4.1 Global Capital Markets History, Market Regimes & Long-Run Asset Class Returns

4.1 Global Capital Markets History, Market Regimes & Long-Run Asset Class Returns

A comprehensive understanding of long-run capital markets history is fundamental for investment consultants. Historical performance data provides empirical boundaries for risk, return, and covariance assumptions, while forward-looking building block models prevent the uncritical extrapolation of past regimes into future portfolio allocations.

---

1. Historical Risk & Return Hierarchy (1926–Present)

Nearly a century of U.S. and global capital market data (Ibbotson-Sinquefield and Dimson-Marsh-Staunton datasets) establishes a consistent structural risk-return spectrum:

$\text{Small-Cap Equities} > \text{Large-Cap Equities} > \text{High-Yield Credit} > \text{Investment-Grade Corporates} > \text{Treasury Bonds} > \text{T-Bills (Cash)}$

The Compounding Drag: Geometric vs. Arithmetic Means

When setting long-term portfolio expectations, consultants must account for the mathematical variance drain between arithmetic and geometric compound returns:

$R_g \approx R_a - \frac{1}{2}\sigma^2$

• Arithmetic Mean ($R_a$): The unweighted average of single-period returns; appropriate for forecasting a single upcoming period's expected return.
• Geometric Mean ($R_g$): The compound annual growth rate (CAGR) over multi-year horizons; appropriate for multi-period wealth projection and long-term financial planning.
• As volatility ($\sigma$) rises, the disparity between $R_a$ and $R_g$ widens significantly. For example, an asset with a 10% arithmetic return and a 20% standard deviation will compound at approximately $10% - 0.5(0.20^2) = 8.0%$.

---

2. The Equity Risk Premium (ERP): Historical vs. Forward-Looking

The Equity Risk Premium (ERP) represents the expected excess return demanded by investors for holding equities over a risk-free benchmark (such as 10-year U.S. Treasuries or 30-day T-bills):

$\text{ERP} = E(R_m) - R_f$

1. Historical Realized ERP (Ibbotson-Sinquefield Approach)

Historically realized excess returns of the S&P 500 over intermediate/long Treasuries have averaged between 4.5% and 6.0% (geometric) and 6.5% and 7.5% (arithmetic) in the U.S.

Limitations of Historical ERP:

• Survivorship & Success Bias: The U.S. was the world's most successful 20th-century economy. Global cross-country studies (Dimson-Marsh-Staunton) indicate non-U.S. developed market historical ERPs averaged a lower 3.5%–4.5%.
• Multiple Expansion Distortion: Significant historical equity returns stemmed from non-repeatable P/E multiple expansion and secular declines in real interest rates.

2. Forward-Looking Supply-Side ERP Models

Supply-side models anchor equity returns to macroeconomic output and corporate cash generation.

Gordon Growth / Constant Growth DDM Approach:

$E(R_e) = \frac{D_1}{P_0} + g$ $\text{ERP} = \left( \frac{D_1}{P_0} + g \right) - R_f$

Where $\frac{D_1}{P_0}$ is the dividend yield and $g$ is long-run sustainable nominal earnings growth (anchored to nominal GDP growth).

Ibbotson-Chen Macroeconomic Supply-Side Model: Decomposes equity returns into fundamental macroeconomic components:

$E(R_e) = (1 + i)(1 + r_{\text{EPSC}})(1 + r_{\text{PE}}) + \text{Yield} + \text{Reinvestment} - 1$

Where:

• $i$ = Expected long-term inflation.
• $r_{\text{EPSC}}$ = Long-run growth in real earnings per share (constrained by real GDP growth and labor productivity).
• $r_{\text{PE}}$ = Expected annualized percentage change in the P/E multiple (assumed to be zero or negative over long horizons if starting from elevated valuations).
• $\text{Yield}$ = Expected dividend and net share repurchase yield.

---

3. Market Regimes & Structural Breaks

Asset class behaviors and correlation structures are non-stationary and depend heavily on macroeconomic regimes.

+-----------------------------------------------------------------------------------------+
|                                HISTORICAL MARKET REGIMES                                |
+-----------------------------+-----------------------------+-----------------------------+
| 1970s: Stagflation          | 1982-1999: Disinflation     | 2008: Global Fin. Crisis    |
| - High Inflation / Low Grow | - Falling Yields & Volcker  | - Credit Crunch & Deflation |
| - Stocks/Bonds fell in real | - Massive Equity/Bond Rally | - Flight to Quality (USTs)  |
| - Commodities & Real Assets | - P/E Multiple Expansion    | - Equities & Credit Plunge  |
+-----------------------------+-----------------------------+-----------------------------+
| 2020-2022: Pandemic & Supply Shock -> 40-Yr Inflation Spike -> Positive Stock-Bond Corr |
+-----------------------------------------------------------------------------------------+

---

4. Capital Market Expectations (CME): The Building Blocks Approach

To construct institutional portfolio allocations, consultants apply the Grinold-Kroner Model for equities and the Yield-plus-Spread model for fixed income.

The Grinold-Kroner Model for Expected Equity Return

The Grinold-Kroner model expresses the expected return of an equity index ($E(R_e)$) as the sum of income return, earnings growth, and repricing:

$E(R_e) = \underbrace{\frac{D}{P} - \Delta S}_{\text{Income Return}} + \underbrace{i + g}_{\text{Nominal Earnings Growth}} + \underbrace{\Delta(P/E)}_{\text{Valuation Change}}$

Where:

• $\frac{D}{P}$ = Expected dividend yield.
• $-\Delta S$ = Net share repurchase yield (percentage reduction in shares outstanding; $-\Delta S > 0$ denotes share buybacks, $-\Delta S < 0$ denotes net equity dilution).
• $i$ = Expected long-term inflation rate.
• $g$ = Real GDP/earnings growth rate.
• $\Delta(P/E)$ = Expected annualized percentage change in the valuation multiple over the forecast horizon.

Numerical Application of Grinold-Kroner

An institutional analyst evaluates the S&P 500 with the following 10-year assumptions:

• Expected Dividend Yield: $1.8%$
• Net Share Repurchases: $1.2%$ per year (share count decreases $1.2%$)
• Long-Term Expected Inflation: $2.2%$
• Expected Real GDP Growth Rate: $2.0%$
• Current P/E is 24.0, with an expected 10-year reversion to historical median of 19.5 (Annualized multiple change $\approx -2.0%$)

$E(R_e) = [1.8\% + 1.2\%] + [2.2\% + 2.0\%] + [-2.0\%] = 3.0\% + 4.2\% - 2.0\% = 5.2\%$

Fixed Income Building Blocks Model

The expected return for a fixed-income asset class ($E(R_b)$) is formulated as:

$E(R_b) = \text{Yield to Maturity (YTM)} + \text{Roll-Down Return} - E(\text{Default Credit Losses}) \pm [-\text{Duration} \times \Delta y]$

• Roll-down return: Capital gain realized as a bond "rolls down" an upward-sloping yield curve as it approaches maturity.
• Credit loss adjustment: $\text{Annual Default Probability} \times (1 - \text{Recovery Rate})$.
• Duration adjustment: Anticipated change in benchmark yields over the investment horizon.