Read on Twitter
The projection used in the report cited here and in other outlets is based on an extremely poor interpretation and use of data. Thread. 1/9 https://www.theguardian.com/environment/2020/sep/09/climate-crisis-could-displace-12bn-people-by-2050-report-warns
2/9 The reported cited by the @guardian, the @Independent and other news sources around the world can be found here, published by @GlobPeaceIndex. Figure 1.1 is what caught my eye. 2/9
http://visionofhumanity.org/app/uploads/2020/09/ETR_2020_web-1.pdf
http://visionofhumanity.org/app/uploads/2020/09/ETR_2020_web-1.pdf
3/9 The figure provides a projection for global displacement due to climate change, reaching above 1.2 billion people by 2050. The 1st alarm bell was the number it cites for currently displaced people. The @UNHCR and @IDMC_Geneva put the stock of IDPs at 46 million, not 379 mil.
4/9 You can download the data the report uses from the @IDMC_Geneva here, under "Total annual new displacements since 2003 (Conflict and violence) and 2008 (Disasters)." Drop it into an excel workbook. https://www.internal-displacement.org/database/displacement-data
5/9 I'm convinced that what @GlobPeaceIndex did was sum the *yearly* number of *new* displaced persons due to conflict or disaster from 2008-2019, and then hit the FORECAST.ETS function in excel for years 2020-2050.
6/9 Do a bit of simple formatting with the resulting numbers, and you can reproduce what looks like the same figure given in the report.
7/9 People move in and out of displacement, and using the annual number of newly displaced persons to project the cumulative number of refugees from climate change is a very poor use of data. If you apply the same function to just the new number of cases each year, you get this.
8/9 That's roughly 27 million displaced persons *in* 2050. That's a lot, but it's not 1.2 billion people big. I'll also say that my 27 million number shouldn't be taken as a serious projection, it's just a proper application of the function in excel.
9/9 People will be displaced by natural disaster and conflict in the future. How many? I'm not sure, but it's unlikely to be 1.2 billion people big. Effective policy development needs to be based on solid analysis, not flimsy projections like the one published in today's papers.
TL;DR - Excel functions aren't methodologies.
