Template-Type: ReDIF-Paper 1.0
Author-Name: Alan Collins
Author-Name: Jingwen Fan
Author-Name: Aruneema Mahabir
Title: Using a LASSO approach to analyse the determinants of measured and ‘natural’ suicide rates
Abstract: A new empirical approach to testing the hypothesis of a (positive) ‘natural’ rate of suicides is considered. 
The clustered variant of the Least Absolute Shrinkage and Selection Operation (LASSO) framework is employed to select the determinants of suicide in the USA 
and a Least Squares Dummy Variable (LSDV) approach is then employed to estimate the ‘natural rate of suicide’ at the state level. Our model suggests that over the 2005-2017 period, 
natural rates for US states ranged from 5.5 (DC) per 100,000 population per year to 18.8 (Montana). Our results support the natural rate hypothesis with a more robust specification. 
We also consider the use of the natural rate measures as a means of evaluating progress in increasingly influential zero suicide initiatives.
File-URL: http://www.ntu.ac.uk/__data/assets/pdf_file/0029/921818/NBSDiscussionPaper-2019-04.pdf   
File-Format: Application/pdf
File-Function: First version, 2019
Creation-Date: 2019-11
Classification-JEL: C23, I12
Keywords: Zero suicide, LASSO, Natural rate of suicides, USA
Handle: RePEc:nbs:wpaper:2019/04