How many nonprofits will shut their doors?
We find ourselves amidst a pandemic-driven recession of uncertain scope and duration. Over the last months, Candid employees have been asked repeatedly, “How many nonprofits are going to shut their doors and go out of business?”
There is no easy way to answer this question. We cannot predict the future. What we can do, however, is offer a set of possible scenarios and comb those scenarios for insight. Over the last two weeks, my colleagues Carol Brouwer and Anna Koob led an effort to analyze the financial future of a set of 315,698 U.S.-based nonprofits by considering 20 possible scenarios.
Below, I’ll explain which organizations we looked at, detail our methodology, and offer some observations. In summary, here’s what we found:
- In our median baseline scenario, we found that 12,042 nonprofits (4%) would close in the absence of a crisis.
- Across nine “realistic” scenarios, a median of 34,472 nonprofits (11%) go out of business.
- In this median case, about 22,000 additional nonprofits (7%) close because of the COVID-19 crisis.
- In our most optimistic scenario, only 8,420 nonprofits (3%) shut their doors.
- In our most dire scenario, we lose 119,517 nonprofits (38%).
These scenarios reveal a range of contradictions. On the one hand, they demonstrate the strength of the nonprofit sector. The majority of nonprofit organizations are positioned to weather this storm. And, indeed, some degree of turnover is healthy to maintain a vibrant sector. But the people served by shuttered nonprofits face practical consequences that cannot be abstracted away.
Context
It will be years until we have comprehensive data about how the recession has affected nonprofit finances. But recent data can suggest a range of possible futures. A survey from LaPiana and Associates shows that 90 percent of US-based nonprofits experienced a reduction in revenue by April 2020, up from 70 percent in March. A CAF America survey of global nonprofits has found that 50 percent expect revenue drops of at least 20 percent. An NFF survey of US-based organizations shows that 75 percent are seeing reduced earned revenue, 50 percent reduced contributions, and 27 percent reduced government revenue.
In response to the crises of 2020, we have seen extraordinary generosity, starting with $11.4 billion to address the COVID-19 crisis and $7.6 billion toward racial equity. A Fidelity Charitable survey reports that 25 percent of donors say they’ll give more.
On the surface, this might seem a contradiction: nonprofits say donors are giving less but donors announce that they’re giving more. But we believe these numbers simply show these crises are affecting organizations in different ways.
On the expense side, there is evidence that nonprofits are cutting costs. Respondents to the LaPiana survey said they had, on average, laid off or furloughed 18 percent of their staffs, with 44 percent expecting further reductions. An analysis by the Center for Civil Society at Johns Hopkins estimated that 1.6 million nonprofit jobs were lost in the U.S. from February through May of this year. At the same time, many organizations are seeing increased need—and thus, greater expense burden.
This is an undoubtedly mixed picture. Accordingly, we used the available data to set the parameters of the scenarios we explore below.
Methodology
Scenario planning involves imperfect—and ultimately arbitrary—assumptions. We have done our best to create a coherent structure for those assumptions. Generally, the scenarios fall into three categories. The first 6 (in yellow green) are “baseline” scenarios that show what might have happened in the absence of the coronavirus crisis. The next 12 (in green) reflect a range of possible impacts of the financial crisis on nonprofits based on the survey data cited above. The final 2 (in gold) are “miracle” scenarios where nonprofits see immense generosity, cost-cutting measures, and an increase in government investment over a short crisis.
We do not know how long this financial crisis will last so the scenarios range from 9 to 36 months. The below chart shows the specific assumptions for all 20 scenarios.
Individual nonprofits face unique financial circumstances. Some rely on earned revenue (e.g., ticket sales at an arts organization), others on donations (e.g., foundation grants to a policy institute). Most have a mix of revenue sources. Each organization brings its own asset base and expenses.
To reflect this financial diversity, our calculations are based on each individual nonprofit’s revenue mix, expense rate, and cash levels. That is, for each of the 20 scenarios, we applied a set of assumptions to each of 315,698 nonprofits. In total, this analysis involved approximately 52 million calculations.
It is important to emphasize that our analysis focused on a specific set of 501(c)(3) charitable nonprofits. Our set of 315,698 does not include foundations, nonprofits not eligible to receive a tax-deductible contribution, or most organizations with revenue under $50,000 a year. In general, this set reflects what most people think of when they think of “charitable organizations” and accounts for the bulk of the financial activity in the social sector. It does not, however, reflect the full scope of the field.
In the methodology notes, we detail several ways that these estimates might underestimate or overestimate the number of bankruptcies. On balance, we believe it is more likely that we are underestimating.
Results
The below chart lays out 20 scenarios for how many US-based nonprofits might go out of business because of the current economic crisis:
Nonprofit loss by issue area
Different issue areas lend themselves to different nonprofit business models. Art museums make and spend money in different ways than health clinics. This variance plays out in our analysis.
Below we share results by issue area. “Baseline average” reflects the median of what we consider the most “realistic” baseline scenarios (#1-3). “Scenario average” is the median of what we consider the most “realistic” COVID-19 scenarios (#7-15). “Net loss” is the difference between “Baseline average” and “Scenario average” and thus reflects how many additional nonprofits might go out of business because of the COVID-19 crisis across scenarios. The “most optimistic scenario” is #19. See the methodology for more information on these definitions and the choices behind them.
We have shortened the descriptive text for those categories marked with an asterisk. The full text for each category can be found on the National Center for Charitable Statistics website.
Conclusion
Nonprofits go out of business naturally. In some cases, that is to be celebrated: they’ve accomplished their mission or discovered what doesn’t work. Under each of our 20 scenarios, most nonprofits survive this crisis, again proving the resilience of the field as a whole.
With that said, the nonprofit community needs to confront the distributed tragedies to come. Candid does not claim to have universal advice for this moment. But there are strategies that could make a difference, such as scenario planning, collaboration, restructuring, and innovative finance. Choices made by government and foundations will matter, and nonprofits have an opportunity to influence those choices.
Nonprofit organizations have agency in this moment, the ability to choose how they respond to the obstacles in front of them. Moments of flux are moments of opportunity. There is the possibility of a miracle: an unprecedented surge of donor generosity or government intervention. In the absence of that miracle, let us see this as a moment for innovation and agility.
Methodological details
- The data on which this report is based comes from 315,698 Form 990 and Form 990-EZ IRS filings of public charities (as defined in subsection 501(c)(3) of the U.S. tax code). We have used the most recent filing available to us, going back as far as fiscal year 2015 in some cases when necessary to maximize the record count and diversity of data.
- Small organizations filing a 990-N were omitted as they do not report financial information. Filings covering substantially more or less than 12 months were removed from the set, as were those reporting negative values for total expenses or total revenue. Additional filters were applied to exclude organizations believed to be defunct or not in good standing with the IRS.
- The data is presented as reported by the filers, other than occasional errors introduced in data assimilation and processing. Numerous records have a total revenue amount that differs from the sum of its parts, although the difference is mostly small. Not all fields map well between form types—the closest available match has been made, sometimes requiring a combination of fields.
- For the baseline in the “net loss” calculations, we use Scenario 2. Scenario 2 is the median of Scenarios 1-3. We chose those scenarios because nonprofits could be expected to cut expenses before going out of business (vs. Scenarios 4-6, which did not include any expense reductions). To be conservative, we chose a 36-month duration.
- The “scenario average” is the median of Scenarios 7-15. These are the nine scenarios that most closely reflect the early data we’ve seen from surveys. We did not include scenarios 16-18 in the scenario average because those reflect situations where total revenues drop but expenses remain unchanged. We believe it is more realistic to model cases where nonprofits reduce expenses in response to reduced revenue. (See note 11d.)
- For the baseline, we chose Scenario 3 because it reflected the longest time horizon (36 months) among those baseline scenarios that assume cautious nonprofit behavior (including some expense cuts).
- In the analysis by issue area we grouped nonprofits into the “Major Codes” of the National Taxonomy of Exempt Entities (NTEE). In some cases we shortened the descriptive text of the NTEE codes. The full text can be found National Center for Charitable Statistics website. Those cases are marked with an asterisk. Over time, we will be shifting our analysis to the taxonomy included in the Philanthropy Classification System.
- By “cash” we mean readily accessible liquid assets. We calculate it as the sum of cash and investments. Some nonprofits have illiquid assets that could be liquidated in the case of a crisis.
- Three hypothetical scenarios of revenue reduction over three time spans were applied to the 501(c)(3) dataset we have been using. Each revenue source was reduced by its respective percentage shown. Total expenses were also reduced (except where noted) to simulate organization cost-cutting measures. The monthly net gain/loss (revenue minus expenses) was calculated and the available cash was divided by this value. Additionally, one scenario of revenue increase and two baselines are shown.
- Organizations are designated “out of business” when their monthly net revenue is negative (loss) and their MTZ (“Months to Zero”) is less than the “MOD” (“Months of Disruption.”) Organizations are also labeled “survived” when their monthly net revenue is positive or zero, regardless of the MOD. (A very small number of organizations have expenses equal to revenue, resulting in zero net revenue.) Note that nonprofits with a monthly net loss with sufficient cash reserves for the scenario period are also treated as survivors.
- There are several ways our analysis might underestimate the number of nonprofits going out of business:a. Most of the reflected revenue reduction estimates we used are lower than what we see from survey results.
b. We chose a baseline number based on 36 months of losses, not 12.
c. For the “scenario average” we used median instead of mean. We made the conservative choice to use median (as opposed to mean) in order to dampen the influence that outliers have on the average.
d. We did not include scenarios 16-18 in the scenario average. (See Note 5.)
e. Both revenues and expenses can be “lumpy.” Cash management and timing issues can force organizations to shut down before they have actually run out of net assets.
f. We did not include depreciation in expenses. Depreciation is not a cash expense so we consider this appropriate for shorter-term analysis. But, as we have noted previously, capital bills do come due. - There are also some ways our analysis could be an overestimate. If the survey results are wrong and donors prove generous in aggregate, it might be that the “miracle” scenarios are more feasible than we thought.
b. Additional government stimulus might buy nonprofits more time.
c. Nonprofits could use debt to buy time to weather the storm.
Vivien Hoexter says:
Dear Jacob,
Thank you (and Candid) for this very helpful analysis. As an advisor to nonprofits and foundations, I have been thinking about this moment and what is to come for my beloved sector. I want as many organizations as should survive, to survive, even if they have to partner or merge with other organizations.
With best regards,
Vivien Hoexter
Jacob Harold says:
Our thanks for the kind words here. We do hope this analysis helps the field as we all try ot figure out how to navigate this moment. Per Stephen's question on nonprofit employment, check out the work of the Johns Hopkins Center for Civil Society Studies here: http://ccss.jhu.edu/publications-findings/?did=517
Steven Saylor says:
Dear Jacob,
The Comstock Foundation for History and Culture would like to commend you on this tremendous report and analysis. We have certainly experienced the impact of Covid-19 having to cancel all our fundraising events. We've cut costs where we could and are exploring new opportunities to boost support.
Wishing you all the best.
Steven Saylor
Executive Director
Comstock Foundation for History and Culture
Restoring the Legacy of a Nation
comstockfoundation.org
Stephen Adams, President Community Foundation of North Central MA says:
Jacob:
This is an exceptionally helpful analysis, made more so by your clear, jargon-free writing. Thanks so much.
Do you know of any databases that use private employer data from the BLS to analyze the nonprofit sector?
Pat Jandacek says:
So very interesting.........and mind-boggling. Just knowing 'all' will not be lost is somewhat comforting to a degree.