Beta

Beta measures how sensitive a stock’s return is to movements in the overall market. A beta above 1 indicates a stronger same-direction response. A beta between 0 and 1 indicates a weaker one. A negative beta indicates an inverse relationship. In valuation, beta serves as a measure of the risk of an investment. Statistically, beta is the slope coefficient of a regression of share returns on the returns of a market index, calculated as the covariance between stock and market divided by the variance of the market.

What it measures is systematic risk, meaning risk that cannot be diversified away. Company-specific risk disappears in a portfolio and is deliberately not captured. In the capital asset pricing model beta feeds directly into the cost of equity, which equals the risk-free rate plus beta multiplied by the market risk premium. A distinction is drawn between levered beta, which contains the capital structure of the company observed, and unlevered beta, which reflects pure business risk.

Under the simplifying assumption of risk-free debt, the conversion follows the relationship under which levered beta equals unlevered beta multiplied by one plus the debt-to-equity ratio weighted by one minus the tax rate. For unlisted companies, which is the normal case in the mid-market, a peer group beta of listed comparables is therefore unlevered, averaged and then relevered at the target capital structure of the company being valued. The estimation parameters matter in practice, because the observation period, return interval and index chosen noticeably change the result. Two to five years on weekly or monthly data are common.

An adjustment towards the market average of one is also frequently applied, because raw historical betas tend to move in that direction over time. For practical use in the mid-market it should be noted that the peer group often consists of markedly larger and more broadly based listed companies whose risk profile does not match an owner-managed business. That is one reason for additional premiums. It should also be checked whether the comparables are genuinely liquid, since thin trading distorts the regression. In practice betas from several sources are therefore compared and outliers excluded before an average is formed.