Calculating DAF Payout and What We Learn When We Do It Correctly

Working Paper: NBER ID: w27888

Authors: James Andreoni; Ray Madoff

Abstract: The tremendous increase in the use of donor-advised funds for charitable donations has led policy-makers to ask if there is sufficient regulation and oversight of DAFs. In the absence of account level reporting, the debate has focused on the average payout rates of DAF sponsoring organizations, which have been reported by the DAF industry to exceed 20%. We show that the industry-preferred method for calculating payout rate overstates the correct payout by more than 50%. We then argue that the flow rate is uninformative unless grounded in the stock of assets held by the DAF sponsor. We suggest a different measure of flow we call the stockpiling rate. Finally, we show that transfers between DAFSs cause DAF grants to be overstated. Reporting transfers separately would allow a more precise estimate of flow.

Keywords: donor-advised funds; payout rates; charitable donations; regulation; financial analysis

JEL Codes: H2; H31; H43; H83


Causal Claims Network Graph

Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.


Causal Claims

CauseEffect
Industry-preferred method for calculating DAF payout rates (G35)Overstates the actual payout by more than 50% (G35)
Flawed denominators that do not include grants made (H81)Inflated payout rates (G35)
New formula (C29)More accurately reflects portion of DAF dollars going to charity (D64)
Application of new formula to 2017 tax returns (H31)Increase in reported payout rates by 53% for all DAF sponsors and by 56% for commercial DAF sponsors (G35)
Flow rate (contributions to grants) (H77)Uninformative measure of DAF performance (C52)
Stockpiling rate (Q31)More meaningful measure of DAF performance (C52)
DAF-to-DAF transfers (H87)Artificially inflate reported grants (H81)
DAF-to-DAF transfers (H87)Overstated payout rates (G35)
Without adjusting for DAF-to-DAF transfers (Y10)Evaluations of DAF performance will remain biased and misleading (C52)

Back to index