Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
Revisiting medium term macro-economic scenarios (1985 — 1995)
generated by a system dynamics model of the New Zealand economy
Robert Y. Cavana
Victoria Management School,
Victoria University of Wellington, PO Box 600, Wellington, New Zealand
Tel: +64 4 4955137; Fax: +64 4 4955253
E-mail: bob.cavana@ vuw.ac.nz
Abstract
This paper revisits the macro-economic modelling and medium term scenarios
undertaken at the New Zealand Planning Council (now disbanded) in the mid 1980's.
The following major reports were published: "A Macro-Economic Model and Scenarios
to 1995" (by E. Haywood & R.Y. Cavana) and "Towards 1995: Patterns of National
and Sectoral Development" (by D. Rose, A. Stroombergen, et al). These reports
discussed the development and use of a macro-economic system dynamics model
(SDMACRO), used to generate trends for the main macro-economic variables, and a
general equilibrium price sensitive sectoral model (JULIANNE), which generated
compatible sectoral and national forecasts of a range of variables for each of 22 sectors
for nominated years. The JULIANNE model used outputs from SDMACRO as
constraints and inputs. This paper provides a brief overview of SDMACRO and its use
at the NZ Planning Council. Also, the reforms of the New Zealand economy that have
taken place since the mid 1980’s will be summarised and a comparison of the
SDMACRO scenarios will be provided against what actually happened over the period
between 1985 to 1995. Finally, the paper discusses some of the more recent
developments that have taken place with the macro-economic model.
Introduction
SDMACRO, a medium-term system dynamics model of the New Zealand (NZ)
economy was developed within the framework of the National Sectoral Programme at
the New Zealand Planning Council. The full documentation and a computer listing of
SDMACRO, together with a discussion of a wide range of model scenarios, are
provided in Haywood and Cavana (1986). SDMACRO was developed to provide likely
trend movements, some 10-15 years into the future, in the key macro-economic
aggregates including gross domestic product, capital formation, population,
employment, exports, imports, and the current account balance.
The general structure of SDMACRO follows, to a degree, that of an earlier medium-
term model of the New Zealand economy (Haywood, 1980). Similarities include the
emphasis on the nation’s external balance as a constraint on economic growth. It was
also recognised that, for medium-term analysis, the degree of disaggregation that could
RY Cavana 1 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
sensibly be modelled or examined was limited. Finally, the emphasis on model
construction was to capture the key economic relationships and to examine the effects of
changes in these relationships.
However, while certain similarities existed to the model outlined in Haywood (1980)
a number of significant differences in approach, content and methodology were evident.
In particular, these were the inclusion of relative prices, additional feedback links, more
complex policy reaction functions, and different exchange rate regimes. In addition, the
model was developed using the system dynamics method with the DYNAMO package
(Pugh, 1976), although many of the behavioural equations have been estimated using
econometric methods. An interesting feature of this model is that it demonstrated the
complementary use of both the system dynamics and econometrics methods in the
construction of a model within a general economics framework. This model has been
discussed at an earlier International System Dynamics Conference held in San Diego,
USA (Cavana & Haywood, 1988).
The main variables, linkages and feedback loops in the model are shown in Figure |
and will be discussed briefly in this paper. The figure illustrates the interactions between
the growth/decay processes (reinforcing or positive feedback loops) and the control
mechanisms (balancing or negative feedback loops) within the NZ macro-economic
system. SDMACRO had been designed to examine the dynamic behaviour generated by
these interactions over a medium term time span (ie 5-10 years).
Employment Rate
Labour Force
Employ-
* he ment Employment
An) Policy _) ~ +
@Y
ey :
Gross Domestic
Product
> Capital Formation
+
eo
Payments +
olicy Imports
Pa eT
of + Projects.
Current Account Domestic Paces
Balance
R * y
+ Relative Prices +
Net Overseas Si. di Fides
Investment Earnings
Figure 1. A simplified causal loop diagram for SDMACRO
Source: Cavana & Haywood, 1988, Fig. 1, p29. (redrawn with Vensim package (Ventana, 2002)).
RY Cavana 2 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
Model Structure
SDMACRO contains 4 major sectors and a number of sub-sectors. These are
summarised in Table 1.
Table 1: SDMACRO Model Sectors
| Sector | Sub-Sector
OuTPuT Gross Domestic Product
Gross Fixed Capital Formation
DEMOGRAPHIC Population
Labour Force
Net Migration
Employment
EXTERNAL Goods Trade [Exports & Imports]
Exports & Imports of Services
Net Overseas Investment Income
PRICE INDICES Domestic Price Indices
[Implicit Gross Domestic Price Deflator]
[Capital expenditure price indices]
Exchange Rate Indices
Export & Import Price Indices
Source: See Appendix A of Haywood and Cavana (1986) for a DYNAMO computer listing of the
model and variable definitions.
The model structure is fully described in section 2 of Haywood and Cavana (1986).
Only a few of the sub-sectors will be described here to give an indication of the nature
and rationale for the structure and equations contained within the model. The
following material is generally taken direct from Haywood and Cavana (1986: 7-22).
Gross Domestic Product
In the early 1980s, most textbooks suggested that the purpose of macro-economic policy
should be to achieve certain objectives. In particular: economic growth, full
employment, price stability and external balance.
However, in practice trade-offs exist between the various objectives. Accordingly,
depending upon the particular conditions that prevail and objective functions of the
nation, different emphasis upon achieving the various objectives will exist.
SDMACRO does not directly include a price stability objective, due to equation
formulation difficulties. However, in its simplest form the model can be operated to
achieve a single objective, eg — balance of payments or a particular employment
objective. Alternatively, the two objectives, balance of payments and employment, can
be combined with specific weights given to each. These two effects, which are
discussed briefly below, are combined as follows in determining the growth rate of gross
domestic product in the model:
RY Cavana 3 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
QRP = WTBP*QRPBP+(-WTPB)*QRPE
QRP = percentage change in real gross domestic product
WTBP = weight given to balance of payments effect on GDP
QRTBP = balance of payments effect on GDP
QRPE = employment policy effect on GDP
It is assumed appropriate policies are adopted to maximise domestic growth
consistent with the desired external and employment objectives. No view on what
specific set of policies should be adopted to achieve this goal is given. In the Base Case
WTBP is given a weighting of 70 percent.
Employment policy effect on GDP: The employment response objective function is
incorporated in the model as the graphical relationship in Figure 2 below which
specifies the desired employment rate in the next year given the existing employment
rate. The desired increase in employment that, via an employment-domestic growth
equation, provides an estimate of the increase in domestic activity that would need to
occur to meet the desired increase in the employment rate in the next period. Full
employment, which is assumed desirable, is specified as 2 percent unemployed.
Figure 2: Employment objective function
Source: Haywood & Cavana, 1986, Fig. 2.1, p8.
RY Cavana 4 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
Balance of payments effect on GDP: The model assumes that GDP growth is dependent
on the relationship between the current account balance and the level of GDP. This is
modelled as the graphical relationship in Figure 3. For example, if the current account
balance as a ratio of GDP was —3 (minus 3) percent, then it is assumed that the desired
rate of growth in the economy is 3 percent. Whether this growth rate is maintained or
even achieved depends upon the relative weight given to the external balance objective,
and secondly, to the various factors determining the current account in following
periods:
Desired ¥ change
in real GDP next
4 period
Ratio of curre:
t balanc:
Figure 3. GDP — Current Account Balance Objective Function
Source: Haywood & Cavana, 1986, Fig. 2.2, p9.
Population, Labour Force and Net Migration
The NZ Department of Statistics estimates of population and labour force, assuming
zero migration, are input into the model as exogenous projections. The effects of net
migration modify the population and labour force projections. Migration is calculated in
the model as the sum of normal net migration flows plus the net inward migration
resulting from an active immigration policy. Normal net migration is assumed to depend
on economic conditions within New Zealand. The result of estimating the relationship
between the five yearly moving average data of net migration and the annual percentage
change of real gross domestic product is:
MIGN = -18262 + 7357*QRP R?=0.77
(5.1) (6.6) 1965-80
MIGN- =_ normal net migration (persons/year)
QRP. = percentage change in real GDP
(t values in brackets)
RY Cavana 5 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
At high levels of employment (ie over 97 percent), it was assumed that an active
immigration policy (see Figure 4) will be introduced by Government to reduce labour
shortages, providing that the balance of payments is at the same time in a satisfactory
position. For example, at an employment rate of 98 percent, it is assumed that an
additional net 5,000 persons are attracted to New Zealand above that associated with the
normal net migration which is related to the growth in domestic economy.
Figure 4. Active Immigration Policy Response Function
Source: Haywood & Cavana, 1986, Fig. 2.3, p15.
Employment
The percentage change in employment in the model is directly related to movements in
domestic output. The estimated relationship is:
EP= 0.522 + 0.435*QRP F
(3.1) (8.0) 1965-80
Exports and Imports of Goods
Three categories of export goods are assumed: traditional (meat, wool and dairy
products), non-traditional (forest products, manufacturing and other primary products),
and those exports associated with the "additional" large-scale projects’. In the base case,
traditional and non-traditional exports are assumed to increase at annual real rates of 1
and 6 percent respectively. The assumed growth rates are modified if relative price
movements between domestic and export prices occur. The price elasticity for
' The large scale projects in New Zealand in the early 1980’s were those capital projects supported by the
Government, and including ammonia-urea, synthetic gasoline, refinery expansion and methanol. These
projects could not readily be explained by movements in domestic activity.
RY Cavana 6 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
traditional exports was assumed to be 0.3 and for non-traditional exports 0.7, giving on
aggregate an overall price elasticity of 0.5.
Various categories of imports of goods are assumed. These are traditional (consumer,
capital and intermediate goods imported), and those imports/import savings associated
with the large-scale projects. In the model the percentage change in traditional imports is
related to the percentage change in real GDP and the difference between the percentage
changes in import prices and domestic prices:
MGTRP = -5.4 + 2.878*QRP - 0.567*(PMP-PDP) R= 0.92
(3.9) (6.9) (3.8) 1965-80
MGTRP = percentage change in traditional imports of goods
QRP. = percentage change in real GDP
PMP percentage change in import price index
PDP = percentage change in domestic price index
Exports and Imports of Services
Exports of services are assumed to be exogenous to the model with a growth rate of 6%
p.a. in the base case. This rate is modified by the relative price movements and a price
elasticity of 0.5. Imports of services are estimated at the historical ratio of 18 percent of
total goods trade.
Net Overseas Investment Income
It is assumed "investment income" credits move 3.5% above world inflation. "Direct
private investment income" debits, i.e. foreign investment in NZ, is assumed to move
directly with the movement in domestic activity as represented by money GDP. "Other
investment income" debits, primarily interest paid on overseas loans by Now
Zealanders, is broken into two categories. Past payments are assumed to represent
interest and will remain at that level adjusted for movements in the exchange rate. The
interest rate on new loans required to cover the current account balance are borrowed at
3 5% above the estimated world inflation rate.
Domestic Price Index
The percentage change in the domestic price index (PDP) consists of two elements -
internally generated inflation (PIP), assumed to be 10% in the base case, and imported
inflation (PMP) These elements are weighted according to their approximate importance
in total domestic economic activity:
PDP = 07 PIP +0 3 PMP
Exchange Rate Index
The model can be set to operate in either a “fixed” or "flexible" exchange rate mode. In
the base case the exchange rate, in its "flexible" mode, adjusts according to relative
differences that emerge between domestic and external prices so that NZ's competitive
position can be retained internationally Also an option exists in the model which
RY Cavana 7 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
introduces a constant "under" or "over” shooting of the assumed exchange rate
adjustment, which causes a gradual depreciation or appreciation of the nation's real
exchange rate over time
Export and Import Price Indices
The external price indices are made up of two components: overseas prices in foreign
currency and the exchange rate In the base case export and import prices, in foreign
currency, are assumed to grow at 6 and 6 5 percent per annum respectively
Model Behaviour
A large number of simulation runs with SDMACRO are reported in Haywood and
Cavana (1986). However, only the base case and the optimistic and pessimistic
scenarios will be presented here. The model commences simulation from a base position
of 31 March 1983. The values for that year are the average of the 1982-4 actual values.
This procedure being adopted to ensure that the model commences its projections from
values that are close to the series medium-term trend values, at that time.
The optimistic scenario assumes that the nation's terms of trade will improve by 0.5%
p.a. (compared with -0.5% p.a. for the base case) and the real annual growth in
traditional exports will be 1.8% (1%), non-traditional exports 8% (6%) and services
8% (6%).
The pessimistic scenario assumes that the nation's terms of trade will decline by 1.5%
p.a. and the annual real growth in traditional exports will be 1%, non-traditional exports
4% and services 4%. In addition, it is assumed that there will be no further "additional"
benefits from the large scale projects (compared with 50% additionality for the base
case) and the level of import substitution will be less than the historical rate (i.e. the
constant coefficient in the import equation above will be -4% rather than -5.4% for the
base case).
Figure 5 displays the model results for the Base Case and optimistic/pessimistic
scenarios for the change in real gross domestic product. Figure 6 illustrates the Base
Case output for the labour force, employment and unemployment levels. Historical data
from 1960 to 1983 are also shown for comparative purposes.
RY Cavana 8 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
Years ended ™
Figure 5. Combined pessimistic & optimistic scenarios: Real GDP
Source: Haywood & Cavana, 1986, Fig. 3.17, p42.
1000 persons)
1,400
1,200
2,000
*Census years
Figure 6. Base Case: Labour force, employment & unemployment
Source: Haywood & Cavana, 1986, Fig. 3.3, p28.
Role of SDMACRO in the National Sectoral Programme
The purpose of the NZ Planning Council's National Sectoral Programme was to provide
a general indication of the likely sectoral and national development path of the New
Zealand economy some five to ten years in the future. The programme, which is
discussed fully in the report of the National Sectoral Working Group (1986), comprised
three major elements:
RY Cavana 9 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
¢ a programme of research and consultations with a range of sectoral organisations in
both the private and public sectors to provide qualitative and quantitative
information for the model runs.
e a system dynamics model, SDMACRO, designed to produce medium term trend
values of the major macro-economic variables.
e a non-linear general equilibrium model, JULIANNE, which generates compatible
sectoral and national projections of a range of variables including output, exports
and employment for each of 22 sectors for nominated future years.
The relationship between the projections generated by SDMACRO and JULIANNE
is illustrated in Figure 7. SDMACRO provided a continuous growth path for selected
macro-economic variables, e.g. GDP, employment and investment up to 1995 from a
base year of 1983, whereas JULIANNE provided "snap shots" of two future time
periods, 1990 and 1995, from a base year of 1982.
Figure 7
Relationship of SDMACHU and JULIANNE Forecasts
average
GDP growth
not to JULIANNE 1990 to 1995
scale) 1920
Input-Output
average Table 7
rowth
base to 1990 ‘ SUCANNE
wan 1995
es Input-Outy
| Lf % average Parable me
growth
SDMACRO base to 1995
annual GDP
: Growth estimates
base: SDMACRO 1983 1990 1995
JULIANNE 1982 Years ended March
Source: National Sectoral Working Group, 1986, Fig. 3.3, p22.
The SDMACRO base run was used with information from the sectoral consultations
to constrain the JULIANNE calibration run. In this run JULIANNE was required to
replicate in 1990 and 1995 SDMACRO figures for employment, the real exchange rate,
the terms of trade, the balance of trade and total exports; to work within a total capital
stock which was compatible with the SDMACRO time profile of capital formation; and
to approximate the investment to GDP ratio in SDMACRO in the nominated years. In
addition, JULIANNE was required to replicate sectoral export growth rates and sectoral
rates of depreciation and technical change as derived from the consultations. The major
results of this calibration run indicated that to secure the level of exports foreseen in the
sectoral consultations the model required substantial increases in export subsidies above
RY Cavana 10 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
the level prevailing in 1981/82 and to secure a level of imports compatible with the
SDMACRO trade balance required the removal of all tariff protection.
However, to produce results more consistent with the policy settings in New Zealand
in the early to mid 1980’s, the JULIANNE current policy run required four major
changes from the calibration run: the model was instructed that no subsidies were to be
paid on exports; for competing imports tariff equivalents were to be reduced to a
maximum of 25 percent and tariffs on non-competing imports were to be reduced to
zero; and to enable the model to determine its own export prices, the SDMACRO
constraints on terms of trade and the real exchange rates were removed. This run
resulted in lower export growth rates and a lower ratio of imports to GDP. In the rerun
SDMACRO was required to replicate JULIANNE'S projections in 1990 and 1995 for
the terms of trade, real exchange rate, and real exports of goods and services. The lower
levels of employment and investment generated by the rerun of SDMACRO were used
to constrain a further rerun of JULIANNE. At this stage the sectoral implications of the
runs outlined above were examined and some further policy experiments with
JULIANNE undertaken.
Comparison of SDMACRO Scenarios with Actual Performance of NZ Economy
The policy environment changed dramatically with the election of the Labour
Government in New Zealand in July 1984. Previously, New Zealand had been a very
tightly regulated economy. This change of Government marked the beginning of a
systematic and comprehensive programme of economic reforms in NZ. These economic
reforms included major changes to international trade policy, monetary and fiscal
policies, industry policy and labour policy. They have been fully discussed in Dalziel &
Lattimore (2001), and Evans et al. (1996). Some of the reforms included:
e ‘Financial market - removal of interest rate and foreign exchange controls; and
floating of the NZ dollar (exchange rate).
¢ Goods market — removal of price and rent freezes, consumer and producer
subsidies, export assistance, import licensing and tariffs; and introduction of
light handed regulation.
e¢ Monetary Policy - introduction of the primary objective of monetary policy to be
price stability, setting the Reserve Bank a target for increases in the Consumer
Price Index of between 0-2 percent (Reserve Bank Act).
e Public Sector Reforms - corporatising and privatising government trading
organizations (e.g. sale of NZ Railways Corporation); reforming Government
Departments; introduction of the Fiscal Responsibility Act.
e Taxation Reforms - introduction of a 12.5% Goods and Services Tax; Fringe
Benefit Taxes; and company and personal income tax cuts.
e Industry Reform — incl. deregulation of the road transport industry; opening up of
domestic air routes to competitors.
e International Trade - establishment of a Trans-Tasman free market under the
terms of the Australia and New Zealand Closer Economic Relations Trading
Agreement.
e Labour Policy — reform of laws governing labour relations; ending compulsory
union membership; and introduction of the Employment Contracts Act.’
(Cavana, 2004, p181)
RY Cavana OT Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
“The decade of radical economic reform between 1984 and 1994 was a very
important period in New Zealand’s economic history. It saw the country’s
transition from being one of the most tightly regulated and controlled economies in
the world to being one of the most liberal market-based economies anywhere”
(Dalziel & Lattimore, 2001, p31)
These reforms, which started with the election of the Labour Government in New
Zealand in July 1984, continued with the election of the National Government in
October 1990. A chronology of the major economic reforms in New Zealand between
1984 to 1995 is provided in Appendix 1.
The economic environment over the period from 1984 to 1995 has been
considerably different to the environment that existed in the early to mid 1980’s when
the modelling work was being carried out at the New Zealand Planning Council.
Nevertheless, some of the policy changes were examined within the National Sectoral
Programme and reported in Haywood & Cavana (1986) and the National Sectoral
Working Group (1986).
Although the policy environment was quite different to that expected, it is useful to
compare some of the SDMACRO generated scenarios with the actual outcomes over
this period. Table 2 summarises a range of comparative figures, including final value
(1995) values and average growth rates over the modelling period (1983 to 1995). By
and large the economy performed at slightly worse than the Base Case on average over
the 10 year period.
However, the policy changes in New Zealand were much more dramatic than
envisaged in the early 1980’s, and together with the share market crash in 1987, the NZ
economy went through a longer than expected ‘adjustment’ period, showing signs of
considerable improvement from about 1993 onwards (see Figures 8-10). This is also
indicated by the real investment to GDP ratio of about 33 percent in 1995, much higher
than even the most positive scenario generated by SDMACRO (see Table 2).
Hence it can be concluded that the SDMACRO macro-economic modelling and
scenarios did provide some useful insights into the emerging nature of the New Zealand
economy over the period from 1985 to 1995.
RY Cavana 12 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
Table 2
Comparisons of SDMACRO Scenarios with the Actual Major Macro-Economic Indicators
SDMACRO Final Values (1995) Actual
Initial Final
Values Base Pessimistic Optimistic Values
(1983) Case Scenario Scenario —(1995)
Real GDP
Value (1983 $000m) 32.0 46.0 37.5 51.8 40.4
Growth Rate, 83-95 (% p.a.) 34 13 44 2.0
Real GDP per capita
Value (1983 $000m) 9.9 12.9 10.9 144 11.4
Growth Rate, 83-95 (% p.a.) - 22 08 3.0 1.2
Real Investment/GDP ratio (%) 22.5 18.7 14.0 25.0 32.7
Population
Number (000) 3,230 3,579 3,426 3,680 3,539
Growth Rate, 83-95 (% p.a.) - 0.9 05 1 08
Total net migration between 1983-95 (000) - 53 (100) 154 (21)
Employment no. (000) 1,300 1,622 1,483 1,707 1,608
Unemployment rate (%) 55 18 49 03 69
Real Exchange Rate 100 99 99 99 na
Terms of Trade 100 94 84 106 118
Real Exports of Goods & Services
Value (1983 $000m) 94 15.4 11.8 19.3 16.4
Growth Rate, 83-95 (% p.a.) - 4.0 19 62 47
Current account balance/GDP ratio (%) (3) (2.4) (4.7) (0.4) (38)
Sources: Haywood & Cavana (1986); actual final values derived from Statistics New Zealand statistics
summarised in Appendix 2.
RY Cavana
Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
+1 4 —
8 70
Figure 8. Comparison of SDMACRO scenarios and actual outcomes: Real GDP growth rates
Source: Haywood & Cavana, 1986, Fig 3.17, p42; and NZ Department of Statistics
Figure 9. Comparison of SDMACRO scenarios and actual outcomes: Unemployment rates
Source: Haywood & Cavana, 1986, Fig 3.18, p43; and NZ Department of Statistics
RY Cavana 14 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
(000 persons /year)
20
20 P
»t/™
-10
=20
|
Figure 10. Comparison of SDMACRO scenarios and actual outcomes: Net migration
Source: Haywood & Cavana, 1986, Fig 3.19, p43; and NZ Department of Statistics
Development of the Three Sector Macro Model
Although the NZ Planning Council was disbanded in the late 1980’s, some work still
continued on both the macro-economic trend model (SDMACRO) and the general
equilibrium model (JULIANNE). The development work was mostly carried out at
Business and Economic Research Ltd (BERL) in Wellington, with funding assistance
from the Foundation for Research Science and Technology, and continuing input from a
number of economists at other institutions, including Victoria University of Wellington.
The development of the macro-economic trend model was principally carried out by
Eric Haywood (reported in Haywood et al. 1993). The macro control model (now called
TRI model), followed the standard approach of modelling the nation's main national
accounting series, but with the difference that output is now divided into three sectors
(Production module) (SDMACRO only had one sector), while growth is achieved via a
set of reaction functions (Control module) contained in the model.
The Production Module
In the TRI model (Haywood, et al. 1993), the economy is divided into three sectors:
primary, secondary and tertiary, with each sector having its own stock of capital and
employment. The overall increase in GDP as estimated elsewhere in the model is split
between the three sectors by a simple routine. The output of the primary sector is
assumed to change at the same rate as do exports of the additional commodities. The
remaining part of the absolute increase in GDP is split between the other two sectors in
fixed proportion.
RY Cavana 15 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
For each sector changes in future demand for labour and capital per unit of output
are driven by exogenously forecast rates of change in capital and labour productivity.
These underlying rates of change are modified in accordance with changes in a capacity
utilisation as proxied by the current and trend values for the rate of growth in sectoral
output. Model requirements of labour and capital per unit of output are also sensitive to
changes in the costs of labour and capital. Differences between sectoral rates of change
in the cost of labour and capital per unit of output are applied through exogenous
elasticities of labour capital substitution. These routines are sufficient to generate
forecasts of total employment and the required stock of capital in each sector.
The Control Module
Growth in the TRI model is determined via various "reaction-functions." The reaction
functions fall into two distinct categories - direct and indirect.
The model contains two direct reaction functions relating to employment and the
balance of payments. It is assumed, for example, that increasing unemployment beyond
a specified level will cause a reaction which will see policies or actions implemented
that will increase domestic activity to increase employment. Exactly what policies or
changes would occur to cause an increase in activity are not specified, merely that
activity will be increased. Offsetting or reinforcing this is the assumption that the
nation's balance of payments position, and thereby the level of net overseas liabilities, is
a direct major determinant of growth. Not an unreasonable assumption in a small open
economy such as New Zealand’s. An improvement/deterioration in the current account
balance is assumed to lead to an increase/decrease in domestic economic activity.
Complementing the above direct reaction functions on the determination of growth
is a set of indirect functions. For example, there is a set of reaction functions dealing
with interest rates, inflation, and the exchange rate, which indirectly impact on the
growth rate. A fuller description of this module is provided in Haywood et al. (1993).
Use of the TRI Model
The new three sector macro-economic trend model (TRI) has been used successfully for
a number of different purposes; including forecasting macro and sectoral trends in the
economy (Haywood ef al. 1993) and forecasting medium term trends in occupational
patterns of employment in New Zealand (Andrews & Rose, 1995). Both of these studies
also involved the use of the general equilibrium model, JULIANNE. The macro trend
model has been used to examine the macro-economic implications of variations in the
net flow of foreign direct investment in New Zealand (Rose, 1996).
Concluding Comments
The macro-economic trend model discussed in this presentation has evolved quite
significantly since its initial development as a FORTRAN model at the New Zealand
Planning Council by Eric Haywood over 20 years ago. It developed into a system
dynamics model in the early to mid 1980’s (also at the New Zealand Planning Council)
and more recently it has been further enhanced and disaggregated into a three sector
model. However, it has been used successfully for a variety of forecasting and scenario
RY Cavana 16 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
modelling projects, which require medium term trend values of the major macro-
economic variables for the New Zealand economy.
In addition, it is still extremely valuable in modelling work to reflect back on the
scenarios and forecasts generated by a model. This is a form of validation, as well as
helping to build improved understanding of the system that is being modelled. In this
way better and more robust models can be built for the future.
The model discussed in this paper (SDMACRO) was used to create macro economic
scenarios and forecasts of the New Zealand economy over the period from 1985 to
1995. This was a period during which considerable economic reform and restructuring
took place. This “revisiting the New Zealand Planning Council’s medium term macro-
economic scenarios” is another step in the process of model development and
refinement. Also it is hoped that it will lead to more insight into the medium to long
term effects of some of the policy changes that have occurred in the New Zealand
economy.
Acknowledgments
T would like to acknowledge all the assistance provided by former colleagues at the NZ
Planning Council and at Business and Economics Research Ltd in Wellington, New
Zealand. In particular, I would like to thank Eric Haywood since this paper is based
mostly on the collaborative work we undertook at the NZ Planning Council in the
1980’s. I would also like to thanks Dennis Rose, Adolph Stroombergen and Bryan
Philpott (now deceased) for their help and support. However, the views and
interpretations within this presentation are the author’s alone.
References
Andrews G, Rose D. 1995. Forecasting the Occupational Pattern of Employment.
Presented at the August 1995 Conference of the NZ Association of Economists,
Lincoln University.
Cavana RY. 2004. A qualitative analysis of reintroducing cabotage onto New
Zealand's coasts. Maritime Policy and Management: An International Journal of
Shipping and Port Research 31(3): 179-198.
Cavana RY, Haywood E. 1988. A System Dynamics Model of the New Zealand
Economy. Proceedings of the 1988 International System Dynamics Conference.
San Diego, USA. pp27-38
Dalziel P, Lattimore R. 2001. The New Zealand Macroeconomy: A Briefing on the
Reforms and their Legacy 4th ed. Oxford University Press: South Melbourne,
Australia.
Evans L, Grimes A, Wilkinson B, Teece D. 1996. Economic reform in New Zealand
1984-95: The pursuit of efficiency. Journal of Economic Literature, XXXIV, 1856-
1902.
Haywood E. 1980. Forecasting the Economy in the Eighties. New Zealand Planning
Council: Wellington (Planning Paper No. 10).
Haywood E, Cavana RY. 1986. A Macro-Economic Model and Scenarios to 1995.
New Zealand Planning Council: Wellington (Planning Paper No. 24).
Haywood E, Nana G, Rose D. 1993. Prospects for Growth Modelling the Economy to
1997. Presented at the August 1993 Conference of the NZ Association of
Economists, Dunedin.
RY Cavana 17, Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
National Sectoral Working Group. 1986. Towards 1995: Patterns of National and
Sectoral Development. New Zealand Planning Council: Wellington (Planning
Paper No. 26).
Pugh III AL. 1976. DYNAMO User's Manual (5" ed.). MIT. Press: Cambridge, MA.
Rose D. 1996. Modelling the Macro Economic Implications of Variations in the Net
Flow of Foreign Direct Investment. Business & Economics Research Ltd:
Wellington, May (Unpublished report prepared for Peter Shirtcliffe).
Statistics New Zealand. 1995. New Zealand System of National Accounts 1995,
Statistics NZ: Wellington. pp. 147-149.
Statistics New Zealand, Key Statistics & Monthly Abstracts, Various Issues, Wellington.
Ventana 2002. Vensim® 5 User’s Guide. Ventana Systems: Harvard, MA.
Appendix 1
Chronology of Major Economic Events/Reforms in New Zealand: 1984 - 1995
1984 Feb-84 Price freeze, imposed in June 1982, lifted. Wage freeze extended until an agreement on
long-term wage fixing procedure is reached.
Jun-84 Supplementary Minimum Payments (SMPs) to farmers removed from end of the 1983/84
season.
Jul-84 Price freeze re-introduced for 3 months.
Jul-84 Labour Government elected.
Jul-84 NZ dollar devalued 20 percent.
Jul/Aug-84 Controls on interest rates and credit growth abolished.
Oct-84 New wage-fixing rules introduce greater flexibility in bargaining and wage freeze, imposed
in June 1982, lifted.
Nov-84 Budget announces intention to eliminate numerous consumer and producer
subsidies/incentives.
Nov-84 Price freeze, re-introduced in July, lifted.
Dec-84 All controls on both outward and inward foreign exchange transactions removed.
Dec-84 Controls on overseas borrowing removed.
Dec-84 New motor vehicle industry plan allows for greater access for imported vehicles and
components.
1985. Feb-85 Compulsory ratio system requiring financial institutions to invest fixed proportions of their
total funds in government and public securities abolished.
Mar-85 NZ dollar floated.
1986 1986 Complete phasing-out of permits for long-distance road haulage.
Progressive removal of imposts on road transport, including excise taxes on fuel and reduced
import tariffs on tyres and trucks.
Jul-86 Government assumes responsibility for major project and producer board debts totalling
$7.2 billion.
Oct-86 Goods and Services Tax (GST) of 10 percent introduced. Compensating reductions in
personal tax and benefit increases accompany the new tax. Some sales duties/taxes also
removed or reduced.
1987 1987 Domestic air routes opened to competitors.
Apr-87 Institutions, including overseas organisations, able to apply for banking licenses.
Apr-87 Ten new state corporations (SOEs) established.
Jun-87 Government announces intention to sell assets to pay off public debt.
Aug-87 Labour Relations Act 1987. Rationalises existing legislation under one statute and reforms
laws governing labour organisations to improve negotiating environment.
Oct-87 World sharemarket crash. NZ sharemarket suffers its biggest ever one day fall.
1988 1988 Remaining price controls on petrol abolished.
RY Cavana 18 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
1989
1990
1991
1992
1993
1994
1995
Mar-88 Sale of NZ Steel to Equiticorp begins the government's programme of stateowned asset
sales,
Apr-88 State Sector Act 1988 comes into force. The Act restructures public sector management,
introducing increased management flexibility for senior public servants along with increased
accountability for performance, and aligns public sector with private sector employment
regulation.
Jul-88 Import licensing ends for most goods.
Aug-88 Australian and New Zealand Prime Ministers sign an agreement to bring forward the date
for establishing a Trans-Tasman free-market to July 1990 under the terms of the Australia-
New Zealand Closer Economic Relations Trading Agreement, Protocols signed to remove
most trade barriers between Australia and New Zealand.
Jul-89 GST increased to 12.5 percent.
Dec-89 Reserve Bank of New Zealand Act passed. As from February 1990 the Act defines the
objective of monetary policy to be the achievement and maintenance of stability in the general
level of prices.
Mar-90 First Reserve Bank policy targets agreement signed. The Minster of Finance and the
Governor of the Reserve Bank reaffirm the objective of price stability and set an annual CPI
increase of 0 - 2 percent by December 1992 as a target.
Oct-90 National Government elected.
Dec-90 New Reserve Bank target agreement signed extending the achievement of the price
stability target to December 1993.
Economic package introduces across-the-board benefit cuts from April1991.
Dec-90 Compulsory union membership ends.
May-91 Employment Contracts Act introduced.
Jul-92 ACC reforms introduced. The scheme’s source of funds are realigned reducing the share
incurred by business and increasing personal contributions. Entitlement redefined, including
the abolition of lump sum payments.
Dec-92 The Government and Reserve Bank sign a new Policy Targets Agreement requiring the
Bank to keep 12-monthly increases in the CPI within the range of 0 - 2 percent.
Jan-93 Sealord Products bought by a Maori/Brierley Investments venture. The Government's
financing of the Maori half share is in exchange for the ending of all Maori claims to
commercial fishing under the Treaty of Waitangi.
Dec-93 Agreement reached in the Uruguay Round of the General Agreement on Tariffs and Trade
(GATT). The agreement will lead to a more open world trading environment through the
agreed process of progressive liberalisation.
Jun-94 The Government records a financial surplus of $755 million for the 1993/94 financial year.
Nov-95 The Royal Assent to the $170 million Tainui Maori land settlement in compensation for
lands confiscated in 1884 is signed.
Source: Statistics New Zealand, 1995, New Zealand System of National Accounts 1995,
Wellington, pp. 147-149.
RY Cavana 19 Revisiting SDMACRO scenarios, 1985-95
Paper for the International Conference of the System Dynamics Society, July 17-21 2005, Boston USA
Appendix 2
Actual Macro-Economic Variables for the New Zealand Economy, 1983 - 1995
Real Real Real NetOverseas Current Gross Ratio
Year Gross Gross Fixed Exportsof Investment Account Domestic CAB to
Domestic Capital Goods & Earnings Balance ~— Product. GDP
Product Formation —_—Services
(1982-83$m) (1982-83$m) (1982-83$m) ($m) (sm) ($m) (%)
1983 31,561 7,827 9,116 (867) (1,945) 31,561 (6.2)
1984 32,422 8,272 9,756 (1,362) (1,992) 35,049 (6.7)
1985 34,022 8,650 10,454 (1,262) (2,700) 39,677 (6.8)
1986 34,284 9,351 10,944 (1,460) (2,845) 45,777 (6.2)
1987 35,005 8,819 11,447 (2,705) (2,878) 55,024 (5.2)
1988 35,471 9,353 12,227 (3,198) (2,506) 62,536 (4.0)
1989 35,212 9,111 12,392 (2,627) (707) 67,228 (1.1)
1990 35,800 9,956 12,126 (2,071) (2,576) 71,406 (3.6)
1991 35,589 9,550 13,053 (2,076) (1,958) 72,962 (2.7)
1992 35,129 8,031 13,879 (4,035) (1,881) 73,030 (2.6)
1993 36,169 8,777 14,031 (2,957) (1,251) 75,220 (1.7)
1994 38,107 10,555 15,285 (3,277) (1,054) 79,999 (1.3)
1995 40,421 13,207 16,365 (4,835) (3,310) 86,304 (3.8)
Terms Consumer Population Net Number Labour — Unemployment
Year of Price Migration Employed Force Rate
Trade Index
(000) (000) (000) (000) (%)
1983 954 488 3,190 3 1,279 4,355 5.6
1984 962 505 3,231 7 1,293 4,371 57
1985 942 572 3,259 (8) na na na
1986 919 647 3,273 (22) 1,545 1,613 42
1987 985 765 3,284 (14) 1,558 4,625 44
1988 1074 834 3,316 (16) 1,529 1,609 5.0
1989 1139 867 3,318 (25) 1,466 1,582 74
1990 4172 928 3,336 (4) 4,471 1,587 73
1991 1095 970 3,373 12 1,464 1,624 99
1992 1081 978 3,416 4 1,450 1,630 144
1993 1125 987 3,452 7 1,475 1,643 10.2
1994 1118 4000 3,491 16 1,532 1,693 95
1995 1125 1040 3,539 22 1,608 1,787 69
Sources: Statistics New Zealand, Key Statistics & Monthly Abstracts, Various Issues, Welington.
Statistics New Zealand, 1995, New Zealand System of National Accounts 1995, Welington.
RY Cavana
20
Revisiting SDMACRO scenarios, 1985-95
CC BY-NC-SA 4.0