Copvrig-llI © IFAC Dmalllic Mod e lling- al1(1
Control of 1'\;tlional Et:onnmies.
l' K, 1'189
SHORT AND LONG RUN PROPERTIES OF
THE BANK OF ITALY
QUARTERLY ECONOMETRIC MODEL
G. Galli, D. Terlizzese and I. Visco
R ('S(,{lrr h Dl'po rt/l/('I/t , BOII{(J d'/ta/ia. RO/l/e . Italy
Abstract. This paper analyzes the long-run equilibrium properties and the short-run
dynamics of the Bank of Italy quarterly econometric model. A demand-determined
frameyork is shoyn to prevail in the short run, yhereas supply behaviour and the
dynamics of stocks (capital and yealth) are progressively more important in the medium
and long run. The stationary state version of the model shares some, though not all,
of the properties of a neoclassical groyth model. The stability of the stationary
state is also taken up, shoying the issue to be sensitive to the dynamic specification
of the Phillips curve.
1. INTRODUCTION
Large scale econometric models are often
criticized because of their lack of transparency.
It is almost impossible, so the critique runs, to
trace the essential causal links among variables
due to the maze of feed-back relationship
implicitly established by the model
specification. As a result, the dynamics of the
economic system portrayed by the model appears as
the outcome of a black box rather than the well
understood consequence of a theory agreed upon.
To answer these critiques a recent line of
research has been developed, trying to "extract"
the core of large scale using both
analytical and numerical tools .
One possible approach is to couple the
original, dynamic model yith its steady state,
static version (the steady state which is
considered is usually a balanced growth path; in
the present paper ye shall consider a stationary
state, i.e. a zero-growth path). The idea is that
the underlying theory, yhich is most often just
an equilibrium theory, comes out most clearly in
the long-run equilibrium, i.e. when the
adjustment processes built into the origin21
specification of the model have come to an end.
The comparative static properties of a
long - run equilibrium version of the model,
however, are in general quite different from its
short-term response. Because prices and yages are
constrained to follow a relatively slow
adjustment path, neither the goods nor the labour
market need be in continuous Yalrasian
equilibrium. The response of quantities to
exogenous shocks or changes in policy variables
will be magnified by the slow speed of adjustment
of nominal variables and relative prices. A good
understanding of the dynamic properties of the
1. See, for example,
(1984), Malgrange (1988).
Deleau -Ma lgrange-Muet
2. Related to this is the important finding that
the elasticity matrix evaluated around a steady-
state growth path is time invariant and therefore
is best suited for the global analysis of the
stability properties of the model . A research
project currently underway in the Research
Department of the Bank seeks to exploit this
finding to gain a better understanding on the
dynamics of the model.
379
model thus requires simulation exercises of a
more traditional sort, performed over a
relatively short horizon.
In the present paper we shall pursue this
line of research, presenting both an analysis of
the stationary state version and of the short run
properties of the quarterly model of the Bank of
Italy . The Bank of Italy model consists of more
than 700 equations, of which about 120 are
stochastic, estimated on quarterly data over the
period 1970-1984. A detailed analysis of the
model specification and the estimation results is
presented in Banca d'Italia (1986). Focusing on
the scaffolding, we remember here that in the
model the usual IS-LM scheme is enriched by
dynamic mechanisms derived both from the
neoclassical growth theory of the sixties and
from the developments of the macro theory of the
seventies. In the short run the output is demand
determined, and the economy operates in a
Keynesian underemployment regime, yith excess
supply both in the labour market (unemployment)
and in the goods market (excess capacity) . The
output time path, however, is driven by capital
accumulation - deriving from investment decisions
by firms - by the response of prices and yages to
disequilibrium and by the evolution of private
yealth, together with the dynamics of the public
debt .
The behaviour of the supply sector is
theoretically conceivable as the result of
optimizing behaviour of monopolistically
competitive firms, which choose their factor
demand taking into account the expected quantity
constraint on the market for their output. Price
decisions are mark- up over costs and the actual
price adjusts sloyly to the desired price . Yage
dynamics is modeled by an expectation-a ugmented
Phillips curve yhich embodies a non-accelerating
rate of unemployment. Consumption decisions are
consistent yith life cycle theory.
The model includes a detailed specification
of the public sector expenditures and revenues;
public debt is part of private wealth. Financial
decisions are modeled in a portfolio frameyork,
extended to alloy for some choices to be
considered as prior claims (consumption, money) .
The flow-of-funds matrix involving the most
important sectors and financial instruments is
reconstructed .
The paper is organized as follows. In the
first section a very simplified, perhaps even
simplistic, version of the model will be