Skip to main content

Advertisement

Log in

Negotiation and executive gender pay gaps in nonprofit organizations

  • Published:
Review of Accounting Studies Aims and scope Submit manuscript

Abstract

This study examines gender pay gaps among nonprofit executives and how compensation negotiability influences these disparities. Using tax return data from IRS Form 990 filings, we find that females earn 8.9% lower total compensation than men in our sample. Further, we observe that settings more conducive to negotiation manifest in larger pay disparities, whereas settings that limit executives’ opportunities to negotiate or that encourage females in particular to negotiate produce smaller gender pay gaps. Our nonprofit setting constrains mechanisms, such as labor force participation rates and risk preferences, that are thought to explain the pay gap, and our results are robust to using a Heckman correction model and matched samples. These findings provide evidence from a large-scale archival dataset of a plausible mechanism for the gender pay gap and point to a potential cost of work environments where negotiations play a larger role in setting compensation.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save

Springer+ Basic
EUR 32.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or Ebook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2

Similar content being viewed by others

Notes

  1. We obtain financial information from Form 990 from the IRS Statistics of Income (“SOI”) database, maintained by the National Center for Charitable Statistics (“NCCS”), and merge this to compensation and executive data reported in the tax return that GuideStar maintains.

  2. For comparison, Carter et al. (2017) report a total compensation gap of 15%.

  3. Balsam and Harris (2014) test for program service revenue reliance and find that NPOs which are more reliant on service revenue are less sensitive to compensation-related donor backlash. We instead test donor reliance, as it is a more direct measure for our research question; however, our results are similar when using program service revenue reliance.

  4. Our methodology does not address whether women choosing to work in the nonprofit labor market differ from those who work for for-profit firms.

  5. Guidestar (2014) provides univariate evidence of a pay gap among nonprofit executives, but this analysis does not address any self-selection concerns or control for differences in NPO or executive attributes.

  6. Aigner and Cain (1977), on the other hand, argue that employer taste preferences are more likely to explain pay disparities in the long run than perceptions of group productivity are.

  7. We acknowledge, however, that NPOs can offer incentive compensation (i.e., bonuses), which could induce some risk-taking behavior that contributes to the gender pay gap.

  8. The application of the non-distribution constraint is a complex area of the tax law that appears subject to some judicial interpretation. For instance, in People of God Community vs. Commissioner (75 T.C. 127 (1980)), the Tax Court found pastor compensation based on a percentage of gross receipts violated the non-distribution constraint. Relatedly, IRS General Counsel Memorandum 39,862 (Nov. 22, 1991) disallowed payment to physicians based on hospital or departmental earnings. Nevertheless, the non-distribution constraint does not explicitly preclude NPOs from paying bonuses based on achieving certain benchmarks related to earnings. Balsam and Harris (2018) provide evidence that profitability is positively associated with bonuses for NPO executives.

  9. We impose the $150,000 threshold because NPOs whose top compensated officer, director, or employee earns less than this amount are not required to report the breakout of their compensation components on Schedule J of Form 990. We also require these NPOs to report at least $10,000 donations, non-missing assets, and positive administrative and program expenses in those years.

  10. Guidestar compiles the executive’s name, title, and reportable W-2 compensation from Part VII of IRS Form 990, which is publicly available but not in machine-readable form. Guidestar infers the executive’s gender using an algorithm based on his or her first name. We follow prior research and use gender to refer to a person’s sex: male or female. However, we recognize that the definition of gender can be used to encompass more than just a dichotomous classification of a person’s sex (Hardies and Khalifa 2018).

  11. Executive directors are coded as CEO. Our results are robust to including executives and key employees with position titles other than CEO, CFO, COO, GC, or marketing officer.

  12. We note that Guidestar was unable to provide the requested detail for every NPO we requested and that 2008 observations were only acquired for calendar year-end NPOs. Our results are similar when limiting our sample to 2009–2012. Our starting point also excludes executives who were not paid or whose compensation was reported as missing.

  13. In untabulated analysis, we follow Newton (2015) and use the log of revenue per employee, fundraising margin, and program expense to assets ratio as alternative measures of NPO efficiency. Our inferences are similar and including these variables as covariates in our main model does not impact our results.

  14. Untabulated multivariate test by position find consistent results with our univariate comparisons in Table 2, Panel B for the CEO and CFO.

  15. Our industry classifications are based on 26 group National Taxonomy of Exempt Entity (“NTEE”) codes; however, we do not have any observations from NPOs in the “unknown” industry category.

  16. Although we are unable to identify executive tenure, we are able to identify executives who were hired within our sample period. This provides us with a control for executives with short tenure. Our results hold when including this additional control in our main analysis.

  17. The short time series of NPO executive data makes it difficult to conduct meaningful changes tests. We only observe 127 changes where the new executive’s gender is different than the predecessor’s. Nonetheless, multivariate analysis of executive changes indicates that compensation is lower when a female executive replaces a male executive, but there is no statistically significant change in compensation when a male executive replaces a female executive.

  18. We use industry rather than NPO fixed effects in the first stage regressions so that we do not drop NPOs whose executives are represented by one gender. We do not include a control for household income in the first stage of our Heckman selection model because it is a state-level variable that is highly correlated (0.514) with our instrument.

  19. Our Heckman results are robust to replacing the missing Congress Ratio of D.C. NPOs with the Congress Ratio from Virginia or to dropping D.C.-headquartered observations.

  20. We use one-to-one matching without replacement. Where there is more than one potential match, the male executive is selected at random within the same industry-year and size quartile, consistent with CEM methodology. Examples of CEM used in the accounting literature include DeFond, Erkens, and Zhang (2016); Gallemore, Gipper, and Maydew (2019); and Maso, Lobo, Mazzi, and Paugam (2020).

  21. In untabulated analysis, our primary result is robust to further matching on position, though this matching severely restricts our sample size. Additionally, our main finding is similar when using a propensity score matched (PSM) sample.

  22. Results are robust to inclusion of a control for average employee pay, indicating that results are not dependent upon resources available to the NPO or to the NPO’s compensation practices in general.

  23. The magnitude of the salary pay gap is approximately 6%. We also test for pay gaps in the other compensation components (bonus, deferred compensation, nontaxable compensation, other reportable compensation). The coefficient on Female is generally significantly negative in each component except for other reportable compensation.

  24. Following Balsam and Harris (2018), we identify same industry using the NTEE/NAICS/SIC crosswalk. Like the aforementioned study, we do not consider same city, as this limits the number of identified for-profit firms in the same labor market. We measure size quartiles separately for the NPO and for-profit firms.

  25. To the extent that past compensation at another organization influences an employee’s current compensation, it is possible that gender pay disparities become entrenched as executives move more freely in the labor market. We attribute the lack of statistical significance on the competition coefficients to the inclusion of NPO fixed effects in our model. When we exclude the NPO fixed effects, the coefficients on these variables are significantly positive, suggesting that labor market competition increases the cost of labor. This competition appears to be fairly static within NPOs during our sample period though.

  26. In untabulated results we find similar results when we use the percentage of program service revenue to total revenue—referred to as “service versus charitable orientation”—as an alternative measure of donation reliance. We also find similar results when using a continuous specification for DonationsReliance.

  27. Our results are robust to using a 0/1 dummy variable for the existence of restricted assets (Yetman and Yetman 2013) and a continuous measure of Restricted Assets.

  28. We note that the coefficients on Donations Reliance and Restricted Assets are statistically insignificant across all specifications. When we run our regression without the NPO fixed effects, these coefficients are negative and statistically significant. We interpret these results as suggesting that donation reliance and asset restrictions do constrain executive compensation but appear to be fairly static within NPOs during our sample period.

  29. Our results are similar to those in Carter et al. (2017), who find that the gender pay gap among for-profit executives decreases 23% with the introduction of female board members. The authors of that study argue that gender-diverse boards with more female directors are less likely to discriminate against females. We cannot rule out that our results are attributable to the same phenomenon.

  30. We note that when performing the Heckman selection model for the subsample of non-CEO executives, the instrumental variable is directionally consistent with our main analysis but is not statistically significant. We include industry fixed effects (based on the NTEE codes presented in Table 2) rather than organization fixed effects in this model, as some NPOs will only show up once per year in this sample. Restricting the sample to NPOs with more than one non-CEO executive and implementing organization fixed effects produces directionally consistent but statistically insignificant results for the interaction of Female*Female CEO. However, the coefficient on Female is still negative and significant.

  31. We limit this analysis to NPOs with at least three unique executives during our sample period. We measure pay dispersion across all years, because requiring at least three executives per NPO-year severely limits our sample size. We are unable to include firm fixed effects in this model because we obtain only one coefficient of variation for each NPO.

  32. We recognize that for-profit firms may also disclose similar compensation policies in their proxy statements, but these disclosures are not available in any of the standard compensation data sets, such as ExecuComp.

References

Download references

Acknowledgements

We appreciate constructive comments from Patricia DeChow (editor), two anonymous referees, Casey Camors (discussant), Mary Ellen Carter, Erica Harris, Rafael Rogo (discussant), and participants at the 2019 AAA GNP mid-year meeting, 2019 AAA Annual Meeting, 2019 Conference on the Convergence of Financial and Managerial Accounting Research, Drexel University Sex and Gender Research Forum, and Wharton Society for the Advancement of Women in Business Academia’s “Women in Business Academia” Conference.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Andrew R. Finley.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

Variable definitions

Variable

Definition

Compensation variables:

Total Compensation

Natural log of total W-2 reportable compensation from the NPO reported on Part VII of Form 990.

Firm characteristic variables:

Female

Set equal to 1 if the executive is a female, and 0 otherwise

NP Competition

Set equal to one if the number of other NPOs in the same city (MSA), industry, and size quartile (following Balsam and Harris 2018) is above the sample median (18), and 0 otherwise.

FP Competition

Set equal to one if the number of for-profit firms in Compustat in the same industry and size quartile (following Balsam and Harris 2018) is above the sample median (100), and 0 otherwise.

Donations Reliance

Set equal to 1 if donor reliance (Contributions / Total Revenue) is above the sample median (0.06), and 0 otherwise.

Restricted Assets

Set equal to 1 if restricted assets to total assets is above the sample median (0.092), and 0 otherwise.

Low Relative Pay

Set equal to 1 if the coefficient on the NPO fixed effect from the compensation model is below the sample median (−0.06), and 0 otherwise.

Female Board

Set equal to 1 if the NPO’s ratio of female board members is above the median, and 0 otherwise

Female CEO

Set equal to 1 if the NPO-year has a female Chief Executive Officer (CEO), and 0 otherwise

Pay Dispersion

Set equal to one if the coefficient of variation on total compensation within an NPO’s executive ranks (i.e., standard deviation/average) is above the sample median (0.035), and 0 otherwise.

Program Spending Ratio

The ratio of Net Program Service Expenses t Net Total Expenses, where Net Program Service Expenses = Total Program Expenses − Compensation allocated to Program Service Expenses and Net Total Expenses = Total Expenses − Compensation Expense (Form 990 [Part IX, Line 25, Column B (i.e. IX.25.B)] − [IX.5.B] − [IX.6.B] − [IX.7.B]) / ([IX.25.A] − [IX.5.A] − [IX.6.A] − [IX.7.A])

Governance Index

Following Newton (2015), the average of four governance sub-indices (Governing Body, Governing Policies, Compensation Policies, Accountability & Transparency), where each sub-index is defined as the ratio of the sum of the indicator variables as a proportion of the total number of possible responses for each firm–year observation where each response is weighted by its annual cross-sectional standard deviation

Board Size

The number of voting members of the governing body [Form 990 Part VI, Section A, Line1a]

No. of Employees

Natural log of the Number of Employees [Form 990 Part I, Line 5], i.e., the number of compensated employees (full-time or part-time) listed on the entity’s W-3, “Transmittal of Wage and Tax Statements”

Total Assets

Natural log of total Assets [Form 990 Part X, Line16, Column.B]

Age

Natural log of the Age of the firm, calculated relative to the Year of Formation [Form 990 Page 1, Item L]

Liquid Assets/Total Expenses

The ratio of Liquid Assets to Net Total Expenses, where Liquid Assets = Cash + Savings and Temporary Cash Investments + Pledges and Grants Receivable, Net + Accounts Receivable, Net, and Net Total Expenses = Total Expenses − Compensation Expense(Form 990 [Part X, Line 1, Column B (i.e. X.1.B)] + [X.2.B] + [X.3.B] + [X.4.B]) / ([IX.25.A] − [IX.5.A] − [IX.6.A] − [IX.7.A])

Government Revenue

The ratio of Government Grants [Form 990 Part VIII, Line 1e] to Total Revenue [Form 990 Part VIII, Line 12, Column A]

Donations Growth

The change in the ratio of Net Contributions to Total Revenue [Form 990, Part VIII, Line 12, Column A] in year t − 1 relative to year t − 2,where Net Contributions = Total Contributions [Form 990, Part VIII, Line 1 h] − Government Revenue [Form 990, Part VIII, Line 1e]

Asset Tangibility

The ratio of Land, Buildings, and Equipment, Net [Form 990, Part X, Line 10c, Column B] to Total Assets [Form 990, Part X, Line 16, Column B]

Commercial Dummy

Set equal to 1 when the ratio of Program Services Revenue to (Program Services Revenue + Donations Revenue) is at least 90%, and 0 otherwise (Form 990 [Part VIII, Line 2 g, Column A i(i.e. VIII.2 g.A)] / [VIII.2 g.A] + [VIII.1 h.A])

Household Income

Natural log of median household income (as reported by the US Census Bureau’s American Community Survey); computed for each state- and fiscal-year pair

Congress Ratio

Percentage of female delegates, relative to total delegates, in the US House of Representatives and Senate from the state in which the NPO is headquartered

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Finley, A.R., Hall, C.M. & Marino, A.R. Negotiation and executive gender pay gaps in nonprofit organizations. Rev Account Stud 27, 1357–1388 (2022). https://doi.org/10.1007/s11142-021-09628-2

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11142-021-09628-2

Keywords

JEL classifications

Navigation