Factor investing is an academically-rooted investment approach based on measurable characteristics — factors — that systematically explain investment returns. Studies since the 1970s have shown that a significant portion of stock returns can be explained by a few fundamental factors. In this article, we examine the 10-year backtest performance, academic foundations, and practical application of momentum factor, value, low volatility, quality, dividend, and machine learning factor portfolios on Borsa Istanbul (BIST).
What is Factor Investing?
Factor investing offers a middle path between tracking traditional market indices (passive investing) and individual stock selection (active investing). In this approach, stocks with academically-proven systematic characteristics are selected through rules-based methodology. The strategy is built on factor exposure rather than individual stock analysis.
The theoretical foundation of factor investing was laid by Eugene Fama and Kenneth French's three-factor model published in 1993. This model demonstrated that stock returns could be explained by market beta, size, and value (book-to-market) factors. In 2015, Fama and French extended this into a five-factor model by adding profitability and investment factors.
Core Factors and Academic Evidence
1. Momentum Factor
momentum factor refers to the tendency of stocks that have generated high returns recently to continue this performance in the short term. Narasimhan Jegadeesh and Sheridan Titman's groundbreaking 1993 study showed that stocks with the highest returns over the past 3-12 months continued to outperform the market in the following 3-12 months.
We use the Sortino ratio for the momentum factor on BIST. The Sortino ratio is calculated by dividing returns only by downside standard deviation — providing a more refined metric that measures loss risk without penalizing upside volatility. The 20 stocks with the highest 21-day Sortino score are added to the portfolio at the start of each month.
2. Value Factor
The value factor is based on the observation that stocks trading below their intrinsic value tend to deliver higher returns over the long term. Sanjoy Basu's 1977 study documented that stocks with low price-to-earnings (P/E) ratios systematically outperformed the market. Fama and French's research showed that stocks with high book-to-market ratios deliver a long-term premium.
We use the quarterly earnings/price (E/P) ratio for the value factor on BIST. Instead of annual (cumulative) E/P, the ratio of the latest quarterly net income to current market cap is calculated. This approach is more responsive to rapidly changing market conditions because it relies on more current earnings information.
3. Low Volatility Anomaly
The low volatility anomaly is a finding that contradicts traditional risk-return relationships: less volatile stocks deliver better risk-adjusted returns over the long term than highly volatile ones. Robert Haugen and Nardin Baker documented this anomaly in 1991; Malcolm Baker, Brendan Bradley, and Jeffrey Wurgler's 2011 study examined the underlying causes.
On BIST, this factor selects the 20 stocks with the lowest 21-day annualized standard deviation. The strategy is suitable for defensive investors and portfolios seeking low drawdown.
4. Low Volatility + Trend (Defensive Momentum)
A weakness of the pure low volatility strategy is that it can select stocks that continue to fall even in bear markets. To solve this problem, we apply a double filter: first, stocks with positive 63-day (about 3-month) momentum are selected, then the 20 with the lowest volatility are taken from this universe. The result: rising but calm stocks.
5. Dividend + Quality
The danger of pure dividend strategies is the 'value trap' — troubled companies with falling prices mathematically showing high dividend yields. To avoid this trap, we apply a quality filter: ROE > 0 (profitable), net margin > 0 (operationally profitable), debt/equity < 3 (reasonably leveraged). The 20 stocks with the highest dividend yield among companies passing these three criteria are selected.
6. ML Ensemble (Machine Learning)
Instead of relying on a single factor, it is possible to combine multiple factors into a single prediction model. In our ML Ensemble portfolio, four different machine learning models — XGBoost, LightGBM, Random Forest, and Ridge Regression — are trained using the walk-forward method. 26 features are used: momentum (21d, 63d, 252d), volatility, Sortino, E/P, P/B, ROE, ROA, debt/equity, revenue growth, and the sector-relative values of all of these.
The ensemble approach reduces the overfitting risk of any single model by averaging predictions from four models. Walk-forward training eliminates lookahead bias because each month uses only historical data.
10-Year Backtest Results
The list below shows the latest 1-year net return, Sharpe ratio, and performance versus BIST 100 for each of the six factor portfolios. All portfolios are automatically rebalanced on the first trading day of each month and consist of 20 stocks.
- Momentum: +120.9% annual return · Sharpe 2.17 · 3.0× above BIST-100
- Dividend + Quality: +55.0% annual return · Sharpe 1.94 · 1.4× above BIST-100
- Value (E/P): +54.5% annual return · Sharpe 1.78 · 2.1× above BIST-100
- ML Ensemble: +52.0% annual return · Sharpe 1.77 · 2.8× above BIST-100
- Low Vol + Trend: +42.8% annual return · Sharpe 2.15 · 1.1× above BIST-100
- Low Volatility: +38.5% annual return · Sharpe 1.93 · 0.9× above BIST-100
Momentum delivers the highest return, while Low Volatility + Trend shows the highest risk-adjusted performance (Sharpe ratio). This highlights the importance of selecting factors based on the investor's risk tolerance.
Factor Weaknesses and Diversification
No single factor delivers the best results in every market condition. Momentum can suffer losses during sudden trend reversals; value experiences extended underperformance periods (1990s dot-com bubble); low volatility lags during rapidly rising markets. For this reason, the academic literature recommends multi-factor portfolios.
The recommended approach for BIST is a 2-3 factor combination: for example, Momentum + Low Volatility + Value combines return potential, risk control, and valuation discipline. You can use the Portfolio Simulator on the Borsafolio platform to build your own multi-factor portfolio.
Backtest Validity: Lookahead Bias and Walk-Forward
Two critical conditions must be met for a backtest result to be reliable:
- No lookahead bias: A backtest must use only data that was available at the time. For example, Q1 financials end at the end of March but are not announced until May; the backtest cannot act as if it knew this data when taking positions at the end of March. In our system, financial data is applied with a 60-day lag.
- Walk-forward training: ML models cannot be trained with future data. Re-training is performed each month using only data available up to that date.
- Transaction costs: Ignoring commissions, slippage, and market impact disconnects the backtest from the real world.
- Survivorship bias: Analyzing only stocks alive today ignores the impact of companies that went bankrupt or were delisted in the past.
Building Your Own Factor Portfolio
The Borsafolio platform helps you put academic factor research into practice for BIST. You can follow these steps:
- Use the BIST Stock Screener to filter stocks by momentum, value, volatility, and quality factors.
- Review each strategy's historical performance and current stock list on the Factor Portfolios page.
- Observe the sector momentum map and valuation information using BIST Radar.
- Analyze your portfolio's correlation structure and risk contributions with the Diversification Score tool.
- Rebalance your portfolio at the start of each month — discipline is the most critical component of factor investing.
Frequently Asked Questions
Is factor investing active or passive?
Factor investing sits between the two. It differs from passive investing because it constructs a portfolio that diverges from the market index. It differs from active investing because decisions are based on systematic rules, not individual stock analysis. This approach is also known as 'smart beta' or 'rules-based active management' in the literature.
Which factor is the best?
There is no single 'best factor.' Momentum offers high returns but requires high turnover. Value is reliable in the long term but demands patience. Low volatility provides stable returns but lags in bull markets. The best approach is a multi-factor combination tailored to your investment goals and risk tolerance.
How often should factor portfolios be rebalanced?
Academic literature and practical application point to monthly rebalancing as the optimal frequency. More frequent (weekly) rebalancing increases transaction costs; less frequent (quarterly) rebalancing causes the factor signal to lose strength. Monthly rebalancing offers the best balance between signal freshness and cost control.
Does factor investing work on BIST?
Yes. 10-year backtest results show that momentum, value, and low volatility factors systematically outperform the BIST 100 index. The emerging market characteristics of the Turkish market — high volatility and the emerging market premium — make factor investing even more effective.
Conclusion
Factor investing is a systematic investment approach supported by more than 50 years of academic research and proven in practice. In an emerging market like Borsa Istanbul, factor portfolios offer significant alpha potential compared to passive index investing. What matters is staying disciplined, avoiding methodological pitfalls like lookahead bias, and thinking long-term. Every factor has periods of short-term underperformance — success comes from trusting the factor and staying loyal to the strategy.
Borsafolio's six factor portfolios offer ready-made solutions backed by academic evidence and optimized for BIST. With automatic monthly rebalancing, transparent methodology, and 10 years of backtest data, you can find all the tools a data-driven investor needs in one place.
Related articles: What Is Factor Investing?, What Is the Momentum Factor?, What Is the Value Factor?, What Is Smart Beta?, What Is Risk Parity?.
