Chapter 7 WORKSHOP 5 The Polak Model: An Application

International Monetary Fund

Jump to Content Jump to Main Navigation

Author: International Monetary Fund

Language:: English

Keywords:: BOOK; exogenous variable; income velocity of money; definitional equation; endogenous variable; regression estimate; Domestic credit; Demand for money; Personal income; Foreign assets

Download PDF (1.1 MB)

Abstract

Since the early 1950s, economists and policymakers have been increasingly preoccupied with the problems of inflation and balance of payments disequilibria. Their preoccupation has led to new approaches to monetary analysis.1 In this period, there has been a gradual evolution of the so-called monetary approach to the balance of payments, the third major approach; the two best-known earlier approaches are the elasticity (neoclassical) approach and the income absorption (neo-Keynesian) approach.2

Since the early 1950s, economists and policymakers have been increasingly preoccupied with the problems of inflation and balance of payments disequilibria. Their preoccupation has led to new approaches to monetary analysis.¹ In this period, there has been a gradual evolution of the so-called monetary approach to the balance of payments, the third major approach; the two best-known earlier approaches are the elasticity (neoclassical) approach and the income absorption (neo-Keynesian) approach.²

It has often been pointed out that each of the three approaches could in principle produce the right answers if it were correctly applied, that is to say, if proper allowance were made for all the repercussions throughout the economy of the change whose effect is being analyzed. It has, however, proved difficult in practice to set forth a suitable framework for use with either the elasticities approach or the income-absorption approach within which the requisite information would be marshaled in a comprehensive and consistent manner. For applied research and background work for policy discussion on balance of payments problems, the monetary approach suggested itself, therefore, as apparently simpler and more manageable than the other approaches. It is based on the postulates of a stable demand function for money and of a stable process through which the money supply is being generated. The demand for money, it is argued, depends on a relatively small number of economic factors, and the effects of economic changes on the demand for money are therefore easy to assess because they can operate only through one or several of these few factors. A similar argument can be made with respect to the determination of the supply of money. By focusing directly on the relevant monetary aggregates, this approach eliminates the intractable problems associated with the estimation of numerous elasticities of international transactions and of the parameters describing their interdependence, which are inherent in other approaches.

The apparent simplicity of the monetary approach to the balance of payments is, however, somewhat deceptive. Even though for many purposes the demand for money can be conveniently expressed as a function of a small number of variables, it is still just as much the resultant of all the influences that come to bear on the economy as are national income and national expenditure. Again, domestic credit creation, which is often taken as being determined exogenously, may in fact be systematically influenced by factors determining the demand for money or by some of the events whose monetary effects are being examined. These considerations do not invalidate the monetary approach; they merely draw attention to the possibility that it will be seen, on further examination, to be not quite so superior in terms of simplicity of application as had first been thought.³

An early exposition of the monetary approach to the balance of payments is contained in a simple, highly aggregative model formulated by J. J. Polak, former Economic Counsellor of the International Monetary Fund.⁴ A considerable volume of literature has since been written on the subject.⁵

The Polak model focuses on the relationship between credit expansion and change in foreign assets. In broad terms, the Polak model can be described as a system of four interdependent endogenous variables—namely, the stock of money (MO), nominal income (Y), imports (M), and change in net foreign assets (ΔNFA)—whose values change if there is a change in value of at least one of the following exogenous variables: exports (X), capital movements (CM), and change in net domestic credit (ΔNDC). Of the exogenous variables, ΔNDC is considered as a policy variable—that is, a variable that can be controlled by the authorities.

In this model, the interest rate plays no role, therefore making the model less complicated and more realistically applicable to developing countries where, because of scarcity of easily marketable financial assets, variations in interest rates do not play a significant role.

Also, because of its simplicity, the few data needed for the application of the model are readily available in most developing countries. Of particular significance is the fact that the Polak model has been estimated for a large number of developing countries with reasonably good results.⁶

In this workshop the parameters of the Polak model are estimated by using statistical data for Kenya. Once the estimates have been made, the model is used to show how to project imports as well as how to find the change in net domestic credit compatible with a preassigned target for the change in net foreign assets.

THE MODEL

The relationships of the Polak model are specified as follows:

\begin{matrix} M_{t} = m Y_{t}, 0 < m < 1 & (1) \end{matrix}

\begin{matrix} Y_{t} = \frac{1}{k} M O_{t}, 0 < k < 1 & (2) \end{matrix}

\begin{matrix} M O_{t} = Δ N F A_{t} + Δ N D C_{t} + M O_{t - 1} & (3) \end{matrix}

\begin{matrix} Δ N F A_{t} = X_{t} + C M_{t} - M_{t} & (4) \end{matrix}

The first relation is a behavioral equation, indicating that nominal imports in period t are a constant fraction m of the nominal income of the same period; the coefficient m is called the (average and marginal) propensity to import. Equation (2) implies a constant proportional relationship between nominal income and money stock;⁷ the coefficient 1/k is the (average and marginal) income velocity of money (Y_t/MO_t). This implies that the increase in income will be equal to the increase in the quantity of money times the income velocity of money. Relationships (3) and (4) are definitional equations (identities). The former states that money in this period (MO_t) is equal to money at the end of the previous period (MO_t-1) plus the change in net foreign assets (ΔNFA_t) and in net domestic credit of the banking system (ΔNDC_t) in this period. The latter defines the change in the net foreign assets of the banking system as being equal to the overall balance of payments surplus or deficit.

The definition of money in the model can be taken either in the broad sense (currency + demand deposits + quasi-money) or in the narrow sense (currency + demand deposits). Adoption of the broad definition implies that MO_t in equation (3) includes quasi-money and that the income velocity of money 1/k is interpreted accordingly.⁸

After combining the two definitional equations to yield the following result:

Δ M O_{t} = X_{t} + C M_{t} - M_{t} + Δ N D C_{t}

and after defining

A_{t} = X_{t} + C M_{t} + Δ N D C_{t}

(that is, the sum of the three exogenous variables, including the policy variable), the model may be written as:

\begin{matrix} M_{t} = m Y_{t}, 0 < m < 1 & {(1)}^{'} \end{matrix}

\begin{matrix} Y_{t} = \frac{1}{k} M O_{t}, 0 < k < 1 & {(2)}^{'} \end{matrix}

\begin{matrix} M O_{t} = A_{t} - M_{t} + M O_{t - 1} & {(3)}^{'} \end{matrix}

The effects on the endogenous variables M, Y, and MO of a change in one or more exogenous variables can be traced in the above equations. Given the constancy of the parameters m and k, a change, in A will affect M through changes in MO and Y. But then a change in M will inversely affect MO and Y, thus generating a second round of effects, this time in the opposite direction, and so on. The net effects can be computed by taking into account all rounds—that is, by obtaining the reduced form of the model (each endogenous variable expressed in terms of exogenous and lagged endogenous variables).

The model could be used for various purposes if there were numerical values (estimates) of the parameters m and k and if the level of the exogenous variables was known or could be estimated. For instance, imports might be projected in the future; the impact on any endogenous variable of a given increase in net domestic credit could be determined, as well as values for the impact, truncated, and overall (or equilibrium) multipliers. Estimates of the short-run and long-run elasticities of imports with respect to exports and other exogenous variables could be computed. Finally, a target regarding income or foreign assets could be set and the change in net domestic credit, as compatible with the preassigned target, could be determined. In short, the estimated model could be used for forecasting and policy purposes.

These possible uses of the Polak model should take into account its limitations. The reliability of the results of any macroeconomic forecasting exercise may vary, depending on (1) whether the specified relationships in the model depict correctly the actual behavior of the economy, (2) whether the statistical observations are free of measurement errors and correspond to the definition of variables involved in the model, and (3) whether the estimation techniques yield reliable estimates.

DEFINITIONS OF VARIABLES AND ESTIMATION OF PARAMETERS

The parameters of a model can be estimated on the basis of historical data (time series) regarding the variables involved in the model. With the aid of established techniques, these data can yield numerical values of the parameters. The choice of techniques depends on the specification of the model, the nature of the available data, and the purpose of the model. However, before choosing a particular technique and applying statistical inference, it is necessary to identify the concept-variables involved in the model with corresponding observations in the statistical world.

Statistical Definitions of Variables

For the purpose of this workshop, the following definitions are adopted:

1. Income (Y) is identified with gross domestic product at market prices.

2. Net foreign assets (NFA) represent net foreign assets of the banking system.

3. Imports (M) comprise payments for goods and nonfactor services.

4. Exports (X) comprise receipts from goods and nonfactor services less net factor payments abroad.

5. Capital movements (CM) cover all items of the balance of payments that are not classified as imports, exports, or reserve movements, so that CM = M − X + ΔNFA.

6. Money (MO) is defined as the sum of money plus quasi-money.

7. Net domestic credit (NDC) is defined as domestic credit extended by the banking system (DC) less other items net (NOI): NDC = DC− NOI.

Statistical Data on Variables

Statistical data corresponding to the above definitions of the variables are presented in Appendix I. These data consist of annual observations expressed in current prices, covering the period 1968–76. For simplicity's sake, data on stock variables (MO, NFA, and NDC) are given for the end of the year. To minimize the calculations involved in the estimation of parameters, Appendix II contains the computed ratios M/Y and MO/Y over the sample period.

Estimation of k and m

The parameters k and m, which are assumed to remain constant over time, can be estimated by calculating for each period the ratios of money/income (MO/Y) and imports/income (M/Y), respectively, and then taking the averages over the sample period (1968–76). This method is used in the present workshop.

Alternatively, these parameters may be estimated using regression analysis. Appendix III contains regression estimates of k and m.

USE OF MODEL FOR FORECASTING AND POLICYMAKING

By combining equations (1)′, (2)′, and (3)′, the reduced form for imports is obtained, as follows:

\begin{matrix} M_{t} = α A_{t} + α β M_{t - 1} & (5) \end{matrix}

where

α = \frac{m}{m + k}, β = \frac{k}{m}

This equation is a dynamic one, as the effects on imports of an increase in one of the components of A will be distributed over an infinite number of periods. This is easily seen by writing

\begin{matrix} M_{t - 1} = α A_{t - 1} + α β M_{t - 2} \\ M_{t - 2} = α A_{t - 2} + α β M_{t - 3} \end{matrix}

and by substituting successively into equation (5).

This then yields

\begin{matrix} M_{t} = α A_{t} + α^{2} β A_{t - 1} + α^{3} β^{2} A_{t - 2} + \dots & (6) \end{matrix}

It may be noted that equation (5), or its equivalent (6), is stable only if αβ is numerically less than unity. In other words, the effects on imports of a unit increase in one of the components of A will diminish from one period to the next. The total effect will be given by

α + α^{2} β + α^{3} β^{2} + α^{4} β^{3} + \dots

α [1 + α β + α^{2} β^{2} + α^{3} β^{2} + \dots]

If αβ is numerically less than unity, this total effect (namely, the long-run multiplier) is equal to

\frac{α}{1 - α β} = \frac{m / (m + k)}{1 - k / (m + k)} = 1

The same result can be obtained if time subscripts are suppressed in equation (5) and rearranged, as follows:

M = α A + α β M

\begin{matrix} M = \frac{α}{1 - α β} A = A & (since \frac{α}{1 - α β} = 1) & (7) \end{matrix}

and

Δ M = \frac{α}{1 - α β} Δ A = Δ A

Forecasting

After estimates of the parameters m and k, and, therefore, of α and αβ have been made, equation (5) or equation (6) can be used to forecast the value of imports in period t. However, from a practical point of view, it is more convenient to use equation (5). All that is needed is the level of A in period t and the value of imports in the preceding period. It should be noted that, once it is desired to project imports beyond the sample period, there arises the problem of knowing the levels of the exogenous variables. It is necessary to estimate or obtain information about their levels separately. Often, their past trend is extrapolated. The problem, of course, does not arise for those exogenous variables which are assumed to be policy (or control) variables. Therefore, equation (5) or equation (6) can provide forecasts of imports that are conditioned on the accuracy of the estimates for exogenous variables and parameters.

Monetary Policy

For analyzing monetary policy problems, equation (5) may be written in the following form:

\begin{matrix} M_{t} = α (X_{t} + C M_{t} + Δ N D C_{t}) + α β M_{t - 1} & (8) \end{matrix}

Equation (8) shows that the authorities, through credit policy, can influence directly the value of imports. For example, a policy of balance of payments equilibrium, equivalent to

\begin{matrix} M_{t} - X_{t} - C M_{t} = 0 & (9) \end{matrix}

will be achieved if credit is set at

\begin{matrix} Δ N D C_{t} = \frac{(1 - α)}{α} (X_{t} + C M_{t}) - β M_{t - 1} & (10) \end{matrix}

Equation (10) is obtained by substituting from equation (8) into equation (9).

Thus, a monetary policy leading to a balance of payments equilibrium consists, according to the Polak model, of maintaining a fixed relationship between the various elements of monetary expansion. This fixed relationship depends on the velocity of circulation of money and the marginal propensity to import.

EXERCISES

1. Estimates of Parameters and Computation of Multipliers and Elasticities
- (a) Using the data provided in Appendix II, make estimates of m and k, and obtain numerical values for α and β.
- (b) What is the impact (first period) multiplier of imports with respect to ΔNDC? What are the two-period and three-period truncated multipliers? What is the long-run or equilibrium multiplier? Are these multipliers different from those with respect to exports?
- (c) The short-run (first period) elasticity of imports with respect to capital flows is given by
  $(e_{M, C M}) t = \frac{Δ M_{t} / M_{t}}{Δ C M_{t} / C M_{t}} = \frac{Δ M_{t} \cdot C M_{t}}{Δ C M_{t} \cdot M_{t}} = α \frac{C M_{t}}{M_{t}}$
  This will vary from period to period as the ratio CM_t/M_t is unlikely to remain constant. A mean elasticity for the sample
  period can be defined by taking the ratio of the averages (CM/M).
  - (i) Compute this mean elasticity.
  - (ii) What will be the long-run elasticity?
  - (iii) Compare the above elasticities with the corresponding ones for exports.
2. Forecasting of Imports
- (a) Given that
  $X_{1977} = 11, 580, C M_{1977} = 2, 282, and Δ N D C_{1977} = 1, 785$
  forecast the value of imports and the change in net foreign assets in 1977.
- (b) Compare the above forecasts with the following outturn figures:
  $M_{1977} = 11, 688 and Δ N F A_{1977} = 2, 174.$
  Given that
  $Y_{1977} = 36, 653 and M O_{1977} = 12, 413,$
  
  analyze the errors of forecast.
3. Monetary Policy
- (a) For the values of exports and capital movements given above, what amount of credit would be compatible with the target of achieving balance of payments equilibrium in 1977?
  $(Δ N F A_{1977} = 0)$
- (b) Suppose that the Kenyan authorities fix a target for net foreign assets at the end of 1977 equivalent to one month of imports in the preceding year
  $(N F A_{1977} = \frac{1}{12} M_{1976}) .$
  What would be the amount of credit expansion compatible with this target?
- (c) Show why
  $Δ N D C_{t} = - Δ N F A_{t}$
  corresponds to a situation of no change in imports, income, and money
  $(Δ M_{t} = Δ Y_{t} = Δ M O_{t} = 0) .$

ISSUES FOR DISCUSSION

1. Comment on the following criticisms of the Polak model:
- (a) The structure of the economy is described in terms of only two fixed ratios and, hence, the model is oversimplified and not adequate for formulating monetary policy.
- (b) The model has an essentially short-run orientation.
2. What are the implications of using the broader definition of money (including quasi-money) in the model?
3. With reference to Appendix II, comment on the stability of the ratios of money/income (MO/Y) and imports/income (M/Y) over the sample period.
4. Discuss in simple economic terms the conclusion in Exercise 3(c).
5. Compare the results of the regressions in Appendix III with those obtained in the above exercises.

APPENDIX I

Kenya: Statistical Data on Variables, 1966–76¹

(In millions of Kenya shillings)

Sources: International Monetary Fund, International Financial Statistics, May 1978; Kenya, Statistical Abstract, 1976; and Kenya, Economic Survey, 1976, 1977, and 1978.

^¹See text for definition of variables.

^²On national accounts basis.

^³Residual variable obtained from the identity CM = ΔNFA + M − X.

^⁴A = X + ΔNDC + CM.

Kenya: Statistical Data on Variables, 1966–76¹

(In millions of Kenya shillings)

t	y	M²	X²	MO	ΔNFA	NDC	ΔNFA	ΔNDC	CM³	A⁴
1968	9,666	2,832	2,389	2.317	789	1,528
1969	10,417	2,926	2,708	2,748	1,256	1,492	467	−36	685	3,357
1970	11,453	3,512	2,996	3,505	1,617	1,888	361	396	877	4,269
1971	12,703	4,470	3,256	3,770	1,201	2,569	−416	681	796	4,735
1972	14,447	4,323	3,568	4,294	1,358	2,936	157	367	912	4,847
1973	16,761	5,036	3,932	5,356	1,558	3,798	200	862	1,304	6,098
1974	20,343	8,676	6,298	5,819	969	4,850	−589	1,052	1,789	9,139
1975	23,343	8,358	6,310	6,814	581	6,233	−388	1,383	1,660	9,353
1976	26,582	9,436	6,468	8,454	1,459	6,995	878	762	1,846	11,076

Sources: International Monetary Fund, International Financial Statistics, May 1978; Kenya, Statistical Abstract, 1976; and Kenya, Economic Survey, 1976, 1977, and 1978.

^¹See text for definition of variables.

^²On national accounts basis.

^³Residual variable obtained from the identity CM = ΔNFA + M − X.

^⁴A = X + ΔNDC + CM.

Kenya: Statistical Data on Variables, 1966–76¹

(In millions of Kenya shillings)

t	y	M²	X²	MO	ΔNFA	NDC	ΔNFA	ΔNDC	CM³	A⁴
1968	9,666	2,832	2,389	2.317	789	1,528
1969	10,417	2,926	2,708	2,748	1,256	1,492	467	−36	685	3,357
1970	11,453	3,512	2,996	3,505	1,617	1,888	361	396	877	4,269
1971	12,703	4,470	3,256	3,770	1,201	2,569	−416	681	796	4,735
1972	14,447	4,323	3,568	4,294	1,358	2,936	157	367	912	4,847
1973	16,761	5,036	3,932	5,356	1,558	3,798	200	862	1,304	6,098
1974	20,343	8,676	6,298	5,819	969	4,850	−589	1,052	1,789	9,139
1975	23,343	8,358	6,310	6,814	581	6,233	−388	1,383	1,660	9,353
1976	26,582	9,436	6,468	8,454	1,459	6,995	878	762	1,846	11,076

Sources: International Monetary Fund, International Financial Statistics, May 1978; Kenya, Statistical Abstract, 1976; and Kenya, Economic Survey, 1976, 1977, and 1978.

^¹See text for definition of variables.

^²On national accounts basis.

^³Residual variable obtained from the identity CM = ΔNFA + M − X.

^⁴A = X + ΔNDC + CM.

APPENDIX II

Kenya: Ratios of Imports to Income (M/Y) and of Money to Income (MO/Y), 1968–76

Source: Appendix I.

Kenya: Ratios of Imports to Income (M/Y) and of Money to Income (MO/Y), 1968–76

t	*M/Y*	*MO/Y*
1968	0.293	0.240
1969	0.281	0.264
1970	0.307	0.306
1971	0.352	0.297
1972	0.299	0.297
1973	0.300	0.320
1974	0.426	0.286
1975	0.358	0.292
1976	0.330	0.296
Sum	2.946	2.598

Source: Appendix I.

Kenya: Ratios of Imports to Income (M/Y) and of Money to Income (MO/Y), 1968–76

t	*M/Y*	*MO/Y*
1968	0.293	0.240
1969	0.281	0.264
1970	0.307	0.306
1971	0.352	0.297
1972	0.299	0.297
1973	0.300	0.320
1974	0.426	0.286
1975	0.358	0.292
1976	0.330	0.296
Sum	2.946	2.598

Source: Appendix I.

APPENDIX III Regression Results, 1968-76

Control Information

Number of Observations: 9
Response Variable: M
Estimation Period: 1968 to 1976

Variable	Coefficient	S.E. Coefficient	t	t Probability	Beta
Y	0.342438	0.149129E-01	22.9625	1.0000	0.8482
S.E. Estimate		783.37914	Durbin-Watson	1.81557
R²		0.985054	RHO	0.521875E-01
Adjusted R²		0.983186	Sum of squares	4909463.0
R		0.992499	F (1,8)	527.278
R² (mean adjustment)		0.910046	F probability	1.000000
Adjusted R² (mean adjustment)		0.910046	Condition number	1.00000

Variable	Coefficient	S.E. Coefficient	t	t Probability	Beta
Y	0.342438	0.149129E-01	22.9625	1.0000	0.8482
S.E. Estimate		783.37914	Durbin-Watson	1.81557
R²		0.985054	RHO	0.521875E-01
Adjusted R²		0.983186	Sum of squares	4909463.0
R		0.992499	F (1,8)	527.278
R² (mean adjustment)		0.910046	F probability	1.000000
Adjusted R² (mean adjustment)		0.910046	Condition number	1.00000

Ordinary Prediction Method Used

Observation Number	Observed	Predicted	Residual
1	2832.00	3310.01	−478.008
2	2926.00	3567.18	−641.180
3	3512.00	3921.95	−409.946
4	4470.00	4349.99	120.007
5	4323.00	4947.21	−624.206
6	5036.00	5739.61	−703.608
7	8676.00	6973.07	1702.93
8	8440.00	8004.15	435.848
9	9436.00	9771.82	−335.819

Observation Number	Observed	Predicted	Residual
1	2832.00	3310.01	−478.008
2	2926.00	3567.18	−641.180
3	3512.00	3921.95	−409.946
4	4470.00	4349.99	120.007
5	4323.00	4947.21	−624.206
6	5036.00	5739.61	−703.608
7	8676.00	6973.07	1702.93
8	8440.00	8004.15	435.848
9	9436.00	9771.82	−335.819

Estimated import equation: M_t = 0.342438 Y_t

Control Information

Number of Observations: 9
Response Variable: MO
Estimation Period: 1968 to 1976

Variable	Coefficient	S.E. Coefficient	t	t Probability	Beta
Y	0.293491	0.527876E-02	55.5985	1.0000	0.9476
S.E. Estimate		277.29476	Durbin-Watson	1.28043
R²		0.997419	RHO	0.158272
Adjusted R²		0.997096	Sum of squares	615139.07
R		0.998709	F (1,8)	3091.20
R² (mean adjustment)		0.980851	F probability	1,000000
Adjusted R² (mean adjustment)		0.980851	Condition number	1.00000

Variable	Coefficient	S.E. Coefficient	t	t Probability	Beta
Y	0.293491	0.527876E-02	55.5985	1.0000	0.9476
S.E. Estimate		277.29476	Durbin-Watson	1.28043
R²		0.997419	RHO	0.158272
Adjusted R²		0.997096	Sum of squares	615139.07
R		0.998709	F (1,8)	3091.20
R² (mean adjustment)		0.980851	F probability	1,000000
Adjusted R² (mean adjustment)		0.980851	Condition number	1.00000

Ordinary Prediction Method Used

Observation Number	Observed	Predicted	Residual
1	2317.00	2836.89	−519.888
2	2748.00	3057.30	−309.300
3	3505.00	3361.36	143.643
4	3770.00	3728.22	41.7791
5	4294.00	4240.07	53.9302
6	5356.00	4919.21	436.791
7	5819.00	5976.36	−157.365
8	6814.00	6860.07	−46.0673
9	8454.00	8375.07	78.9302

Observation Number	Observed	Predicted	Residual
1	2317.00	2836.89	−519.888
2	2748.00	3057.30	−309.300
3	3505.00	3361.36	143.643
4	3770.00	3728.22	41.7791
5	4294.00	4240.07	53.9302
6	5356.00	4919.21	436.791
7	5819.00	5976.36	−157.365
8	6814.00	6860.07	−46.0673
9	8454.00	8375.07	78.9302

Estimated money-income relationship: MO_t = 0.293491 Y_t

M. W. Holtrop, “Method of Monetary Analysis Used by De Nederlandsche Bank,” Staff Papers, Vol. 5 (February 1957), pp. 303–15; Paolo Baffi, “Monetary Analysis in Italy,” Staff Papers, Vol. 5 (February 1957), pp. 316–22; Ralph A. Young, “Federal Reserve Flow-of-Funds Accounts,” Staff Papers, Vol. 5 (February 1957), pp. 323–41; Earl Hicks, Graeme S. Dorrance, and Gerard R. Aubanel, “Monetary Analyses,” Staff Papers, Vol. 5 (February 1957), pp. 342–433; and Harold M. Knight, “A Monetary Budget,” Staff Papers, Vol. 7 (October 1959), pp. 210–23.

Of several attempted reconciliations of the elasticity and absorption approaches, particular mention may be made of S. C. Tsiang, “The Role of Money in Trade Balance Stability: Synthesis of the Elasticity and Absorption Approaches,” American Economic Review, Vol. 51 (December 1961), pp. 912–36.

International Monetary Fund, The Monetary Approach to the Balance of Payments (Washington, 1977), pp. 3–4.

The model was originally presented in J. J. Polak, “Monetary Analysis of Income Formation and Payments Problems,” Staff Papers, Vol. 6 (November 1957), pp. 1–50. It was subsequently modified and applied to several countries, and the results were reported in J. J. Polak and Lorette Boisson-neault, “Monetary Analysis of Income and Imports and Its Statistical Application,” Staff Papers, Vol. 7 (April 1960), pp. 349–415. A more recent version of the model is contained in J. J. Polak and Victor Argy, “Credit Policy and the Balance of Payments,” Staff Papers, Vol. 18 (March 1971), pp. 1–24. All three of the articles cited here appear in The Monetary Approach to the Balance of Payments (see footnote 3).

In addition to the writings cited in the preceding footnotes, the following may be of interest: Jacob A. Frenkel and Harry G. Johnson (eds.), The Monetary Approach to the Balance of Payments (London: Allen and Unwin; and University of Toronto Press, 1976); Michael Mussa, “A Monetary Approach to Balance of Payments Analysis,” Journal of Money, Credit and Banking, Vol. 6 (August 1974), pp. 333–51; Arthur B. Laffer, “Monetary Policy and the Balance of Payments,” Journal of Money, Credit and Banking, Vol. 4, Part I (February 1972), pp. 13–22; and Robert A. Mundell, Monetary Theory: Inflation, Interest and Growth in the World Economy (Pacific Palisades, California: Goodyear Publishing Company, 1971).

Reference may be made to the following studies: J. Marcus Fleming and Lorette Boissonneault, “Money Supply and Imports,” Staff Papers, Vol. 8 (May 1961), pp. 227–40; Lorette Boissonneault and Joseph O. Adekunle, “Monetary Analysis of Imports and Income: Further Investigations” (unpublished paper, December 1968); and S. N. Kimaro, “The Polak Model: Empirical Evidence from Selected African Countries,” East African Economic Review, Vol. 7 (December 1975), pp. 53–107.

Nominal income is taken over a given period, while money stock corresponds to the end of that period.

The original Polak model, as well as the subsequent version, follows the second approach, which implies that quasi-money is exogenous. When choosing between the two definitions of money, the main consideration is whether the assumption of constant income velocity of money is more realistic following the first or the second approach.

Foresee

Observation Number	Observed	Predicted	Residual
1	2317.00	2836.89	−519.888
2	2748.00	3057.30	−309.300
3	3505.00	3361.36	143.643
4	3770.00	3728.22	41.7791
5	4294.00	4240.07	53.9302
6	5356.00	4919.21	436.791
7	5819.00	5976.36	−157.365
8	6814.00	6860.07	−46.0673
9	8454.00	8375.07	78.9302

Topics

Countries

Series

About

Resources

Abstract

THE MODEL

DEFINITIONS OF VARIABLES AND ESTIMATION OF PARAMETERS

Statistical Definitions of Variables

Statistical Data on Variables

Estimation of k and m

USE OF MODEL FOR FORECASTING AND POLICYMAKING

Forecasting

Monetary Policy

EXERCISES

ISSUES FOR DISCUSSION

APPENDIX I

APPENDIX II

APPENDIX III Regression Results, 1968-76

Within Same Series

Other IMF Content

Other Publishers Content

Asian Development Bank

Inter-American Development Bank

International Labour Organization

Nordic Council of Ministers

The World Bank

Share Link

Table of Contents

Sections

Figures

Metrics

Foresee