Inside SMART Equity: active ETFs seeking alpha beyond traditional factors

Traditional investment factors have helped investors navigate equity markets for decades. The familiar measures of value, quality, momentum and growth provide a durable framework for understanding why some stocks behave differently from others.

Yet a residual component of excess return, often associated with information advantages and modelling skill, remains less well explained and less systematically captured by traditional factor definitions.

Our Systematic Equity team has spent decades developing ways to capture more of that information, using established quantitative disciplines alongside newer AI techniques and alternative data to uncover less obvious sources of alpha.

Now, our SMART Equity ETFs are making these insights readily investable. Here’s how they work.

The team has engineered four SMART Equity signals, each designed to capture sources of information traditional factors may miss:

1. Investor positioning: what professional investors are doing. 
2. Style interaction: which stock characteristics matter most in different market segments.
3. Earnings sentiment: what company management communication can reveal.
4. Advanced momentum: which market trends, dynamics, and patterns matter most. 

Together, these signals integrate behavioural, fundamental and market-based information into a systematic framework for identifying potential sources of alpha.

1. Investor positioning

Share prices are influenced not only by company fundamentals, but also by how investors are positioned and how that positioning changes over time.

Short selling offers one way to observe investors’ conviction on a stock’s prospects. An investor expecting a company’s share price to fall may borrow shares and sell them in the market, leaving an observable footprint in the securities-lending market.

Our Investor Positioning signal identifies patterns in potential short-selling activity across companies. Rather than relying on what professional investors say, it learns from what they actually do.

This signal can build a picture of how professional investors are positioned and what their behaviour may imply about individual companies. Those patterns can be compared across company peer groups and translated into a clear positioning signal.

2. Style Interaction

Financial markets are driven by complex investment decisions, and the characteristics associated with future stock performance can interact differently across market segments too.

SMART Equity’s Style Interaction signal uses a Random Forest machine-learning model to reflect these complex relationships. Compare it to assessing someone’s fitness: age, diet or exercise habits may each tell you something in isolation, but considering several measures together gives a more complete and helpful picture.

Random Forest applies a similar principle to stocks. Its many decision rules examine companies from different angles, including valuation, profitability, and earnings momentum.

Some rules can prove more useful for certain types of companies like growth companies, while others may be more informative for deep-value opportunities. So rather than relying on a single rule or indicator, the model aggregates insights across many independent rules, effectively using a “wisdom-of-the-crowd” approach to identify relationships that prove more persistent over time.

Analysing style interaction helps identify which combinations of familiar investment characteristics matter most in different market environments. That can give SMART Equity a more adaptive edge in stock selection as market conditions change.

3. Earnings sentiment

When companies report results, the numbers they release are only part of the story. Company leaders explain what is going well, where challenges lie, and how they see the future. Analysts ask questions that often draw out additional insight.

Much of this information is forward-looking and may contain clues about future performance. But analysing it consistently across thousands of companies is difficult to do at scale using traditional research methods.

Our Earnings Sentiment signal uses natural language processing to analyse thousands of company earnings calls each earnings season. Advanced language models convert words and sentences into numerical representations that capture meaning and context, helping the model to distinguish constructive communication from cautious or defensive language.

Those assessments can then be compared across companies in the same sector or region.

The aim is to detect shifts in management tone and confidence that conventional financial data may not capture as readily, turning corporate conversation into another systematic input for stock selection.

4. Advanced Momentum

Momentum describes a familiar market tendency: stocks that have performed well often continue to perform well, and those that have lagged tend to keep lagging.

The principle is simple; applying it is not. There is no single definition of what makes a “winner”. A stock may look strong over 12 months but weak over a shorter period, while different market conditions can favour different momentum patterns.

Our research has shown that momentum does not act in isolation, but rather interacts with other stock trading patterns such as volatility, trading volume or liquidity. These interactions can amplify and dampen a trend, making it difficult to capture with simple rules.

To address this complexity, SMART Equity uses an Advanced Momentum signal based on deep-learning neural networks to analyse market trends and price dynamics more flexibly.

One way to picture the process is as a layered research team. At the first level, a large group of junior analysts each focuses on information such as past price movements, volatility or trading activity, and forms an initial view on a stock. These views are then passed to a second level, comparable to senior analysts, who combine and interpret these inputs, placing greater emphasis on what is most relevant and reliable and allowing more complex relationships to emerge.

The resulting signal estimates the probability that a particular stock will outperform its peers. In other words, Advanced Momentum doesn’t try to simply identify a trend, but to judge which trends appear most meaningful in the market conditions surrounding them.

The four SMART signals do not operate independently. They are aggregated into a single composite AI signal, bringing their different perspectives together into one assessment of each stock’s relative attractiveness. That composite signal then becomes the key input to portfolio construction.

It works through a portfolio optimiser which increases exposure to stocks with more favourable signals. But it does not simply buy the highest-scoring companies: it also keeps tracking error, style exposures, sector and regional positioning and portfolio turnover within defined limits. The aim is to give the stock-selection ideas enough weight to matter while preserving the broad characteristics expected from a core equity holding.

SMART Equity remains stock-led rather than style-led. It doesn’t begin by deciding whether, say, value or growth should lead at any particular moment. Instead, the signals identify the stocks the process finds relatively attractive, with whatever traditional style exposure emerges largely a consequence of those choices rather than a set investment thesis.

The result is a broadly diversified portfolio that uses active stock selection to seek incremental alpha, while keeping risk tightly controlled so the strategy remains recognisably core.¹ The AI signals can adapt as market relationships change, while portfolio managers remain accountable for the final portfolio.

Investing in equities is about understanding both companies and markets: how businesses evolve, how investors behave and how trends develop over time. SMART Equity applies systematic research and advanced technology to that task, using four complementary signals to identify insights that traditional investment factors may not fully capture.

By combining those signals in a disciplined portfolio process, SMART Equity seeks to turn that richer information set into better-informed stock selection and incremental alpha, while keeping the portfolio recognisably core.

That is the substance behind SMART Equity: differentiated signals and systematic investment expertise, delivered through flexible, transparent active ETFs.

Get active with AllianzGI.

1) Diversification does not guarantee a profit or protect against losses.