There are three primary approaches to forecasting fixed-income returns:
1. Discounted Cash Flow (DCF) * Concept: Uses the present value of expected cash flows. It is the most precise method for individual securities and supports asset allocation. * Yield to Maturity (YTM): A reasonable first approximation of expected return. * Horizon Effects: * If Investment Horizon < Duration: Reinvestment risk dominates (rising rates = higher return). * If Investment Horizon > Duration: Capital/Price risk dominates (rising rates = lower return). * If Investment Horizon = Macaulay Duration: Capital gain/loss and reinvestment effects roughly offset.
2. The Risk Premium (Building Block) Approach * Formula: Expected Return = Short-term Default-free Rate + Term Premium + Credit Premium + Liquidity Premium. * Short-term Default-free Rate: Usually the closest central bank policy rate. * Term Premium: Compensation for duration risk. Positively related to the slope of the yield curve. * Credit Premium: Compensation for default risk and expected level of losses. Credit spreads are the primary indicator. * Liquidity Premium: Baseline estimate is the yield spread between the highest-quality issuer (sovereign) and the next highest (e.g., agency).
3. Equilibrium Models * Uses models like Black-Litterman to ensure consistency across asset classes.
Example: Forecasting Return Based on YTM Scenario: A bond portfolio has a YTM of 1% and a Macaulay duration of 4.89. Yields rise by 200 bps immediately. Outcome: Over a 5-year horizon (close to duration), the initial capital loss from the rate hike is offset by higher reinvestment income, resulting in a realized return close to the initial 1% YTM.
Emerging market (EM) debt carries standard fixed-income risks plus specific country risks.
1. Economic Risk (Ability to Pay) * Refers to the country’s ability to service debt. * Warning Signs/Ratios: * Deficit-to-GDP > 4%. * Debt-to-GDP > 70–80%. * Real growth rate < 4%. * Current account deficit > 4% of GDP. * Foreign exchange reserves < 100% of short-term debt.
2. Political/Legal Risk (Willingness to Pay) * Refers to property rights, laws, and political stability. * Key Issues: Sovereign immunity (difficulty enforcing claims), corruption, capital controls, and unstable coalition governments.
1. Grinold-Kroner Model (DCF Approach) This model decomposes expected return into cash flow, growth, and repricing components.
Example: Grinold-Kroner Calculation Inputs: Dividend yield = 2.25%; Share repurchases = 1% (so \(\% \Delta S = -1\%\)); Earnings growth = 6%; P/E expansion = 0.25%. Calculation: \(2.25\% + (6.0\% - (-1.0\%)) + 0.25\% = 9.5\%\).
2. Singer-Terhaar Model (Equilibrium/Risk Premium Approach) Adjusts CAPM for market imperfections (segmentation vs. integration). * Steps: 1. Calculate risk premium assuming full integration (Global CAPM): \[RP_i^G = \beta_{i,GM} \times RP_{GM} = \rho_{i,GM} \times \sigma_i \times \left( \frac{RP_{GM}}{\sigma_{GM}} \right)\] 2. Calculate risk premium assuming full segmentation: \[RP_i^S = \sigma_i \times \left( \frac{RP_i^S}{\sigma_i} \right)\] (Note: The term in parentheses is the Sharpe ratio of the segment). 3. Take the weighted average based on degree of integration (\(\phi\)): \[RP_i = \phi RP_i^G + (1 - \phi) RP_i^S\].
EM equities face similar risks to EM debt (fragile economies, political instability) but with different claim structures. * Key Risks: * Corporate Governance: Weak standards may allow insiders to misuse assets or dilute minority shareholders. * Transparency: Poor accounting standards may hide true financial health. * Property Rights: Risk of nationalization or expropriation. * Contagion: EM markets are often less integrated, meaning local factors dominate, but they are highly susceptible to “hot money” flows and crises.
1. Valuation Metrics * Capitalization (Cap) Rate: Defined as Net Operating Income (NOI) / Property Value. It is the standard valuation metric. * Formula for Expected Return: \[E(R_{re}) = \text{Cap Rate} + \text{NOI Growth Rate} - \% \Delta \text{Cap Rate}\] (Note: Similar to Grinold-Kroner but without share change).
2. Risk Premiums * Liquidity Premium: Highly significant for direct real estate due to inability to sell quickly. Estimated at 2–4%. * Cycle Sensitivity: Real estate is pro-cyclical. Cap rates typically rise (values fall) when interest rates rise, but counter-cyclical credit spreads can mitigate this.
3. Data Issues * Smoothing: Appraisal-based data smoothes volatility. True volatility is higher than reported. * De-smoothing Formula: \[R_t = (1 - \lambda)r_t + \lambda R_{t-1}\] * \(r_t\) is true return; \(R_t\) is reported return. * True variance is significantly higher: \(Var(r) = \frac{1+\lambda}{1-\lambda} Var(R)\).
1. Purchasing Power Parity (PPP) * Concept: Exchange rates adjust to equalize the price of goods (inflation) across countries. * Application: Poor for short-term; useful for long-term fair value. Exchange rate changes should equal inflation differentials.
2. Capital Flows & Risk Premiums * Concept: Capital seeks the highest risk-adjusted return. Exchange rates adjust to balance expected returns. * Formula (Uncovered Interest Parity + Premiums): \[E(\% \Delta S_{d/f}) = (r_d - r_f) + (Term_d - Term_f) + (Credit_d - Credit_f) + (Equity_d - Equity_f) + (Liquid_d - Liquid_f)\]. * Overshooting: When a currency offers higher returns, it may appreciate immediately (“overshoot”) and then be expected to depreciate.
3. Carry Trade * Borrowing in low-rate currencies to invest in high-rate currencies. Empirically profitable, contradicting Uncovered Interest Parity (UIP).
1. Sample Statistics * Simple but subject to sampling error. Requires \(Observations > 10 \times Assets\).
2. Factor Models * Uses common factors to estimate covariance. Reduces the number of parameters needed but introduces bias if the model is mis-specified.
3. Shrinkage Estimation * Weighted average of Sample Matrix and Target Matrix (e.g., Factor Model). Increases efficiency (reduces Mean Squared Error).
4. ARCH Models (Time-Varying Volatility) * Captures “volatility clustering” (periods of high/low volatility persist). * ARCH(1) Formula: \[\sigma_t^2 = \gamma + \alpha \sigma_{t-1}^2 + \beta \eta_t^2\] * Current variance depends on previous variance (\(\sigma_{t-1}^2\)) and the most recent shock (\(\eta_t^2\)).
Recommendations are based on macro trends: