The flexibility of corporate payouts vis-à-vis capital investment: some UK evidence

This study investigates the relative flexibility of payouts vis-à-vis investment in the UK, motivated by concerns regarding this market's distinct payout characteristics and limited relevant research. It addresses the information gap related to the use of conditional mean estimations and examines firm behavior across the investment distribution.

The sample is an unbalanced panel of 6,173 firm-year observations, from 271 non-financial firms in the FTSE-All Share Index, during 1990–2019. Estimation methods include pooled- ordinary least squares (OLS) and firm fixed-effects regressions as well as unconditional quantile regressions with firm fixed effects.

For the “average” firm results show a negative relationship between share repurchases and investment, amplified in the presence of financial constraints and growth opportunities. Quantile regressions analysis reveals heterogeneous firm behavior as this relationship becomes stronger in successive quantiles of the investment distribution and disappears at the upper/lower extremes. Results suggest that UK firms exploit the inherent flexibility of share repurchases to facilitate investment. However, this flexibility appears irrelevant to firms with extremely high/low investment, characterized by significant differences in growth opportunities, cash flows and external financing cost. Dividends and investment are independent across the investment distribution, underlining the rigidity of dividends in the UK.

To the best of our knowledge, this is the first study to investigate the relative flexibility of payouts vis-à-vis investment in the UK, using firm-level financial data and at points other than the conditional mean. Its value lies in that it shows that share repurchases facilitate rather than impede investment and thus do not corroborate relevant concerns by economists and policymakers. Additionally, by utilizing a relatively new methodology it uncovered heterogeneous firm behavior across the investment distribution suggesting that conditional mean estimations should be applied with caution at least for highly heterogeneous samples.