The Stock Market Valuation of Human Capital Creation

How do you recognise companies that create value through their employees?

Matthias Regier and Ethan Rouen explore whether publicly disclosed personnel expenses (PE) capture future value in their paper The Stock Market Valuation of Human Capital Creation.

They link past PE to current operating income to compute the personnel expenditure future value (PEFV) and examine its pricing, its influence on analysts' forecasts, and its ability to predict future stock returns.

Their main conclusions include:

• A long–short strategy based solely on PEFV produces annualised abnormal returns of 6.5% for value–weighted portfolios and 3.5% for equal–weighted portfolios.
• When PEFV is combined with current personnel expense scaled by total assets, abnormal returns rise to an annualised range of 7.8% (value–weighted) and 4.8% (equal–weighted).
• These results suggest human capital is not properly valued by financial markets, both considering historical human capital investments and future investment opportunities.
• While current PE is priced negatively, the future value component (PEFV) commands a positive pricing coefficient, as if investors recognise the potential of human capital creation but not future benefits in full.
• Analyst forecast errors increase with higher PEFV, implying that analysts systematically underestimate future earnings for firms with substantial human capital investments.
• Abnormal returns based on PEFV decay over time, from 7.8% in the first year to 5.1% in the second year, and become statistically insignificant by the third year.

This research suggests active investors can integrate PEFV metrics into their models to identify undervalued firms. For instance, adjusting earnings forecasts to account for human capital creation may yield superior risk–adjusted returns and contribute to a more robust evaluation of intangible assets.

As a limitation, the PEFV measure is derived from total personnel expense, which includes both cost and investment components, which can introduce measurement error.

The optimal lag structure also varies by industry, and assuming that past personnel expense translates directly into future operating income may oversimplify the dynamics of human capital creation.