…it was just someone using Excel on a laptop who was highlighting cells for a formula and released his index finger from the left-clicky button of his mouse too soon.
Writes Irish stand-up comedian Colm O’Regan for BBC Magazine, in his piece about “The mysterious powers of Microsoft Excel”. As you will likely have guessed, his article was inspired by spreadsheet blunders in Reinhart and Rogoff’s 2010 Growth in a Time of Debt paper. Read more
Statistical modelling. It goes everywhere.
The Monkey Cage points us to this 2004 paper on strategic voting in papal elections (there was still no sign of white smoke at pixel time on Wednesday): Read more
OK – Lisa’s the modelling maven and she’s promised a longer post on this paper tomorrow…
“‘The Formula That Killed Wall Street’? The Gaussian Copula and the Material Cultures of Modelling” is a recommended read (H/T Tracy Alloway). It’s not really about anything killing Wall Street — more a combination of economic sociology and an oral history stretching back to January 2007. The authors even got in touch with David X. Li. Read more
Guest post by Emanuel Derman
Foods with equal deliciousness should sell for equal prices per ounce. Read more
Hark — the standard deviation devils sing (again).
As Reuters columnist John Kemp pointed out yesterday, recent swings in the commodities complex have produced some impressive probabilities figures. The kind you can wheel out in dinner party conversation. For instance, front-month Brent crude futures sank almost $12 per barrel (or over 9 per cent) on Thursday, leading the market down from over $120 at the start of the day to under $110. Read more
How big the fall in world trade in 2008/2009?
THIS big: Read more
To understand Andrew Haldane’s latest — all you have to do is glance at these charts.
One is regulatory bank capital, the other is a market-based signal of bank solvency: Read more
Perhaps it’s not too astounding a finding…
But a Federal Reserve staff working paper by Dobrislav P. Dobrev and Pawel J. Szerszen has found that using historical high frequency data to forecast equity returns is far more effective than using general daily or monthly data. Read more