Beta

Three expressions of beta are algebraically identical and each is testable. The covariance form divides the covariance of asset and market returns by the variance of market returns. The correlation form multiplies the correlation between asset and market returns by the ratio of the asset's standard deviation to the market's standard deviation. The regression form treats beta as the slope coefficient when excess asset returns are regressed on excess market returns, the line often called the security characteristic line.

The correlation form is the most instructive. An asset can be extremely volatile yet carry a modest beta if its correlation with the market is weak, and a moderately volatile asset can carry a high beta if it moves in lockstep with the index. Beta therefore captures only the portion of volatility that co-moves with the market; the residual variance from the regression is diversifiable and earns no risk premium in equilibrium.

Practical estimation requires choices, and each choice moves the answer. The analyst selects a market proxy, a return frequency and an estimation window: five years of monthly data and two years of weekly data are both common conventions and can produce materially different slopes for the same security. Thin trading, mergers, changes in leverage and shifts in business mix all make a historical slope a noisy forecast of future sensitivity.

Because estimated betas tend to drift toward one over time, practitioners often report adjusted betas that blend the raw regression slope with a value of 1.0. For private companies or divisions, analysts take a peer-group beta, strip out the peers' financial leverage to obtain an asset or unlevered beta, then relever it at the target's own debt-to-equity ratio and tax rate. Portfolio beta is simply the market-value-weighted average of the component betas, which makes it easy to shift exposure by reallocating weights.

Beta is the pricing input in the CAPM, so it feeds directly into the cost of equity used in dividend discount models, free cash flow to equity valuations and the equity component of WACC. A one-tenth error in beta can shift a terminal value materially, which is why capital budgeting disputes so often reduce to arguments about the correct comparable set.

Beta is equally a risk-management tool. A manager who wants to reduce market exposure without liquidating positions can short index futures in a quantity determined by the ratio of target to current portfolio beta applied to portfolio value; the same arithmetic converts cash into synthetic equity exposure. In performance evaluation, beta defines the benchmark line against which Jensen's alpha and the Treynor ratio are measured, and factor models such as Fama-French retain a market beta alongside size, value and other factor loadings.

Expect to compute beta from a covariance and a market variance, from a correlation and two standard deviations, or by backing it out of a CAPM expected return when the risk-free rate and market risk premium are given. Expect also to compute a portfolio beta from weights and then price that portfolio on the security market line.

Interpretation matters as much as arithmetic. Be ready to explain why an asset plotting above the security market line is underpriced, why only systematic risk is compensated once a portfolio is well diversified, and why total risk measured by standard deviation and systematic risk measured by beta rank assets differently. Also be prepared to critique the model: beta is estimated with error, the true market portfolio is unobservable, and a single-factor specification omits documented sources of return that multifactor models attempt to capture.