Whether you are presented with some research claim or about to do some research yourself with financial data, there are certain biases you need to watch out for if you want to get accurate results. Not only that, when presented with financial research, you need to look into the little details, to see if the claim you are presented with is valid.
The most “popular” bias is called the survivorship bias. Here is a quick example: Suppose we want to check the claim that small caps companies have higher returns than large cap companies. We examine historical data, to further check the claim with our statistical models, and we might get to the conclusion, that small caps outperform large caps by 2% (This is a made up figure). There are a few problems though. When we are examining historical data of past prices, we would only examine companies that did not go bankrupt in the process, hence the name survivorship bias.
It is more likely that small caps companies would go bankrupt than large cap companies. It is also likely that the small caps companies who did not go bankrupt, had high returns. So in essence we took the best companies out there which result in higher returns for the small caps companies.
Another bias is the time intervals bias. This simply means that we take, consciensly or not, historical data in which different time intervals yield different results. This is actually done all the time consciensly, when hedge funds companies market their funds. They conviniently take a time interval in which they have the highest returns. If you are examining an hypothesis on past data, it’s always good to see what happens if you change the time intervals.
If you liked this post, buy me a beer
1 Response to “Financial Biases - What You Need to Watch Out For”