Risk
Risk shapes model design
We study risk across the model portfolio, including market exposure, concentration, liquidity and implementation assumptions.
The model portfolio is the unit of risk
Models can appear independent while favouring the same shares or factors, and correlations can change. Research therefore examines the combined model portfolio, including covariance, concentration, netting, trading costs, financing, margin and liquidity assumptions.
Shared risk
Correlations, overlapping positions and factor exposure.
Model weighting
Simulated weights under exposure, liquidity and capacity assumptions.
Combined model
Testing netted model positions and possible hedges for residual market exposure.
Model weights are assessed for incremental contribution and sensitivity to implementation assumptions.
Dimensions of risk
Model exposure, concentration, liquidity and implementation assumptions are examined together.
Market exposure
Beta, gross and net exposure, and leverage: how much the portfolio as a whole moves with the market.
Concentration
Position, strategy and common-factor risk, tested under different exposure limits and position constraints.
Liquidity
Tradability, financing, borrow and capacity: whether positions can be traded, funded and, for short positions, borrowed at the intended size.
Implementation assumptions
Trading costs, margin and borrow assumptions, including where simulated execution may differ from live conditions.
Research reviews record changes in model behaviour and the reasons for revisions.
Risks that remain
Market-neutral does not mean risk-free. These are general strategy risks; a future product would have additional risks defined by its legal structure and terms.
Losses in either market direction
A market-neutral portfolio loses money when its long positions underperform its short positions, whether markets are rising or falling, or when costs exceed the difference between them.
Short-selling risks
A shorted share can rise sharply, for example in a short squeeze, and losses on a short position are theoretically unlimited. Borrowed shares can be recalled, and the cost of borrowing can rise.
Changing correlations
Relationships between strategies, positions and factors can change, particularly in stressed markets, reducing the diversification the portfolio is designed to achieve.
Model risk
Systematic models can be wrong, can stop working, or can rely on relationships that no longer hold. Data can be incomplete or inaccurate, and historical evidence may not repeat.
Leverage, financing and counterparties
Long/short portfolios rely on margin and financing. Leverage can magnify losses, and the failure of a broker or other counterparty could cause loss.
Operational disruption
Failures of systems, data or processes can interrupt trading or cause errors.
This list is not exhaustive. See Important information.
Models under review Evidence on the record
Research and live validation examine portfolio exposure, concentration, liquidity, actual execution and costs. Live real-money trading lets us compare model assumptions with observed conditions. Changes in behaviour prompt investigation, and decisions to revise, retest or retire a model are documented. This website does not offer portfolio management to clients.
Questions about risk?
Questions about the research
Read the research FAQs or contact us about the models and their limitations.