Spatial matters are of critical importance when considering socio-economic (and other) phenomena
(see, for example, Bockstael 1996; Weinhold 2002), as well as because of their implications for
policymaking (Lacombe 2004). To account for the presence of spatial structures that influence
(positively or negatively) observable economic entities, such as unemployment or trade, calls for a
rigorous and systematic assessment of their impact and extent. Spatial autocorrelation (SA) represents
the correlation, computed among the values of a single georeferenced variable, that is attributable to the
geographic proximity of the objects to which the values are attached. Introduction of the SA concept is,
of course, a departure from the classical assumption of independence of observations constituting a
single variable. SA also complements the concept of temporal autocorrelation, which has been
extensively studied and dealt with in time-series econometrics. SA measures are used to quantify the
nature and degree of the spatial correlation within a variable, or to test the assumption of independence
or randomness. From a statistical analysis viewpoint, spatial correlation patterns are problematic, since
they make standard statistics, such as correlation coefficients or ordinary least squares (OLS) estimates,
potentially inappropriate.
http://www.advancedphotoshop.co.uk/show_profile.php?username=chrisben
http://www.mathisfunforum.com/profile.php?id=151509
http://www.lasvegasextremes.com/users/chrisben
http://www.dandyid.org/id/90848
This paper aims to provide an assessment of how important spatial effects are in explaining
unemployment levels in Germany, and, particularly, to show that these (or, more precisely, a subset of
these) patterns are consistent over time. The definition of stable and recognizable spatial patterns
enables one to observe systematic differences in regional unemployment. Such findings can have
implications for policy evaluation and strategic planning. This paper presents analyses carried out by
means of a semi-parametric ‘spatial filtering’ technique, described in Griffith (2003), which is based on
the decomposition of geographic weights matrices. In our analysis, these matrices are defined for 439
German districts, according to both topological and distance-based criteria – such as shared boundaries
or centroid distance – and economic flows. In this regard, journey-to-work flows are employed as a
proxy for economic linkages.
Kosfeld and Dreger (2004) investigate spatial patterns of German regional labour markets, for the
period 1992–2000. However, their approach involves computing spatial filters for each year within the
framework of a spatial seemingly unrelated regression (SUR) model. Our approach differs from theirs
in that we focus on the search for a set of spatial filters that are significant and consistent over time, and
therefore can be employed for the entire time period considered (that is, 1996–2002). Also, we employ
data at a finer level of disaggregation (439 districts versus 180 regions), which enables a more detailed
analysis of the underlying spatial patterns.
http://www.musikcube.com/punbb/profile.php?id=23519
http://www.libervis.com/user/chrisben
http://www.doublefine.com/forums/member/68450/
http://community.katc.com/_cheap-backlinks/blog/5161128/23348.html
http://board.muse.mu/member.php?u=366414
http://www.cmstorm.com/forum/member.php?u=26166
(see, for example, Bockstael 1996; Weinhold 2002), as well as because of their implications for
policymaking (Lacombe 2004). To account for the presence of spatial structures that influence
(positively or negatively) observable economic entities, such as unemployment or trade, calls for a
rigorous and systematic assessment of their impact and extent. Spatial autocorrelation (SA) represents
the correlation, computed among the values of a single georeferenced variable, that is attributable to the
geographic proximity of the objects to which the values are attached. Introduction of the SA concept is,
of course, a departure from the classical assumption of independence of observations constituting a
single variable. SA also complements the concept of temporal autocorrelation, which has been
extensively studied and dealt with in time-series econometrics. SA measures are used to quantify the
nature and degree of the spatial correlation within a variable, or to test the assumption of independence
or randomness. From a statistical analysis viewpoint, spatial correlation patterns are problematic, since
they make standard statistics, such as correlation coefficients or ordinary least squares (OLS) estimates,
potentially inappropriate.
http://www.advancedphotoshop.co.uk/show_profile.php?username=chrisben
http://www.mathisfunforum.com/profile.php?id=151509
http://www.lasvegasextremes.com/users/chrisben
http://www.dandyid.org/id/90848
This paper aims to provide an assessment of how important spatial effects are in explaining
unemployment levels in Germany, and, particularly, to show that these (or, more precisely, a subset of
these) patterns are consistent over time. The definition of stable and recognizable spatial patterns
enables one to observe systematic differences in regional unemployment. Such findings can have
implications for policy evaluation and strategic planning. This paper presents analyses carried out by
means of a semi-parametric ‘spatial filtering’ technique, described in Griffith (2003), which is based on
the decomposition of geographic weights matrices. In our analysis, these matrices are defined for 439
German districts, according to both topological and distance-based criteria – such as shared boundaries
or centroid distance – and economic flows. In this regard, journey-to-work flows are employed as a
proxy for economic linkages.
Kosfeld and Dreger (2004) investigate spatial patterns of German regional labour markets, for the
period 1992–2000. However, their approach involves computing spatial filters for each year within the
framework of a spatial seemingly unrelated regression (SUR) model. Our approach differs from theirs
in that we focus on the search for a set of spatial filters that are significant and consistent over time, and
therefore can be employed for the entire time period considered (that is, 1996–2002). Also, we employ
data at a finer level of disaggregation (439 districts versus 180 regions), which enables a more detailed
analysis of the underlying spatial patterns.
http://www.musikcube.com/punbb/profile.php?id=23519
http://www.libervis.com/user/chrisben
http://www.doublefine.com/forums/member/68450/
http://community.katc.com/_cheap-backlinks/blog/5161128/23348.html
http://board.muse.mu/member.php?u=366414
http://www.cmstorm.com/forum/member.php?u=26166
No comments:
Post a Comment