Fluctuating selection at work: Estimating the firms’ fitnesses from time series of their market caps and forecasting their future shares

Abstract

Abstract Evolutionary economics has been useful to explain the nature of the processes of innovation, as well as providing useful heuristics for applied research. However, quantitative tests are rare and in general fail to capture the observed dynamics of firms in real markets. A main problem is how to estimate the fitnesses of companies. We present a quantitative implementation of fluctuating selection to explain and forecast the dynamics of firms’ market capitalizations (MC). We start by proposing a recipe to estimate the financial fitness of a company from its MC time series, which is completely analogous to the Malthusian fitness of evolutionary biology. Using Malthusian fitnesses in terms of MC implies that all companies, through their stocks, irrespective of their industry or sector, are competing for investors’ money. These fitnesses, in turn, allow to forecast the companies’ proportions of total MC through a natural selection equation. We validate the method using the daily MC of public-owned Fortune 100 companies across 2000–2021. This Fluctuating Selection from Market Capitalization (FSMC) formula is particularly accurate in forecasting the MC proportions of companies. It outperforms the geometric random walk stochastic benchmark for all forecasting instances across validation periods ranging from a month to a year. JEL codes: B52, C51, C52, C53, C63, G17