at the World Economic Forum, Davos, and in the UK Parliament. In 2013, he was awarded tenure at the Wharton School, and in 2021, he was named MBA Professor of the Year by Poets&Quants. Edmans writes regularly for the Wall Street Journal, the Financial Times, and Harvard Business Review. His first book, Grow the Pie, was a Financial Times Book of the Year. He is a Fellow of the Academy of Social Sciences.
Edmans applied the "Page 99 Test" to his new book, May Contain Lies: How Stories, Statistics, and Studies Exploit Our Biases―And What We Can Do about It, and reported the following:
Page 99 of May Contain Lies: How Stories, Statistics, and Studies Exploit Our Biases – And What We Can Do About It explains how to test a hypothesis. Here, the hypothesis is to test whether frequent traders earn higher returns than investors who buy shares and then leave their portfolio untouched. It says:Visit Alex Edmans's website.You then test the hypothesis. For this, you’d ideally like the trading records of every single trigger-happy investor. That’s impossible, so the second step is to gather a sample. What’s critical is that the sample is representative, not selected – it captures a broad mix of traders rather than pre-screening them on some criterion, such as whether they volunteered to share their record or had an account for five years (both of which would skew the sample to more successful investors). That’s similar to how you’d sample a cake by cutting it vertically so that your slice contains the icing, sponge, filling and base, rather than splitting it horizontally and skimming off only the icing. The extensive compilation of excitable shareholders is known as the test sample – you’re testing whether it performs better.The Page 99 Test works well because this page highlights one of the key messages of the book: that combating misinformation goes beyond just checking the facts. Even if the facts are 100% accurate, they may be misleading if they aren’t representative: the exception that doesn’t prove the rule. In the chapter leading up to page 99, I describe a YouTuber who brags about how much money he made day-trading. But even if he’s telling the truth and not exaggerating his profits, this doesn’t mean that day trading sets you on the road to riches. You have a selected sample: only the day traders that got lucky parade their success. There could be hundreds of other day traders who lost their shirt, but you’ll never hear about them.
Step three is equally critical – to find a control sample that doesn’t have the input. The high returns to fidgety investors might be nothing to do with the input (frequent trading) but just because the market went up. So you need to find out how much was earned by buy-and-hold investors who didn’t trade at all. Step four is to calculate the average output across the two samples, which gives you the 11.4% and 17.9%.
You’re tempted to conclude that frequent trading lowers returns, but there’s one final step. Even if frequent trading has no effect on profits, it could still underperform due to luck.
Page 99 is part of Part II of the book, which takes the reader through the Ladder of Misinference, the four missteps we make when we interpret information. The first misstep is mistaking a statement for fact, when it may not be accurate: for example, it may be quoted out of context. The second misstep is confusing facts for data, when they may not be representative. This is the misstep addressed by page 99. The third misstep is mixing up data for evidence, when it may not be conclusive: such as a correlation without causation. The fourth misstep is misinterpreting evidence for proof, when it may not be universal: it may not apply in different contexts. The Ladder helps the reader to navigate the minefield of misinformation out there, to think smarter, sharper, and more critically, to make better sense of the world and take better decisions.
--Marshal Zeringue
Posts
