Download PDF

The East Asian Meltdown: A Dynamic Computable
General Equilibrium Analysis
Elena Ianchovichina, Robert McDougall and Thomas Hertel∗
American Agricultural Economics Association Meetings
Nashville, Tennessee†
August 8-11, 1999
∗
Ianchovichina: Assistant Professor, Department of Agricultural Economics, Kansas State University, 342
Waters Hall, Manhattan, KS 66506-4011, USA. McDougall: Deputy Director, Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, 1145 Krannert Building, IN 47907, USA.
Hertel: Professor and Director, Center for Global Trade Analysis, Department of Agricultural Economics,
Purdue University, 1145 Krannert Building, IN 47907, USA.
†
Copyright 1999 by Ianchovichina, McDougall and Hertel. All rights reserved. Readers may make verba-
tim copies of this document for non-commercial purposes by any means, provided that this copyright notice
appears on all such copies.
1
Abstract
The paper proposes a new disequilibrium approach to modeling international capital
mobility. Key to this approach are errors in investors’ assessments of potential returns
to capital – such as those recently observed in Asia. We use the model to study dynamic
adjustment of North American farm and food industries to a marginally deeper, longer
crisis in East Asia.
Key words: International capital mobility, adaptive expectations, East Asian crisis,
North America, agriculture, farm and food industries
2
1
Introduction
The Asian financial crisis, which began in 1997 and which is still playing itself out, resulted
in a heightened awareness of the need to understand global economic linkages. Virtually
every country in the world has registered concern about the crisis’ potential impact on
their economy. The range of modeling tools available for the analysis of this economic
phenomenon extend from multi-region partial equilibrium models to global macroeconomic
models. Multi-region partial equilibrium and single-region general equilibrium trade models,
while appropriate for some types of analyses, cannot adequately capture sectoral and regional
linkages, respectively. Static, multi-region general equilibrium models (Hertel and Tsigas,
1997) overcome these limitations. However, these models offer little macroeconomic detail,
do not produce time path results, and do not distinguish between short-run and longer term
outcomes. Therefore, these models are of limited use in studying economic phenomena such
as the East Asian crisis. Macroeconomic models, on the other hand, are designed to predict
the impact of financial phenomena. These models, however, lack the sectoral and regional
detail necessary to study the economywide implications of the crisis. Given that the effects of
the East Asian would be felt over the course of several years, and that these effects will differ
across industries and economies, the study of this type of economic phenomena warrants a
dynamic general equilibrium approach.
Despite the fact that a number of dynamic multi-region applied general equilibrium
(AGE) models are in use today, only very few of them incorporate international capital
mobility adequately. The GREEN model of Burniaux et al. (1992) and Mercenier’s intertemporal multi-country, multi-sector model (1995) do not permit international capital
mobility. The G-Cubed model of McKibbin and Wilcoxen (1998) overcomes this limitation.
1
The rational expectations assumption and the equilibrium nature of the G-Cubed model,
however, are inconsistent with the disequilibrium nature of the East Asian crisis, which
revealed errors in investors’ expectations about the region’s economic growth.
This study contributes to the existing literature by offering a new disequilibrium approach
to modeling international capital mobility. With this approach instead of changing the
data to fit the model, as required in models with rational expectations, the theory of the
model accommodates empirical facts as observed in the current world economy. Key to our
disequilibrium approach is a new investment theory of adaptive expectations that emphasizes
errors in investors’ assessments of potential returns to investment – such as those recently
observed in Asia. In addition, this practical approach to modeling international capital
mobility permits a simple recursive-solution procedure. This greatly facilitates its use with
highly disaggregated models of the world economy.
Our goal in this paper is to study the impact of a marginally deeper, longer crisis in East
Asia on North American farm and food industries. For this purpose, we introduce the new
disequilibrium theory of investment into an existing static global AGE model, GTAP (Hertel
and Tsigas, 1997). The resulting, dynamic global AGE model uses the new investment
theory, while preserving other features of GTAP. This model can, therefore, be implemented
by adding minimum additional data to the publicly available GTAP data base (McDougall,
1997).
2
The model
A serious limitation of forward–looking intertemporal models, in the context of modeling the
East Asian crisis, is the assumption of investors’ perfect foresight of returns to capital. The
2
developments in Asia suggest that investors have not foreseen correctly returns to capital.
The crisis-affected regions in Asia experienced a severe credit crunch as investors withdrew
investment from the region, acknowledging errors in their expectations and adjusting these
expectations in a downward direction. In deference to this empirical reality, this paper
acknowledges the importance of errors in investors’ expectations and introduces a novel
investment theory of adaptive expectations.
The introduction of international capital mobility involves explicitly keeping track of
regional capital stocks, which accumulate through time giving rise to medium-run growth
effects, referred to by Baldwin (1989) as accumulation effects. Since the reallocation of
capital across regions affects aggregate regional welfare and wealth, to obtain valid welfare
results the model keeps track of capital ownership.
Equity in domestic enterprises and equity in enterprises located in foreign regions comprise regional wealth. Over time regional savings, which are a fixed share of income, augment
regional wealth. Due to an absence of data on bilateral investment flows, we do not model
bilateral equity ownership, instead we assume that regional households invest abroad via
a global investment trust. The global trust collects the savings that the regional households have chosen to invest in foreign regions and invests these funds on their behalf. The
investment theory determines how the trust allocates funds across regions.
Since regional households earn income, not from the capital stock they harbor, but from
the capital stock they own, the model takes into account separately of capital and wealth
accumulation by region. The total equity of the domestic economy consists of equity in
domestic enterprises owned by domestic investors and equity in domestic enterprises owned
by foreign investors.
We assume that each region specializes in its own assets. The composition of regional
3
wealth and domestic capital change as needed to be consistent with the level of regional
savings and investment determined elsewhere in the model. We do not distinguish between
debt and equity investment. All foreign funds are used for the purchase of physical investment
goods, which are then added to the existing stock of physical capital.
In each region, there is an investment firm owned both by domestic and foreign residents.
The firm provides capital to all industries in a region and earns rent on the capital services
K it supplies. Domestic and foreign residents invest in the region through the investment
firm; i.e., they buy capital goods, I. Following the adjustment cost models of Lucas (1967),
Treadway (1969) and Uzawa (1969), we assume that the investment process is associated
with waste of some of the purchased investment goods. This waste occurs either during or
before installation.
The investment firm maximizes intertemporal profits:
max
Z ∞
t
Φ
[q (s)K(s) − P (s)I(s) − π (s)
2
e
e
e
!
I(s)
I(s)]e−Rs (s−t) ds
K(s)
s.t. K̇(t) = I(t) − δ K(t).
(1)
(2)
Variables q, P , and π are the rental price of capital, the price of raw capital goods, and
the price of capital goods, respectively; superscript e denotes expectations; delta is the
coefficient of depreciation. We adopt Uzawa’s approach by assuming that the per unit cost
of installation is a linear function of investment, depends on the rate of investment I/K and
on the parameter Φ, which is constant with respect to time.
Under myopic expectations of the investment firm about the price of capital, the firstorder conditions of the problem of the investment firm lead to the following expression for
4
the gross rate of return (shown in proportionate changes, denoted with a “hat”: b):1
d
rd
+ δ = A(q̂ − P̂ ) − αI/K,
(3)
where r is net return to capital, and A = 1/(1 + .5Φ(I/K)2 ) ≤ 1. In the special case where
A = 1, there are no adjustment costs (Φ = 0). More generally, A is a function of the
investment-capital ratio and adjustment costs, and α is the elasticity of the rate of return
with respect to the investment-capital ratio.
It can be shown that adjustment costs alone cannot satisfactorily determine capital allocation in a multi-region model. A theory, specifying supply of investment funds is required.
Next, we propose such an investment theory of adaptive expectations.
In each region there is a target (gross) rate Rt . The target rate of return equals the global
rate of return that clears the global market for capital. To reflect inter-regional differences
in returns due to different investment risks, we adjust the target rate of return with a regionspecific risk premium.2 Investment supplied to each region is such as to achieve some required
def
rate of growth Γ = Ṙ/R in the rate of return:
dΓ = Λ(R̂t − R̂e ).
(4)
where Λ is a parameter determining the speed of adjustment in the expected rate of return
towards the target rate and Re is the expected rate of return.3 Since investors typically
expect to derive returns over some considerable period of time, they are concerned not only
with the rate of return at the moment of investing funds, but also with the rate of return
through the life of the asset. Therefore, it is the expected return in future periods, Re , not
the actual return to which investors respond.
1
For convenience, from here onward, we omit the time index and the index on regions.
2
In the absence of risk premia, the target rate is uniform across regions.
3
The operators d and b denote a change and a proportionate change, respectively.
5
To determine the regional composition of investment, we impose equality between the
actual and required rates of growth in the rate of return. Thus, Γ varies inversely with the
actual rate of growth in the capital stock and directly with the normal rate of growth in the
capital stock Ω:
c
I
I
dΓ = −φ
K
K
!
!
− dΩ ,
(5)
where φ is the elasticity determining the response of the actual rate of return to changes in
the capital stock.
The normal rate of growth in the capital stock is this rate of growth in the capital stock
that allows the rate of return to remain constant through time. If there is a discrepancy
between the estimated and the normal rate of growth in capital stock, Ω will change as
specified by the following equation:
!
c
R
a
dΩ = η K̂ +
− Ωdt .
φ
(6)
This relationship is such that the normal rate of return adjusts towards the estimated rate
c /φ at a speed determined by the parameter η.
of growth in capital stock K̂ + R
a
Investors’ expectations are ”sticky” or ”sluggish.” When the observed rate of return
changes, investors are unsure whether this change is transient or permanent. They adjust
their expectations of future rates of return only with a lag. At first investors make a small
adjustment, then if the change in the actual rate persists, they make further changes in
expectations, until eventually the expected rate conforms to the observed rate:4
R̂e = −φ(K̂ − Ωdt) − µlog
Re
dt.
Ra
(7)
Equations (4), (5), (6) and (7) comprise the investment theory of adaptive expectations
and determine regional supply of investment funds. These equations, together with equation
4
Ianchovichina (1998) provides complete derivations of these equations.
6
(3), determine regional investment in the dynamic multi-region model.
The equilibrium in the model is characterized by equality between constant expected,
target and actual rates of return, and constant normal rate of growth in the capital stock,
which equals zero in equilibrium:
Re = Rt = Ra ,
(8)
Ṙe = Ṙt = Ṙa = 0
(9)
Ω = 0, Ω̇ = 0.
(10)
These conditions imply in turn constant investment-capital ratio. Jointly with the assumption of myopic expectations of the investment firm about the price of capital, we have the
same conditions characterizing the equilibrium of the model as those characterizing the
equilibrium solution to the problem of the investment firm. The mechanism of adaptive
expectations ensures that the model moves towards this equilibrium over time. We can also
show that this equilibrium is a stable one (Ianchovichina et al., 1999).
3
Empirical performance: economic crisis in East Asia
This section examines the consequences of a deeper, longer crisis in Korea, Indonesia, Philippines, Malaysia and Thailand (East Asia 5) for North American farming and food processing
industries, both in the short-run (1999) and in the long-run (2010). Our baseline, based on
a macroeconomic scenario of the World Bank (1998), portrays the recent Asian crisis and
captures the rise in unemployment; the impairment of regional capital markets; the sharp
increase in trade balances in East Asia 5 and Japan accompanying the crisis; the high expectations of potential returns in East Asia before 1997, and the subsequent drop in expectations
as investors realign their expectations with the actual returns in the region (e.g., Figure 1).
7
0.25
0.2
Levels
0.15
0.1
0.05
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
0
-0.05
Figure 1: Errors in Expectations: Thailand
We then compare this baseline to an alternative simulation of a longer, deeper Asian crisis
resulting in a percentage point lower annual GDP growth rates in East Asia 5 than the
baseline. The crisis is deeper due to further impairment of regional capital markets, a larger
decline in employment, and a continued decline in investment in the East Asia 5 relative
to the baseline over the near term (1999). The crisis is longer due to slower employment
recovery and subsequent lower factor productivity in the five East Asian economies relative
to the baseline over the longer term (2010).
Marginally lower growth of the five Asian economies due to a deeper, longer crisis, results
8
50
0
US$ Million
-50
-100
-150
-200
Others
S. Asia
Africa
S. America
Thailand
Philippines
Malaysia
Indonesia
China
Taiwan
Korea
Japan
W. Europe
N. America
-250
Figure 2: Farm Products: 1999 Import Volume Changes, Relative to Baseline,
Due to a Marginally Deeper, Longer Asian Crisis
in additional capital outflows (a cumulative increase of US$ 189 billion relative to the base
case by 2010) from the region. Investors, realizing that the crisis in Asia is more serious
than previously anticipated, redirect their funds away from East Asia 5 and towards North
America. Additional inflows of investment into North America put downward pressure on
interest rates. This is good news for net debtors in the United States and Canada, be they
farmers borrowing to finance their operations; or the government financing its debt.
In the near term (1999), a deeper, longer crisis in the East Asia 5 reduces farm imports
in the five Asian regions (Figure 2). Overall world trade in farm products also falls. The
9
crisis slows down income growth and the transition from the labor-intensive farm and natural
resource-based production to the capital-intensive manufacturing in East Asia 5 (Figure 3).
As a consequence of the slowdown in industrial transformation, farm production does not
tend to fall as much as industrial output in East Asia 5, relative to the baseline. This
further limits the scope for increased exports of farm products from North America. Indeed,
the model predicts a small increase in farm exports from the five East Asian economies
and a decline in farm exports from all other regions due to the deeper, longer crisis. The
negative impact on farm exports from North America is especially strong given the large
proportion of farm commodities exported to East Asia 5. The North American farm trade
balance deteriorates by US$ 289 million in 1999, while farm output declines by about US$
236 million in the same year relative to the base case (Figure 4).
The effect of a deeper, longer crisis in East Asia 5 on this region’s food exports will be
negative in the near term due to limited availability of capital and lower incomes in East
Asia 5 (Figure 5). North America, in contrast, undergoes an economic expansion, fueled
by foreign capital inflows and cheaper capital, which raises incomes and stimulates private
consumption. These increases in consumption offset the decline in North American food
exports. Thus, while the North American farm sector suffers losses, the region’s food sector
expands in the face of a deeper, longer crisis in East Asia 5 (Figure 4).
In the longer term (2010), provided that the Asian crisis continues to impair the industrial
development of the five East Asian economies, the farm sector in North America and its
exports are expected to contract further relative to the base case. However, the situation in
the case of processed foods is very different. Since the regional food industries in East Asia 5
rely more heavily on capital than the farm sector, the decline in capital stock and productivity
in the five Asian economies diminishes their comparative advantage in food manufacturing.
10
3
2
Services
Heavy Mnfc.
Machinery
Light Mnfc.
Apparel
Textiles
Resources
-1
Food
0
Farming
Percent change from base
1
-2
-3
-4
Korea
Indonesia
Malaysia
Philippines
Thailand
Figure 3: Change, Relative to Baseline, in the Composition of Value Added in
1999 Due to a Marginally Deeper, Longer Asian Crisis, East Asia 5 Countries
11
3000
2500
2000
US$ Millions
1500
1000
500
0
Imports
Exports
Output
-500
-1000
-1500
Farm Sector 1999
Farm Sector 2010
Food Sector 1999
Food Sector 2010
Figure 4: North American Farm and Food Sectors: Volume Changes, Relative to
Baseline, Due to a Marginally Deeper, Longer Asian Crisis
12
20
0
-20
US$ Million
-40
-60
-80
-100
-120
-140
Others
S. Asia
Africa
S. America
Thailand
Philippines
Malaysia
Indonesia
China
Taiwan
Korea
Japan
W. Europe
N. America
-160
Figure 5: Food Products: 1999 Export Volume Changes, Relative to Baseline,
Due to a Marginally Deeper, Longer Asian Crisis
13
This opens the opportunity for more exports from other regions. For example, while in 1999
the longer, deeper crisis leads to a small negative change in the North American trade balance
for processed foods, by 2010 this change turns around, and amounts to a positive US$ 1,867
million. The change in North American food output in 2010 is US$ 2,632 million compared
to the base case. It dwarfs the decline in farm output in North America in 2010 (Figure 4).
Thus, the food and farm sectors, viewed as a combined entity, actually experience higher
long-run output as a result of the deeper, longer crisis. However, the beneficiaries are those
North American food manufacturers presently competing with food processors in Asia, not
farmers.
4
Concluding remarks
The feature that distinguishes the dynamic multi-region multi-sector model, proposed in this
paper, from other dynamic global AGE models is the disequilibrium approach to modeling
capital mobility. Central to this new approach is the assumption about the adaptive expectations of investors supplying capital to the regions. This assumption, compatible with a
simple recursive solution procedure, ensures the stability of the model over the long-run. In
addition, the investment theory of adaptive expectations brings realism into the analysis of
economic phenomena such as the East Asian crisis. The model observes errors in investors’
expectations about potential returns, while offering flexibility in accommodating to empirical data. Our analysis suggests that, in the long run, the benefits to North American food
manufactures will far outweigh the losses to North American farmers from a deeper, longer
crisis in Asia. Net debtors will also cheer a deeper, longer crisis in East Asia as it continues
to put downward pressure on interest rates in North America.
14
5
References
Arndt, C., T. Hertel, B. Dimaranan, K. Huff, and R. McDougall (1998), “China in 2005:
Implications for the Rest of the World,” Journal of Economic Integration.
Baldwin, R. (1989), “The Growth Effects of 1992,” Economic Policy, 4(9), pp. 247-283.
Burniaux, J., G. Nicoletti and J. Olivera-Marins (1992),“GREEN: A Global Model for Quantifying the Costs of Policies to Curb CO2 Emissions,” OECD Economic Studies 19, OECD.
Hertel, T. W. and M. E. Tsigas (1997), “Structure of GTAP,” In Hertel, T.W., (editor),
Global Trade Analysis Modeling and Applications, Cambridge University Press, Chapter 2,
pp. 13-73.
Ianchovichina, E. I. (1998), “International Capital Linkages: Theory and Application in A
Dynamic Computable General Equilibrium Model,” Ph.D. Dissertation, Purdue University.
Ianchovichina, E. I., R. A. McDougall and T. W. Hertel (1999), “A Disequilibrium Model
of International Capital Mobility,” paper prepared for the Second Annual Conference on
Global Economic Analysis, Denmark, June 20-22.
Lucas, R. E. (1967), “Optimal Investment Policy and the Flexible Accelerator,” International Economic Review, 8(1), pp. 78-85.
McDougall, R. A. (ed.) (1997), Global Trade, Assistance, and Protection: The GTAP 3
Data Base, Chapter 3, Purdue University.
McKibbin, W. and P. Wilcoxen (1998), “The Theoretical and Empirical Structure of the
G-CUBED Model,” Economic Modelling (forthcoming).
Mercenier, J. (1995), An Applied Intertemporal General Equilibrium Model of Trade and
Production With Scale Economies, Product Differentiation and Imperfect Competition,”
C.R.D.E. and University of Montreal.
15
Treadway, A. (1969), “On Rational Entrepreneurial Behavior and the Demand for Investment,” Review of Economic Studies, 3(2), pp. 227-39.
World Bank (1998), “Global Economic Prospects and the Developing Countries,” The International Bank of Reconstruction and Development/ The World Bank, 1818 H Street, NW,
Washington, D.C. 20433 U.S.A.
Uzawa, H. (1969), “Time Preference and the Penrose Effect in a Two Class Model of Economic Growth,” Journal of Political Economy, 77, pp. 628-652.
16