What Is a Quantitative Fund?

Quantitative investing is an investment process in which securities are chosen based on defined rules.

Conventional active management involves a team doing security-specific research: modeling company financials, comparing industry peers, and assessing competitive advantages in picking the best stocks. While these approaches surely contain some rigid elements, humans make the final call, which embeds a qualitative flexibility in the decision-making process.

Quantitative investing has some creativity and flexibility, but it’s just in choosing, arranging, and replacing data and inputs, not in choosing the actual stocks. A team hunts for the best inputs so a statistical model can spit out the most attractive ideas with these criteria. The aim is to minimize human judgment (and potential bias) when choosing stocks.

How does a quantitative fund choose its data and criteria? Managers sift through a forest of data, seeking the slivers that best indicate outperformance. These slivers then become criteria in the model that screens securities.

They identify the best criteria through back-testing data. In back-testing, they isolate a criterion they believe could outperform, such as stocks with low price/earnings ratios, over set a time period, like five years. They’ll then simulate how a portfolio with this criterion fares against the broader market over the five-year period.

If the back-test indicates the criterion helped returns, the team will test the variable in other market environments and data sets to ensure the relationship isn’t an isolated incident.

What types of data and criteria can quantitative managers use? Most quantitative strategies will incorporate a company's financial statement data. This can include metrics like net income and ratios like net margin or price/earnings.

They can also gauge sentiment about the economy or a business. Potential data points include gross domestic product growth or the dispersion of analysts’ earnings estimates for next quarter.

Finally, these strategies might use unstructured data, or variables outside of conventional financial analysis. Satellite imaging or credit card data might contain consumer trend insights and help with estimating company sales, for instance.

What are the ways these funds differentiate themselves? Quantitative strategies might choose different metrics within similar data. One team might use price/book, another might use price/earnings, and a third could use both.

Two strategies could use identical criteria but differ in how they use it. Do they compare a stock’s price/earnings to sector peers, industry peers, both, or neither? Are chosen criteria emphasized in sectors or industries that show stronger relationships? The criteria can also differ by time period: One team might gauge a stock’s price/earnings over three months while another looks at six months.

Plus, a criterion’s value can deteriorate over time, so managers regularly replace decaying ones for promising ones.

How does strategic beta fit into this? Strategic beta replicates a benchmark. But the benchmark is weighted by a factor other than market cap; usually it's tilted toward stocks whose financial metrics are associated with an investing style such as value (low price/book), momentum (price movements), or growth (increasing revenue).

Strategic-beta funds seek to earn returns you would expect from a specific investing style. Each style has common and well-known factors, so tilting a portfolio toward one is easy and inexpensive. However, this also makes strategic-beta funds simpler in construction.

Pure active quantitative strategies, contrarily, can choose investments outside their benchmark. This oftentimes results in more complex (and expensive) models that choose stocks through sophisticated algorithms or unique information like unstructured data.

How do quantitative strategies manage risk? Quant funds often use an optimizer: a separate model component that keeps sector and position sizes in check.

Imagine if a model spit out 10 technology stocks as the highest-ranking choices. Buying them all would spike the fund’s technology weighting, exposing it to sector risk if the sector dipped. An optimizer creates constraints--such as maximum sector weightings--to limit risks like these.

Quantitative strategies are often less concentrated. Because the model has defined rules, it’s easy to evaluate a wide security universe and buy the ones that look attractive, unlike a human research team that might have resources to deeply cover only a fraction of the market.

What are unique challenges to quantitative funds? Most quantitative funds struggle in rapidly shifting markets, as past relationships might be less meaningful in different, future environments. For instance, many fared poorly in the fourth quarter of 2018 and the first quarter of 2019 when stocks plummeted and subsequently rose.

There are also some strategic risks and challenges quant funds can face, including:

• Survivorship Bias: past data not accounting for businesses that no longer exist. Some trends might look different if the data included bankrupt companies.
• Data Mining: identifying a criterion because you actively sought it out rather than it being a legitimate excess-return source. You can often find a relationship in data just by looking hard enough, even if it's not truly there.
• Overfitting: when a statistical model anchors its assumptions too heavily on past data, to the point that these trends might not hold in the future.

Quantitative strategies also trade frequently as the model’s rankings constantly shift. This means the portfolio is constantly changing to reflect the best ideas, which results in above-average trading costs. These costs make it more difficult for funds to add value after fees.

How do I invest in quantitative strategies? Quant funds are available in virtually any investing style (large-cap growth, small-cap value, and so on).

Below is a list of six quantitative strategies that we like.