CONVERGENCE OF LEAST-SQUARES LEARNING TO A NONSTATIONARY EQUILIBRIUM

被引：8

作者：

EVANS, GW

HONKAPOHJA, S

机构：

[1] UNIV HELINSKI,DEPT ECON,BOX 54,SF-00014 HELSINKI,FINLAND

[2] UNIV EDINBURGH,DEPT ECON,EDINBURGH EH8 9JY,SCOTLAND

[3] ACAD FINLAND,HELSINKI,FINLAND

来源：

ECONOMICS LETTERS | 1994年 / 46卷 / 02期

关键词：

D O I：

10.1016/0165-1765(94)90007-8

中图分类号：

F [经济];

学科分类号：

02 ;

摘要：

Linear rational expectations models can have explosive equilibria. For a linear model with multiple AR(1) solutions we show that least squares learning can converge to an explosive AR(1) equilibrium if a stability condition is met.

引用

页码：131 / 136

页数：6

共 8 条

[1] RATIONAL-EXPECTATIONS EQUILIBRIA, LEARNING, AND MODEL-SPECIFICATION [J].

BRAY, MM ;

SAVIN, NE .

ECONOMETRICA, 1986, 54 (05) :1129-1160

[2]

EVANS GW, IN PRESS INT EC REV

[3]

EVANS GW, 1993, UNPUB EC DYNAMICS LE

[4]

EVANS GW, 1993, IN PRESS ECONOMETRIC

[5]

EVANS GW, 1991, STICERD TE90212 LSE

[6] ANALYSIS OF RECURSIVE STOCHASTIC ALGORITHMS [J].

LJUNG, L .

IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1977, 22 (04) :551-575

[7] CONVERGENCE OF LEAST-SQUARES LEARNING-MECHANISMS IN SELF-REFERENTIAL LINEAR STOCHASTIC-MODELS [J].

MARCET, A ;

SARGENT, TJ .

JOURNAL OF ECONOMIC THEORY, 1989, 48 (02) :337-368

[8]

Marcet A., 1989, EC COMPLEXITY CHAOS, P119

← 1 →