A Random Walk Down the Factor Zoo

Since the 1990’s, factor-based quant strategies have continued to increase in popularity.

The standard style factors, such as size, value, momentum, growth, beta, etc, have all become widely known, and investment practitioners have employed them in their own variations of factor portfolio-based strategies.

That said, the dissemination of the so-called "factor zoo" has not come without its own set of issues.

For one, given the wide variety of factors popularized in both academia as well as the sell-side literature, it’s hard to know which factors to invest in.

Often investors rely on historical backtests of factor performance.

But given the large number of factors and their variations, could it be that what looks like a significant historical backtest may have simply been due to random chance?

After all, if we try enough random permutations of factors and their variations, we may find what looks to be a legitimate signal, but that this signal performs poorly out-of-sample into the future.

At AlphaLayer, we strive to build strategies with a high level of statistical rigor and while a backtest may look good it is only real-world performance that matters.

So how can we better understand factor significance? And as a result, increase the probability of delivering better factor-based strategies?

Read the full article on AlphaLayer's Substack