Systematic investing doesn't eliminate risk. It manages it differently | by making risk decisions explicit and pre-specified, rather than ad hoc and emotional. This module covers the main risk types that systematic equity investors face and the controls that actually help.
The key distinction: risk management in systematic investing is almost entirely done at the strategy design stage, not at the trade execution stage. By the time you're looking at a portfolio and deciding whether to trade, almost all risk decisions should already be encoded in the rules.
The six risks that actually matter
Stop-losses in systematic strategies
Stop-losses are a common risk management tool | automatically selling a stock when it falls X% from purchase price. For discretionary investors, they're useful. For systematic factor strategies, they're usually counterproductive.
Here's why: in a momentum strategy, you're already scheduled to sell a stock when it drops below the top-30 rank on the next rebalance date. A stop-loss effectively adds a discretionary override | "sell this stock before the scheduled rebalance if it falls 15%." This introduces:
- Whipsaw risk | you sell on a 15% dip, the stock recovers and re-enters the top-30 before the rebalance, you've crystallised a loss unnecessarily
- Higher turnover and costs | more trades, more costs, more tax events
- Inconsistency with the backtest | your backtest probably didn't include stop-losses. Adding them in live trading means you're running a different strategy than what was tested
The systematic strategy's rebalancing mechanism IS the stop-loss. If a stock falls far enough to drop out of the top 30 by rank, it gets sold on the next rebalance. This is a cleaner, less expensive, and more consistent approach than arbitrary percentage-based stop-losses.
Drawdown limits | the one legitimate override
There is one scenario where overriding a systematic strategy is genuinely rational: when the portfolio drawdown has exceeded your pre-specified personal limit.
The logic: you invested money in a systematic strategy based on the assumption that you could hold through a 35% drawdown (based on the historical maximum). If the strategy draws down 45% | beyond what was historically experienced and beyond what you modelled | this is information that wasn't in the backtest. The strategy may be behaving differently than expected.
The correct protocol: Before deploying any systematic strategy, decide: "What drawdown, sustained for how long, would cause me to exit?" Specify this in writing, in advance. Example: "If the strategy draws down more than 40% from peak and remains below -30% for more than 6 months, I will reduce allocation by 50% and reassess." Having a pre-specified rule means this decision is also systematic | not made in panic at the bottom.
What NOT to do: Override the strategy every time it underperforms the market for a few months. This is not risk management | it's performance chasing disguised as prudence. The most common way investors destroy the value of a systematic strategy is by adding discretionary overrides during normal drawdown periods that are well within the strategy's historical parameters.
Allocation sizing as risk management
The most powerful risk management lever is often the simplest: how much of your total investable assets do you put into any single systematic strategy?
- A ₹50 lakh portfolio with 100% in a momentum strategy is very different from 50% in momentum + 50% in a fixed income fund
- The right allocation depends on your investment horizon, income stability, and actual risk tolerance | not your stated risk tolerance
- A common framework: allocate no more than you could watch fall 50% without needing to sell for living expenses
RupeeCase shows live risk metrics for any active strategy: current drawdown from peak, days in drawdown, sector concentration, largest single-stock weight, and daily Value-at-Risk estimate. The system alerts you if any constraint (sector cap, stock cap) is approaching its limit before the next scheduled rebalance. Risk management is a dashboard feature, not an afterthought. Available at invest.rupeecase.com.
Glossary
- Concentration risk
- Risk from excessive capital in a single stock, sector, or factor. Controlled through position caps and sector limits.
- Liquidity risk
- The risk that a position cannot be exited without significantly moving the price. Controlled by limiting position size to a fraction of the stock’s average daily traded volume.
- Model risk
- Risk that the strategy’s signal or backtest is incorrect, due to overfitting, data errors, or look-ahead bias. Controlled through out-of-sample testing and economic rationale.
- Drawdown limit
- A pre-specified threshold for portfolio loss from peak, beyond which the investor will reduce allocation or reassess. The only legitimate basis for a systematic strategy override.
- Average Daily Volume (ADV)
- The average daily traded volume of a stock. Position sizes should typically not exceed 5 to 10% of ADV to maintain adequate liquidity.
Sources & further reading
- → SEBI, Portfolio Manager Regulations (risk management framework)
- → NSE India, Market Data (liquidity monitoring)
- → Taleb, N.N. (2007). The Black Swan. Random House. (Fat-tail risk and model limitations)
- → Litterman, R. (2003). Modern Investment Management. Wiley. (Risk management frameworks)
🏅 Path 4 Assessment, 30 Questions
Test your knowledge across all 5 modules. Pass 21/30 (70%) to unlock your certificate.
This assessment covers everything in Path 4: Portfolio Construction, Modern Portfolio Theory, position sizing, rebalancing, performance measurement, and risk management for systematic Indian equity portfolios.
Questions are drawn from all five modules. You need 21 correct answers out of 30 to pass. You can retry as many times as you like.
↓ Enter your name and email below to unlock your downloadable certificate.
Quick check, Module 4.5
Parametric VaR Calculator
Estimates the worst expected daily loss at a given confidence level, assuming normal returns. Useful as a back-of-envelope; real Indian markets are fatter-tailed than the normal model.