Factors: Scoring Stocks

Score every stock

A factor is a way of scoring stocks: the value factor looks at how cheap they are, the momentum factor at recent strength, the quality factor at how steady their earnings are. Rank by score, split into five groups, and compare their average future returns.

Only works across a basket

Individual stock results are mostly down to chance, and factor signals are weak. The information coefficient (IC) is usually only 0.02–0.1, but hold enough stocks and a small edge shows up reliably.

Factors can stop working

The value factor lagged in US stocks for most of the 2010s, and the momentum factor can crash when the market reverses sharply. Holding several factors and keeping turnover low is steadier than betting on just one.

Key takeaways

• A factor is a trait used to score stocks, such as value (cheap), momentum (recently strong) or quality (stable profits).
• Split stocks into five groups (quintiles) by factor score and compare the average returns of the top and bottom groups to see whether the factor works.
• The information coefficient (IC) measures the correlation between factor rankings and future return rankings. Single stocks are very noisy, so factors only work across a "basket".
• Factors can stop working for years at a time, so holding several factors is steadier than betting on one.

Try it yourself

You can do these exercises in this lesson's interactive simulation:

1. Score by momentum (default 500 stocks): do the five groups' returns rise step by step? What are the long-short spread and the IC?
2. Switch to scoring by “Random”, then click “New batch of stocks” a few times: does the long-short spread jump between positive and negative?
3. Go back to momentum, cut the number of stocks to 50 and click “New batch of stocks” a few times: does the IC swing more or less than with 500?
4. Lower signal strength to 0.3 and raise the number of stocks to 2,000: is the IC still positive? Do the five groups still rise step by step?

Quiz and explanations

1. When a factor works, the average returns of the five groups usually look like…

Answer: Rising step by step from the lowest-scoring group to the highest.The higher the group's score, the higher its average return: a monotonic relationship.

2. Why does factor investing buy a basket of stocks rather than just the single top scorer?

Answer: Single stocks are too noisy; factors only work on average.Random events in individual stocks far outweigh the factor signal; only after diversifying does the signal show up.

3. A factor has an IC of about 0.05. Which statement is right?

Answer: The signal is weak, but spread across many stocks and periods it can still be valuable.An IC of 0.05 isn't bad in practice: weak each time, but accumulated across hundreds of stocks and many periods it can form a steady edge.

Terms in this lesson

• factor (Factor): A trait used to score stocks and explain differences in returns, such as value, momentum, quality or size.
• information coefficient (IC) (Information Coefficient (IC)): The correlation between factor-score rankings and next-period return rankings, measuring a factor's predictive power; above 0.05 is already useful.
• momentum factor (Momentum): The tendency of stocks that have done well recently to keep outperforming in the near term.
• quality factor (Quality): The tendency of companies with steady profits, low debt and strong cash flow to do better on average over the long run.
• quintile (Quintile): Splitting stocks into five equal groups from lowest to highest score to see how a factor relates to returns.
• value factor (Value): The tendency of stocks that are cheap relative to profits or book value to do better on average over the long run.