FIRM-LEVEL DETERMINANTS OF THE PRICE–EARNINGS RATIO: PANEL EVIDENCE FROM TEN U.S. MARKET-CAP EQUITIES (2005–2026)

Abstract

This thesis investigates the firm-level determinants of the Price–Earnings (P/E) ratio using a balanced quarterly panel of ten U.S. mega-capitalisation equities — Apple, Microsoft, NVIDIA, Amazon, Berkshire Hathaway, Walmart, Eli Lilly, Exxon Mobil, Johnson & Johnson and Alphabet — observed over 84 quarters from the second quarter of 2005 through the first quarter of 2026, yielding 840 firm-quarter observations collected from the Macrotrends database. Three estimators — pooled OLS, fixed effects, and random effects — are compared and adjudicated through the Breusch–Pagan Lagrange multiplier test (LM = 239.55, p < 0.001) and the Hausman specification test (χ²(6) = 53.83, p < 0.001). Both tests decisively identify the fixed-effects estimator with firm-clustered robust standard errors as the consistent and efficient specification. The preferred estimates reveal three central findings: the price-to-sales and price-to-book multiples enter the P/E equation with strongly significant positive coefficients, reflecting a common latent growth factor shared across valuation multiples; the return on equity enters with a significant negative coefficient, consistent with Penman’s (1996) persistence–transience argument and Basu’s (1977) value-premium finding; and leverage and liquidity variables lose all explanatory power once firm fixed effects are introduced. The results imply that roughly four-fifths of observed cross-firm dispersion in P/E multiples is attributable to time-invariant firm characteristics rather than to balance-sheet fundamentals, with direct implications for relative valuation practice.