The graph of both losses and normalised losses starts in 1990. The graph trend implies normalised losses would have been higher historically. Is the insurance data in earlier decades prior to 1990 reliable enough to say cover the last century? How far into the past can the methodology be pushed?
The graph of both losses and normalised losses starts in 1990. The graph trend implies normalised losses would have been higher historically. Is the insurance data in earlier decades prior to 1990 reliable enough to say cover the last century? How far into the past can the methodology be pushed?
Great Q. Here is why I start in 1990, based on guidance from Munich Re about their data:
"According to Munich Re, since the mid-1990s, ‘there has been a distinct improvement in the reporting of overall losses’ (Munich Re Group, 2006). Thus, the analysis presented here utilizes the Munich Re dataset from 1990, recognizing that there may be some degree of under-reporting in earlier years of the time series (Mohleji & Pielke, 2014)."
The graph of both losses and normalised losses starts in 1990. The graph trend implies normalised losses would have been higher historically. Is the insurance data in earlier decades prior to 1990 reliable enough to say cover the last century? How far into the past can the methodology be pushed?
Great Q. Here is why I start in 1990, based on guidance from Munich Re about their data:
"According to Munich Re, since the mid-1990s, ‘there has been a distinct improvement in the reporting of overall losses’ (Munich Re Group, 2006). Thus, the analysis presented here utilizes the Munich Re dataset from 1990, recognizing that there may be some degree of under-reporting in earlier years of the time series (Mohleji & Pielke, 2014)."
https://doi.org/10.1080/17477891.2018.1540343