Exploring the spatial pattern of mental health expenditure
Please note: this is a legacy publication from CPEC (formely PSSRU at LSE).Journal of Mental Health Policy and Economics 8 4 205-217
Available online: 10 December 2005
Abstract
There are marked variations in mental health expenditure across localities, but it is possible that those variations are related in ways not usually explored in empirical research. The paper examines a hitherto ignored source of variation in spending: the underlying spatial pattern of either unobservable risk factors or genuine policy interdependence. Exploratory data analyses test for the existence of spatial interaction and spatial structures of data. Drawing hypotheses from these initial exploratory analyses, a spatial autoregressive model is then specified and confronted with data to explore the sources of spending variation between municipalities. The estimation results when specifying a spatial structure are compared with those from a classical (‘non-spatial’) model. There are some strong regional patterns in mental health expenditure that should be factored into allocation mechanisms and performance monitoring. Spatial econometrics methods can contribute to our understanding of why costs of ostensibly similar services vary from one locality to another.
There are marked variations in mental health expenditure across localities, but it is possible that those variations are related in ways not usually explored in empirical research. The paper examines a hitherto ignored source of variation in spending: the underlying spatial pattern of either unobservable risk factors or genuine policy interdependence. Exploratory data analyses test for the existence of spatial interaction and spatial structures of data. Drawing hypotheses from these initial exploratory analyses, a spatial autoregressive model is then specified and confronted with data to explore the sources of spending variation between municipalities. The estimation results when specifying a spatial structure are compared with those from a classical (‘non-spatial’) model. There are some strong regional patterns in mental health expenditure that should be factored into allocation mechanisms and performance monitoring. Spatial econometrics methods can contribute to our understanding of why costs of ostensibly similar services vary from one locality to another.