ForecastingStocks

CAPM explains an asset's return with a single factor: the market. Arbitrage Pricing Theory (APT), proposed by Stephen Ross in 1976, generalizes this to many sources of risk.

The core idea

APT argues that an asset's return is driven by its exposure to several systematic factors, each carrying its own risk premium:

expected return = risk-free rate
                + beta1 times (premium of factor 1)
                + beta2 times (premium of factor 2)
                + ... and so on

The factors are not fixed by the theory. Common choices are surprises in inflation, industrial production, interest-rate shifts, credit spreads and broad-market moves. Each beta measures how sensitive the asset is to that particular factor.

The "arbitrage" in the name

APT rests on a single, powerful assumption: if two portfolios have identical factor exposures, they must offer the same expected return. If they did not, traders would buy the cheap one and short the expensive one for a risk-free profit, and that arbitrage would force the prices back into line. So no such gap can persist.

A small illustration

Risk-free rate = 3 percent
Factor 1 (market) beta = 1.0, premium = 5 percent
Factor 2 (inflation surprise) beta = -0.4, premium = 2 percent
    expected return = 3 + 1.0(5) + (-0.4)(2) = 7.2 percent

The negative inflation beta lowers the expected return: this asset cushions inflation shocks, so investors accept slightly less for holding it.

CAPM versus APT

CAPM is APT with exactly one factor. APT is more flexible and more realistic — it is the intellectual ancestor of modern multi-factor models like Fama-French. But that flexibility is also its weakness: the theory does not tell you which factors to use or how many. You must choose them yourself, and a model with badly chosen factors can fit the past beautifully while predicting the future poorly.