Q: What is the history of the fund?
The Northern Small Cap Value Fund was launched in April 1994. Our focus in the small-cap space is twofold: using a quantitative approach, we seek reasonably priced small-company stocks to invest in and, at the same time, seek to identify lower-quality companies to avoid.
In our experience, successful companies are successful in unique ways, while those that are unsuccessful tend to be unsuccessful in similar ways. In 2002, a year after I became the fund’s primary portfolio manager, we devised two stock selection models that rank companies using specific winning and losing characteristics.
Our objective is to generate the greatest return for the risk taken – and to do so consistently, and in a disciplined manner, with a tight view on our style exposure. Having a team of experienced individuals in place over the long term has been useful in pinpointing intended risks and steering clear of unintended ones.
Q: Are there any challenges specific to the small-cap investment process?
In the small-cap space, being careful about the obstacles to achieving results is as important as the opportunities and potential results themselves. A chief obstacle is transaction cost; the lower liquidity of small-cap stocks makes trading in them more expensive. As a result, we have designed our process and models to reflect this.
Understanding the data availability and quality in small cap is also necessary because it differs from large cap. The risks differ as well. While there is opportunity for big returns in small cap, there are equally great risks in many cases – including the risk of the company going bust.
Q: What core principles drive your investment philosophy?
We believe that investors should be compensated for the risks they take. That principle guides this fund and our larger “Engineered Equity®” platform at Northern Trust Asset Management. A focus on quality is a related principle we follow, and a big component of the fund’s track record.
“Design matters” is the other key principle. We believe the characteristics of low-quality companies tend to be much more predictive of under-performance than characteristics of high-quality companies in predicting outperformance. Our stock-selection methodology is designed to help us evaluate such quality characteristics, and our portfolio construction process sorts out companies with low-quality characteristics.
Having the details, truly understanding the risks we want to take or to avoid, and being compensated appropriately for risk underlies everything we do.
Q: What is your investment strategy?
Our approach is bottom-up with a rules-based, disciplined process. We believe this is a risk-efficient way to identify stocks that meet our objectives.
We have two models focused on stock selection. Both models are binary by design, meaning each signal is rated either good or bad. Within each sector, these models identify the companies with the highest number of good signals – in other words, the most positive companies versus the least positive.
The first model looks at traditional signals: valuation, momentum, and earnings. It also identifies the top, middle, and bottom within each sector and assigns buy, hold, and sell ratings. We use it to look for reasonably priced but profitable small companies.
The second model looks at quality using three major signals: quality of earnings, sources of financing, and margins. These are important because we believe underperforming companies share the characteristics of declining margins, increased need for outside capital (either debt or equity) and lower quality of earnings. Once this model identifies companies, we avoid purchasing them and reduce their exposure if already in the portfolio.
Putting the two models together, we buy from the “buys” and sell from the “sells.” The highest-rated companies found by the first model become buys unless they also appear on the distressed list generated by the second model. When the models send a mixed message on a company, ranking it high on the first and low on the second, the company goes into our hold bucket. A more consistent signal that ranks low on both models puts a company into our sell category.
Q: When do you update your model?
The model is run quarterly and is timed relative to the availability of quarterly data. We are very careful about the timeliness and precision of the data.
Our model is designed to reflect durability, as well as our concern about transaction cost, and we want it to work with relatively low turnover and seek a high level of consistency within the scores themselves. Confirming that our model gets in the portfolio without a whole lot of trading means we look for long-term signals.
Q: What is your research process?
We use a quantitative approach designed specifically for small caps. For example, fundamental data on small-cap companies tends to take a long time to get into the market and prices tend to move faster. Our first model, which looks at traditional signals of valuation, momentum, and earnings, designed to avoid value traps by finding companies where the fundamental data may not have kept up with the price.
Another way we perform our analysis, in a space where research and data is limited, is to focus on avoiding lower-quality companies that tend to have higher debt levels, lower margins and less cash flow. When analyzing these, we take a deep dive, examining nine individual components within earnings quality, sources of financing and margin categories.
All of our factors are designed to reduce portfolio turnover (and trading cost) and manage risk relative to our performance benchmark, as well as to identify potential stock mispricing in a systematic way.
Q: What characteristics are common among the stocks in your portfolio?
It is composed of more positive earning stocks and cash flow than the benchmark in general –we have a modestly higher capitalization and an aversion to low-quality stocks. We generally invest in companies with higher return on equity (ROE), return on assets (ROA), higher margins operating on that margin, and lower debt; our portfolio does not have big biases on price-to-earnings (P/E), price-to-cash flow, or price-to-book.
This goes back to our desire to generate long-term performance from stock selection and quality biases rather than from taking big bets on sectors or styles.
Q: How is the portfolio constructed? What is your benchmark?
The portfolio construction process is straightforward since the buy and sell disciplines are part of our models.
Every stock in the universe is scored in our model and described as a buy, hold, or sell based on the relative rank within sectors. To be more specific, the buys are the top 30% within each sector, the sells the bottom 30%, and the holds the broad 40% in the middle. However, if a company shows up as a low-quality score, buys become holds and the holds become sells.
With this list of buys, we target plus or minus 2% at the sector level relative to our benchmark, the Russell 2000® Value Index, and plus or minus 2% relative to the individual stock weight within the benchmark. The idea is to be relatively tightly constrained so our value-add comes from stock selection rather than sector selection.
We do have some biases depending on the model output. For example, the portfolio has been underweight in real estate investment trusts (REITs) for some time and overweight industrials and information technology because they have tended to have more high-quality stocks.
Sector weight differences from the benchmark tend to be a residual of the bottom-up process rather than the identification of individual sectors.
Our construction process helps us stay disciplined and focused on the fund’s long-term objectives. The models are run on a quarterly basis and we do trade around the refreshes, but mainly our trading is patient and slow-moving throughout the quarter.
The portfolio is highly diversified with 400-600 holdings. Such diversification is a natural risk-management tool, and helps manage transaction costs.
Q: How do you define and manage risk?
We view the entire investment process through the lens of compensated risk, devising models, constructing the portfolio, and managing the fund to reflect the systemic and idiosyncratic risk inherent to the market.
Our focus on quality and long-term holdings is also an important part of what we do, and having lower volatility has several risk-management benefits. We seek to have a favorable downside risk profile as well as overall excess return.
Finally, we aim for high information ratio (IR), which measures excess return for the amount of active risk we take relative to the benchmark. We gauge success not on excess return alone but on how we get that return by taking compensated risks and avoiding uncompensated risks. Our goal is to maintain an IR that ranks competitively within the industry. It is part of our overall culture here at Northern Trust Asset Management.
