Arbitrage Pricing Theory

The Arbitrage Pricing Theory (APT) is a multi-factor framework for estimating cost of equity, developed by Stephen Ross as an alternative to the single-factor CAPM. Rather than assuming cost of equity depends only on market beta, APT says it depends on multiple systematic risk factors: interest rate risk, inflation risk, industry risk, etc. The theory is elegant, but in practice, identifying and measuring the factors is subjective.

The concept

CAPM assumes one risk factor (market risk) explains returns.

APT assumes multiple risk factors explain returns. The cost of equity equals the risk-free rate plus the sum of factor-based risk premiums.

Cost of equity = Risk-free rate + (Factor 1 loading × Factor 1 premium) + (Factor 2 loading × Factor 2 premium) + … + (Factor n loading × Factor n premium)

Unlike CAPM, APT does not specify which factors matter; it merely says multiple factors should be considered.

Potential factors

Different factors might explain equity returns:

Macroeconomic factors:

• Unexpected inflation
• Unexpected changes in interest rates
• Unexpected changes in economic growth
• Unexpected changes in oil prices

Market factors:

• Market risk premium (equivalent to CAPM beta)
• Size premium
• Value premium

Firm-specific factors:

• Earnings surprises
• Analyst forecast revisions
• Momentum

Building an APT model

1. Identify relevant factors. Which risks affect the company? If it is a bank, interest-rate risk is critical. If it is a consumer staple, economic-growth risk might be secondary.

2. Estimate factor loadings. How sensitive is the stock to each factor? Run regressions of historical returns against factor returns.

3. Estimate factor premiums. What is the expected excess return for each factor?

4. Sum. Add all factor contributions to get total cost of equity.

Advantages of APT

Flexibility. You are not locked into CAPM’s single beta. You can include factors that are relevant to the company.

Theoretical soundness. APT is built on arbitrage principles: if two portfolios have the same risk exposures, they should have the same expected return. The theory is elegant.

Can be empirically richer. Multi-factor models (like Fama-French) naturally fit within APT’s framework.

Challenges in practice

How many factors? One factor (CAPM) is too few. But 10 factors might be too many. The choice is arbitrary.

Which factors? APT doesn’t say. You must choose. Different analysts choose differently, yielding different cost-of-equity estimates.

Measuring factor premiums. While CAPM’s market risk premium can be estimated from historical equity returns, premiums for other factors are harder to pin down. What is the expected premium for “unexpected inflation”?

Data limitations. For some factors, long-term historical data is unavailable. For a new risk (e.g., climate change), there is no history.

Overfitting. If you fit factor loadings using historical data, you might find spurious factors that don’t persist.

APT vs. Fama-French

Fama-French is essentially a practical implementation of APT. It identifies specific factors (market, size, value, profitability, investment) that empirically explain returns.

APT is the theoretical framework; Fama-French is an applied model.

In practice, practitioners use Fama-French (or variations) rather than building a custom APT model, because Fama-French’s factors are well-researched and stable.

When APT is useful

Specialized industries. If your company is in a niche industry (e.g., gold mining, telecom infrastructure), standard factors might miss key risks. A custom APT model with commodity-price or regulatory factors might be more appropriate.

Multi-national companies. A company with major exposure to multiple countries faces currency risk, country-specific risk, etc. A multi-factor APT model can model these.

Hedging applications. Portfolio managers use APT-like models to understand and manage their exposures to various risk factors.

Practical adoption

In industry. Most practitioners use CAPM, Fama-French three-factor, or build-up methods. Pure APT is less common.

In academia. APT is widely taught and used in research. It is the theoretical foundation for understanding multi-factor models.

In practice for valuations. When analysts use multiple factors, they are implicitly using an APT-like framework, even if they don’t call it APT.

Building a practical APT for a valuation

Rather than deriving factors from economic first principles, a practical approach is:

1. Start with Fama-French three-factor (market, size, value).
2. Add factors specific to your company (e.g., oil prices if energy-exposed, interest rates if a bank, regulatory risk if regulated).
3. Estimate or assume premiums for these additional factors.
4. Calculate cost of equity.

This is more practical than pure APT while capturing industry-specific risks.

See also