國立高雄大學圖資館 |

Language: English

Back

Exploiting sequential learning to es...

Greenstreet, David L.

Exploiting sequential learning to estimate microeconomic productivity dynamics in strategic settings.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Exploiting sequential learning to estimate microeconomic productivity dynamics in strategic settings.
Author:	Greenstreet, David L.
Description:	137 p.
Notes:	Adviser: Kai-Uwe Kuhn.
Notes:	Source: Dissertation Abstracts International, Volume: 67-07, Section: A, page: 2671.
Contained By:	Dissertation Abstracts International67-07A.
Subject:	Economics, General.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3224892
ISBN:	9780542786921

Exploiting sequential learning to estimate microeconomic productivity dynamics in strategic settings.
Greenstreet, David L.

Exploiting sequential learning to estimate microeconomic productivity dynamics in strategic settings. - 137 p.

Adviser: Kai-Uwe Kuhn.

Thesis (Ph.D.)--University of Michigan, 2006.

Because the sequential learning estimator relies upon belief updating rather than optimal strategic behavior or first-order conditions, it may be used under any industry structure and with any type of productivity dynamics. Consequently, estimated productivity beliefs can be used in second stage modeling of the determinants of firms' actions. Productivity dynamics, innovation effort, and other strategic behaviors can be compared across industries, under differing industry states, and with varying external economic conditions and public policies.

ISBN: 9780542786921Subjects--Topical Terms:

212429
Economics, General.

Exploiting sequential learning to estimate microeconomic productivity dynamics in strategic settings.
LDR:03613nmm _2200301 _450 001 180608
005 20080111103752.5
008 090528s2006 eng d
020 $a 9780542786921
035 $a 00311633
040 $a UMI $c UMI
100 0 $a Greenstreet, David L. $3 264184
245 1 0 $a Exploiting sequential learning to estimate microeconomic productivity dynamics in strategic settings.
300 $a 137 p.
500 $a Adviser: Kai-Uwe Kuhn.
500 $a Source: Dissertation Abstracts International, Volume: 67-07, Section: A, page: 2671.
502 $a Thesis (Ph.D.)--University of Michigan, 2006.
520 # $a Because the sequential learning estimator relies upon belief updating rather than optimal strategic behavior or first-order conditions, it may be used under any industry structure and with any type of productivity dynamics. Consequently, estimated productivity beliefs can be used in second stage modeling of the determinants of firms' actions. Productivity dynamics, innovation effort, and other strategic behaviors can be compared across industries, under differing industry states, and with varying external economic conditions and public policies.
520 # $a Firms are assumed to learn from the same data available to the econometrician by updating productivity beliefs after observing each period's production experience. The difference between predicted and realized productivity in the next period is a forecast error and therefore independent of both exit and factor inputs. The resulting sequential updating process is equivalent to a Kalman filter. The estimation algorithm simulates a panel of establishment filters. Parameterized initial beliefs are identified from cross-sectional, within-cohort variation and a rational expectations assumption.
520 # $a The dynamics of establishment-level productivity include both persistent shocks and random noise. Therefore firms are uncertain about their own persistent productivities and must form productivity beliefs in order to make decisions about exit, innovation, investment, and inputs. Extending an idea from Jovanovic's (1982) dynamic industry model, this uncertainty is exploited in a sequential learning estimator of production functions and productivity dynamics for unbalanced panels.
520 # $a The sequential learning estimator is used to estimate Cobb-Douglas and constant elasticity of substitution (CES) production functions from annual panels of Chilean manufacturers in four selected industries. Standard errors are smaller than those from other estimators that address endogeneity. Estimated returns to scale, the relation between factor elasticities and cost shares, and the industry ranking of imputed mark-ups are all plausible. Estimates of productivity dynamics strongly confirm persistent, nonpermanent shocks to idiosyncratic productivity. Estimated productivity beliefs are positively related to investment in three out of four industries. The relation to exit hazard is negative, as expected, in all four industries. An implication of sequential learning estimator assumptions, that firms' selected actions should be independent of productivity information revealed in subsequent periods, is accepted in seven of eight validity tests.
590 $a School code: 0127.
650 # 0 $a Economics, General. $3 212429
650 # 0 $a Economics, Commerce-Business. $3 212577
690 $a 0501
690 $a 0505
710 0 # $a University of Michigan. $3 212464
773 0 # $g 67-07A. $t Dissertation Abstracts International
790 $a 0127
790 1 0 $a Kuhn, Kai-Uwe, $e advisor
791 $a Ph.D.
792 $a 2006
856 4 0 $u http://libsw.nuk.edu.tw:81/login?url=http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3224892 $z http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3224892