ifdp · July 31, 1989

A Forward-Looking Multicountry Model: MX3

Abstract

This is paper discusses the theoretical structure and empirical properties of MX3, a multicountry macroeconometric model with rational expectations. MX3 is a medium-sized quarterly model of the United States, Japan, and West Germany. The primary objective of the model is to analyze the effect of fiscal and monetary rules on national economies in an international context. By incorporating rational expectations into almost all of the model's behavioral equations, MX3 takes a large step toward addressing the "Lucas critique" of model-based policy analysis.

Board of Governors of the Federal Reserve System International Finance Discussion Papers Number 359

August 1989

A FORWARD-LOOKING MULTICOUNTRY MODEL: MX3

Joseph E. Gagnon

NOTE: International Finance Discussion Papers are preliminary materials circulated to stimulate discussion and critical comment. References in publications to International Finance Discussion Papers (other than an acknowledgement that the writer has had access to unpublished material) should be cleared with the author.

ABSTRACT

This paper discusses the theoretical structure and empirical properties of MXi, a multicountry macroeconometric model with rational expectations. MX3 is a medium-sized quarterly model of the United States, Japan, and West Germariy. The primary objective of the model is to analyze the effect of fiscal and monetary rules on national economies in an international context. By incorporating rational expectations into almost all of the model's behavioral equations, MX3 takes a large step toward addressing the "Lucas

critique" of model-based policy analysis.

A FORWARD-LOOKING MULTICOUNTRY MODEL: MX3

Joseph E. Gagnon! INTRODUCTION

MX3 is a medium-sized macroeconomic model of the United States, Japan, and West Germany. In MX3, quarterly econometric models of each country are linked by trade and capital flows. To close the system, data from the four next largest industrial economies are aggregated as a proxy for the rest of the world (ROW), and are modeled as a fourth country in Mx3 7 Each country block in MX3 has 11 behavioral equations, 21 identities, 4 government policy rules, and 2 exogenous variables. The scale of MX3 is thus considerably smaller than the Federal Reserve Board's Multicountry Model (MCM). (The MCM has app:oximately 170 equations per country block.) This paper presents a theoretical description of MX3 and discusses its empirical implementation

and estimation.

1. The author is a staff economist in the Division of International Finance. This paper represents the views of the author and should not be interpreted as reflecting the views of the Board of Governors of the Federal Reserve System or other members of its staff.

I would like to thank Gwyn Adams for outstanding research assistance. I would also like to thank Sean Craig, Neil Ericsson, David Gordon, William Helkie, Dale Henderson, David Howard, Eric Leeper, Jaime Marquez, John Taylor, Ralph Tryon, and participants in the Division's Monday Workshop for helpful comments and suggestions.

2. One avenue for future research is to extend the ROW sector by collecting data from other countries and building separate models for blocs of similar countries. Possible country groupings include the rest of the OECD countries, the OPEC countries, the newly industrialized countries (NICs), the non-oil developing countries, and the socialist countries. It would also be of interest to model each of the seven largest economies separately.

The structure of MX3 is in many ways similar to traditional Keynesian macro models. Economic agents are separated into four groups- -households, producers, traders, and governments. Each of the main aggregates in the national income accounts is associated with the decision rule of one or more of these groups. For example, households determine aggregate consunption and producers determine aggregate investment. >

MX3 differs from traditional large-scale quantitative macro models in three important dimensions. The first, and most obvious, difference is that expectations are rational and forward-looking rather than backward- ooking. MX3 imposes "rational expectations" in the sense that unobserved expectations are set equal to the model's own prediction of the future .* Only in the past few years have modelers begun to introduce rational expectations into empirical macro models. Two notable examples are John Taylor's multicountry model and the International Monetary Fund's MULTIMOD .> MX3 builds upon the work of these two forerunners.

The second innovation of MX3 lies in its treatment of lags in the structural relations. In MX3, the behavioral equations contain only one lagged dependent variable and no other lagged variables. (The appearance of

a lagged dependent variable in the decision rule is a general result of

3. Even though the profits of producers and traders revert to households, the decisions of producers and traders are not directly coordinated with the decisions of households. A general equilibrium in the model is achieved only through the incentives given by market interest rates and prices.

4. Because it is not feasible to compute true expectations in a large stochastic nonlinear model, the expectations variables are solved under the assumption that future disturbances are identically zero, i.e. the model solution enforces certainty equivalence. This procedure introduces an approximation error. Simply put, the model solves nonlinear functions of expectations when the theory calls for expectations of nonlinear functions.

5. See Taylor [1988] and Masson, et. al. [1988] for a description of these models.

optimizing behavior with costly adjustment.) Higher-order dynamics in the behavior of any individual time series are assumed to reflect the transmission and equilibration of shocks throughout the entire system of equations. In other words, a system of several first-order equations typically gives rise to time series behavior of individual variables that is higher than first order. This research takes the view that the apparent significance of lagged variables in much empirical work can be traced to misspecification of the estimation equation and, in particular, to the lack of a good measure of expected future variables.

The third, and perhaps most significant, difference between MX3 and traditional models concerns the long-run properties of the model. MX3 is designed to exhibit the qualities of an optimal stochastic growth model in the long run. The ultimate sources of growth in this economy are exogenous increases in labor force and technology. MX3’s parameters are carefully restricted to ensure that changes in government policy and.permanent shocks

to supply are consistent with steady-state growth paths.

OBJECTIVES

The primary objective of this project is to develop a simulation model for analyzing fiscal and monetary policy. By allowing expectations to react endogenously to changes in policy rules, MX3 takes a large step toward addressing Lucas’ [1976] critique of model-based policy analysis.

The essence of the Lucas critique is that the "structural" equations of most macro models really are not capturing stable decision rules of economic

agents. Instead, these equations are better characterized as reduced forms

that combine the interactions of policymakers and private agents. Lucas demonstrated that one would not expect such a reduced form relationship to hold constant in the face of a change in the policymakers’ behavior.

Lucas’ prescription for macro modelling is to consider the decision problem for each class of economic agents. Lucas argued that for a wide range of decisionmaking environments, agents base their actions on expectations of future variables as well as the realizations of current and past variables.° Only when modellers have correctly identified the optimal decision rules and information sets of each class of agents can they hope to gauge the effects of different policy rules accurately.

Unfortunately, a fully satisfactory analysis of macroeconomic dynamics based on optimizing behavior has yet to be developed, and it is likely to be years away for models of the scale of MX3. The strategy behind MX3 is to build a tractable model now by appealing heuristically to the structural equations that might result froma suitably specified set of agents, tastes, and technologies. There are three guiding assumptions: First, in the absence of shocks, the economy approaches a perfectly competitive, steadystate growth path. Second, in the face of shocks, agents must undertake costly adjustments. Third, the different classes of agents--consumers, ‘producers, traders, and governments--do not coordinate their decisions except through market prices and interest rates.

Many of the structural equations of MX3 are based on the Euler equation

decision rules that characterize optimal behavior with quadratic adjustment

6. "Rational expectations" embody a simplifying assumption that ignores any learning process by agents about the nature of the economy or the shocks that have ocurred recently. Under rational expectations, agents know the true stochastic structure of the economy, including the policy rules in effect.

costs, / The decision variable is a function of its own past and the expected future discounted sum of the forcing variables. The coefficients on these explanatory variables are typically constrained to ensure an eventual return to an optimal growth path. The speed of adjustment to the steady state can be freely estimated.

A. second objective of MX3 is to learn more about the world economy through estimation and testing of the model. Ideally, all the private sector behavioral and government policy equations should be estimated simultaneously using a technique such as full-information maximum likelihood (FIML) .2 Unfortunately, the computational requirements for FIML in all but the smallest rational expectations models are prohibitive.

YX3 was therefore estimated using instrumental variables techniques. One acvantage of estimating each equation separately and using instruments for current and future endogenous independent variables is that one need not

specify the exact form of the government policy rules before estimating the

. : : 9 private sector behavioral equations.

7. See, for example, Sargent [1978].

8. The advantages of FIML are especially important in the context of ratioral expectations models because future expectations in the equations being estimated can be solved directly by the model's own structure. Moreover, the implied cross-equation restrictions of rational expectations

can be tested, both jointly across all equations and individually in particular equations.

9. The treatment of expectations during the estimation of MX3 thus differs from the treatment of expectations during simulation. In order to simulate the model all equations must be specified, including the policy equations. Because estimation of all the equations simultaneously (FIML) is too expensive, the parameters of MX3 were estimated equation by equation, using instrumental variables for the future expectations.

THEORETICAL STRUCTURE

Overview

MX3's fundamental structure is that of a stochastic growth model with Cobb-Douglas technology, perfectly competitive firms, and long-lived utility-maximizing households. In MX3, households and firms rationally forecast future income and real interest rates when making their consumption and investment plans. Growth in the model is driven exogenously by growth in the labor force and in technology.

With Cobb-Douglas technology and perfect competition, capital.’s share of total output is given by the exponent on capital in the production function. The capital-output ratio equates the returns to capital. with the cost of capital, which is in turn dependent on the real rate of interest. The real interest rate serves to equilibrate consumption and investment at the level of output given by the production function.

While it would be possible to build a model of the economy with only the simple relationships described above, such a model would not be able to explain the short- to medium-run dynamics evident in the data. The transmission of shocks throughout the economy is almost certainly influenced by adjustment costs, gestation lags, and delays in the assimilation of new information. These characteristics of the economic environment may prevent markets from behaving competitively in any given period, and yet market forces may move the economy to a competitive outcome over a longer horizon.

Only recently have economists begun to enrich the dynamics of growth models by solving the decision problems of agents with costs of adjustment or gestation lags. At present, this work has yielded only rudimentary

models that require the assumption of continuously competitive market

clearing in order to obtain a solution. Extending these models rigorously to allow for monopolistic competition and endogenous entry of new firms is a task beyond the scope of this project.

The structure of MX3 reflects the view that economic theory in its present state yields clearer insights about the long-run behavior of the economy than about short-run dynamics. The approach taken by MX3 is to enforce a competitive steady state in the long run, but to allow (heuristically) for imperfect competition and costly adjustment in the short run. In several instances, the model's dynamics are inspired by optimal decision rules in the face of convex adjustment costs. These decision rules determine the control variable as a function of its previous value and the discounted expected future sum of the forcing variables. However, with the exception of consumption, the structural equations of the model are not

derived from the maximization of specific objective functions.

Markets and Agents

Each country is composed of four different types of economic agents. Producers in each country produce a homogeneous good that is differentiated from the goods produced in other countries. Productive capacity is modeled by a Cobb-Douglas function in the capital stock and the labor force. Total production can deviate temporarily from capacity production, but these deviations will be associated with equilibrating price movements.

Traders do not utilize capital and labor; they are modeled as pure arbitragers. Domestic traders purchase goods from domestic producers to sell to foreigners. This trade is characterized by significant costs of transportation and adjustment that prevent the continuous equalization of

prices across countries. The preferences of households, producers, and

governments for foreign goods relative to domestic goods jointly determine the demand curve faced by foreign traders selling into the domestic market.

Households maximize utility from discounted future consumption subject to their budget constraint. Households own the firms that produce and trade goods, and the net income earned by these firms passes directly to the households. The notional labor supply of each household is constant, but actual labor supplied may fluctuate as output fluctuates around capzecity. (The model essentially enforces equal capacity utilization of capitel and labor.)

Governments determine the level of the monetary base and real government spending. The government budget constraint determines the level of bonds outstanding. Tax rates are modeled with an ad hoc adjustment mechanism to ensure that the ratio of bonds to taxable income returns to an exogenous target level. The target level of government debt and the speed of adjustment to that target may be considered as additional policy instruments of government.

Financial markets determine the levels of interest rates and exchange rates. These financial markets represent the combined behavior of the four sectors in the model. Production technology and the labor force are modeled as exogenous to the rest of the economy.

Appendix 1 (attached to this paper) presents a simplified overview of a typical country model in MX3 and lists the data mnemonics used in the paper. Appendix 2 (not attached, but available upon request) provides a detailed

listing of the equations in MX3; it also documents the model database.

Consumption In the absence of liquidity constraints, adjustment costs, and

information lags, the representative household consumes a constant fraction

10

of its wealth. Wealth is defined as the discounted sum of expected future

disposable income.

- -i 1. C= B =YD (b+ 2+ a-may, ote - PPA i]

——1

4 4

In equation 1, RL, i is the nominal interest rate at time t on a risk-

free bond maturing after i periods; TAU, i is the average tax rate on the

interest from such a bond; DPA, , is the average rate of inflation of the

t,i

domestic absorption deflator between period t and period t+i; and A is a risk premium. !+ C is total private consumption and YD is private disposable income. RL, DPA, and A are all divided by four to convert annual rates to

quarterly rates of discount.

Ceteris paribus, higher levels of current or future income lead to

higher current consumption; higher interest rates reduce current

10. This consumption relation can be derived for infinitely-lived households with time-separable, logarithmic utility.

11. The premium A has two components. The largest component derives from the fact that private rates of return typically exceed the rate of return on government bonds. This excess return may represent a risk premium, and it has an average value of 6 percentage points in the United States. (See Mehra and Prescott [1985].) The economics profession has made little progress to date in explaining this risk premium or its fluctuations. In MX3 it is assumed to take a constant value of 6 percent.

The second component of A is the probability that the representative consumer will not survive until the following year. The probability of death leads all consumers to discount the future at a faster rate than the market rate of interest. (See Blanchard [1985].) In MX3 the probability of death for the representative consumer is assumed to be 2 percent per annum, which implies that the representative consumer expects to live for 50 more years.

consumption. In practice, however, positive shocks to income will]. tend to raise interest rates via the money demand equation, with ambiguous results for current consumption.

There are two modifications of equation (1) that may or may not be important in modeling consumption. First, a fraction of consumers: may be liquidity constrained, so that they simply consume their current clisposable income. 14 Second, the non-liquidity-constrained consumers may adjust slowly

to shifts in wealth by smoothing consumption from period to period due to an

aversion to sharp changes in their spending habits. > 2. C. =~ Cy, + Cy,- 3. Cie = ayD,. ~ A L DPA, .) > 4. Cy = bC,,+°(1-b)(1l-a)B = YD Ae + _ + (1-TAU .)® tii - t i| t 2t : t+i t,i i=0 4 4 4 12. For a discussion of the empirical magnitude of liquidity corstraints, see Hall [1988] and Poterba and Summers [1987]. Given the asymmet:iry between

ability to borrow and ability to save, it may be more descriptive to call these consumers myopic.

13. These households also should be forecasting movements in wealth in order to smooth consumption optimally. It is easy to show that forecastable movements in wealth over short horizons are extremely small under a broad range of environments. Therefore, the current value of wealth is a close approximation to its expected value over the near horizon. For more on habit-formation and slow adjustment in consumption, see Nason [1969].

Even if individual consumers adapt their spending plans rapidly, there will be a lag between the date their plans are made and the date the transactions are recorded. This lag will vary depending on the individual

plans: a European vacation may wait until summer, while a new car may be purchased quickly.

-ll-

According to equation (2) total consumption consists of the sum of consumption by liquidity-constrained consumers, C); and consumption by slowly-adjusting, unconstrained consumers, Co. Liquidity-constrained consumers simply consume their current disposable income, and the parameter, a, represents the share of disposable income earned by liquidity-constrained consumers. The remaining consumers adjust slowly toward the target level of consumption; b is the lagged adjustment parameter. Equations (3) and (4) can be substituted into (2) to yield a simplified expression for total consumption. If liquidity constraints and consumption smoothing are

unimpertant (a=0 and b=0) equation (5) reduces to equation (1).

5. C. = aYD, + i (ae - aD, _1]

- “i + (1-a)(1-b)B z D3 (1 +24 ci-tau, ,)Beta - PPAt i| . i- 4 car

In order to eliminate the infinite sum of future variables in equation (5) consider the following transformation of equation (4) using the term

structure relation that is presented in a later section (equation (36)).

6. Cor = bCo. 41 ~ i-l -1 + (l-a)(1-b)B 3 4¥YD.., * I (2 +44 c-tau_,.)PSt4y - PPAceit1] . t+1 : t+j i=0 j=0 4 4 4 The one-period interest rate, RL. 1? has been abbreviated to RS, as has the

one-period inflation rate, DPA,. This relationship also holds in the

subsequent period.

- 12 -

7. bec

Coed = Poe

re) i-l -1 + (l-a)(1-b)B 5 {ea * ae + : + (1-TAU,, 5 Ses - Parsi +1 } i=l j-

By dividing both sides of equation (7) by (1 + “+ (1-Tau,) Sse - or e+1) 4 4 mn

and subtracting equation (7) from equation (6), it is easy to show that

8. [2 + b/(2 + 4 + (1-TAU,) Se - Pact) co. = bCy, 4 4 4

4

PPAven) + (1-a)(1-b)p¥D,.

A RS +C (2 +" + (1-TAU,) ot - 2t+1 i tz z

Once again, the liquidity-constrained consumers are described by equation (3). Combining equations (2) and (3) with equation (8) yields a description of aggregate consumption that relies on expectations of only one

future period.

A RS DPA 9. C= b(Se-1- a? 4)/(1 + b/(2 + : + (1-TAU,) _ - “c1)}

A RS DPA + (Seat a¥D.43)/[1 +b +t i + (1-TAU,)—£ - Pett) + (a + (1-a)(1-b) 8) ¥D,. Fixed Investment

The model's investment equation is essentially neoclassical. In the

long run, the returns to capital should equal the cost of capital:

- 13- 10. ((1-Tav)Rs + (1-TAU)S - (1-TAU)DPA + | * K = a@(1-TAU)GDP.

The first three terms in the brackets on the left-hand-side of equation (10) represent the cost of holding a unit of capital for one period. The inte:‘est charge, RS, is reduced by the tax rate, TAU, because firms are allowed to deduct interest expense from their taxes. Similarly, the depreciation, 6, is also tax deductible. The inflation rate, DPA, represents a capital gain to the firm, so it reduces the cost of holding capittal. However, because the ability to deduct future depreciation from the firm's taxes is based on historical nominal cost rather than current value, inflation today increases the firm's future real tax liability. Finally, the model allows for a constant risk premium, 7, needed to induce agents to hold capital instead of risk-free government bonds.

The right-hand-side of equation (10) represents the returns to capital. With a Cobb-Douglas production function and competitive markets, capital’s share of output is simply the exponent on capital, a, in the production function. These returns are reduced by the average tax rate.

If it is costly to adjust the capital stock, even perfectly competitive markets are not sufficient to enforce equation (10) continuously. Some slowness in the adjustment process will generally be optimal. Equation (12) describes investment as a process that adjusts slowly to deviations between the desired and the actual capital stock. One explanation for slow adjustment of investment is that many capital-spending projects require multi-period commitments of a stream of investment that is costly to

change. !4 On the other hand, costly adjustment also provides an incentive

14. See Kydland and Prescott [1982].

-14-

for forward-looking behavior. A convenient way to capture both of these effects is to include a lagged dependent variable and expected future values of the target variables in the decision rule. Equation (12) does this without abandoning the long-run relationship in equation (10) and it

introduces only two new parameters. 11. cc, = ((1-Tav,) * (RS + 6 - DPA...) + n).

12. IF. = cIF, |

+ (l-c)(1-d) = a [ (a(1-ta0,, .)60P,,5/CCe,5) - (1-5)K

i~0 ces}:

13. K. = (1 - 6/4)K. + IF /4.

Equation (11) describes the one-period holding cost of capital. The term inside the inner set of brackets in equation (12) can be interpreted as the equilibrium capital stock in the absence of adjustment costs, as given by equations (10) and (11) .}5 The second term in the brackets is the capital stock carried over from the previous period. Equation (13) is the

perpetual inventory identity which defines the evolution of the capital

stock, 16

15. Equation (12) presents a causal relationship between expected output and desired capital. Future output is not exogenous, however, since it is affected by the amount of capital installed in the current period. Due to decreasing marginal returns to capital in the production function, there will be a unique combination of capital and output that satisfy. equation (12) in the steady state.

16. All stock variables refer to quantities at the end of the period. Because all flow variables are expressed at annual rates, they must be divided by four for purposes of stock accumulation. This rule applies to the capital stock, government bonds, and net foreign assets.

- 15 -

As in the case of consumption, it is possible to write equation (12) without the infinite sum of future variables. First, rewrite equation (12)

in serms of lag and lead operators (L and F): -1 14, (1-cL)IF, = (1-dF) (2(1-TaU,) GDP, /cc, - (1-6)K,_4)- Multiplying both sides of equation (14) by (1-dF) yields the following:

15. (l+ced) IF, = cIF + dIF.

t-l. 1

+ (1-¢) (1-d) (a(1-TAU, )GDP,/CC, - (1-8)K, 4).

Inventory Investment

Producers are assumed to hold inventories to adjust to expected and unexpected changes in demand. Thus, the net change in the,stock of inventories responds negatively to current output and positively to expected future output. The cost of holding inventories is the short-term real

interest rate.

16. Il, = €g + e,GDP.., - e GDP. - e3(RS_- DPA, 41) -

Export Prices

Traders are modeled as imperfectly competitive arbitrageurs who buy goods in their home country and sell them in a foreign country. The price of these exports reflects output prices at home and abroad. Export prices

are constrained to be homogenous of degree one with respect to output

- 16 -

prices. The foreign output price, PGNPW, is a weighted average of foreign prices converted to domestic currency at the exchange rate, E. The weights, w,, are fixed according to the average share of domestic exports destined for country i over the estimation period. The superscripts index individual countries. The home country is normalized at zero and its superscript is suppressed.

Over time, export prices in all developed economies have fallen relative to aggregate prices. This phenomenon is most likely due to faster technological progress in tradables than nontradables and it is modeled here

as a simple function of time.

17. log(PEX, ) = && + 8 log(PEX, |) + 8 log(PGNPW, ) + (1-8) -&5) log(PGNP ) - &3¢.

18. log(PGNPW,) = w,log(E-*pcnP!) + w,log(E2*PGNP2)

7 508 t OB ETE GNE gr OB, t

3 3 + (1-w, -wy)log(E,*PGNP_ ).

Export Volumes

Unlike many other econometric models, MX3 allows for different cyclical and secular demand elasticities in trade. The cyclical demand for exports depends on weighted foreign absorption relative to foreign productive capacity, AW/CAPW, and the price of domestic exports relative to the price of foreign exports, PEX/PIM. The secular demand for exports depends on the level of worldwide production capacity, CAPTOT. MX3 thus incorporates the assumption that long-run growth in trade is due as much to supply-side as to demand-side factors. (In equations (20) and (21) the weights, W;, are the

same as those used to compute PGNPW in equation (18).)

- 17 -

19. log (EX, ) = ho + hy log(EX, 4) + h, log (AW, /CAPW, )

+ h,log(PEX,/PIM,) + h,log(CAPTOT,).

I

1 2 3 20. log(AW, ) w, log(A,) + wo log(A,) + (1-w, -w,)log(A,).

1 2 3 21. log(CAPW, ) = w, log(CAP.) + w,log(CAP.) + (1-w,-w,)log(CAP_).

2

1 22. CAPTOT, = CAP. + CAP. + CAP.

+ CAPS. Import, Volumes and Prices

Because each country’s exports are the imports of the other countries, it would not be theoretically consistent to model imports and import prices independently of exports and export prices. MX3 thus estimates equations that describe the share of a country’s exports that are destined for each other country. These share equations incorporate the global trade balance identity. Country i's total imports are computed in both nominal and real terms by adding up the fraction of each other country’s nominal and real exporzs that are destined for country i. (The aggregate import price for country i is the ratio of nominal to real imports.)

Ideally, there should be two sets of export share equations: one set for nominal exports and one set for real exports. The allocation of nominal and real exports across trading partners need not be identical because the price of exports to different trading partners need not be identical. Unfortunately, on a bilateral quarterly basis only nominal trade shares are available. Both nominal and real imports in MX3 are computed using the same

share weights of exports.

As an alternative to modelling the export side, it would be possible to estimate behavioral equations for imports and import prices and use import share equations to compute exports and export prices. The former strategy is adopted by MX3 for two reasons. First, the assumption that nominal and real trade shares move together is more realistic for exports than for imports, as long as export prices are more closely correlated with the exporter’s price level than the importer's. Second, it is econome=rically easier to model the effect of relative prices on nominal export shares than on nominal import shares. An increase in the price level of one trading partner relative to another will unambiguously increase the share of nominal exports to that country by encouraging both higher prices and quantiities. However, an increase in the price level of one trading partner relative to another will have offsetting price and quantity effects on the share of nominal imports from that country.

The allocation of each country’s exports among its trading partners is modeled via a system of equations that captures the effects of changing relative prices while forcing the shares to sum to unity for each exporter, !” SHR1 refers to the share of country O exports destined for

country 1. sHRot refers to the share of country 1 exports destined for

country 0.

= . 17. In theory one also might want to capture the effects of relative

absorption and relative capacity, but empirically these effects were insignificant.

- 19 -

1

t

23. SHR1 = vy + Ty) o*SHRL, 4 + 7, )*1og[E t

+PGNP?/E;*PGNP?|

1

+ 7, y¥1og(Et

1.3 3 *PGNPE/EE*PCNP?)

1 1,,2 2 24. SHR2/= Ve + Too *SHR2, 4 - 7, p#log[Et*PGNPL/Ee*PCNP?|

2 2.3 3 + 1, ¥log{E-¥PoNP?/E>*PCNP?)

25. SHR3 = (1-4, -¥) - T *SHR1 | - T *SHR2

10 20 1

1 1,,3 3 - 7 y*1og [Et*PoNP)/E>+PGNP

2 2.3 3 :| - T, y#log(Ep*PoNPL/E?*PoNP®)

li pyl 2 2 3 3

26. IM, ~ SHRO | *EX) + SHRO|,*EX) + SHRO | *EX, .

1 1 loioyl 3 3

3 3 = * * 27. PIM. & PEX. SHRO | *EX +... + E,*PEX?*SHRO,+EX>| / IM, .

Finally, net foreign assets are the sum of previous current account

surp..uses. The currency denomination of all international assets is assumed

to be U.S. dollars, and the return on these assets is equal to the treturn on

U.S. government bonds. = * - * 28. NFA, (2 + RS, /4)NFAL 1 + PEX, *EX/4 PIM, *IM,/4.

Capacity

Capacity output, CAP, is given by a Cobb-Douglas production furiction. The labor force, L, and production technology, Q, are exogenous. The rate of capacity utilization, CU, is simply the ratio of domestic output to domestic capacity. In this model capacity denotes the sustainable, equilibrium level of output given the values of K, L, and Q, and the preferences of workers and managers. As discussed below, it is possible for

the economy to operate above or below "capacity" at any given time. fo4 = * 29. CAP. Qe Ke. 30. CU, = GDP, /CAP..

Prices

The model abstracts from the labor market in its description of aggregate price behavior; in other words, it treats workers’ wages as just additional prices in the system. The model therefore does not rely on

movements in the real wage to explain output fluctuations. 18 Instead, MX3

18. The traditional Keynesian explanation of the business cycle relied on countercyclical real wages caused by sticky nominal wages: during periods of high demand, firms would charge higher prices, thus reducing the ireal wage and encouraging more employment and output. The seminal work of Dunlop [1938] as well as recent studies by Bils [1985] and Roberts [1987] all conclude that the real wage is nearly constant over the business cycle.

- 21-

posits an expectations-augmented Phillips curve to explain price adjustment. In the model, the rate of inflation accelerates when output is above capacity or when output is expected to be above capacity in the future. Similarly, inflation decelerates when output is below capacity or when output is expected to be below capacity in the future. The model also is

characterized by a significant degree of inertia in the inflation rate. = - * 31. DPGNP Po*DPGNP. 4 + (1 Po) DPGNP 44 + P,log(cu.).

A Phillips curve that is both backward- and forward-looking, like equation (31), can be justified as a rough approximation to a model of staggered price contracts. 1? According to models of staggered contracts, firms and workers set nominal prices for a predetermined period of time and agree to supply whatever quantity is demanded during the contract period. If the contracts last for more than one period, lagged adjustment will be introduced into the inflation process. Because firms and workers try to predict conditions over the life of their contracts, there will also be a forward-looking element to price behavior.

When output equals capacity, CU = 1 and log(CU) = 0. In this state of full employment there is no tendency for inflation to accelerate or decelerate, according to equation (31). The real side of the MX3 model can therefore be in equilibrium at any constant inflation rate of the price level.

The absorption deflator is an average of the GNP deflator and the

export and import deflators. It is solved from the nominal GNP identity.

19. See Taylor [1980]. The dynamics induced by staggered contracts are more complex than those of equation (31).

- 22 -

The rate of inflation of output prices and the rate of inflation of

absorption prices are defined in annual rates by simple identities.

= - * * - * 32. PA *AL PGNP, *GNP PEX EX. + PIM, IM, RS. NFA, _1-

33. DPGNP = 4*(PGNP,-PGNP, ,)/PGNP. |.

34. DPA, = 4*(PA,-PA,_1)/PA,_1-

Exchange rate

The basic exchange rate equation is motivated by open interest rate parity. The difference in nominal rates of return, RS, across countries is exactly matched by the expected movement of nominal exchange rates. g, 20 There are three exchange rates in the model; U.S. dollars are the numeraire.

Equation (35) presents a typical exchange rate equation.

1

35. (EL - E41) / EY = (RS. - RS.) / 4.

Term Structure

The model incorporates the pure expectations theory of the term structure of interest rates. Long rates are a function of expected future short rates over the term to maturity. Because the other behavioral equations of the model use this term structure relation to simplify their expressions,

long-term interest rates are not needed to solve MX3. The implicit equation

DN : : : 20. It is possible to augment the equation to include either a constant risk premium or a variable risk premium that depends on the ratio of foreign

to domestic bonds (portfolio balance). See Dooley and Isard [1982] and Frankel [1983].

- 23 -

for the interest rate on a bond maturing in i periods is given below. The one-period interest rate, RL, L? has been abbreviated to RS ..

36. (1 + RL, > - (1 +RS,)(1 + RS (1 + Rs

t+1) t+i-L°

Money Demand - Money Supply

{t is possible to model either the short-term interest rate, RS, or the monetary base, MB, as the instrument of monetary policy. One of the main purposes of MX3 is to analyze the effect of different monetary policies on the overall economy. In the simplest case, one may consider a monetary policy that sets a constant growth rate for the monetary base.

The public demands real money balances, MB/PA. The absorption deflator and domestic absorption appear in the money demand equation on the assumption that cash balances are held to support spending. Nominal interest rates adjust to ensure that the public willingly demands the

quantity of money supplied.

37. MBL = mMB, 4.

38. log(MB, /PA,) = To + r,log(MB. _j/PA,_4) + rolog(A,) +r RS .. Fiscal Policy

Real government spending is denoted by G. Nominal tax revenues, TAX, equal the tax rate, TAU, times taxable income, TI. Taxable income is net national product plus interest on government bonds, B. The stock of government debt is given by the cumulation of past budget deficits minus

revenues from money creation.

- 24 - = *

= * - 40. TI PGNP | GNP. 6K

* * r PA, + RS,*B,_

t-1 1°

41. Be = (2 + RS_/4) Bey + (PAL *G,. - TAX, /4 - (MB. - MB. 1).

Based on equations (39)-(41) it would appear that governments are free to choose values of G and TAU independently and without constraints. However, when private agents form their expectations of future fiscal policy, they recognize that the government must satisfy its budget constraint (equation (41)) at every future date. Thus, an intertemporal budget constraint implicitly restricts the future paths of G and TAU. If the government is not allowed to default on its obligations, the national debt cannot grow so large that interest payments on the debt exceed the government’s ability to raise revenues. Assuming a positive interest rate

and a fixed rate of money growth, this feasibility condition places an upper

bound on the ratio of government debt to taxable income.71 21. When there are no liquidity constraints, risk premia, or finite

-horizons (i.e., a=0 and A=0 in the consumption equation) optimizing behavior places the following restrictions on expected future fiscal policy:

fo 0]

= - * - 42. By Fan {(7% e574 Par Cey 3/4 + MB MB 45-1)

/ i (2 + RS 44/4}.

i=1

j 43. limB.., / I (1 +RS_,./4) = 0.

Placing bounds on the ratio of debt to income is sufficient to ensure that (42) and (43) hold, provided that the nominal rate of interest exceeds the growth rate of nominal income.

(Footnote continues on next page)

- 25 -

In order to compute expectations of future fiscal policy in a manner consistent with the government’s intertemporal budget constraint, MX3 models the tax rate, TAU, as a reaction function that gradually adjusts to return the ratic of bonds to taxable income to some exogenous target value, BRATIO. The variable TBAR represents the tax rate necessary to return the bond to income ratio to its target, BRATIO, in one period. The actual tax rate, TAU, adjusts part of the way toward TBAR in.each period. 2? If desired, automatic stabilizers in the tax system could be added to equation (44). As with the monetary base, government spending and the bond ratio are left unspecified. However, in order to solve the model, some policy rule must be assumed to describe the future behavior of G and BRATIO. In the simplest

case, G prows at a constant rate and BRATIO is constant. 44. TAU, = wIAU, 4 + (1-w)TBAR, .

= * - 4* - 45. TBAR, (S. PA. + RS_*BL_, - 4*(MB, MB, .;))/TI,

- BRATIO, + B._j/TI,.

(Footnote continued from previous page)

If the rate of interest is smaller than the growth rate of income, a bounded debt-to-income ratio will still ensure feasibility of future fiscal policy, but it does not guarantee that equations (42) and (43) will hold. However, if households have finite lives or if they are risk averse, then (42) and (43) are no longer necessary for optimality. See Abel, et. al. [1987].

22. The adjustment parameter, w, is always bounded between zero and one. However, depending on the remaining parameters of the model, large values of w may not adjust the tax rate quickly enough to ensure stability.

- 26 -

47. BRATIO, = Z.

Accounting Identities

48. A. = Cc, + IF. + Il, + GC.

49. GDP. - A, + EX. - IM, .

50. GNP. = GDP, + RS .*NFA,_j/PGNP..

= * - - 51. YD, PGNP | GNP /PA, 6 TAX, /PA,

Kee + RS,*B,_1/PA, - (“Bet B, 1} *DPA,/PA,.

ESTIMATION

The equations to be estimated are consumption, fixed investment, inventory investment, export volumes, export prices, export shares, production capacity, price adjustment, and money demand. Since many of the variables in these equations are nonstationary, they must undergo appropriate transformations in order to eliminate heteroscedastic ~esiduals. For most equations, the relationships are estimated in logarithmic form. In

other cases, all the nonstationary variables are divided by a smoothly-

growing variable with which they are presumed to be cointegrated.?? 23. The presumed cointegrating relationships are that consumption and

investment grow proportionately with output in the long run. These relationships have not been tested with the MX3 data set because of the

short sample that is available. However, they are implied by the theory of the previous section.

The data available for estimation are 48 quarterly observations from 1976:1 to 1987:4. The data are expressed at annual rates. All equations are estimated over the maximum possible range after allowing for necessary lags and leads. The consumption, inventory investment, and fixed investment regressions were run in RATS 3.0. The remaining regressions were run in TROLL 13.0.

Wherever practical, MX3 uses common parameter estimates from pooled regressions. In some cases there are theoretical reasons for expecting a common parameter. In other cases a common parameter was imposed only if the

unrestricted estimates were not significantly different statistically.

Consumption

The. consumption equation (9) is highly nonlinear and it contains expectations of future variables. Hansen and Singleton [1982] develop a generalized method of moments (GMM) procedure for the estimation of nonlinear equations with future expectations. The procedure is based on the orthogonality condition between future disturbances and past information. This orthogonality condition is an implication of rational expectations: agents should use all available information, of which the instruments are a subset, in order to compute the expectations of future variables th-* concern them. Any deviation between their expectation and the subsequent realizat:ion of a variable ought to be orthogonal to all information that was availab’e at the time they formed their expectation.

In order to scale for growth over time, both consumption and disposable income “in equation (9) have been divided by production capacity. The equation was estimated for each country individually and in a pooled

regress:.on for all countries together. The instruments used were a

- 28 -

constant, one lag of consumption, one lag of disposable income, current government spending, a lagged interest rate, a lagged inflation rate, and lagged real money balances.

The liquidity-constraints parameter, a, was not significantly different from zero in any regression. After setting a=0, the equation was reestimated. None of the estimated parameters in any of the single country regressions ever deviated from the pooled estimates by more than two standard deviations. Moreover, the standard errors in the pooled regression were uniformly smaller than the standard errors in the individual country regressions. Therefore, the model takes the coefficients from the pooled regression for every country. These results are shown below:

52. 11 + b/f1 + 4 + (1-Tau_yBSe ~ PPALs iy Ic. = bc 4 t —— t t-l

4 4

+ Cuya/(2 +S + a-tauyBSe - PPAtea) + ppv, 4 ns ls t

b= 0.852 B = 0.00985 A= 0.08 J= 1.50 (2) (0.059) (0.00101) n.a.

The Hansen-Singleton J-statistic, which tests the orthogonality implications of rational expectations, is not significant at conventional levels. Formal tests of parameter constancy were conducted both over time and across countries. These tests are described in Andrews and Fair [1988]. The test for parameter constancy over time splits the sample into t:wo equal subsamples and tests whether the parameters estimated in each subsemple are significantly different from each other. This test was not significant at

the 10 percent level. There are four tests for parameter constancy across

countries. Each test compares the parameters estimated in a single country to the pooled estimates. None of these four tests was significant at the

5 percent level. The ROW estimates were significantly different at the 10 percent level, however.

In addition to tests of parameter constancy, the consumption equation was reestimated after incorporating a constant term and it was also reestimated after incorporating a lag of disposable income. In neither case was the extra term significant at the 5 percent level.

Returning to the estimated equation, one may interpret the economic significance of the coefficients b and 8. The estimate of b implies that consumers adjust to new circumstances at the rate of 15 percent per quarter. In other words, after a shock to permanent income, consumption adjusts 47 percent of the way to its new long-run level in the first year. The estimate of 6 implies that in steady-state, households consume 1 percent of

their wealth per quarter.

Inventory Investment

“he inventory equation (16) was estimated via GMM. The inventory investment and GDP series were first divided by capacity. The instruments were a constant, a lagged growth rate of GDP, current government spending, a lagged real interest rate, and a lagged growth rate of the monetary base. The equation was estimated on individual countries as well as pooled across countries. In every case the restriction e) = & could not be rejected. The estimated value of e, varies considerably across countries, but in every case the associated standard deviation is quite high. The pooled estimate lies approximately in the middle of the range. Results from the pooled

regression are presented here:

- 30 -

53. II, =e, + 2) (PPL - GDP.) - e,(RS_- DPA, +1): @9y = 0.0073 e, = 0.483 e, = 0.078 J= 5.35 (4) (0.0011) (0.570) (0.035)

The J-statistic is significant at the 10 percent, but not the 5 percent, level. Parameter constancy tests were conducted across time and across countries as in the case of consumption, and none of the tests was

significant at the 10 percent level.

Fixed Investment

The fixed investment equation (15) was also estimated using GYM. Once again, the series IF, GDP, and K were divided by capacity before estimation. The instruments were a constant, two lags of investment, a lagged capital stock, a lagged real interest rate, a lagged growth rate of the moretary base, and a lagged inflation rate.

The discount factor d is fixed at 0.97. This value was calibrated empirically as follows: The rate at which firms discount the future sequence of desired capital stocks ought to be related to the real discount rate and the rate of depreciation of capital. The sum of the average real after-tax interest rate, the Mehra-Prescott risk premium, and the measured rate of depreciation has averaged about 12 percent per annum, or 3 percent per quarter, historically.

The rate of depreciation, 6, and the share of output accruing to capital, a, were estimated independently. The remaining parameters of the fixed investment equation are the lagged adjustment, c, and the risk premium, x. Attempts at estimation were unsuccessful, as the lag

coefficient was approximately unity. In the end, the lag coefficient c has

- 31 -

been constrained at 0.95, and the risk premium, 7, has been estimated

independently for each country. 54. (1+ed)IF, = cIF, | + dIFL,, + (1-c) (1-4) (aGDP,/cC, - (1-6)K, 4).

where ce. = (RS, + 6 - DPA ) +n / (1-TAU,).

t+1

Germany Japan ROW USA ww (x100) 4.784 ; 6.707 6.707 6.796 (0.784) (0.951) (0.888) (1.621) c 0.95 0.95 0.95 0.95 n.d. n.a, n.a,. n.a. d 0.97 0.97 0.97 0.97 n.a. n.a. nea. na. J (x2) 7.75 0.01 27.3 35.4

eee

The J-Statistic is significant at the 1 percent level in the ROW and U.S. regressions. The test for parameter constancy is significant at the

1 percent level in every country.

Capital Stock

The depreciation rate of capital was estimated via ordinary least squares (OLS). The regressand is the series defined by KL -IF,. The regressor is Keeq: This regression estimates the fraction of capital that

; 4 aaa ‘ survives after one quarter.” The quarterly depreciation rate is the

24. Equation (55) is very nearly an identity, and it is treated as such in the model. In practice, statistical agencies estimate the capital stock at a disaggregated level, using different depreciation rates for each type of capital. If the proportion of investment in each type of capital good were

(Footnote continues on next page)

- 32 -

fraction that decays in one quarter, or unity minus the estimated coefficient. The annual depreciation rate is approximately four times the quarterly rate.

55. K. - IF, = 6*K, where 6 = 4 * (1-6).

t 1? Germany Japan ROW USA

§ 0.9879 0.9924 0.9836 0.9845 (0.0000) (0.0003) (0.0001) (0.0001)

F 0.0483 0.0304 0.0658 0.0623

R? 0.999 0.998 0.996 0.999

D-W 1.34 0.50 0.28 1.13 —_

Production Function

Under perfect competition, the exponent on capital in a Cobb-Douglas production function is equivalent to the fraction of output that accrues to the owners of capital. The value of a used in MX3 differs slightly between the four country blocks. It is estimated by taking the average fraction of after-tax GDP that is composed of capital consumption allowances and operating surplus, according to the OECD's National Accounts over the period

1976-1987. The values of a in Germany, Japan, and the United States are,

(Footnote continued from previous page) constant over time, then the aggregate depreciation rate would be constant and equation (55) would hold identically.

- 33 respectively, 0.35, 0.39, and 0.32. The value of @ in ROW has been arbitrarily fixed at the value estimated for Germany.

Technology, Q, is assumed to follow a log-linear time trend. This trend is estimated as the fitted value of the following OLS regression.

56. (Log(aDP,) - alog(K, 4) - (1-a) log(L,_1)) = [ + pt.

Germany Japan ROW USA

r 2.1990 0.3792 1.6272 2.0944 (0.0051) (0.0024) (0.0035) (0.0080) » 0.00116 0.00231 0.00140 0.00079 (0.00019) (0.00009) (0.00013) (0.00029) Rr? 0.45 0.94 0.72 0.14 D-W 0.41 _ 0.55 0.18 0.15 ne ea 2 © > > Se ©; © PO

Output Price

It proved impossible to obtain sensible estimates of the expectationsaugmented Phillips curve for the GNP deflator. Consequently, the coefficients on lagged and lead inflation have been arbitrarily set at 0.5 each, and the sensitivity of inflation to capacity utilization has been deternined by simulation trials to yield a reasonable responsiveness of prices to aggregate demand.

Since the process of price adjustment is central to understanding the transmission of monetary policy to the rest of the economy, the lack of a well-estimated structural price equation is a serious flaw in the MX3 model as it exists currently. The first step in the next stage of development

must be to consider alternative price adjustment mechanisms and estimation

- 36 -

Germany Japan ROW USA hy -0.612 -0.301 -0.373 -1.645 (0.192) (0.172) (0.187) (0.318) hy 0.775 0.937 0.864 0.476 (0.069) (0.033) (0.067) (0.101) hy 0.688 0.418 0.268 1.929 (0.223) (0.223) (0.112) (0.297) hy -0.260 -0.107 -0.120 -0.372 (0.075) (0.036) n.a. (0.092) hy, 1.000 1.000 1.000 1.000 nia. nia. nia. n.a. R2 0.97 0.99 0.98 0.96 D-W 2.02 2.00 2.05 1.75 __

59. log(PEX, ) = & + 8, log(PEX, 1) + B,log(PGNPW, )

+ (1-g)-B,)log(PGNP,) + B,log(PEX, |/PEX, 4) - 8, (1-g,)t.

Germany Japan ROW USA B 0.015 0.005 0.008 0.003 (0.002) (0.006) (0.003) (0.002) a 0.734 0.824 0.985 0.962 (0.050) (0.063) (0.037) (0.032) B, 0.076 0.120 0.000 0.000 (0.013) (0.046) nua. n.a. B3 0.408 0.136 0.527 0.728 (0.092) (0.123) (0.118) (0.119) b,, -0,0025 -0,0025 -0.0025 -0.0025 (0.0002) (0.0002) (0.0002) (0.0002) R2 1.00 0.93 1.00 1.00

D-W 1.67

1.09

1.27 1.70 __ neem ne

- 37 -

A likelihood ratio test for parameter constancy was run by reestimating the entire model over two equal subsamples. The test rejected parameter

constancy at the 1 percent level.

Export Shares

The export shares for each country are estimated as systems of equations so that the cross-equation restrictions on the parameters Ty can be imposed. Because the share equations sum to unity, as do the share data, the last equation in the system is omitted from the estimation since its residual is a linear combination of the other residuals. Estimation is by

FIML, treating the relative prices across countries as exogenous variables.

1

t

60. SHR1,~ ¥) + T))*SHRL_ | + 1 #Log(z .

*PGNPL/E,*PCNP S|

1 1.3 3 + 1 *log[etePonPteesponp?)

1 1.2 2 61.. SHR2)= Vo + To9*SHR2, 1 - 1 p#log[BExPcNPL/E?*PcNP?|

2 253 3 + T,#log(BC#PoNP2/EesPoNP?)

62. SHR3 = (1-¥) -¥,) - Ty *SHR1 4 - Ty 9*SHR2. oy

1 1.3 3 2 2.3 3 . 7, j*log(Et¥PoNP/E2+PcnP?) - 1, y*log[Ep*PGNP? /E>¥pcNP?)

- 38 -

German Japan ROW USA

¥, -0.003 0.007 0.137 0.063 (0.002) (0.004) 0.062) (0.014) v, 0.014 0.034 0.074 0.006 (0.005) (0.003) 0.031) (0.017) To 0.900 0.876 0.705 0.166 n.a. (0.006) (0.090) (0.166) Too 0.850 0.900 0.778 0.711 (0.059) n.a. (0.079) (0.118) Ty 0.000 - 0,002 0.041 0.005 (0.001) (0.003) (0.013) (0.004) T, 0.003 0.000 0.003 0.018 (0.001) na. (0.017) (0.007) T53 0.019 0.044 0.000 0.009 (0.007) (0.012) oe (0.010) RS 0.91 0.75 0.90 0.46 RS 0.91 0.97 0.88 0.68

In the case of German export shares, 1 and 2 refer to Japan and the United States, respectively. In the case of Japanese export shares, 1 and 2 refer to Germany and the United States. In the case of ROW, 1 and 2 refer to Germany and the United States. Finally, for the United States, 1 and 2 refer to Germany and Japan, respectively.

Two of the lag coefficients had to be constrained to avoid es:cimating unit roots. Two of the relative price coefficients were restricted from

taking the wrong sign. Together, these four restrictions could be rejected

- 39.

at the 1 percent level. A test for parameter constancy also rejected

constant parameters in favor of a break at midsample at the l percent level.

Money Demand

The money demand equations were estimated by two-stage least squares. The instruments are a constant, current government consumption, and one lag each of real money balances, the nominal interest rate, and total

absorption.

63, log (MB. /PA,) = 1) + vr log(MB, _4/PA,_1) + rolog(A,) + r4RS.. Germany Japan ROW USA ry -3.286 -1.805 -0.885 -1.395 (0.814) (0.746) (0.272) (0.190) r, 0.678 0.766 0.768 0.776 (0.081) (0.094) (0.063) (0.038) r., 0.477 0.271 0.139 0.193 (0.116) (0.110) (0.039) (0.026) re, -0.670 -0.840 -0.600 -0.390 (0.210) (0.270) (0.120) (0.050) R“ 0.99 0.99 0.96 0.99 D-W 1.82 2.31 2.25 1.74 cme SPY 20 PY 2

The German equation was estimated via Cochrane-Orcutt in order to correct for serial correlation. The estimated autocorrelation coefficient was 0.63. In order to check for homogeneity of the effects of the real interest rate and the expected inflation rate on money demand, the lagged

rate of inflation was added to the instrument list and the current rate of

- 40 -

inflation was added to the list of regressors. The estimated coefficient on the inflation rate was significant at the 5 percent level in the German equation, but not in the other countries’ equations. A Chow test for parameter constancy (see Fair [1987]) rejected parameter constancy at the 5 percent level for the ROW equation only. The U.S. equation failed at the 10

percent, but not the 5 percent, level.

Monetary and Fiscal Policy

The coefficients of the monetary base and real government consumption equations are "estimated" by the average rate of growth of these variables over the period 1976-87. The (presumed) constant target ratio of bonds to income, BRATIO, is simply the actual value of the ratio of bonds to taxable income in the fourth quarter of 1987. The coefficient w in the tax adjustment equation has been arbitrarily fixed at 0.95.

Obviously, the simple policy rules described here are not very realistic. One of the most attractive features of MX3 is the ability to consider alternative policy rules, both as descriptions of past behavior and as proposals for future policy. By allowing private expectations to fully incorporate the implications of a particular policy rule, we hope to obtain a more accurate characterization of the economy's behavior under that rule,

at least in the long run.

SIMULATION PROPERTIES

This section presents the dynamic response of the model to a simple

monetary shock and a simple fiscal shock. The monetary shock consists of a

2 percentage point increase of the U.S. monetary base in the first quarter of 1988. This shock has a permanent effect on the monetary base due to the simple growth rate rule for monetary policy. The second shock increases U.S. government consumption by 1 percent of total U.S. productive capacity in the first quarter of 1988. This shock also has a permanent effect through the simple fiscal spending rule.

In order to highlight the effect of expectations in MX3, each of these shocks is implemented in two different ways. In the "surprise" scenarios, the monetary or fiscal shock is first announced in the quarter of implementation, 1988:1. In the "anticipated" scenarios, the government announces its intention to change monetary or fiscal policy in 1986:1, eight quartiers before the planned implementation.

The results of these simulations are presented in terms of percentage deviations from the baseline path, except for real net exports, which are presented as deviations from the baseline path in percentages of baseline capacity output. The baseline path uses actual values through the end of 1987 Beginning in 1988 real variables increase at a constant 3 percent rate, prices increase at a 4 percent rate, nominal variables increase at a 7 percent rate, and interest rates and exchange rates are fixed at their 1987:4 values. Residuals are computed for each equation to keep the model on the baseline path. In other words, when the model is simulated with the residuals it tracks the baseline path. In the shock simulations the model is solved with the baseline residuals in addition to the monetary or fiscal

shock. The use of baseline residuals allows us to isolate the effect of the

- 42 policy shock under consideration. ”>

The first simulation is a surprise increase of the monetary base by 2 percent effective 1988:1. Because MX3 incorporates the long-run neutrality of real activity with respect to money, we know that prices must eventually rise by 2 percent and output must return to its baseline value. Figure 1 demonstrates that the model does perform as expected in the long-run. The domestic price, UPGNP, rises steadily from its baseline level to a maximum value of 2.5 percent over baseline in 1990:4. The price level then drops to a value 1.8 percent above baseline in 1993:3 before gradually approaching its long-run equilibrium of 2 percent above baseline.

Movements in the weighted foreign price level, UPGNPW, primarily reflect movements in exchange rates. Consistent with a long-run equilibrium, we expect that foreign prices will be unaffected by a domestic monetary expansion and that exchange rates will rise proportionally to the increase in the monetary base. The weighted foreign price level in domestic currency should therefore rise by 2 percent in the long-run. Figutre 1 shows that UPGNPW jumps almost 2 percentage points in the first quarter before falling sharply by 1 percentage point in the following quarter. UPGNPW remains at about 1 percent above baseline over the next three quart:ers before climbing back up to--and temporarily overshooting--its long-run value. The dynamic behavior of UPGNPW is primarily explained by lower

nominal interest rates in the United States during the first two quarters

25. An alternative approach is to use the baseline implied by the model with future residuals set at their expected value of zero. This procedure is much more computationally intensive. Moreover, to a first approximation, the effect of the policy shocks relative to baseline is unaffected by which baseline is chosen. If the model were linear, the effect of policy shocks would be completely independent of the baseline

- 43 - Figure 1

Surprise Monetary Shock

(Percent Deviation from Baseline)

UPGNPW

0.5

1Oo+

0.5

1992 1995 1986 1989

NETX *

1Oo+

0.5

1 1992 1995 1986 1989

“In percent deviation from baseline capacity output.

1992

1992

0.5

lo+

0.5

1 1995

lot

0.5

1 1995

- 44 -

and higher nominal interest rates in the United States over the following 12 quarters. Because of the open interest rate parity equations, movements in exchange rates are completely explained by interest rate differentials after the impact of the shock.

UGDP jumps about 0.7 percent for the first three quarters before returning gradually to its baseline value. There is a very slight, damped oscillation of UGDP about baseline. The minimum value is reached “n 1991:2 at -0.12 percent of baseline. Net exports increase by 0.15 percent of potential GDP in the first quarter and decline very slowly thereafter.

Next we consider the behavior of MX3 in the face of an anticipated monetary shock. (See Figure 2.) It is important to remember that traditional macro models cannot consider such an experiment since they do not incorporate future expectations.

The domestic price level responds in almost exactly the same manner as before, except that everything is shifted eight quarters earlier. By the time the monetary base jumps in 1988:1, the price level has already risen by 2.1 percent and it continues to overshoot its equilibrium value, reaching a peak of 2.6 percent above baseline in 1989:1.

Weighted foreign prices--and the exchange rates--do not jump up on the announcement of future monetary policy. Rather, UPGNPW rises steacily from its baseline value to a first peak of 2.3 percent in 1988:1. The subsequent dynamics are basically a damped version of the behavior of UPGNPW under the surprise monetary shock.

The stimulative effect of the anticipated monetary shock is smaller in Magnitude, but more persistent, than the effect of the surprise shcck. Despite the presence of rational expectations, there is still a small spike

in UGDP in the quarter of impact, 1988:1, when UGDP jumps to 0.5 percent

- 45 - Figure 2 Anticipated Monetary Shock

(Percent Deviation from Baseline)

UPGNP

2.5

1.5

0.5

1o+

0.5

1 1986 1989 1992 1995

UGDP

0.5

lOo+

1 1986 1989 1992 1995

UPGNPW

1986

NETX *

1986

*In percent deviation from baseline capacity output.

1989

1989

1992

1992

1995

0.5

lo+

0.5

2.6

1.5

0.5

1Oo+

0.5

1 1995

- 46 -

above baseline. A cyclical trough is reached in 1989:4 at -0.16 percent of baseline output. Net exports rise steadily to a peak of 0.18 percent of baseline capacity output in 1988:1 and fall slowly back to baseline.

The other shock to be presented is a permanent increase in zovernment consumption by 1 percent of capacity output. Figure 3 shows the behavior of the same four variables in response to a surprise fiscal shock. The domestic price, UPGNP, rises to a peak 0.5 percent above baseline in 1989:3 before falling back to baseline in 1992:1. Beginning in 1993:1 UPGNP gradually rises to a value 0.3 percent above baseline, where it remains permanently. UPGNPW initially drops 0.6 percent below baseline. In subsequent periods it quickly returns to oscillate about the baseline. What is not evident in Figure 3 is that UPGNPW eventually rises to a level about 0.3 percent above baseline after several more years. UGDP jumps 0.7 percent above baseline in the initial period, but it quickly drops below the baseline and it gradually settles at about 0.3 percent below baseline. Finally, net exports drop about 0.1 percent of baseline capacity output by the second period and then begin a gradual return to baseline.

In order to understand the long-run effects of this fiscal shock, we must review the consumption and investment relations in MX3. The increased government consumption necessitates an eventual rise in the tax rate in order to maintain the ratio of bonds to income. The higher tax rate reduces disposable income, and thus consumption, by an amount equal to the rise in government spending. Thus, to a first approximation, government spending fully crowds out private consumption, leaving all other variables unaffected in the long run. This analysis ignores a secondary effect of higher taxes,

however. The secondary effect works through the investment equation.

- 47 -

Figure 3

Surprise Fiscal Shock (Percent Deviation from Baseline)

UPGNP

1.5

0.5

lo+

0.5

1 1986 1989 1992 1995

UGDP

0.5

lOo+

0.5

1 1986 1989 1992 1995

UPGNPW

1986

NETX *

1986

“In percent deviation from baseline capacity output.

1989

1989

1992

1992

1995

0.5

1o+

0.5

1.5

0.5

lot

0.5

1 1995

According to equation (10) an increase in the tax rate will reduce the long-

run desired capital stock. 2°

It is the secondary effect that explains the permanent drop in UGDP evident in Figure 3. With lower absorption and a fixed monetary base, UPGNP must rise to reduce the level of real money balances. UPGNPW must also rise in the long-run to equilibrate relative prices. When relative prices across countries return to their baseline values and U.S. absorption equals U.S. production capacity, both real and nominal trade flows will reequilibrate. During the transition period the United States runs a real trade deficit, but the favorable terms of trade help to minimize both the nominal trade deficit and the associated decline in net foreign assets.

It is interesting to note that even in the short run, fiscal policy has very little expansionary effect in MX3. Both consumption and investment drop immediately when government spending rises. This crowding out occurs because consumers and investors know that tax rates will rise in the future in order to satisfy the government's intertemporal budget constraint. The strength of this crowding out is somewhat surprising in light of the fact that private agents in MX3 discount the future much faster than the real rate of interest on government bonds. Blanchard {1985] showed that such a wedge between private and government discount rates can reduce crowding out of fiscal policy.

Figure 4 shows that the announcement of a future fiscal expansion is strongly contractionary. Domestic output and domestic prices both drop

steadily until 2 quarters before the period of implementation. This

26. This effect hinges on the assumption that the risk premium is constant in after-tax terms. If the risk premium is constant before taxes, then the tax rate does not affect the desired capital stock.

- 49 - Figure 4

Anticipated Fiscal Shock (Percent Deviation from Baseline)

UPGNP 2 UPGNPW 15 1 0.5 + NS ° 0.5 1986 1989 1992 1995 1986 1989 UGDP 3 NETX * 1.5 0.5 + oO 0.5 1986 1989 1992 1995 1986 1989

*In percent deviation from baseline capacity output.

1992

1992

1.5

0.5

1Oo+

0.5

1 1995

0.5

1o+

0.5

1 1995

contraction results from the forward-looking behavior of investors and consumers, who foresee a long-run decline in desired capital anc. disposable income. After implementation, the expansionary effect of the ariticipated policy is only about two-thirds that of the surprise policy. The long-run

effects are the same for both surprise and anticipated fiscal expansions.

CONCLUSION

This paper has presented the theoretical structure, empirical implementation, and simulation properties of the MX3 model. The structure of MX3 represents a significant step toward incorporating more economic theory in macroeconometric models. In particular, agents are assumed to have rational expectations; short-run dynamics are constrained to. resemble behavior under costly adjustment; and the economy moves toward a competitive steady state in the long run.

The implementation of MX3 has been largely successful, but more work clearly remains to be done. Probably the first areas for further work are the price adjustment and fixed investment equations. Eventually, it would be desirable to expand the geographic coverage of the model so that it captures a more accurate description of global feedbacks to domestic policies.

While the simulations presented in this paper are of some interest for the insights they provide on the properties of MX3, the proposed shocks and the associated policy rules are not wholly satisfactory. It is clearly the case that monetary and fiscal instruments do not evolve exogenously with

respect to the rest of the economy. Rather, the monetary and fiscal

authorities must be responding in some manner to the shocks that originate in the rest of the economy. They also may respond to evidence on the effect their policies are having on the economy. A realistic policy rule must therefore include some reaction of the policy instruments to information on the economy that the authorities have available at the time the policy instruments are set.

One experiment that is particularly attractive--although it is not explored in this paper--is to consider the dynamic properties of the model under alternative policy rules with stochastic simulations. The objective is to discover the macro policy rules that are best able to stablize output, inflation, or other target variables in the face of shocks similar to those that typically occur. It is particularly important to use a rational expectations model when searching over alternative policy rules if one believes that the private sector will eventually learn about any new rule

and alter its behavior accordingly.

- 52 -

REFERENCES

Abel, Andrew B., N. Gregory Mankiw, Lawrence H. Summers, and Richard J. Zeckhauser, (1987) "Assessing Dynamic Efficiency: Theory and Evidence," National Bureau of Economic Research, Working Paper No. 2097.

Andrews, Donald W.K., and Ray C. Fair, (1988) “Inference in Nonlinear Econometric Models with Structural Change," Review of Economic Studies, 55, pp. 615-640.

Bailey, Victor B., and Sara R. Bowden, (1985) Understanding United States Foreign Trade Data, (Washington: U.S. Department of Commerce).

Bils, Mark, (1985) "Real Wages over the Business Cycle: Evidence from Panel Data," Journal of Political Economy, 93, pp. 666-689.

Blanchard, Oliver J., (1985) "Debt, Deficits, and Finite Horizons," Journal of Political Economy, 93, pp. 223-247.

Chouraqui, J.C., B. Jones, and R.B. Montador, (1986) "Public Debt in a Medium-Term Perspective," OECD Economic Studies, No. 7, pp. 103-153.

Dooley, Michael, and Peter Isard, (1982) "A Portfolio-Balance Rational- Expectations Model of the Dollar-Mark Exchange Rate," Journal of International Economics, 12, pp.257-276.

Dunlop, John T., (1938) "The Movement of Real and Money Wage Rates," Economics Journal, 48, pp. 413-434.

Fair, Ray C., (1987) "International Evidence on the Demand for Money," National Bureau of Economic Research, Working Paper No. 2106.

Frankel, Jeffrey A., (1983) "Monetary Portfolio-Balance Models of Exchange Rate Determination," in Jagdeep S. Bhandari and Bluford H. Patnam

(eds.) Economic Interdependence and Flexible Exchange Rates,

(Cambridge, Massachusetts: MIT Press).

Ghysels, Eric, and Alastair Hall, (1988) "A Test for Structural Stability of Euler Conditions Parameters Estimated Via the Generalized Method of Moments Estimator," unpublished paper, Universite de Montreal.

Goldfeld, Stephen M., and Daniel E. Sichel, (1987) "Money Demand: The Effects of Inflation and Alternative Adjustment Mechanisms," The Review of Economics and Statistics, 69, pp. 511-515.

Hall, Robert E., (1988) "Intertemporal Substitution in Consumption, " Journal of Political Economy, 96, pp. 339-357.

Hansen, Lars P., and Kenneth J. Singleton, (1982) "Generalized Instrumental Variables Estimation of Nonlinear Rational Expectations Models," Econometrica, 50, pp. 1269-1286,

Kydland, Finn E., and Edward C. Prescott, (1982) "Time to Build and Aggregate Fluctuations," Econometrica, 50, pp. 1245-1369.

- 53 -

Lucas, Robert E., (1976) "Econometric Policy Evaluation: A Critique," Journal of Monetary Economics, Supplement, pp. 19-46.

Masson, Paul, Steven Symansky, Richard Haas, and Michael Dooley, (1988) "MULTIMOD: A Multi-Region Econometric Model," IMF Working Paper No. 88/23, International Monetary Fund.

Mehra, Rajnish, and Edward C. Prescott, (1985) "The Equity Premium: A Puzzle," Journal of Monetary Economics, 15, pp. 145-161.

Nason, James, (1989) "Permanent Income, Current Income, Consumption, and Changing Tastes," unpublished paper, Federal Reserve Board.

Poterba, James M., and Lawrence Summers, (1987) "Finite Lifetimes and the

Effects of Budget Deficits on National Savings," Journal of Monetary Economics, 20, pp. 369-391.

Roberts, John M., (1987) "Two Studies of Price and Marginal Cost in U.S. Manufacturing Industry," Ph.D. Dissertation, Stanford University.

Sargent, Thomas J., (1978) "Estimation of Dynamic Labor Demand Schedules

under Rational Expectations," Journal of Political Economy, 86, pp. 1009-1044.

Taylor, John B., (1980) "Aggregate Dynamics and Staggered Contracts," Journal of Political Economy, 88, pp. 1-23.

_. (1986) “Improvements in Macroeconomic Stability: The Role of Wages

and Prices," in Robert J. Gordon (ed.) The American Business Cycle:

Continuity and Change (Chicago: University of Chicago Press) pp. 639- 669,

_, (1988) "The Treatment of Expectations in Large Multicountry Econometric Models," in Ralph C. Bryant, Dale W. Henderson, Gerald Holtham, Peter Hooper, and Steven A. Symansky (eds.) Empirical

Macroeconomics for Interdependent Economies (Washington, DC: The Brookings Institution) pp. 161-182.

World Economic Outlook, (1988) (Washington, DC: International Monetary Fund) April.

- 54 -

APPENDIX 1: SIMPLIFIED COUNTRY MODEL

I. Data Definitions

A

AW

B BRATIO Cc

CAP CAPTOT CAPW CC

CU DPA DPGNP

GDP GNP IF II IM

MB NFA PA PEX PGNP PGNPW PIM

RS SHRx

Absorption

Weighted Foreign Absorption National Debt

Target Ratio of National Debt to Taxable Income Private Consumption

Capacity Output

World Capacity Output

Weighted Foreign Capacity Output Cost of Capital

Capacity Utilization Rate Inflation Rate (Absorption) Inflation Rate (GNP)

Exchange Rate

Export Volume

Government Consumption

Gross Domestic Product

Gross National Product

Gross Fixed Investment

Inventory Investment

Import Volume

Net Capital Stock

Labor Force (exogenous)

Monetary Base

Net Foreign Assets

Absorption Deflator

Export Deflator

GNP Deflator

Weighted Foreign GNP Deflator Import Deflator

Production Technology (exogenous, estimated) Short-term Nominal Interest Rate

Share of Total Exports Destined for Country x

- 55 -

TAU Tax Rate

TAX Tax Revenues

TBAR Tax Rate Required to Hit Target Debt Ratio TI Taxable Income

YD Disposable Income

II. Private Sector Demand

Private Consumption 1.C= (vp, RS, DPA, TAU) Fixed Investment | 2. IF = IF (GDP, CC, TAU, R) Inventory Investment 3. II = 11 (AcDP, RS, DPA) Money Demand 4. MB = MB(A, RS) * PA

IIl. Aggregate Supply

GNP Deflator (Inflation Rate) 5. DPGNP = P(DPGNP, cu)

Capacity Output 6. CAP = F(Q, K, L)

IV. Exchange Rate and Trade

Exchange Rate 7. aEt = RS - rs! Expo::t Volume 8. EX = EX (AW/CAPW, PEX/PIM, CAPTOT) Export Price 9. PEX = PEX(PGNP, PGNPW, t}

Expoxt Shares 10. SHRI = siRi(etepene", ...,z°«ponp’]

V. Monetary and Fiscal Policy

Money Supply Government Consumption Target Ratio of National Debt

Tax Rate

VI. Identities and Definitions

Imports

Import Prices

Absorption

Gross Domestic Product Gross National Product

Disposable Income

Absorption Deflator

Inflation Rate (Absorption)

Inflation Rate (GNP)

- 56 -

ll.

12.

13. 14. 15.

16.

17.

18.

19.

20.

21.

22.

23.

24. 25.

SHR2 = sur2[et*pne?, ... e*xpenr’|

SHR3 = 1 - SHR1 - SHR2

AMB = m AG = g BRATIO = b

TAU = 0.95 * TAU_, + 0.05 * TBAR

1

IM = sHRO! «Ex! + SHRO?*EX? + SHRO°*EX?

1 1 1

PIM (surolen *PEX *EX” + ,

SHRO°*E?*PEX?*EX?] / IM

+

A=C+IF+II+G

GDP = A + EX - IM

GNP

GDP + RS*NFA_,/PGNP

YD = PGNP*GNP/PA - §*K_, - TAX/PA

+ (RS - DPA)*B_|/PA - RS*MB_,/PA

PA*A = PGNP*GNP - PEX*EX + PIM*IM

- * RS*NFA |

DPA = APA

DPGNP = APGNP

- 57 -

Taxable Income 26. TI = PGNP*GNP - §*K_,*PA + RS*B | Tax Revenues 27. TAX = TAU*TI Equilibrium Tax Rate 28. TBAR = (c*PA + RS* (By - MB_1)) / TI

- BRATIO + B_,/TI_,

Cost of Capital 29. CC = (1-TAU) * (RS + 6 - DPA) +x Capacity Utilization 30. CU = GDP/CAP Total World Capacity 31. CAPTOT = CAP + cap! + cap? + CAP? Weighted Foreign Absorption 32. AW = [at}oh [7]? [?}? Weighted Foreign Capacity 33. CAPW. = [car ret * (car? 2 |e * [car?}? Weighted Foreign Prices 34. PGNPW = [rowel ra [rone? wet]?

34.53] 03

* [powp3s E?| Capital Stock 35. K= (1 - §)*K | + IF Government Bonds 36. B= (1 + RS)*B_y + PA*G - TAX - RS*MB_)

Net Foreign Assets 37. NFA = (1 + RS)NFA_} + PEX*EX - PIM*IM

Note: Superscripts. denote foreign country variables. Variables with hats are future expectations. A denotes the percentage rate of change of a variable. 6 is the rate of depreciation of fixed capital. m is the risk premium for holding capital instead of government bonds. wl-w3 are fixed weights that sum to unity.

APPENDIX 2: MODEL LISTING AND DATABASE*

This appendix is divided into four main parts. The first three parts comprise a complete listing of the current version of the MX3 model: the equation listing, the variable and equation cross-reference table, and the variable definitions. The final part is a detailed description of the MX3 database.

The equation listing is grouped by country and then by sector. In this way, each country model is presented as though it were a separate

macroeconometric model using the following format:

1. Private Sector Demand

2. Aggregate Supply

3. Exchange Rate and Trade

4. Monetary and Fiscal Policy

5. Identities and Definitions

The sectors and variables in each country model are preceded by the first letter of the country name. For example, "G.1" is the first sector of the German country model (private sector demand). Likewise, "GGNP" is German gross national product. Although equations are grouped within sectors, equations are numbered consecutively throughout the whole model. Each equation is reported with its coefficient values and labeled with the associated left-hand side variable. (Estimation results and test statistics

are reported in the text of the paper).

1. Special thanks are due to Gwyn Adams for preparing this appendix.

The cross-reference table gives the number of each equation in which each variable in the model appears. It is grouped by endogenous and exogenous variables. The variable listing gives each variable alphabetically with a short definition. The variable listing is also grouped by endogenous and exogenous variables.

The variable naming convention uses the first letter of the variable name to indicate the country. The middle portion of the name describes the variable and the presence of a trailing "V" or "W" indicates a nominal value or fixed weighted value respectively. The existence of a final "_ERR" is used to indicate the error term for the equation describing the indicated endogenous variable. For example, JGDEBTV is current Japanese government debt, and GIF_ERR is the residual term in the German fixed investment equation. Also, the bilateral trade shares are represented in either a "XijS" or "XijS3" format. Both represent the share of country i’s exports

destined for country j. The trailing 3 indicates the export share with ROW

for G-3 countries.

GERMAN MODEL

1. GC: Private consumption expenditure - 1980 prices GC/GCAP = .852 * GC(-1)/GCAP(-1)/(1 + .852 /(1 + .2 + (1 - GTAU) * GRS/400 - GDPA(1)/400)) + GC(1)/GCAP(1)/(1 + .852 + .02 + (1 - GTAU) * GRS/400 - GDPA(1)/400) + (1 - .852 ) %* .00985 * GYD/GCAP + GC_ERR

2. GIF: Total fixed investment - 1980 prices

(l + .9& %* .97 ) * GIF/GCAP = .95 * GIF(-1)/GCAP(-1) + .97 * GIF(1)/GCAP(1) + (1 - .$5 )* (1 - .97 3 * (€ .35 * GCU * 100/(GRS + 4.83 - GDPA(1) + 4.78 /(1 - GTAUJ) - (1 - 4.83 7100) * GK(-1)/GCAP) + GIF_ERR

3. GII: Inventory investment - 1980 prices

GII/GCAP = .00733 + .483 * (GCU(1) - GCU) - .000778 * (GRS - GDPA(1)) + GII_ERR

4. GMB: Monetary base

LOG(GMB/GPA ) = -3.2863 + .6776 * LOG(GMB(-1)/GPA(-1)) - .0067 * GRS + .4771 * LOG(GA) + GMB_ERR

G2. AGGREGATE SUPPLY

5. GPGNP: Gross national product deflator - 1980=100.00

(GPGNP/GPGNP(-1) - 1) * 400 = .5 * (GPGNP(-1)/GPGNP(-2) - 1) * 400 + (1 - .5 J ® (GPGNP(1)/GPGNP - 1) * 400 + 10 * LOG(GCU) + GPGNP_ERR

6. GCAP: Total capacity output

GCAP = GQ * GK(-1)* * .35 * GLF(-1)**(1 - .35 ) + GCAP_ERR

G3. EXCHANGE RATE AND TRADE

7. GER: Spot exchange rate - US$/DM

(GER(1)/GER - 1) * 400 = URS - GRS + GER_ERR

8. GXGSNI: Exports - NIA basis - 1980 prices LOG( GXGSNI ) = -.612 + .775 * LOG(GXGSNI(-1)) + .688 * LOG(GAW/GCAPW) - .26 * LOG(GPXGSNI/GPMGSNI) + (1 - .775 ) * 1 %* LOG(CAPTOT) + GXGSNI_ERR

9. GPXGSNI: Export deflator - NIA basis - 1980 prices LOG( GPXGSNI ) -015 + .734 * LOG(GPXGSNI(-1)) + .076 * LOG(GPGNPW) + (1 - .734 - .076 ) * LOG(GPGNP) + .408 * LOG(GPXGSNI(-1)/GPXGSNI(-2)) - .0025 * (1 - .734 ) * TIME + GPXGSNI_ERR

10. XGUS: Share of German exports destined for US

XGUS = .01¢ + .85 * XGUS(-1) + O * LOG(UPGNP/( JER * JPGNP)) + .019 * LOGCUPGNP/(RER * RPGNP)) + XGUS_ERR

11. XGJS: Share of German exports destined for Japan

XGJS = -.003 + .9 ® XGJS(-1) - O * LOG(UPGNP/( JER * JPGNP)) + .003 * LOG{ JER * JPGNP/(RER * RPGNP)) + XGJS_ERR

12. XGRS3: Share of German exports destined for ROW

XGRS3 = 1 - .014¢ - -.003 - .85 * XGUS(-1) - .9 * XGUS(-1) - .019 * LOG(UPGNP/(RER * RPGNP)) - .003 * LOG(JER * JPGNP/(RER * RPGNP)) + XGRS3_ERR

G4. MONETARY AND FISCAL POLICY

13.

14.

15.

1é.

GRS: 3-month Treasury Bill rate

GMB = 1.0155 * GMB{-1) + GRS_ERR

GG: Real government purchases - 1980 prices

GBRATIO: Target ratio of government bonds to taxable income

GBRATIO = 0.262 + GBRATIO_ERR

GTAU: Actual income tax rate

GTAU = .95 * GTAUC-1) + (1 - .95 ) * GTBAR + O * LOG(GCU) + GTAU_ERR

-7-

GS. IDENTITIES, AND DEFINITIONS

17. GMGSNI: Imports - NIA basis - 1980 prices

GMGSNI = (XUGS * UXGSNI * 0.9169 + XJGS * JXGSNI * 4.424 + 1.181 * XRGS3 % RXGSNI * 1.173)/0.5505 + GMGSNI_ERR

18. GPMGSNI: I:mport deflator - NIA basis - 1980=100.00

GPMGSNI = (XUGS * UPXGSNI * UXGSNI + XJGS * JER * JPXGSNI * JXGSNI + 1.181 % XRGS3 * RER * RPXGSNI * RXGSNI)/(GMGSNI * GER) + GPMGSNI_ERR

19. GA: Absorption

GA = GC + GIF + GG + GII + GA_ERR

20. GGDP: Gross domestic product - 1980 prices

GGDP = GA + GXGSNI - GMGSNI + GGDP_ERR

21. GGNP: Gross national product - 1980 prices

GGNP = GGDP + URS * GNFAV(-1)/(GER * GPGNP) + GGNP_ERR

22. GYD: Disposable income - 1980 prices

GYD = GPGNP * GGNP/GPA - &.83 *% GK(-1)/100 - GTAXV * 100/GPA + (GRS - GDPA) * GGDEBTV(-1}3/GPA —- GMB(-1} * GRS/GPA + GYD_ERR

23. GPA: Gross domestic product deflator ~ 1980=100.00

GPA = (GPGNP * GGNP - GPXGSNI * GXGSNI + GPMGSNI * GMGSNI - URS * GNFAV(-1)/GER)/GA + GPA_ERR

24. GDPA: Rate of inflation of absorption prices

GDPA = (GPA/GPA(-1) - 1) * 400 + GDPA_ERR

25. GDPGNP: Rate of inflation of output prices

GDPGNP =: (GPGNP/GPGNP(-1) - 1) * 400 + GDPGNP_ERR

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

GTIV: Nominal taxable income

GTIV = GPGNP * GGNP/100 - 4.83 * GK(-1) * GPA/10000 + GRS * (GGDEBTV(-1) - GMB(-1))/100 + GTIV_ERR

GTAXV: Nominal tax revenues

GTAXV = GTAU * GTIV + GTAXV_ERR

GTBAR: Equilibrium tax rate

GTBAR = (GG * GPA + GRS * GGDEBTV(-1))/(GTIV * 100) - GBRATIO + GGDEBTV( ~-1)/GTIV 4 GTBAR_ERR

GCU: Capacity utilization rate

GCU = GGDP/GCAP + GCU_ERR

GAW: Trade weighted foreign absorption

GAH = UA®*0.0767 * JAX*0.0134 * RAX*0.9099 + GAW_ERR

GCAPH: Trade weighted foreign total capacity output

GCAPH = UCAP**0.0767 * JCAP**0.0134 * RCAP**0.9099 + GCAPW_ERR

GPGNPH: Trade weighted foreign gross national product deflator - 1980=100.00

GPGNPH = (UPGNP/0.8572 )**0.0767 * (JPGNP * JER/4.424)**0.013¢ % (RPGNP x RER/1.173 )**0.9099 * 0.5505/GER + GPGNPW_ERR

GK: Total net capital stock

GGDEBTV: Current total government debt

GGDEBTV = (1 + GRS/400) * GGDEBTV(-1) + GPA * GG/4G00 - GTAXV/4 - GMB(-1) x GRS/400 + GGDEBTV_ERR

GNFAV: Net foreign assets

GNFAV = (1 + URS/400) * GNFAV(-1) + (GPXGSNI * GXGSNI * GER - GPMGSNI x GMGSNI * GER)/400 + GNFAV_ERR

JAPANESE MODEL

36. JC: Private consumption expenditure - 1980 prices JC/JCAP = .852 * JC(-1L)J/JCAP(-1)/(1 + .852 7(1 + .02 + (1 - JTAU) * JRS/400 - JDPA(1)/400)) + JCCLI/JCAP(1)I/(1 + .852 + .02 + (1 - JTAU) * JRS/400 - JDPA(1)/400) + (1 - .852 ) * .00985 * JYD/JCAP + JC_ERR

37. JIF: Total fixed investment - 1980 prices

(1 + .95 * .97 J) * JIF/JCAP = .95 * JIF(-1)/JCAP(-1) + .97 * JIFC1)I/JCAP(1) + (1 - .95 }*% (1 - .97 3} ® € .39 * JCU * 100/(JRS + 3.04 - JDPA(1) + 6.71 /(1 - JTAU)) - (1 - 3.04 7100) * JK(-1)/JCAP) + JIF_ERR

38. JII: Inventory investment - 1980 prices

JII/JCAP = .00733 + .483 * (JCU(1) - JCU) - .000778 * (JRS - JDPA(1)) + JII_ERR

39. JMB: Monetary base

LOG( JMB/JPA) = -1.8045 + .766 * LOG(JMB(-1)/JPA(-1)) - .0084 * JRS + .2707 * LOG(JA) + JMB_ERR

- 10 -

J2. AGGREGATE SUPPLY

40. JPGNP: Gross national product deflator - 1980=100.00

(JPGNP/JPGNP(-1) - 1) * 400 = .5 * (JPGNP(-1)/JPGNP(-2) - 1) * 400 + (1 - .5 ) * (JPGNP(1)/JPGNP - 1) * 400 + 10 * LOG(JCU) + JPGNP_ERR

G1. JCAP: Total capacity output

JCAP = JQ * JK(-1)% * .39 ® JLF(-1)%*(1 - .39 ) + JCAP_ERR

-ll-

J3. EXCHANGE RATE AND TRADE

42. JER: Spot exchange rate - US$/DM

(JERC(1)/JER - 1) * 400 = URS - JRS + JER_ERR

43. JXGSNI: Exports - NIA basis - 1980 prices LOG( JXGENT ) = -.301 + .937 * LOG(JXGSNI(-1)) + .418 * LOG(JAW/JCAPNW ) - .107 * LOG( JPXGSNI/JPMGSNI) + (1 - .937 ) * 1 * LOG(CAPTOT) + JXGSNI_ERR

44. JPXGSNI: Export deflator - NIA basis - 1980 prices LOG( JPXGSNI ) = .005 + .824 * LOG(JPXGSNI(-1)) + .12 * LOG(JPGNPH) + (1 - .824¢ - .12 ) % LOG(JPGNP) + .136 * LOG( JPXGSNI(-1)/JPXGSNI(-2)) + (1 - .824 ) * -.0025 * TIME + JPXGSNI_ERR

45. XJUS: Shere of Japanese exports destined for US

XJUS = .034¢ + .9 * XJUS(-1) + .002 * LOG(UPGNP/(GER * GPGNP )) + .044 % LOG(UPGNP/(RER * RPGNP)) + XJUS_ERR

46. XJGS: Shere of Japanese exports destined for Germany

XJGS = .007 + .876 * XJGS(-1) - .002 * LOG(UPGNP/(GER * GPGNP )) + O * LOG(GER * GPGNP/(RER * RPGNP)) + XJGS_ERR

47. XJRS3: Share of Japanese exports destined for ROW

XJRS3 = 1 - .034¢ - .007 - .876 * XJGS(-1) - .9 * XJUS(-1) - .044 * LOG(UPGNP/(RER * RPGNP)) - 0 * LOG(GER * GPGNP/(RER * RPGNP)) + XJRS3_ERR

- 12 -

J4. MONETARY AND FISCAL POLICY

48. JRS: 3-month Treasury Bill rate

JMB = 1.0178 * JMB(-1) + JRS_ERR

49. JG: Real government purchases - 1980 prices

JG = 1.0085 * JG(-1) + JG_ERR

50. JBRATIO: Target ratio of government bonds to taxable income

JBRATIO = 0.283 + JBRATIO_ERR

51. JTAU: Actual income tax rate

JTAU = .95 * JTAU(-1) + (1 - .95 ) * JTBAR + O * LOG(JCU) + JTAU_ERR

-13-

J5. IDENTITIES AND DEFINITIONS

52. JMGSNI: Imports - NIA basis - 1980 prices

JMGSNI = (XUJS * UXGSNI * 0.9169 + XGUS *% GXGSNI x 0.5505 + 1.355 * XRJS3 % RXGSNI * 1.173)/4.424 + JMGSNI_ERR

53. JPMGSNI: Import deflator - NIA basis - 1980=100.00

JPMGSNI = (XUJS * UPXGSNI * UXGSNI + XGUS * GER x GPXGSNI * GXGSNI + 1.355 % XRJS3 * RER * RPXGSNI * RXGSNI)/(JMGSNI * JER) + JPMGSNI_ERR

54. JA: Absorption

JA = JC + JIF + JG + JII + JA_ERR

55. JGDP: Gross domestic product - 1980 prices

56. JGNP: Gross. national product - 1980 prices

57. JYD: Disposable income - 1980 prices

JYD = JPGNP * JGNP/JPA ~ 3.04 * JK(-1)/100 - JTAXV * 100/UPA + (URS - JDPA) x JGDEBTV(-1)/JPA - JMB(-1) * JRS/JPA + JYD_ERR

58. JPA: Gross domestic product deflator - 1980=100.00

JPA = (JPGNP * JGNP - JPXGSNI * JXGSNI + JPMGSNI x JMGSNI - URS * JNFAV(-1)/JER)/JA + JPA_ERR

59. JDPA: Rate of inflation of absorption prices

JDPA = (JPA/JPA(-1) - 1) * 400 + JDPA_ERR

60. JDPGNP: Rate of inflation of output prices

JDPGNP = (JPGNP/JPGNP(-1) - 1) * 400 + JDPGNP_ERR

él.

62.

63.

64.

65.

66.

67.

68.

69.

70.

- 14 -

JTIV: Nominal taxable income

JTIV = JPGNP * JGNP/100 - 3.04 * JK(-1)} * JPA/10000 + JRS * (JGDEBTV(-1) - JMB(-1)})/100 + JTIV_ERR

JTAXV: Nominal tax revenues

JTAXVY = JTAU * JTIV + JTAXV_ERR

JTBAR: Equilibrium tax rate

JTBAR = (JG ® JPA + JRS * JGDEBTV(-1))/(JTIV * 100) - JBRATIO + JGDEBTV( -1)/JUTIV + JTBAR_ERR

JCU: Capacity utilization rate

JCU = JGDP/JCAP + JCU_ERR

JAW: Trade weighted foreign absorption

JAN = UA®*0.2954 * RAX*0.6637 * GAX*0.0409 + JAW_ERR

JCAPW: Trade weighted foreign total capacity output

JCAPH = UCAPX*0.2954 %* RCAPX*0.6637 * GCAP%*0.0409 + JCAPW_ERR

JPGNPH: Trade weighted foreign gross national product deflator - 1980=100.00

JPGNPWH = (UPGNP/0.8572 )**0.2954 * (RPGNP * RER/1.173 )**0.6637 * (GPGNP x GER/0 .5505 )**0.0409 * 4.424/JER + JPGNPW_ERR

JK: Total net capital stock

JGDEBTV: Current total government debt

vGDEBTV = (1 + JRS/400) * JGDEBTV(-1) + JPA * JG/400 - JTAXV/G - JMB(-1) x JRS/400 + JGDEBTV_ERR

JNFAV: Net foreign assets

JNFAV = (1 + URS/G00) * JNFAV(-1) + (JPXGSNI * JXGSNI * JER - JPMGSNI * JMGSNI * JER)/400 + JNFAV_ERR

- 15 -

ROW MODEL

71. RC: Private consumption expenditure - 1980 prices

RC/RCAP = .852 * RC(-1)/RCAP(-1)/(1 + .852 /(1 + .02 + (1 - RTAU) * RRS/G00 - RDPA(1)/400)) + RC(1)/RCAP(1)/(1 + .852

+ .02 + (1 - RTAU) * RRS/G00 — RDPA(1)/400) + (1 - .852 ) * .00985 * RYD/RCAP + RC_ERR

72. RIF: Total fixed investment - 1980 prices

(1 + .95 *% .97 ) * RIF/RCAP = .95 % RIF(-1)/RCAP(-1) + .97 ® RIF(1)J/RCAP(1) + (1 - .95 )* (1 - 1.97 3 * ( .35 % RCU ¥ 1OO/(RRS + 6.58 - RDPA(1) + 6.71 /(1 - RTAU)) - (1 - 6.58 7100) * RK(-1)/RCAP) + RIF_ERR

73. RII: Inventory investment - 1980 prices

RII/RCAP = .00733 + .483 * (RCU(1) - RCU) - .000778 * (RRS - RDPA(1)) + RII_ERR

74. RMB: Monetary base

LOG( RMB./’RPA ) = -.8852 + .7683 * LOG(RMB(-1)/RPA(-1)) - .006 * RRS + .139 * LOG(RA) + RMB_ERR

-16-

R2. AGGREGATE SUPPLY

75. RPGNP: Gross national product deflator - 1980=100.00

CRPGNP/RPGNP(-1)} - 1) * 400 = .5 % (RPGNP(-1)/RPGNP(-2) - 1) * 400 + (1 - .5 ) % (RPGNP(1)/RPGNP - 1) * 400 + 10 * LOG(RCU) + RPGNP_ERR

76. RCAP: Total capacity output

RCAP = RQ * RK(-1)% * .35 % RLF(-1)%*(1 - .35 } + RCAP_ERR

-17-

R3. EXCHANGE RATE AND TRADE

77. RER: Spot exchange rate - US$/DM

CRERC1)/RER - 1) * 400 = URS - RRS + RER_ERR

78. RXGSNI: Exports - NIA basis - 1980 prices LOG! RXGSNI ) = -.373 + .864¢ * LOG(RXGSNI(-1)) + .268 x LOG( RAW/RCAPW ) - .12 * LOG(RPXGSNI/RPMGSNI) + (1 - (864) x 1 * LOG(CAPTOT) + RXGSNI_ERR

79. RPXGSNI: Export deflator - NIA basis - 1980 prices LOG( RPXGSNI ) = .008 + .985 * LOG(RPXGSNI(-1)) + 0 * LOG(RPGNPH) + (1 - - 985 - © ) * LOG(RPGNP) + .527 * LOG(RPXGSNI(-1)/RPXGSNI(-2)) + (1 - .985 ) * -.0025 * TIME + RPXGSNI_ERR

80. XRUS3: Share of ROW exports destined for US

XRUS:3 = .074 + .778 * XRUS3(-1) + .0¢1 * LOG(UPGNP/(GER x GPGNP )) + O * LOG(UPGNP/( JER * JPGNP)) + XRUS3_ERR

81. XRGS3: Share of ROW exports destined for Germany

XRGS3 = .137 + .705 * XRGS3(-1) - .041 * LOG(UPGNP/I(GER x GPGNP ) ) + .003 * LOG(GER * GPGNP/( JER * JPGNP)) + XRGS3_ERR

82. XRJS3: Share of ROW exports destined for Japan

XRJS3 = 1 - .074 - .137 - .778 * XRUS3{-1) - .705 x XRGS3(-1) - 0 * LOG(UPGNP/(JER * JPGNP)) - .003 * LOG(GER * GPGNP/{(JER * JPGNP}) + XRJS3_ERR

83.

84.

85.

86.

- 18 -

- MONETARY AND FISCAL POLICY

RRS: 3-month Treasury Bill rate

RMB = 1.0297 * RMB(-1) + RRS_ERR

RG: Real government purchases - 1980 prices

RG = 1.005 * RG(-1) + RG_ERR

RBRATIO: Target ratio of government bonds to taxable income

RBRATIO = 0.569 + RBRATIO_ERR

RTAU: Actual income tax rate

RTAU = .95 * RTAU(-1) + (1 - .95 ) * RTBAR + O * LOG(RCU) + RTAU_ERR

- 19 -

R5. IDENTITIES AND DEFINITIONS

87. RMGSNI: ‘Imports - NIA basis - 1980 prices

RMGSNI = (XURS3 * UXGSNI * 0.9169/1.3 + XGRS3 * GXGSNI * 0.5505/1.183 + XJRS3 * JXGSNI * 4.424/1.243)/1.173 + RMGSNI_ERR

88. RPMGSNI: Import deflator - NIA basis - 1980=100. 00

RPMGSNI = (XURS3 * UPXGSNI * UXGSNI/1.3 + XGRS3 * GER * GPXGSNI x GXGSNI 71.183 + XJRS3 * JER * JPXGSNI * JXGSNI/1.243)/(RMGSNI * RER) + RPMGSNI_ERR

89. RA: Absorption

RA = RC + RIF + RG + RII + RA_ERR

90. RGDP: Grass domestic product - 1980 Prices

91. RGNP: Gross national product - 1980 Prices

RGNP = RGDP + URS * RNFAV(-1)/(RER * RPGNP) + RGNP_ERR

92. RYD: Disposable income - 1980 prices

RYD = RPGNP * RGNP/RPA - 5.58 * RK(-1)/100 - RTAXV * 100/RPA + (RRS - RDPA) * RGDEBTV(-1}/RPA - RMB(-1) x RRS/RPA + RYD_ERR

93. RPA: Gros: domsetic product deflator - 1980=100.00

RPA = '\RPGNP * RGNP - RPXGSNI * RXGSNI + RPMGSNI * RMGSNI - URS * RNFAV(-1)/RER)/RA + RPA_ERR

94. RDPA: Rate of inflation of absorption prices

RDPA = (RPA/RPA(-1) - 1) * 400 + RDPA_ERR

95. RDPGNP: Rate of inflation of output prices

RDPGNP =: (RPGNP/RPGNP(-1) - 1) * 400 + RDPGNP_ERR

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

- 20 -

RTIV: Nominal taxable income

RTIV = RPGNP * RGNP/100 - 6.58 * RK(-1) % RPA/10000 + RRS * (RGDEBTV(-1) - RMB(-1))/100 + RTIV_ERR

RTAXV: Nominal tax revenues

RTAXV = RTAU * RTIV + RTAXV_ERR

RTBAR: Equilibrium tax rate

RTBAR = (RG * RPA + RRS * RGDEBTV(-1))/(RTIV * 100) - RBRATIO + RGDEBTV( -1)/RTIV + RTBAR_ERR

RCU: Capacity utilization rate

RCU = RGDP/RCAP + RCU_ERR

RAW: Trade weighted foreign absorption

RAW = UA**0.4593 % JAX*0.1956 * GA¥*0.3451 + RAW_ERR

RCAPW: Trade weighted foreign total capacity output

RCAPH = UCAP*®*0.4593 * JCAP**0.1956 * GCAP**0.3451 + RCAPW_ERR

RPGNPH: Trade weighted foreign gross national product deflator - 1980=100.00

RPGNPH = (UPGNP/0.8572 )**0.4593 % (JPGNP * JER/4.424)**0.1956 * (GPGNP * GER/0 .5505 )**0.3451 * 1.173/RER + RPGNPW_ERR

RK: Total net capital stock

RK = (1 - 6.58 7400) * RK(-1) + RIF(-1)74 + RK_ERR

RGDEBTV: Current total government debt

RGDEBTV = (1 + RRS/G00) * RGDEBTV(-1) + RPA * RG/400 - RTAXV/4 - RMB(-1) * RRS/400 + RGDEBTV_ERR

RNFAV: Net foreign assets

RNFAV = (1 + URS/400) % RNFAV(-1) + (RPXGSNI % RXGSNI * RER - RPMGSNI * RMGSNI * RER)/400 + RNFAV_ERR

-21-

U.S. MODEL

106. UC: Private consumption expenditure - 1982 prices

UC/UCAP = .852 *% UC(-1)/UCAP(-1)/(1 + .852 /(1 + .02 + (1 - UTAU) * URS/400 - UDPA(1)/400)) + UCC L)/UCAP(1)/(1 + .852 + .02 + (1 - UTAU) * URS/400 - UDPA(1)/400) + (1 - .852 ) * .00985 * UYD/UCAP + UC_ERR

107. UIF: Total fixed investment - 1982 prices

(1 + .,95 * .97 ) % UIF/UCAP = .95 * UIF(-1)/UCAP(-1) + .97 % UIF(1)/UCAP(1) + (1 - .95 )*% (1 - .97 3% ( .32 % UCU ® 100/(URS + 6.23 - UDPA(1) + 6.8 /(1 - UTAU)) - (1 - 6.23 7100) * UK(-1)/UCAP) + UIF_ERR

108. UII: Inventory investment - 1982 prices

UIT/UCAP = .00733 + .483 * (UCU(1) - UCU) - .000778 * (URS - UDPA(1)) + UITI_ERR

109. UMB: Monetary base

LOG( UMB,'UPA ) = -1.395 + .776 *% LOG(UMB(-1)/UPA(-1)) - .0039 * URS + .1926 * LOG(UA) + UMB_ERR

- 22 -

U2. AGGREGATE SUPPLY

110. UPGNP: Gross national product deflator - 1982=100.00

(UPGNP/UPGNP(-1) - 1) * 400 = .5 * (UPGNP(-1)/UPGNP(-2) - 1) * 400 + (1 - .5 ) % (UPGNP(1)/UPGNP - 1) * 400 + 10 * LOG(UCU) + UPGNP_ERR

111. UCAP: Total capacity output

UCAP = UQ * UK(-1)% * .32 *% ULF(-1)%*(1 - .32 ) + UCAP_ERR

-~ 23 -

U3. TRADE

112. UXGSNI: Exports - NIA basis - 1982 prices LOG(UXGSNI ) = -1.645 + .476 * LOG(UXGSNI(-1)) + 1.929 * LOG(UAW/UCAPW) - .372 * LOG(UPXGSNI/UPMGSNI) + (1 - .476 ) * 1 * LOG(CAPTOT) + UXGSNI_ERR

113. UPXGSNI: Export deflator - NIA basis - 1980 prices L'3G(UPXGSNI ) = .003 + .962 * LOG(UPXGSNI(-1)) + O * LOG(UPGNPW) + (1 - .962 - O ) * LOG(UPGNP) + .728 * LOG(UPXGSNI(-1)/UPXGSNI(-2)) + (1 - .962 ) * -.0025 * TIME + UPXGSNI_ERR

114. XUGS: Share of US exports destined for Germany

XUGS = .063 + .166 * XUGS(-1) + .005 * LOG(GER * GPGNP/( JER * JPGNP)) + .018 * LOG(GER * GPGNP/(RER * RPGNP)) + XUGS_ERR

115. XUJS: Share of US exports destined for Japan

xUJS = .006 + .711 * XUJS(-1) - .005 * LOG(GER * GPGNP/( JER * JPGNP)) + .009 * LOG( JER * JPGNP/(RER * RPGNP)) + XUJS_ERR

116. XURS3: Share of US exports destined for ROW

XURS3 = 1 - .063 - .006 - .166 * XUGS(-1) - .711 * XUJS(-1) - .009 * LOG(JER * JPGNP/(RER * RPGNP)) - .018 * LOG(GER * GPGNP/(RER * RPGNP)) + XURS3_ERR

117.

118.

119.

120.

- 24 -

- MONETARY AND FISCAL POLICY

URS: 3-month Treasury Bill rate

UMB = 1.0195 % UMB(-1) + URS_ERR

UG: Real government purchases - 1982 prices

UG = 1.0067 * UG(-1) + UG_ERR

UBRATIO: Target ratio of government bonds to taxable income

UBRATIO = 0.317 + UBRATIO_ERR

UTAU: Actual income tax rate

UTAU = .95 * UTAU(-1) + (1 - .95 ) % UTBAR + 0 * LOG(UCU) + UTAU_ERR

- 25 -

121. UMGSNI: Imports - NIA basis - 1982 prices

UMGSNI = (XGUS * GXGSNI * 0.5505 + XJUS * JXGSNI * 4.424 + 1.294 * XRUSZ % RXGSNI * 1.173)/0.9169 + UMGSNI_ERR

122. UPMGSNI: Import deflator - NIA basis - 1982=100.00

UPMGSNI: = (XGUS * GPXGSNI * GXGSNI * GER + XJUS * JPXGSNI * JXGSNI * JER + 1.294 * XRUS3 * RPXGSNI * RXGSNI * RER)/UMGSNI + UPMGSNI_ERR

123. UA: Absorption

UA = UC + UIF + UG + UII + UA_ERR

124. UGDP: Gross domestic product - 1982 prices

UGDP = UA + UXGSNI - UMGSNI + UGDP_ERR

125. UGNP: Gross national product - 1982 prices

UGNP = UGDP + URS * UNFAV(-1)/UPGNP + UGNP_ERR

126. UYD: Disposable income

UYD = UPGNP * UGNP/UPA - 6.23 % UK(-1)/100 - UTAXV * 100/UPA + (URS - UDPA) * UGDEBTV(-1)/UPA - UMB(-1) * URS/UPA + UYD_ERR

127. UPA: Gross domestic product deflator - 1982=100.00

UPA = (UPGNP * UGNP - UPXGSNI * UXGSNI + UPMGSNI * UMGSNI - URS * UNFAV(-1))/UA + UPA_ERR

128. UDPA: Rate of inflation of absorption prices

UDPA = (UPA/UPA(-1) - 1) * 400 + UDPA_ERR

129. UDPGNP: Rate of inflation of output prices

UDPGNP = (UPGNP/UPGNP(-1) - 1) * 400 + UDPGNP_ERR

- 2 -

130. UTIV: Nominal taxable income

UTIV = UPGNP * UGNP/100 - 6.23 * UK(-1) * UPA/10000 + URS * (UGDEBTV(-1) - UMB{-1))/100 + UTIV_ERR

131. UTAXV: Nominal tax revenues

UTAXV = UTAU * UTIV + UTAXV_ERR

132. UTBAR: Equilibrium tax rate

UTBAR = (UG * UPA + URS * UGDEBTV(-1))/(UTIV * 100) - UBRATIO + UGDEBTV( -1)/UTIV + UTBAR_ERR

133. UCU: Capacity utilization rate

UCU = UGDP/UCAP + UCU_ERR

134. CAPTOT: Total world capacity

CAPTOT = UCAP * 0.8572 + GCAP * 0.5505 + JCAP * 4.424 + RCAP * 1.173 + CAPTOT_ERR

135. UAW: Trade weighted foreign absorption

UAN = GA**0.0463 * RA**0.8532 * JA*®*0.1005 + UAW_ERR

136. UCAPW: Trade weighted foreign total capacity output

UCAPH = GCAP**0.0463 * RCAP**0.8532 %* JCAPxx0.1005 + UCAPW_ERR

137. UPGNPH: Trade weighted foreign gross national product deflator - 1980=100.00

UPGNPH = (GPGNP * GER/0.5505 )**0.04¢63 * (RPGNP * RER/1.173 )**0.8532 * (JPGNP * JER/4.424)**0.1005 * 0.8572 + UPGNPW_ERR

138. UK: Total net capital stock

UK = (1 - 6.23 /400) * UK(-1) + UIF/4 + UK_ERR

139. UGDEBTV: Current total government debt

UGDEBTV = (1 + URS/400) * UGDEBTV(-1) + UPA * UG/400 - UTAXV/4 - UMB(-1) * URS/400 + UGDEBTV_ERR

- 27 -

140. UNFAV: Net foreign assets

UNFAV = (1 + URS/400) * UNFAV(-1) + (UPXGSNI %* UXGSNI - UPMGSNI x UMGSNI 3/400 + UNFAV_ERR

VARIABLE

CAPTOT GA

GAW GBRATIO GC

GCAP GCAPW GCU GDPA GDPGNP GER

GG GGDEBTV GGoOP GGNP GIF GII

GK

GMB GMGSNI GNFAV GPA GPGNP

GPGNPW GPMGSNI GPXGSNI GRS GTAU GTAXV GTBAR GTIV GXGSNI GYD

JA

JAW JBRATIO Jc

JCAP JCAPW JCU JDPA JDPGNP JER

JG JGDEBTV JGDP JGNP JIF JII

JK JMB JMGSNI JNFAV JPA JPGNP

JPGNPW JPMGSNI JPXGSNI JRS JTAU JTAXV JTBAR JTIV JXGSNI JYD

RA

RAW RBRATIO RC

RCAP RCAPW RCU RDPA RDPGNP RER

RG RGDEBTV RGDP RGNP RIF

RII

- 28 -

CROSS REFERENCE LIST OF ENDOGENOUS VARIABLES AND EQUATIONS

| EQUATION NUMBER

i! NUE MORENO LO

78

112 23

16 22

23 137

34 26

26 23 24 22 35 35

27

35 55

55 89

72

86 92

18 137 104 104

96

134 65

29

29 24

32

33 35

26 23

52 58

41

64 59

32

68 70

61 40

70 42

90

73

99 94

32

100

66

35

28 25

88& 22

53 100

64

42

63 44

88 57

70 93

76

45

135

101

45

34¢ 26

122 26

87 135

101

53

69 56

122 61

87 135

99

46

134¢

46

45

28

88

134¢

56

57

63

88

134

47

136

47

46

34¢

121

136

58

58

69

121

136

53

53

47

122

67

60

122

67

67

67

70

61

77

80

80

80

80

88

81

81

81

81

91

82 88

82 102

82 88

82 102

93 102

102

114

102

114

105

114

115

114

115

114

115

116

115

116

115

VARIABLE | EQUATION NUMBER

RK RMB RMGSNI RNFAV RPA RPGNP

RPGNPH' RPMGSNI RPXGSNI RRS RTAU RTAXV RTBAR

UBRATIO uc

UCAP UCAPW UucU UDPA UDPGNP UG UGDEBTV UGDP UGNP UIF

UIT

UK

UMB UMGSNI UNFAV UPA UPGNP

UPGNPW UPMGSNI UPXGSNI URS

UTAU UTAXV UTBAR UTIV UXGSNI UyD XGJS XGRS3 XGUS XJGS XJRS3 XJUS XRGS3 XRJS3. XRUS3 XUGS XUJS XURS3

53 109

106

120 126

139 139

130 130 130 127 128

32 140 112

132 131

53 53

122 47

122 82

122 116 116

103 104 105

96 45

93 77

78 123

107

133 128

138 139 140

130 45

113 139

87

98 46

105 92

90 12¢

108

132 46

127 140

104¢ 47

122 96

93 127

111

139 47

140 58

112

- 29 -

67 75 79

98 104

105 121 122

133 134

67 80 81

70 77 91

124 127 140

91 92 93 95 96 114 115 116

82 102 110 113 125 126 127 129

93 105 106 107 108 109 #125 126

- 30 - CROSS REFERENCE LIST OF EXOGENOUS VARIABLES AND EQUATIONS

VARIABLE | EQUATION NUMBER

CAPTOT_ERR GA_ERR GAW_ERR GBRATIO_ERR GC_ERR GCAP_ERR GCAPW_ERR GCU_ERR GDPA_ERR GDPGNP_ERR GER_ERR GG_ERR GGDEBTV_ERR GGDP_ERR GGNP_ERR GIF_ERR GII_ERR GK_ERR

GLF

GMB_ERR GMGSNI_ERR GNFAV_ERR GPA_ERR GPGNP_ERR GPGNPW_ERR GPMGSNI_ERR GPXGSNI_ERR GQ

GRS_ERR GTAU_ERR GTAXV_ERR GTBAR_ERR GTIV_ERR GXGSNI_ERR GYD_ERR JA_ERR JAW_ERR JBRATIO_ERR JC_ERR JCAP_ERR JCAPH_ERR JCU_ERR JDPA_ERR JDPGNP_ERR JER_ERR JG_ERR JGDEBTV_ERR JGDP_ERR JGNP_ERR JIF_ERR JII_ERR JK_ERR

JLF

JMB_ERR JMGSNI_ERR JNFAV_ERR JPA_ERR JPGNP_ERR JPGNPW_ERR JPMGSNI_ERR JPXGSNI_ERR

JQ

JRS_ERR JTAU_ERR JTAXV_ERR JTBAR_ERR JTIV_ERR JXGSNI_ERR JYD_ERR RA_ERR RAW_ERR RBRATIO_ERR RC_ERR RCAP_ERR RCAPW_ERR RCU_ERR RDPA_ERR RDPGNP_ERR RER_ERR RG_ERR RGDEBTV_ERR RGDP_ERR RGNP_ERR RIF_ERR RII_ERR RK_ERR

VARIABLE | EQUATION NUMBER

RLF

RMB_ERR RMGSNI_ERF: RNFAV_ERR RPA_ERR RPGNP_ERR RPGNPW_ERF. RPMGSNI_ERR RPXGSNI_ERR

RQ

RRS_ERR RTAU_ERR RTAXV_ERR RTBAR_ERR RTIV_ERR RXGSNI_ERR RYD_ERR TIME

UA_ERR UAW_ERR UBRATIO_ERR UC_ERR UCAP_ERR UCAPW_ERR UCU_ERR UDPA_ERR UDPGNP_ERR UG_ERR UGDEBTV_ERR UGDP_ERR UGNP_ERR UIF_ERR UII_ERR UK_ERR

ULF

UMB_ERR UMGSNI_ERR UNFAV_ERR UPA_ERR UPGNP_ERR UPGNPW_ERR UPMGSNI_ERR UPXGSNT_ERR

UQ URS_ERR UTAU_ERR UTAXV_ERR UTBAR_ERR UTIV_ERR UXGSNI_ERR UYD_ERR XGJS_ERR XGRS3_ERR XGUS_ERR XJGS_ERR XJRS3_ERR XJUS_ERR XRGS3_ERR XRJS3_ERR XRUS3_ERR XUGS_ERR XUJS_ERR XURS3_ERR

44

79

113

-31-

MNEMONIC |

CAPTOT GA

GAW GBRATIO GC

GCAP GCAPW GCU GDPA GDPGNP GER

GG GGDEBTV GGDP GGNP GIF

GII

GK

GMB GMGSNI GNF AV GPA GPGNP GPGNPW GPMGSNI GPXGSNI GRS GTAU GTAXV GTBAR GTIV GXGSNI GYD

JA

JAW JBRATIO Jc JCAP JCAPW JCU JDPA JDPGNP JER

JG JGDEBTV JGDP JGNP JIF

JII

JK

JMB JMGSNI JNFAV JPA JPGNP JPGNPW JPMGSNI JPXGSNI JRS JTAU JTAXV JTBAR JTIV JXGSNI JYD

RA

RAW RBRATIO RC

RCAP RCAPW RCU RDPA RDPGNP RER

RG RGDEBTV RGDP

RGNP RIF

- 32 -

ALPHABETICAL LIST OF ENDOGENOUS VARIABLES FOR MODEL

EQUATION }

DEFINITION

Total World Capacity

Absorption

Trade weighted foreign absorption

Target ratio of government bonds to taxable income Private consumption expenditure - 1980 prices Total capacity output

Trade weighted foreign total capacity output Capacity utilization rate

Rate of inflation of absorption prices

Rate of inflation of output prices

Spot exchange rate - US$/DM

Real government purchases - 1980 prices Current total government deb

Gross domestic product - 1980 prices

Gross national product - 1980 prices

Total fixed investment - 1980 prices Inventory investment - 1980 prices

Total net capital stock

Monetary base

Imports - NIA basis - 1980 prices

Net foreign assets

Gross domestic product deflator - 1980=100.00 Gross national product deflator - 1980=100.00 Trade weighted foreign gross national product deflator - 1980=100.130 Import deflator - NIA basis - 1980=100.00 Export deflator - NIA basis - 1980 prices 3-month Treasury bill rate

Actual income tax rate

Nominal tax revenues

Equilibrium tax rate

Nominal taxable income

Exports - NIA basis - 1980 prices

Disposable income - 1980 prices

Absorption

Trade weighted foreign absorption

Target ratio of government bonds to taxable income

Private consumption expenditure - 1980 prices Total capacity output Trade weighted foreign total capacity output

Capacity utilization rate

Rate of inflation of absorption prices

Rate of inflation of output prices

Spot exchange rate ~ US$/DM

Real government purchases - 1980 prices Current total government debt

Gross domestic product - 1980 prices

Gross national product - 1980 prices

Total fixed investment - 1980 prices Inventory investment - 1980 prices

Total net capital stock

Monetary base

Imports - NIA basis ~- 1980 prices

Net foreign assets

Gross domestic product deflator - 1980=100.00 Gross national product deflator - 1980=100.00 Trade weighted foreign gross nationa. product deflator - 1980=100.)0 Import deflator - NIA basis - 1980=100.00 Export deflator - NIA basis - 1980 prices 3-month Treasury bill rate

Actual income tax rate

Nominal tax revenues

Equilibrium Tax Rate

Nominal taxable income

Exports - NIA basis - 1980 prices

Disposable income - 1980 prices

Absorption

Trade weighted foreign absorption

Target ratio of government bonds to taxable income Private consumption expenditure - 1980 prices Total capacity output

Trade weighted foreign total capacity output Capacity utilization rate

Rate of inflation of absorption prices

Rate of inflation of output prices

Spot exchange rate - US$/DM

Real government purchases - 1980 prices Current total government debt

Gross domestic product - 1980 prices

Gross national product - 1980 prices

Total fixed investment - 1980 prices Inventory investment - 1980 prices

Total net capital stock

Monetary base

Imports - NIA basis - 1980 prices

Net foreign assets

MNEMONIC

UPXGSNI URS UTAU UTAXV UTBAR UTIV UXGSNI UYD xGJS XGRS3 XGUS XxJGS XJRS3 XJUS XRGS3 XRJS3 XRUS3 XUGS XUJS XURS3

| EQUATION |

- 33 -

DEFINITION

Gross domsetic product deflator - 1980=100.00 Gross national product deflator ~ 1980=100.00 Trade weighted foreign gross national Product deflator - 1980=100.00

Import deflator - NIA

basis - 1980=100.00

Export deflator - NIA basis - 1980 Prices 3-month Treasury bill rate

Actual income tax rate

Nominal tax revenues

Equilibrium tax Rate

Nominal taxable income

Exports - NIA basis - 1980 prices

Disposable income - 1980 prices

Absorption

Trade weighted foreign absorption

Target ratio of government bonds to taxable income Private consumption expenditure - 1982 prices Total capacity output

Trade weighted foreign total capacity output Capacity utilization rate

Rate of inflation of absorption prices

Rate of inflation of output prices

Ueal government purchases - 1982 prices

Current total

government debt

Gross domestic product - 1982 prices Gross national product - 1982 prices

Total fixed investment - 1982 prices Inventory investment - 1982 prices

Total net capital stock

Monetary base

Imports ~ NIA basis - 1982 prices

Net foreign assets

Gross domestic product deflator - 1982=100.00 Gross national product deflator - 1982=100.00 Trade weighted foreign gross national product deflator - 1980=100.00 Import deflator - NIA basis - 1982=100.00 Export deflator - NIA basis - 1980 prices 3-month Treasury bill rate

Actual income tax rate

Nominal tax revenues

Equilibrium tax rate

Nominal taxable income

Exports - NIA basis - 1982 prices

Disposable income

Share Share Share Share Share Share Share Share Share Share Share Share

German exports destined for Japan German exports destined for ROW German exports destined for US Japanese exports destined for Germany Japanese exports destined for ROW Japanese exports destined for US ROW exports destined for Germany ROW exports destined for Japan ROW exports destined for US

US exports destined for Germany US exports destined for Japan

US exports destined for ROW

- 34 -

ALPHABETICAL LIST OF EXOGENOUS VARIABLES FOR MODEL

MNEMONIC | DEFINITION CAPTOT_ERR Residual term in CAPTOT equation GA_ERR Residual term in GA equation GAW_ERR Residual term in GAW equation GBRATIO_ERR Residual term in GBRATIO equation

__ERR Residual term in GC equation GCAP_ERR Residual term in GCAP equation GCAPW_ERR Residual term in GCAPW equation GCU_ERR Residual term in GCU equation GDPA_ERR Residual term in GDPA equation GDPGNP_ERR Residual term in GDPGNP equation GER_ERR Residual term in GER equation GG_ERR Residual term in GG equation GGDEBTV_ERR Residual term in GGDEBTV equation GGDP_ERR Residual term in GGDP equation GGNP_ERR Residual term in GGNP equation GIF_ERR Residual term in GIF equation GII_ERR Residual term in GII equation GK_ERR Residual term in GK equation

GLF Labor force GMB_ERR Residual term in GMB equation GMGSNI_ERR Residual in GMGSNI equation GNFAV_ERR Residual in GNFAV equation GPA_ERR Residual in GPA equation GPGNP_ERR Residual in GPGNP equation GPGNPW_ERR Residual in GPGNPW equation GPMGSNI_ERR Residual in GPMGSNI equation GPXGSNI_ERR Residual in GPXGSNI equation GQ Production technology GRS_ERR Residual in GRS equation GTAU_ERR Residual in GTAU equation GTAXV_ERR Residual in GTAXV equation GTBAR_ERR Residual in GTBAR equation GTIV_ERR Residual in GTIV equation GXGSNI_ERR Residual in GXGSNI equation GYD_ERR Residual in GYD equation JA_ERR Residual term in JA equation JAW_ERR Residual term in JAW equation JBRATIO_ERR Residual term in JBRATIO equation JC_ERR . Residual term in JC equation JCAP_ERR Residual term in JCAP equation JCAPW_ERR Residual term in JCAPW equation JCU_ERR Residual term in JCU equation JDPA_ERR Residual term in JDPA equation JDPGNP_ERR Residual term in JDPGNP equation JER_ERR Residual term in JER equation JG_ERR Residual term in JG equation JGDEBTV_ERR Residual term in JGDEBTV equation JGDP_ERR Residual term in JGDP equation JGNP_ERR Residual term in JGNP equation JIF_ERR Residual term in JIF equation JII_ERR Residual term in JII equation JK_ERR Residual term in JK equation JLF Labor force JMB_ERR Residual term in JMB equation JMGSNI_ERR Residual in "MGSNI equation JNFAV_ERR Residual in JNFAV equation JPA_ERR Residual in JPA equation JPGNP_ERR Residual in JPGNP equation JPGNPW_ERR Residual in JPGNPW equation JPMGSNI_ERR Residual in JPMGSNI equation JPXGSNI_ERR Residual in JPXGSNI equation JQ Production technology JRS_ERR Residual in JRS equation JTAU_ERR Residual in JTAU equation JTAXV_ERR Residual in JTAXV equation JTBAR_ERR Residual in JTBAR equation JTIV_ERR Residual in JTIV equation JXGSNI_ERR Residual in JXGSNI equation JYD_ERR Residual in JYD equation RA_ERR Residual term in RA equation RAW_ERR Residual term in RAW equation RBRATIO_ERR Residual term in RBRATIO equation RC_ERR Residual term in RC equation RCAP_ERR Residual term in RCAP equation RCAPW_ERR Residual term in RCAPW equation RCU_ERR Residual term in RCU equation RDPA_ERR Residual term in RDPA equation RDPGNP_ERR Residual term in RDPGNP equation RER_ERR Residual term in RER equation RG_ERR Residual term in RG equation RGDEBTV_ERR Residual term in RGDEBTV equation RGDP_ERR Residual term in RGDP equation RGNP_ERR Residual term in RGNP equation RIF_ERR Residual term in RIF equation

|

RII_ERR Residual term in RII equation

MNEMONIC

RK_ERR

RLF

RMB_ERI RMGSNI_ERI2 RNFAV_ERR RPA_ERR RPGNP_ERR RPGNPW_ERI RPMGSNI_ERI RPXGSNI_ERR RQ

RRS_ERR RTAU_ERR RTAXV_ERR RTBAR_ERE RTIV_ER#t RXGSNI_ERR RYD_ ERR

!

UAW_ERF!: UBRATIO_ERF: UC_ERF: UCAP_ERR: UCAPW_ERF: UCU_ERR UDPA_ERR UDPGNP_ERR UG_ERR UGDEBTV_ERR UGDP_ERR UGNP_ERR UIF_ERR UII_ERR UK_ERR

ULF

UMB_ERR UMGSNI_ERR UNFAV_ERR UPA_ERR UPGNP_ERR UPGNPW_ERR UPMGSNI_ERR UPXGSNI_ERR

UQ URS_ERR UTAU_ERR UTAXV_ERR UTBAR_ERR UTIV_ERR UXGSNI_ERR UYD_ERR XGJS_ERR XGRS3_ERR XGUS_ERR XJGS_ERR XJRS3_ERR XJUS_ERR XRGS3_ERR XRJS3_ERR XRUS3_ERR XUGS_ERR XUJS_ERR XURS3_ERR

|e |

- 35 -

DEFINITION

Residual term in RK equation

Labor force

Residual term in RMB equation

Residual in Residual in Residual in Residual in Residual in Residual in Residual

RMGSNI equation RNFAV equation RPA equation RPGNP equation RPGNPW equation RPMGSNI equation

in RPXGSNI equation

Production technology

Residual in Residual in Residual in Residual in Residual in Residual in Residual in Time trend

RRS equation RTAU equation RTAXV equation RTBAR equation RTIV equation RXGSNI equation RYD equation

Residual term in UA equation Residual term in UAW equation Residual term in UBRATIO equation Residual term in UC equation Residual term in UCAP equation Residual term in UCAPW equation Residual term in UCU equation Residual term in UDPA equation Residual term in UDPGNP equation Residual term in UG equation Residual term in UGDEBTV equation Residual term in UGDP equation Residual term in UGNP equation Residual term in UIF equation Residual term in UII equation Residual term in UK equation

Labor force

Residual term in UMB equation

Residual in Residual in Residual in Residual in Residual in Residual in Residual

UMGSNI equation UNFAV equation UPA equation UPGNP equation UPGNPW equation UPMGSNI equation UPXGSNI equation

in Production technology

Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in Residual in

RRS equation UTAU equation UTAXV equation UTBAR equation UTIV equation UXGSNI equation UYD equation XGJS equation XGRS3 equation XGUS equation XJGS equation XJRS3 equation XJUS equation XRGS3 equation XRJS3 equation XRUS3 equation XUGS equation XUJS equation XURS3 equation

Construction of the Database

Most data are taken from the OECD's Quarterly National Accounts. The trade shares are computed using data from the IMF's Direction of Trade Statistics. The exchange rate, interest rate, monetary base, and labor force are obtained directly from national sources.

All data are seasonally adjusted at annual rates.” Interest rates are expressed in percents, not decimals.> The data are expressed in billions of local currency units, except for data from Italy and Japan, which are expressed in trillions. Real quantities are expressed at 1982 prices for U.S. data, 1981 prices for Canadian data, and 1980 prices for all other data. Price deflators are equal to 100 in the base year.“ Exchange rates are the number of U.S. dollars required to purchase a unit of the currency in question, except for the cases of Italy and Japan, where the exchange rates are the number of U.S. dollars required to buy 1000 lire or yen.

The newly revised Italian national accounts data are not available prior to 1980. MX3 has spliced the old series onto the new series for the years 1976-1979.

The United States government consumption series differs from all other MX7 government consumption series because it includes public gross fixed capital formation. Thus, in the United States gross fixed capital formation

refers to private investment only.

2. Most of the data are available only on a seasonally adjusted basis. When the data are not available seasonally adjusted, we have adjusted them using Census X-11 as implemented by the SEASAQ command in TROLL.

3. For expositional purposes, the text assumes that interest rates are expressed in decimals. This convention allows for simpler notation.

4. For expositional purposes, the text assumes that deflators equal 1 in the base year.

- 37 -

‘To create the ROW national accounts data, series from Canada, France, Italy and the United Kingdom were multiplied by their respective purchasing power parity (PPP) exchange rates in 1980 and then summed. The only exception is the Canadian data which are also rescaled from a 1981 base year to 1980. The PPP exchange rates were obtained from OECD National Accounts.

Factor payments abroad and factor receipts from abroad are available on an anrual basis only, and they were interpolated to yield quarterly figures. The nominal factor receipts and payments were deflated by either PGNP or PGDP, depending on availability, a procedure which ensures that PGNP will be identical to PGDP. For countries that report quarterly GDP, net factor receipts were added to obtain GNP. For countries that report quarterly GNP, factor payments were subtracted from reported imports and factor receipts were subtracted from reported exports. GDP in these countries is obtained by subtracting net factor receipts from GNP.

The share of U.S. exports to each trading partner is obtained by dividing nominal bilateral exports to that partner by total nominal U.S. exports. Japanese and German export shares are computed in the same manner. The rest of world (ROW) bilateral exports are the residual exports from those countries other than the United States, Japan, and Germany. Thus, ROW's export share to the United States is the total exports of ROW countries to the United States divided by total ROW exports to the G-3. For the purpose of computing trade shares, ROW includes all countries reported on the IMF Direction of Trade Statistics.

All trade weighted series are geometric averages and use the average

bilateral trade shares over the period 1976-1987 as weights. The

construction of each of these series is described in the text.

- 38 -

The short-term interest rate is the 3-month Treasury bill rate waere available. In France it is the 3-month interbank rate. In Japan it is the 2-month Treasury bill rate.

Where not explicitly available, the monetary base is computed as the sum of all currency outside the central bank plus deposits held by private banks as reserves at the central bank.

In certain countries the labor force is computed as the sum of employment and unemployment. In France the labor force is reported o1 an annual basis and has been interpolated.

The capital stock series have been constructed by interpolating annual net capital stock series.” The primary source is the OECD's Flows and_ Stocks of Fixed Capital, 1960-1985. Missing components of these series have been approximated using estimates of gross capital stocks from the OECD's Sectorial Database and the ratios of net to gross capital for similar series in other countries as reported in Flows and Stocks of Fixed Capital.

Capital stocks after 1985 were extrapolated by cumulating fixed investment less depreciation at the estimated depreciation rate.

Except for Germany and Japan, the outstanding stock of-governmen: debt has been computed from a benchmark value by cumulating the public seczor deficits (at a quarterly rate) in successive quarters. The benchmark values are net public sector debt stocks at yearend 1982 and were obtained f--om

OFCD Economic Studies, No. 7, 1986, pp. 103-153. For Germany and Japan

public sector debt series were obtained from national sources.

5. The net increment to the capital stock over the year was allocated to

each quarter in a manner proportional to the measured flow of gross fixed investment in that quarter.

- 39 -

Similarly, the stock of net foreign assets was computed by cumulating the current account balances over time. The benchmark values for these series are yearend 1982 from the IMF’s World Economic Outlook, April 1988, pp. 88-90. The current account balance is the sum of nominal net exports

and nominal net factor receipts.

Cite this document
APA
Joseph E. Gagnon (1989). A Forward-Looking Multicountry Model: MX3 (IFDP 1989-359). Board of Governors of the Federal Reserve System, International Finance Discussion Papers. https://whenthefedspeaks.com/doc/ifdp_1989-359
BibTeX
@techreport{wtfs_ifdp_1989_359,
  author = {Joseph E. Gagnon},
  title = {A Forward-Looking Multicountry Model: MX3},
  type = {International Finance Discussion Papers},
  number = {1989-359},
  institution = {Board of Governors of the Federal Reserve System},
  year = {1989},
  url = {https://whenthefedspeaks.com/doc/ifdp_1989-359},
  abstract = {This is paper discusses the theoretical structure and empirical properties of MX3, a multicountry macroeconometric model with rational expectations. MX3 is a medium-sized quarterly model of the United States, Japan, and West Germany. The primary objective of the model is to analyze the effect of fiscal and monetary rules on national economies in an international context. By incorporating rational expectations into almost all of the model's behavioral equations, MX3 takes a large step toward addressing the "Lucas critique" of model-based policy analysis.},
}