Welcome to Relative Value Methods!
Welcome to one of the most practical and quantitative areas of the CAIA Level II curriculum! In traditional directional investing, we ask: "Is this asset going up or down?" In relative value (RV) methods, we shift perspective entirely: "Is this asset mispriced relative to a closely related asset or benchmark?"
By taking offsetting long and short positions, relative value strategies seek to extract abnormal returns from price discrepancies while neutralizing broad market exposure. In this chapter, we will master the mechanics of statistical pairs trading across equities, commodities, fixed income, and currencies, followed by a rigorous breakdown of the unique risks facing relative-value market-neutral funds.
1. Core Principles of Relative Value
In traditional investing, managers pursue absolute value. In contrast, relative value investing exploits pricing inefficiencies between economically or statistically linked securities.
The Neighbor Analogy: Imagine two identical houses on the same street. House A is priced at \$500,000 and House B at \$440,000. You do not need a forecast for nationwide home prices to conclude that House B is relatively cheap compared to House A. An RV trader buys House B and shorts House A, profiting when the pricing spread closes.
Key Terminology:
Mean Reversion: The foundational assumption that prices, yields, or spreads will revert toward their long-term historical mean or equilibrium relationship.
Convergence: The process whereby the valuation spread narrows back to fair value, generating profit for the trade.
Divergence: The risk that a pricing spread widens further after entry, imposing interim mark-to-market losses.
2. Statistical Pairs Trading in Equities
Statistical pairs trading identifies two equities that historically track each other due to shared business models, cost drivers, or fundamental factors (such as two major credit card networks or beverage producers).
The Two Primary Approaches:
1. Distance Method: Standardizes historical price series (normalized to start at 1.0) and measures the Euclidean distance between them. When the distance exceeds a specified threshold (e.g., two historical standard deviations), the diverged pair is traded.
2. Cointegration and Regression Method: Unlike simple correlation (which measures co-movement in returns), cointegration tests whether a linear combination of two non-stationary price series produces a stationary spread with a constant mean and variance: \(S_t = P_{A,t} - \beta P_{B,t}\). Here, \(\beta\) represents the cointegrating hedge ratio estimated via ordinary least squares (OLS).
Trading Mechanics and the Z-Score:
Traders track the standard deviation of the spread using a rolling Z-score:
\(Z = \frac{S_t - \mu_S}{\sigma_S}\)
Where \(S_t\) is the current spread, \(\mu_S\) is the historical mean spread, and \(\sigma_S\) is the historical standard deviation.
- If \(Z \ge +2.0\): The spread is unusually wide. The trader shorts Stock A and purchases \(\beta\) shares of Stock B.
- If \(Z \le -2.0\): The spread is unusually narrow or negative. The trader buys Stock A and shorts \(\beta\) shares of Stock B.
- Exit Rule: The position is closed when the spread reverts to its mean (\(Z = 0\)) or hits a stop-loss boundary.
3. Spread-Based Relative Value in Commodities
Commodity relative value strategies trade physical spreads rather than directional commodity price movements. Because physical commodities entail storage, transportation, and processing, spreads are tied to structural economic relationships.
Key Commodity Spread Strategies:
Calendar (Intra-Market) Spreads: Trading contracts of different maturity months within the same commodity (e.g., buying near-month crude oil futures and selling distant-month futures). When markets are in backwardation (spot price above futures, reflecting near-term scarcity), the near contract trades at a premium. In contango (futures above spot due to financing and storage costs), near contracts trade at a discount.
Processing Spreads (Inter-Commodity): Capitalizing on input-versus-output margins. Classic examples include the crack spread (buying crude oil futures against selling refined gasoline and heating oil futures) and the crush spread (buying raw soybeans against selling soybean meal and soybean oil).
Locational and Quality Spreads: Exploiting geographic or grade price differentials, such as the Brent versus West Texas Intermediate (WTI) crude oil spread, reflecting regional supply bottlenecks, refining suitability, and pipeline infrastructure.
4. Relative Value in Fixed Income and Currency Markets
Fixed income and foreign exchange markets offer deep, liquid, and mathematically structured relative value opportunities.
Fixed Income Relative Value:
Yield Curve Plays (Butterfly Trades): Constructing a duration-neutral portfolio combining the short "wings" and a middle "body" maturity (e.g., long 2-year and 30-year Treasuries versus short 10-year Treasuries). This allows traders to isolate and profit from changes in yield curve curvature without bearing directional interest rate risk.
Swap Spreads and Basis Trades: Trading the spread between interest rate swaps and sovereign bond yields, or between sovereign bonds and sovereign futures, capturing small market imbalances through leveraged repo financing.
Currency Market Relative Value:
Covered and Uncovered Interest Parity Spreads: Exploiting persistent pricing dislocations between spot rates, forward contracts, and short-term interbank interest rates across borders.
FX Carry Trades: Borrowing in currencies with low policy interest rates (funding currencies) to invest in currencies with high yields (target currencies), hedged or partially paired across correlated currency baskets to achieve relative-value stability.
Cross-Currency Basis Swaps: Monetizing deviations from covered interest parity (CIP) when global demand for dollar funding creates a premium in the cross-currency basis.
5. Risks of Relative-Value Market-Neutral Strategies
While marketed as "market-neutral," relative value strategies carry distinctive structural and operational risks. Because typical spread discrepancies are small, RV managers use significant leverage to achieve target institutional returns, amplifying exposures.
1. Divergence and Gap Risk: Mean reversion is an empirical tendency, not a contractual certainty. Spreads can widen substantially further before converging, creating large unrealized mark-to-market losses.
2. Financing and Margin Call Risk: Leveraged positions require collateral. If spreads widen sharply, clearing brokers and repo counterparties issue margin calls. If the fund runs out of unencumbered cash, it faces forced liquidation at the worst possible prices (the classic failure mode of Long-Term Capital Management).
3. Model and Cointegration Breakdown: Historical statistical relationships can break down permanently due to structural shifts—such as regulatory changes, technological disruption, corporate restructurings, or mergers. A spread that appears "cheap" may simply be pricing a permanent fundamental deterioration.
4. Liquidity and Short-Squeeze Risk: Short legs in equities or off-the-run bonds may become hard to borrow or suffer from recalls. A short squeeze can force premature closure of the short position, leaving the portfolio directionally exposed.
5. Crowded Trades and Synchronization Risk: Because many quantitative relative-value funds deploy similar statistical filters, an initial sell-off by one large player can trigger cascading stop-losses across the entire strategy, evaporating market liquidity.
Summary Review:
- Objective: Extract pure alpha from mispricings between related assets while remaining beta-neutral.
- Applications: Equity pairs (distance & cointegration), commodity spreads (calendar, processing, location), and fixed income/currency parity trades.
- Critical Vulnerabilities: High leverage, divergence risk, margin-call-induced forced liquidation, and structural breakdown of historical statistical relationships.