Are Corporate Payouts Abnormally High in the 2000s?

Well-trod ground new NBER research finds no evidence that buybacks reduce investment though there is a negative relationships btw payouts and investment, firms in the top 10% of payout distribution had higher returns than firms in bottom 90%.

“….In this paper, we show that payouts in the 2000s are sharply higher than from 1971 to 1999, whether we look at constant dollar aggregate payouts or firm-level payouts. The increase in payouts comes about both through an increase in payout rates and through an increase in funds available for payouts. Thirty seven percent of the increase in aggregate constant dollar payouts is explained by an increase in constant dollar aggregate operating income while sixty-three percent of the increase is explained by an increase in the payout rate…. The companies in our sample pay out almost $10 trillion from 2000 to 2017 - gross payouts are $10,823 billion and net payouts, which net out equity issues by repurchasing firms, are $9,862 billion. In the 2000s, annual aggregate real payouts average roughly three times their pre-2000 level. Aggregate net payouts as a percentage of aggregate corporate assets also increase substantially, averaging 2.7% from 1971 to 1999 versus 4.1% from 2000 to 2017. When we examine the ratio of aggregate net payouts to aggregate operating income, we see a similar increase; the average ratio increases from 18.9% to 32.4%. The year 2018 stands out, as aggregate net payouts as a percentage of aggregate corporate assets are 6.2% and aggregate payouts as a percentage of operating income are 48.4%. We show that payouts increase both because firms’ capacity to pay increases and because their willingness to pay increases. Specifically, in the aggregate, 38% of the increase in payouts is due to the fact that firms earn more and 62% is explained by the fact that firms have a higher payout rate. To explain the increase in payouts, it is therefore essential to explain why the payout rate increases. … Dividends average 14.4% of operating income from 1971 to 1999 and 14% from 2000 to 2017. In contrast, net repurchases average 4.8% of operating income before 2000 and 18.3% from 2000 to 2017….The aggregate level of payouts in the 2000s can be explained well using a model available in the literature estimated on data from 1971 to 1999 to predict payouts in the 2000s conditional on actual determinants of payouts in the 2000s. For instance, this model estimated with data from 1971 to 1999 overpredicts payouts in 2017 by 6%. As documented in Kahle and Stulz (2017), firm characteristics change substantially over time. Specifically, firms in the 2000s are larger and older than firms over the period 1971-1999. Larger and older firms have higher payouts, so changes in firm characteristics imply an increase in payouts and payout rates….We find similar results for aggregate net payouts. In 2018, ten firms account for more than 25% of the aggregate net payouts in our sample. Fewer than 10% of firms make 90% of the aggregate net payouts. For example, before 2000, on average, the net payouts of 338 firms amount to 90% of the net payouts of all firms. In the 2000s, 90% of net payouts come from the top 248 payers. Another way to see that aggregate net payouts reflect the actions of few firms is to look at the net payouts of firms who pay out more than $100 million in 2017 dollars. In 2018, these firms represent 21.7% of the firms in our sample…To investigate how much of the increase in payouts can be explained by changes in firm characteristics, we estimate payout rate models on data from 1971 to 1999 that relate payouts to firm characteristics. We then use these models to predict payout rates in the 2000s given actual firm characteristics. Models estimated from 1971 to 1999 predict an increase payout rates for the 2000s because of changes in firm characteristics, but they do not predict an increase as large as the actual increase. For the firms for which we have adequate data to predict the payout rate, the payout rate increases from 12.9% to 20.2% from before the 2000s to the 2000s. Changes in firm characteristics explain 71% of that increase.Models estimated with data from 1971 to 1999 are less successful in estimating the increase in payout rates for firms that have payouts. For these firms, the payout rate increases by 10.8%. Of this increase, changes in firm characteristics explain 49% of the increase. However, the predicted average payout rate for firms with payouts is always below the actual payout rate in the 2000s. We investigate whether the underprediction can be explained by the tax law changes that affect multinationals and by cross-market arbitrage. We find that both tax law changes and cross-market arbitrage explain part of the large prediction errors of our models. The question this paper tries to answer is whether payouts in the 2000s are abnormal. The answer is that much of the increase in payouts can be explained by known determinants of payouts. However, these models do a better job in explaining the evolution of aggregate payouts than they do in explaining the payouts of individual firms. A plausible explanation is that aggregate payouts reflect the payouts of the largest payers. These payouts may be more predictable over time. There is evidence that part of the increase in the payout rate can be explained by the fact that firms are more sensitive to determinants of payouts in the 2000s. In other words, if a firm’s payout rate is positively related to a firm characteristic before 2000, it is more strongly related to that firm characteristic in the 2000s. An increase in the sensitivity of payouts could be a positive development if it means that firms are less likely to hoard funds internally that could be invested more profitably outside the firm. Alternatively, however, such an increase could be problematic if it means that firms are more reluctant to take advantage of valuable internal investment opportunities. However, while our study does not provide tests that would establish or reject a causal relation between capital expenditures and payouts, it is noteworthy that even if one were to argue that firms decreased capital expenditures to increase payouts, this effect would explain little of the increase in payouts given the weak relation between capital expenditures and payouts….”