larger. Eventually, economic agents must turn towards more robust expectation
formation technologies as the loss in terms of utility resulting from large errors
of perception outweighs information gathering and processing costs. However,
this puts the economy on a convergent path and again the economy approaches
a potential rest point. When the economy is sufficiently close to a rest point
agents turn again towards adaptive expectation formation and the cycle repeats
itself. As a consequence the economy oscillates between two distinctive regimes:
a convergence and a divergence regime. Whenever, the economy is in the con-
vergence regime agents form expectations in an adaptive manner this, however,
pushes the economy into the divergence regime. Moreover, in the divergence
regime agents form expectations in a nearly perfect foresight manner. This,
however, introduces stability and pushes the economy back into the conver-
gence regime. The cycle repeats itself and results in permanent and endogenous
fluctuations in aggregate variables.
The paper is written is an explicit macroeconomic context. Specifically, the
paper develops a general equilibrium model that comprises the Diamond [14]
OLG model, the Blanchard and Kiyotaki [5] monopolistic competition para-
digm, and the Brock and Hommes [7] framework of expectation selection as the
three key ingredients. In the model the presence of monopolistic competition
allows the composition of aggregate demand to affect aggregate variables, i.e.,
the model exhibits multiplier characteristics. In particular, the higher the frac-
tion of income saved the lower the output and vice versa. In the context of the
paper economic agents form assessments of their future income streams. An ex-
pectation of favorable conditions in the future leads through the intertemporal
smoothing channel to low savings and in turn, through the multiplier effect, to
high output, while an expectation of low future income increases savings and
leads through the multiplier effect to low output today. Expectations, however,
are not assumed to be exogenous. Quite the opposite the process of expecta-
tion formation is a part of the equilibrium and any shifts in expectations are
equilibrium outcomes.
Moreover, the paper shows that the fact that expectations need not settle on
a specific value even in the long run can be deeply rooted in economic properties
of macroeconomic systems. Observe that equilibrium outcomes depend in part
on the amount of resources available. In particular, future income depends
on the future capital stock. Specifically, a higher level of capital stock implies,
other things equal, a higher level of income. Furthermore, a high level of income
in the future implies, ceteris paribus, low savings at the present. However,
savings determine future capital stock. Therefore, low savings result in low
capital stock and hence in a low income in the future. This basic mechanism
illustrates that expectations of economic agents expecting favorable conditions
in the future and saving little must necessarily be invalidated when the future
actually arrives. The paper shows that if the ex post discrepancy is large enough
then the existence of this basic natural negative feedback mechanism can lead
to endogenous fluctuations in aggregate demand.
In addition, the paper delivers several predictions with the respect to pol-
icy prescriptions. It shows that the movements in aggregate output obtained
32