CAPM simplifies risk assessment by assuming a single risk factor, the market portfolio, to determine an asset's expected return . It relies on systematic risk measured by beta, which captures a security's response to market movements . In contrast, APT accounts for multiple risk factors, allowing for deviations in expected returns due to various economic influences. It does not specify these factors a priori, making it flexible but also more complex, as it requires identification of relevant risk factors which influence stock returns and are not directly observable .

The CML in CAPM represents the risk-return tradeoff for efficient portfolios, specifically those lying on the best possible line from the risk-free rate to the market portfolio on the efficient frontier. It indicates the highest possible returns for a given level of risk . In contrast, the SML depicts the expected return of individual securities as a function of their beta, reflecting the additional return needed to be compensated for taking on additional systematic risk, stemming from a security's volatility relative to the market . Unlike CML, SML applies to individual securities, regardless of portfolio efficiency .

The arbitrage-free equilibrium price mechanism in APT assumes that if two otherwise identical assets have different prices, arbitrageurs will exploit these discrepancies until equilibrium is restored, whereby the prices adjust to eliminate arbitrage opportunities . This mechanism rests upon the law of one price, suggesting that risk-adjusted returns should be equal across securities when arbitrage opportunities don't exist, ensuring that the expected returns align with risk factors without permitting unexploited price advantages . Unlike CAPM, APT recognizes multiple risk factors, thus facilitating this balance in a more complex market environment .

Implementing APT in practice is fraught with difficulties mainly due to the need to identify the relevant factors influencing security returns, which are not well-specified a priori . The diversity and unpredictability of potential factors, such as macroeconomic influences, interest rates, or inflation, complicate the model’s application. Additionally, the expected risk premiums and sensitivities (factor loadings) must be estimated, which requires extensive data and complex statistical techniques. These challenges stand in contrast to CAPM's focus on the single market factor, which is more straightforward but less flexible in considering various risk dimensions .

The Single Index Model attributes total risk of a security into market risk, which is captured by beta, and unique risk, which is associated with the security's specific characteristics. The model splits the security's return into a market-related component influenced by the index, and a unique component that is independent . This risk attribution implies that effective portfolio management can focus on diversifying unique risk, as market risk is systematic and cannot be eliminated. By constructing a diversified portfolio, managers can minimize unique risks and rely on the model to analyze the impact of market movements on potential returns .

Portfolio beta is often considered more reliable than individual security beta because the aggregation of multiple securities tends to stabilize the beta, reducing the impact of company-specific events and estimation errors that can drastically affect the beta of individual securities . Portfolio betas capture the overall systematic risk exposure of a combinatory set of assets, providing a smoother representation of risk relative to the market index . Moreover, changes in one security's circumstances do not disproportionately affect the overall portfolio beta, making it less volatile over time .

A security's beta in CAPM is estimated using the market model, which assumes a linear relationship between the security's returns and the market returns, represented as Ri = αi + βi RM + ei . Estimation requires past data to regress security returns against market returns, typically using stock indices as proxies. However, beta estimates may vary over time due to changes in a company's situation and are not stationary. Additionally, estimation involves error, and differing observation periods further contribute to variability. Portfolio betas tend to be more stable than individual security betas .

CAPM is based on several key assumptions: all investors utilize the same information to create efficient frontiers, have identical one-period time horizons, can lend or borrow at the risk-free rate, and are not affected by transaction costs, personal taxes, or inflation. No single investor can influence stock prices, and capital markets are in equilibrium . These assumptions streamline the model's application, but they also limit its practicality, since in reality, investors might possess different information, face transaction costs, and be influenced by taxation and inflation .

Estimating the expected market return under CAPM involves considering historical market returns, taking into account the long-term perspectives, and expected growth rates of the economy. Since the expected market return is not directly observable, it is often approximated through past market indices returns and adjusted for expected future conditions . Analysts must also consider economic forecasts and potential changes in risk premiums driven by market dynamics, investor sentiment, and macroeconomic factors that could influence future market performance . These estimates are critical for determining the appropriate risk premiums used in calculating security expected returns .

The Single Index Model simplifies the understanding of security covariance and risk decomposition by linking a security's return to the returns on a common index, such as the S&P 500, with a focus on a unique part and a market-related part. This model suggests that securities covary together only because of their common relationship to the market index, allowing covariances to depend solely on market risk . This contrasts with the multi-index models or the full variance-covariance method of Markowitz, which require consideration of multiple factors or relationships, making them more complex .