defined as:

NETFINGAPt=[Pt(Gtc+GtI)+Rt+et(D0t+D2t)nPtvvtXt(20)bPtXttMetPt*M.MtvdYtetTt]ΔDCRGtDOMNONBANKtet(OFFINFLt+AMORTt)+etENCAPt1/

Debt module

Three loan categories are permitted in the debt module for foreign financing of the government deficit, GBtL, where L = I, II, III. Each category is defined by the proportion of total borrowing that takes place in that category, ZtL, the grace period and maturity of loans in that category, GRAL and MATL, respectively, and the time path of the interest rate in that category, itL. The assumption is made that disbursement and amortization payments for each loan occur entirely in the middle of the year and interest payments are assumed to be made on the last day of each six-month period.

The gross borrowing in each category is determined by the net financing gap, the proportion of total borrowing in that category, and the interest rate:

GBtL=NETFINGAPtZtL+GBtL(112itL)(21)

The total gross official borrowing required to close the net financing gap is:

GBt=ΣL=IIIIGBtL(22)

Total interest payments on official borrowing in the current year, D1t, are determined by:

D1t=1etΣL=IIII12itLGBtL(23)

Amortization payments in each loan category for each year, s, beyond the grace period and up to maturity, ENAMORTSL, t, are assumed to be spread evenly over the life of the loan:

ENAMORTSL, t=GBtLMATLGRAL(24)for(st)GRALand(st)MATL

Total endogenous amortization payments in year t, ENCAPt, are calculated by summing over each loan category the amortization payments due for each loan contracted in a prior year and then summing over the loan categories:

ENCAPt=1etΣL=IIIIΣn=1t=1ENAMORTtL, n(25)

Interest payments on loans contracted prior to the current year, D2tL, are calculated for each type of loan according to equation (23), which effectively breaks down the calculation of interest payments into six-monthly intervals. For example, since amortization payments are assumed to be made in the middle of the year and interest payments are made at the end of each six month period, the interest due in the third year on a loan of 100 received in the second year, with a maturity of five years, no grace period, and an interest rate of 5 percent would be 9, the sum of 5 in the first half of the third year and 4 in the second half:

D2L=Σs=2t1[12iL(GBsLΣn=1t1ENAMORTnL, s)(26)+12itL(GBsLΣn=1t1ENAMORTnL, s)]

Total interest payments in year t (>2) on loans incurred prior to t are calculated by summing over the three loan categories:

D2t=ΣL=1IIID2tL(27)

External sector

The current account balance, CABt, is defined as exports less imports plus foreign transfers and net factor income to the private sector and official transfers less interest payments on official external debt:

CABt=Pt*EEtvPt*MMtv+Rt*+Tt(D0t+D1t+D2t)(28)

The capital account balance, CAPt, is defined as the sum of exogenously specified scheduled disbursements of official loans from abroad, OFFINFLt; scheduled amortization payments on these loans, AMORTt; other net inflows, OTHINFLt; and endogenous government borrowing (net), GBt/et + ENCAPt; and private capital flows responding to disequilibrium in the domestic money market, ENCAPPVTt:

CAPt=OFFINFLt+AMORTt+OTHINFLt+(GBt/et+ENCAPt)+ENCAPPVTt(29)

The overall external balance or the change in net official foreign assets, Ft - Ft-1, is equal to the sum of the current and capital accounts:

FtFt1=CABt+CAPt(30)

The stock of gross official reserves, GRFt, can then be computed as the sum of net official foreign assets, Ft, and official foreign liabilities, including liabilities to the Fund, UFOFt:

GRFt=Ft+UFOFt(31)

Monetary sector

The supply of money, Mts, is determined by the monetary survey accounting identity with the components of domestic credit, credit to government, DCRGt, credit to the private sector, DCRPt, and credit to state enterprises, DCSEEt, being determined exogenously, net foreign assets of commercial banks assumed to remain at the base year (B) level, and other items (net), OINt, determined as a constant proportion of the stock of money (equal to the proportion in the base year) adjusted for valuation changes in net foreign assets:

Mts=etFt+NFADMBt+DCRGt+DCRPt+DCSEEtOINt(32)
OINt=OINBMBsMts+(etet1)Ft1(33)NFADMBt=NFADMBB

The (log of the) demand for real money balances, MtD/Pt is specified to be a function of (the log of) real GDP and the rate of inflation; any discrepancy between the demand for and supply of money leads to a gradual outflow or inflow of capital ENCAPPVTt, through the capital account of the balance of payments, the size of the flow being determined by an exogenously set parameter, Λ:

Ln(MtDPt)=M¯D+k12Ln(GDPtPGDPt)+k13(PtPt1Pt1)(35)
ENCAPPVTt=[Mt1dMt1s]1et(36)

III. Estimates of Behavioral Equations for Turkey

This section presents regression estimates of the parameters of the consumption, investment, export, import, capacity output, and money demand functions for Turkey. The functional forms and parameter estimates of the behavioral relationships presented were chosen from among a number of estimated alternatives on the basis of their statistical properties. The flexibility of the model allows, however, for the relatively straightforward substitution of user-defined alternatives. The notation used for all variables is the same as in the previous section, except where specifically noted. The numbers in parentheses below coefficient estimates are their t-statistics. Also presented for each equation are the R2, the Durbin-Watson statistic, the significance level of the Q Statistic, and the sample period of estimation. 10/ All equations were estimated by ordinary least squares. All price indices have 1986 as a base and therefore all real series employed below are in 1986 prices.

1. Consumption

In equation (37) disposable income is decomposed into trend and transitory components by regressing disposable income on a constant and time trend. The fitted values are referred to as YDRPt, and represent the ‘trend’ or ‘permanent’ component in real disposable income, while the residuals are referred to as YDRTt, or the ‘transitory’ component. In equation 38 consumption is regressed on the permanent and transitory components of real disposable income:

YDRt=15088.63(31.90)+1039.98(19.99)t(37)R¯2=0.97D.W.=0.86Q%=0.004197286
Ct=1240.612(2.63)+0.892(45.00)YDRPt+0.753(6.85)YDRTt(38)R¯2=0.99D.W.=1.59Q%=0.58197286

2. Investment

Real private investment (in log form) is estimated as a function of lagged values of (the log of) real private investment and (the log of) the real value of the change in domestic credit extended to the private sector by the banking system. It is interesting to note that the coefficients on lagged values of investment alternate in sign in equation (39) which implies cyclical behavior in investment expenditures. 11/

Ln(It)=1.97(1.32)+1.46Ln(6.29)(It1)0.77Ln(3.79)(It2)(39)+0.076(1.10)Ln(ΔDCPtPt)R¯2=0.78Q%=0.32197486

3. Exports

In a small open economy, such as Turkey, the level of exports may plausibly be assumed to be supply determined. The (log of the) real value of exports of goods and nonfactor services is estimated as a function of the (log of the) relative price of exports and (the log of) real GDP. 12/ The numerator in the relative price term is the domestic currency price of exports. The variable employed was the implicit exports deflator. The wholesale price index is used as the proxy for the price of the home good.

Ln(Etv)=6.14(1.25)+2.18(3.38)Ln(etPt*EPt)+1.414(3.01)Ln(GDPtPGDPt)(40)R¯2=0.81D.W.=1.17Q%=0.28197386

4. Imports

The volume of imports is measured here by the real value of imports of goods and nonfactor services. The (log of the) factor share of imports, mt, in equation (41) is estimated as a function of the (log of the) relative price of imports and a dummy variable. The domestic currency price of imports is measured by the implicit imports deflator. 13/ In Turkey the level of imports has been severely constrained by the lack of sufficient foreign exchange reserves. A dummy variable, set equal to unity for 1978, 79, 80 and zero otherwise is added to represent this constraint.

Ln(mt)=1.80(70.6)0.316(5.69)Ln(PtMPt)0.274(5.2)DUMt(41)R¯2=0.79D.W.=1.22Q%=0.45197286

5. Capacity output

Output is defined as the sum of real GDP and the volume of imports. Since capacity output is in general a function of the capital stock of the economy, which is in turn determined by the level of investment, capacity output is estimated by regressing (the log of) output on real investment and a time trend in equation (42).

Ln(Xt)=7.074(9.71)+0.026(5.98)t+0.358(4.19)Ln(It+GtI)(42)R¯2=0.93D.W.=1.4Q%=0.13197286

6. Money demand

Real money demand (in log form) is estimated as a function of the rate of inflation and (the log of) real GDP in equation (43). 14/ The deflator for real money balances, and hence the relevant price index for measuring the rate of inflation is defined as the wholesale price index.

Ln(MtsPt)=2.755(2.28)0.368(4.33)(PtPt1Pt1)+1.15(9.79)Ln(GDPtPGDPt)(43)R¯2=0.89D.W.=1.10Q%=0.83197386

IV. Properties of the Model

While the structure of the model is fairly simple its properties can only be determined by carrying out simulation exercises. In this section the model properties are examined by running simulations for discrete changes in some of the exogenous and policy variables. Both impact and five year effects on key macroeconomic aggregates are reported in Table 1 while graphs showing the paths over five years of changes in these aggregates are presented in Appendix I. The simulation results are central to understanding the interrelations among the macroecomic variables of interest. The exercises reported in this section examine the effects on four key macroeconomic aggregates of changes in three exogenous variables (world prices of exports and imports and the domestic price of value added) and seven policy variables (the exchange rate, excise and direct tax rates, the import duty rate, the flow of credit to the private sector, and real government consumption). The model linkages that give rise to the reported results are also discussed.

Table 1.

Summary of Effects of Changes in Exogenous and Policy Variables

(In percent)

article image

GDP deflator.

In U.S. dollars. Negative figures represent a reduction in deficit.

Negative figures represent a reduction in deficit.

Increases are one percent above baseline values except as noted.

Rates are one percentage point above baseline values.

Flow increased by 10 percent above baseline values.

1. Price of exports

For an increase in the price of exports of one percent in each year above the baseline, Table 1 indicates that real GDP and the GDP deflator will rise above the baseline values in both the short-and medium-term. The effect on GDP is 1.3 percent at impact and 1.6 percent after 5 years, while the GDP deflator is higher by 0.2 percent in the initial period, and by 1.2 percent after 5 years. The current account balance improves from a deficit equivalent to one percent of GDP in the initial year being eliminated (Table 7 and Chart 1) and a deficit equivalent to 6 percent of GDP in the fifth year being reduced by 25 percent. The public sector deficit is reduced by almost 8 percent in the initial year and 4 percent after 5 years.

The increase in the relative price faced by exporters stimulates the volume of exports. The real return per unit of exports, moreover, rises, expanding private sector incomes and raising consumption demand and output. Since output is produced jointly by domestic and imported factors of production, imports also rise with the change in the real resource balance being determined by the relevant elasticities.

Since output rises, government revenue from excise taxes increases as does revenue from direct taxes, reflecting the higher level of nominal private incomes. In addition, revenue from import duties increases owing to the higher volume of imports. On the government expenditure side, since the price of the home good is unaffected, as neither the price of value added nor the price of the imported factor change, nominal government expenditures remain unchanged, resulting in a smaller public sector deficit and a reduced external financing gap, which, in turn, lowers interest costs associated with new borrowing.

While the external current account improves due to the reduction in the resource gap as well as lower interest payments, capital inflows fall owing to the improvement in the fiscal accounts. On a net basis, official reserve loss is reduced from the baseline case, raising the stock of money; money demand, however, increases by a larger amount, reflecting the improvement in real GDP. Money market disequilibrium leads to a reduction in capital outflows in the following period.

2. Price of imports

A one percent increase in the price of imports reduces real GDP, raises the GDP deflator marginally and increases the external current account and public sector deficits. Real GDP declines by 0.4 percent in the initial year and by 1.1 percent in the fifth year. The current account deficit deteriorates by about $100 million in the first year and by $500 million by the fifth year (Chart 2) while the public sector deficit grows by TL40 million in the initial year and by TL 400 million after 5 years.

Higher import prices lead producers to substitute the domestic factor of production for the foreign, reducing the volume of imports; per unit of output, the share of the domestic factor of production will rise. This, however, results in a positive income effect on output as private sector incomes and thus nominal aggregate demand rise. The expansion in output will tend to raise imports. On the other hand, an increase in the price of imports will raise the price of the home good (although the substitution effect will mitigate this effect), lower real disposable income and reduce real aggregate demand, GDP, and exports. The net result of these effects is a reduction in the volume of imports but a rise in their foreign currency value.

3. Price of value-added

An increase of one percent in the price of value-added reduces real GDP, raises the GDP deflator, and causes a deterioration in the external current account and the public sector’s overall budgetary position. Real GDP falls by almost one percent in the initial year and by 2.5 percent after 5 years while the GDP deflator rises by 0.8 percent initially and by 5 percent after 5 years. The external current account deficit deteriorates by $175 million initially and by almost $800 million after 5 years (Chart 3) while the fiscal position worsens by TL100 billion initially and by TL800 billion after 5 years.

Producers substitute the imported factor of production for the domestic, resulting in a reduction of the domestic factor share per unit of output. Whether or not nominal private sector income declines or rises depends on the proportional decline in the real share of the domestic factor compared to the proportional rise in the price of value-added. Reflecting the higher price of value-added, the price of the home good rises, reducing real income, GDP and exports. Despite the shift to the imported factor per unit of output, import volume falls, but by less than the drop in exports, resulting in a deterioration in the trade and current account balances. The fiscal deficit worsens as tax revenues fall along with output and income and nominal government expenditures rise reflecting the higher price of the home good.

4. Exchange rate

An increase of one percent per year in the Turkish lira price of foreign currency over the period results in a rise in real GDP and the GDP deflator, and an improvement in the current account and overall public sector balances. Real GDP rises by 1 percent in the initial year and 2 percent in the fifth year while the GDP deflator rises by 0.2 percent and 0.7 percent, respectively.

The devaluation stimulates the supply of exports, expanding output; moreover, the return to exporters at the baseline volume of exports rises in terms of domestic currency, raising nominal incomes. On the other hand, the domestic currency price of imports rises, increasing the price of the home good and reducing real incomes. On balance, real incomes rise over the period. As a consequence of the increase in exports and despite a rise in imports, the external current account deficit is reduced by almost 70 percent in the initial year and by about 20 percent in the fifth year.

Both tax revenues and expenditures in the fiscal accounts rise as a result of the devaluation. Direct taxes increase, reflecting the rise in nominal incomes while excise taxes and import duties increase as a result of the rise in output and imports, respectively. On the expenditure side, since real government consumption and investment are exogenously determined, their nominal value rises as the price of the home good increases; the domestic currency equivalent of foreign interest payments on previously incurred debt also rises. The net result is a 4 percent reduction in the fiscal deficit in the initial year, declining to 2.5 percent at the end of the period.

5. Excise tax

An increase in the excise tax rate of one percentage point reduces real GDP, leaves the GDP deflator virtually unchanged, raises the external current account deficit in the initial year but lowers it in subsequent years, and reduces the fiscal deficit. Real GDP declines by about 2 percent in both the initial and fifth years. The current account deficit grows by about $30 million in the first year but is reduced by about $200 million in the fifth year (Chart 5) while the fiscal deficit is reduced by TL300 billion and TL800 billion in the first and fifth years, respectively. Since the share of output going to doemstic factors of production is reduced by a rise in the excise tax, real private incomes fall, in this case by 3 percent, resulting in a decline in real consumption, real GDP, and real output. Reflecting the decline in real output, both import and export volumes fall, with the latter drop exceeding the former in the initial year but the opposite holding in subsequent years. The magnitude of the reduction in foreign interest payments stemming from a lower fiscal deficit fails to fully offset the deterioration in the balance of exports and imports of goods and non-factor services in the initial year. The improvement in the fiscal deficit results from higher excise tax revenues (16 percent in the initial year), which offset lower revenues from other sources due to the decline in output, and a reduction in borrowing and related interest payments in subsequent years.

6. Direct tax

An increase in the income tax rate of one percentage point reduces real GDP and the fiscal deficit, leaves prices virtually unchanged, and increases the external current account deficit in the initial year while reducing it in subsequent years. Real GDP declines by 1.8 percent initially and by 1.6 percent after 5 years and the current account deficit after rising by about 10 percent in the initial year declines by 4 percent after 5 years.

As real disposable income of the private sector declines as a consequence of the higher income tax rate, consumption demand falls, reducing real output and exports; the demand for imports also falls but by less than the reduction in exports in the initial year, resulting in a deterioration in the resource balance and the current account deficit. Subsequently, import volumes decline by more than export volumes.

The fiscal deficit falls substantially, by 1 percent of GDP in the initial year as higher direct tax revenues more than compensate for a reduction in revenues from excise taxes and import duties. While the direct tax revenue base, gross private sector income, declines, the higher direct tax rate results in an increase in revenues from this source.

7. Import duty

An increase in the import duty of one percentage point reduces real GDP marginally, leaves the GDP deflator unchanged, and improves both the external current account and the fiscal balance. Real GDP falls by about 0.1 percent in both the initial and fifth years, the current account improves by about 11 percent initially and 2 percent in the fifth year, and the fiscal deficit is reduced by about 3 percent and 2 percent in the first and fifth years, respectively.

The rise in the import duty rate reduces the share of output attributable to domestic factors, as was also the case for the increase in the excise tax rate. The reduction in the share of output going to the domestic factor results in a decline in real disposable income, real consumption and real GDP. Import and export volumes both fall, with the drop in imports dominating and, as a result, the external current account balance is improved. A decline in foreign interest payments stemming from the improvement in the fiscal position also helps to reduce the current account balance. The overall fiscal deficit is reduced as lower direct and indirect tax collections are more than offset by higher revenues from import duties.

8. Flow of credit to the private sector

An increase of ten percent in the flow of credit to the private sector raises real GDP, leaves the GDP deflator virtually unchanged, and improves the external current account and fiscal balances. Real GDP rises by 0.2 percent initially and by 2 percent in the fifth year, the current account deficit is reduced by 1.7 percent and 0.1 percent in the first and fifth years, respectively, and the fiscal deficit is lowered by 0.8 percent and 5.6 percent in the initial and fifth years, respectively.

Real private investment, a positive function of real domestic credit, expands, stimulating real output and GDP. Export volume responds to the increase in GDP as does import volume, with the former effect slightly outweighing the latter. Government revenues rise along with output, reducing the fiscal deficit. As GDP rises, the demand for money increases but lags the growth of the money stock resulting from the rise in credit. Hence, private capital outflows rise in the years following the initial change in the flow of credit, from a 5 percent increase in the second year to 27 percent by the fifth year.

9. Government consumption

A one percent increase in real government consumption raises real GDP, leaves the GDP deflator unaffected, improves the external current account in the initial year but causes a deterioration thereafter, and adds to the fiscal deficit. Real GDP increases initially by only 0.2 percent but rises by 1.2 percent after 5 years. The current account improves by about 2 percent initially but deteriorates after the fifth year by 1.2 percent, and the fiscal balance deteriorates by TL15 billion and TL300 billion in the first and fifth years, respectively (Chart 9).

As output expands the supply of exports increases and the demand for imports required for the production of additional output rises. As revenues from all sources are higher, the overall fiscal position deteriorates by about one-half of the rise in government consumption.

V. Conclusions

The objective of this paper was to present a small macroeconomic model which could be used in the design of growth-oriented adjustment programs. While the model was applied to a particular country, Turkey, its structure was designed to be sufficiently flexible to allow it to be modified for a wide range of developing countries. The model is able to evaluate the short- and longer-run impacts of a variety of economic policy measures on major macroeconomic targets and a distinguishing feature of the model is the explicit linkage between exogenous and policy variables and macroeconomic targets. The model encompasses a broad spectrum of variables of interest to the policy-maker which can be changed by the user to represent alternative policy packages. External shocks may also be easily evaluated. Fiscal policies made explicit in the model include policies with respect to direct and indirect tax rates and import duties and policies with respect to government consumption and investment. Credit policies, exchange rate policies and wage policies are also made explicit. The external debt module allows the public sector to incur alternative types of debt with respect to interest rates, grace periods and maturities. Moreover, the model was written in a programming language which allows not only exogenous and policy variables to be easily changed but behavioral relations as well, a feature that is often missing from financial programming-type models.

While the model contains many desirable features, such as internal consistency, variable factor contributions to output, demand and supply effects on domestic prices and dynamic money market adjustment, it clearly could be improved, particularly in the areas of interest rate effects, government domestic non-bank financing relationships, and wealth effects. Lag structures are for the most part omitted and expectations are not taken into account. Of course, the model is subject to the Lucas critique that the behavior of economic agents reflects the prevailing pattern of policies as well as expectations of the future path of policies so that changes in policies may alter the structure of relationships among variables in a way that cannot be predicted from historical relationships. 15/ The importance of these and other possible shortcomings will, of course, depend on the use to which the model is put; nevertheless, from an operational perspective its flexibility with respect to both economic and programming structure should enhance its potential to provide a framework for growth-oriented adjustment programs.

APPENDIX I

CHART 1
CHART 1

TURKEY: HIGHER EXPORT PRICES1

Citation: IMF Working Papers 1989, 024; 10.5089/9781451920901.001.A001

1 One percent above baseline.
CHART 2
CHART 2

TURKEY: HIGHER IMPORT PRICES1

Citation: IMF Working Papers 1989, 024; 10.5089/9781451920901.001.A001

1 One percent above baseline.
CHART 3
CHART 3

TURKEY: HIGHER PRICE OF VALUE ADDED1

Citation: IMF Working Papers 1989, 024; 10.5089/9781451920901.001.A001

1 One percent above baseline.
CHART 4
CHART 4

TURKEY: DEVALUATION1

Citation: IMF Working Papers 1989, 024; 10.5089/9781451920901.001.A001

1 One percent above baseline.
CHART 5
CHART 5

TURKEY: HIGHER EXCISE TAX RATE1

Citation: IMF Working Papers 1989, 024; 10.5089/9781451920901.001.A001

1 Baseline plus One percent.
CHART 6
CHART 6

TURKEY: HIGHER IMPORT DUTY RATE1

Citation: IMF Working Papers 1989, 024; 10.5089/9781451920901.001.A001

1 Baseline plus One percent.
CHART 7
CHART 7

TURKEY: HIGHER DIRECT TAX RATE1

Citation: IMF Working Papers 1989, 024; 10.5089/9781451920901.001.A001

1 Baseline plus One percent.
CHART 8
CHART 8

TURKEY: INCREASED FLOW OF PRIVATE SECTOR CREDIT1

Citation: IMF Working Papers 1989, 024; 10.5089/9781451920901.001.A001

1 10 percent above baseline.
CHART 9
CHART 9

TURKEY: HIGHER GOVERNMENT CONSUMPTION1

Citation: IMF Working Papers 1989, 024; 10.5089/9781451920901.001.A001

1 One percent above baseline.

APPENDIX II: User’s Guide to the Program

This appendix describes in some detail the layout of the computer program and contains the standard tables for the baseline scenario.

1. Loading program into Lotus

Insert the diskette on which the program is written into drive B. In Lotus, get into command mode by pressing the SLASH key. Choose FILE, then RETRIEVE. Now press ESCAPE twice to get out of the default directories and type B:MODEL and press the carriage return key. The file will take a few seconds to load.

2. Layout of spreadsheet

The spreadsheet consists of a set of tables. They are listed in Table 2 in the order in which they appear in the spreadsheet and they can each be examined by scrolling through the worksheet. The first table is comprised of the parameters estimated for the behavioral equations. Each parameter value is accompanied by an explanatory comment. The next table consists of lagged values of variables needed by the program. For example, since investment depends on lagged values of investment, it is necessary for the program to know the lagged values of investment. This table is followed by tables for exogenous variables and policy variables. These two tables provide the essential exogenous user inputs used to generate forecasts. For example, the user may project a different path for the price of exports than the one in the baseline scenario. Or, the user may wish to make alternative assumptions regarding the behavior of government expenditures or average tax rates, etc. The exogenous and policy variables tables are followed by a table of equations which, though not essential, may help the user better understand how the model and program work. Each equation is accompanied by an explanatory comment. It should be noted that the table of equations includes “dummy” or additional variables used to store intermediate values, for switches to control program execution, etc. The table of equations is followed by a convergence indicator table, explained below, and the debt module. Next are the output tables. These include the national income accounts at current and constant prices, balance of payments in foreign currency, public sector finances, and monetary flows.

Table 2.

Layout of Spreadsheet

article image

3. Changing exogenous and/or policy variables

To illustrate how to change a policy variable, suppose that instead of the 5 percent devaluation of the nominal exchange rate in year one of the baseline scenario we wished to have a 15 percent devaluation. By scrolling, locate the row for the exchange rate in the policy variables table. Move the cursor to year one (second column). Note that the value expected by the program is an actual number for the exchange rate. One can now either (i) directly type in the number 1.15 since the exchange rate in the base year is 1, or (ii) put in a simple formula, expressing a 15 percent increase in the nominal exchange rate over the previous year.

4. Solution of model

Notice that the equations of the model are simultaneous and the cells containing the equations reference each other. The model is solved by iterating, using a Gauss-Seidel convergence criterion. The F9 function key is the Lotus recalculate key. Pressing the F9 key once causes each cell in the worksheet to be recalculated a number of times. The number of iterations (recalculations) can be set to be between 1 and 50, and in the program has been set at 35. In general, pressing the F9 key once or twice will be sufficient for convergence. The easiest way to check whether or not the model has converged is to scroll to the convergence indicator table, then press the F9 key once and examine the table. The values in the first row labelled “S-I” should be “small” and not change between iterations. All other entries in the table should be approximately (i.e., at least to the second decimal place) zero.

5. Data normalization

All prices in the model, including the nominal exchange rate, are expected to be set to 1 for the base year. All real variables are therefore, in base year prices. For foreign currency denominated inputs this implies that they have to be entered in base year foreign currency prices. For example, suppose we know that amortization payments on outstanding official external debt were US$1742.99 million for the base year and US$2145.45 million for the following year. The exchange rate for the base year was 674.51 Turkish lira per U.S. dollar. The numbers to be entered in the program therefore are 1175.67 = (1742.99).(674.51)/1000 for the base year and 1446.15 = (2145.45) (674.51)/1000 for year one. The division by a thousand is simply to reflect the fact that the balance of payments entries in Turkey in foreign currency are measured in millions of U.S. dollars and in domestic currency in billions of Turkish lira.

6. The baseline scenario

Exogenous variables and parameters are shown in Table 3, the paths of policy variables are set out in Table 4, and the outcome of the baseline scenario is contained in Tables 5 to 9.

7. Definitions of variables and parameters

a. Exogenous variables and parameters
article image
Table 3.

Exogenous Variables and Parameters

article image

World Economic Outlook, Based on projections of developments in partner countries.

Assumption. Approximately 50 percent of historical growth rate from 1972-86.

Average value from 1972-86.

IMF, Treasurer’s Department.

World Bank, World Debt Tables.

LIBOR projected by World Economic Outlook.

Table 4.

Path of Policy Variables

article image

Assumed.

Historical annual average growth rate for the period 1972-86.

Average tax rates in 1986.

Table 5.

National Accounts at Current Prices

(In billions of Turkish liras)

article image
Table 6.

National Accounts at Constant Base Year Prices

(In billions of Turkish liras)

article image
Table 7.

Balance of Payments

(In billions of U.S. dollars) 1/

article image

Based on a base year exchange rate of LT1 = US$1; actual US$ values may be obtained by dividing entries by the actual base year LT/US$ exchange rate.

Table 8.

Public Sector Finances

(In billions of Turkish liras)

article image
Table 9.

Monetary Accounts

(In billions of Turkish liras)

article image
article image
b. Policy variables
article image
c. Endogenous variables
article image
article image
*

The authors are grateful to Jeffrey Davis, Ichiro Otani, John Clark and Duncan Ripley for helpful comments and to Leilei Aung and Maxy Pinto for excellent assistance.

1/

See Guitian, Manuel, “Fund Conditionality Evolution of Principles and Practices, “International Monetary Fund, Pamphlet Series, No. 39 (Washington, D.C., 1981), p. 25.

2/

Khan, Mohsin S., Peter Montiel, and Nadeem U. Haque, “Adjustment with Growth: Relating the Analytical Approaches of the World Bank and the IMF,” World Bank, Development Policy Issues Series Discussion Paper (October 1986).

3/

Khan, Mohsin and Peter J. Montiel, “Growth-Oriented Adjustment Programs: A Conceptual Framework”, IMF, unpublished (July 1988).

4/

For an example of such a model, see Wong, C.H., and 0. Petterson, “Financial Programming in the Framework of Optimal Control,” Weltwirtschlaftliches Archiv (Kiel), Vol. 115 (1979), pp. 20-37.

5/

These receipts represented 4.4 percent of GDP at factor cost in Turkey in 1986. The activities that generate this income include placing deposits with the domestic banking system which earn interest, providing services for which user fees are charged, and the production of other goods and services.

6/

Since there is no distinction in production between exports and goods sold on the home market, an adjustment needs to be made to incomes of domestic producers for differences in the domestic currency price received by producers from selling abroad and at home. This difference is denoted here as “export profits” and equals the difference (in domestic currency) between the price of exports and the price of the home good times the volume of exports.

7/

Net factor income includes interest payments on private external debt although there is no explicit link in the projection years between such payments and the amount of private external capital inflows.

8/

See equation (5) for the determination of the share of imports in output, and equation (16) for the determination of the share of the domestic factor.

9/

AMORTt enters the equation with a positive sign since it is defined as a negative number.

10/

The Q statistic provides a test statistic for the presence of autocorrelation in the error term. The null hypothesis that the error terms are independent is tested. A ‘low’ significance level of the Q statistic implies one may reject the null hypothesis that the estimated error terms are generated by a white noise process.

11/

The coefficient of a real interest rate variable in the investment function was either not significant or had the wrong sign. An after tax real lending rate might, however, produce the expected sign of the coefficient.

12/

An alternative formulation of the export function might be to substitute real output, Xt, for real GDP.

13/

This import price index is adjusted to maintain compatibility with the computer program so that 1986 = 1.15, to reflect the average 15 percent import duty in 1986.

14/

Interest rates were tried as an explanatory variable without success.

15/

Lucas, R.E. Jr., “Econometric Policy Evaluation: A Critique” in Karl Brunner and Allan Meltzer, eds., The Phillips Curve and Labor Markets, Amsterdam: North Holland, 1976.

A Simulation Model for Financial Programming
Author: International Monetary Fund