Multi-factor investing blends distinct return drivers—momentum, quality, value, low volatility, size and carry—into one portfolio, then risk-budgets and executes them with discipline. This article explains our high-level framework (“Alpha Suite”) for India—signal design, portfolio construction, and real-time risk—without revealing proprietary recipes or data weights.
Why multi-factor matters in India
Indian equities are diverse: sector skews, family-owned promoters, varying accounting quality, periodic liquidity shocks, and derivatives depth in the F&O segment. A single style can underperform for long stretches. A multi-factor approach smooths cycles by combining complementary edges and sizing risks consistently across regimes (bull, range-bound, stress).
What exactly is a multi-factor model?
A multi-factor model estimates each stock’s expected excess return as a weighted combination of standardized signals:
Composite Scorei = ∑k=1Kwk⋅zi,k\text{Composite Score}_i \;=\; \sum_{k=1}^{K} w_k \cdot z_{i,k}
- zi,kz_{i,k}: standardized (z-scored) signal for stock i on factor k
- wkw_k: factor weight (can be static, risk-parity, or IC-weighted)
Expected Active Return (portfolio level):
E[Ractive] = wfactors⊤πpremia\mathbb{E}[R_\text{active}] \;=\; \boldsymbol{w}_\text{factors}^\top \boldsymbol{\pi}_\text{premia}
where π\boldsymbol{\pi} are long-run factor premia (estimated from Indian data).
Inside the “Alpha Suite”: a high-level blueprint (no IP disclosed)
1) Signal Layer (what to buy/sell)
We design robust, India-aware signals that survive transaction costs and corporate actions.
- Momentum: 6–12-month total return, excluding the most recent month; cross-sectional rank.
- Quality: profitability (ROCE/ROE), accruals, earnings stability; governance red-flags filtered.
- Value: earnings/operating yield, free-cash-flow yield; sector-neutralized to avoid cyclicals bias.
- Low Volatility: inverse of realized volatility/beta (e.g., 252-day); scaled to control concentration.
- Size/Liquidity: penalize micro-cap or low-turnover names to keep the strategy executable.
- Carry (F&O): futures basis/roll yield for NIFTY stocks and liquid midcaps; disciplined filters.
Standardization: robust z-scores using median/MAD for outlier resistance:
zi,k=xi,k−median(x⋅,k)1.4826×MAD(x⋅,k)z_{i,k}=\frac{x_{i,k}-\text{median}(x_{\cdot,k})}{1.4826\times \text{MAD}(x_{\cdot,k})}
Orthogonalization: de-beta and sector-neutralize signals so each factor targets distinct risk.
2) Portfolio Construction Layer (how much to hold)
The objective is to transform signals into positions that are profitable and well-behaved.
- Rank-to-weight mapping: monotone functions (e.g., piecewise-linear) to convert scores to raw weights.
- Risk budgeting across factors: equal-risk-contribution or IC-weighted:
wk∝IC^kσkw_k \propto \frac{\widehat{\text{IC}}_k}{\sigma_k}
where IC^k\widehat{\text{IC}}_k is the rolling information coefficient and σk\sigma_k the factor’s volatility.
- Constraints:
- Stock/sector caps (e.g., ≤ 5% per stock, ≤ 25% per sector)
- Turnover budget (bps/day) to contain costs
- Liquidity screens (e.g., %ADV limits)
- Leverage, gross/net exposure brackets for long-only or long-short sleeves
- Volatility targeting:
positiont←positiont×σ∗σ^t\text{position}_t \leftarrow \text{position}_t \times \frac{\sigma^*}{\widehat{\sigma}_{t}}
where σ∗\sigma^* is target vol and σ^t\widehat{\sigma}_{t} is recent realized vol.
3) Execution & Real-Time Risk Layer (turning theory into P&L)
- Execution algos: VWAP/TWAP/POV with slippage guards and FOK/IOC logic for illiquid names.
- Cost models: impact ∝ spread × participation\sqrt{\text{participation}}; brokerage, STT, stamp duty, GST accounted.
- Intraday risk: hard limits for exposure, single-name drawdown, sector shocks; live kill-switches.
- Compliance: SEBI/Exchange API usage, pre-trade checks, and full audit trails.
How our factors reflect India’s market structure (examples)
- Promoter quality matters: Quality and governance filters often prevent value traps.
- Event cadence: Results season, index rebalances, and block deals can temporarily distort prices—momentum and carry harvest these dislocations with controls.
- Sector cycles: Banking/NBFCs vs IT vs manufacturing rotate leadership; sector-neutral signals avoid unintended macro bets.
- Derivatives depth: Basis and options-implied metrics enrich carry/risk signals for NIFTY/BANKNIFTY heavyweights.
Measuring skill: the scorecard that actually matters
- CAGR: (Ending NAVStarting NAV)1/n−1\left(\frac{\text{Ending NAV}}{\text{Starting NAV}}\right)^{1/n}-1
- Volatility (annualized): σann=stdev(returns)×252\sigma_\text{ann}=\text{stdev}(\text{returns})\times \sqrt{252}
- Sharpe (ex-RF): Sharpe=μσ\text{Sharpe} = \frac{\mu}{\sigma}
- Information Ratio (vs benchmark): IR=μactiveσactive\text{IR}=\frac{\mu_\text{active}}{\sigma_\text{active}}
- Max Drawdown: peak-to-trough loss percentage
- Calmar: CAGR/∣MaxDD∣\text{CAGR}/|\text{MaxDD}|
- Hit Rate: % of months with positive return
- Turnover: 0.5 × sum of absolute weight changes (monthly)
- Capacity: P&L degradation beyond a assets under management; monitored via cost-slippage curves
- IC / ICIR: cross-sectional corr(signal, next-period return) and its stability
Regime awareness: why factor blending beats factor timing
Instead of “timing” factors (often noisy in India), we blend them with regime-aware risk:
- Detect regime: realized vol (e.g., NIFTY 30-day), credit spreads, rupee trend, macro event flags.
- Adjust risk, not views: keep signal ranks; dial exposure via volatility targeting and tighter stock/sector caps in stress.
- Protect compounding: max portfolio drawdown triggers de-risking ladders (e.g., 5%/8%/12%).
Transparent, but not revealing the cookbook
We don’t publish: proprietary data cleaning rules, stock-level signals, exact factor weights, decay half-lives, turnover budgets, or execution venues. We do share our philosophy, controls, and what outcomes we hold ourselves accountable to: lower drawdowns than single-style funds, efficient turnover, and consistent IR vs a sensible benchmark (e.g., NIFTY 200 TRI or a custom sector-neutral index).
Illustrative portfolio outcomes (conceptual)
- Long-only core: Quality + Momentum + Low-Vol sleeve over NIFTY 500, sector-neutral, liquidity-aware.
- Enhanced index: Tracking-error target (e.g., 3–5%) versus NIFTY 200 TRI, focusing on tax efficiency and low turnover.
- Market-neutral satellite: Dollar-neutral long/short (top- vs bottom-decile composite scores) for absolute return.
(Exact results vary by data, costs, and constraints; the above is a framework, not performance claims.)
Compliance & governance note (India-specific)
- SEBI: Clear segregation across RA/RIA/PMS/Algo Provider roles; investor communication avoids promissory language.
- Auditability: Backtests with timestamped universes, survivorship-bias controls, and reproducible run IDs.
- Risk & review: Model change logs, quarterly factor drift reviews, exception reporting for slippage and limit breaches.
FAQs
Q1. Is multi-factor the same as Smart Beta?
Smart Beta is usually rule-based and long-only with transparent indices. Multi-factor in our Alpha Suite also governs risk, turnover, and execution, and can include long-short and carry components.
Q2. Do factors stop working in India once popular?
Edges compress, but process quality—clean data, neutralization, cost control, risk discipline—often matters more than discovery novelty.
Q3. Can I time factors instead of blending them?
Factor timing is difficult and noisy. We prefer regime-aware risk, letting diversified signals compound.
Q4. What benchmark should I compare against?
For long-only, a style-appropriate TRI (e.g., NIFTY 200 TRI) or a sector-neutral custom index. For market-neutral, look at absolute metrics (Sharpe, Calmar, drawdowns).
Key takeaways
- Blend complementary factors (momentum, quality, value, low-vol, carry) rather than betting on one style.
- Engineer risks, not just returns: sector neutrality, volatility targeting, turnover and costs are first-class citizens.
- Stay regime-aware: change exposure, not your philosophy.
- Measure what matters: IR, drawdowns, capacity, and after-cost performance.
Disclaimer: This is educational content for Indian investors. It is not investment advice or a solicitation. Markets involve risk, including possible loss of capital.