Analysis notebook: preliminary examination of the impact of DQ changes

December 29, 2025: Data updated to be current as of 30 November, 2025. Update adds 393 filings. No material changes to interpretations.

November 20, 2025: Data has been updated to be current as of 31 October, 2025. This has resulted in material changes to some interpretations - in the process of working through the new data.

Note: This is a work in progress and the estimates generated are neither final, nor guaranteed to be accurate. Please do not circulate or cite without permission - contact dlasby@imaginecanada.ca for permissions / more detail.

To do list:

• brief description of the changes to the DQ, including summary of consultation process

• come up with some sort of imputation approach to deal with missing non-charitable assets data

  • alternatively, at least characterize the effects of missing data¹

• generate basic data files

• top line trend data

  • disbursement patterns
  • non-charitable assets

• cross-sectional DQ shift data

  • aggregate qualifying disbursements
  • distributions of qualifying disbursements
  • relationship between qualifying disbursements and expenditures overall
  • see what changes there are with grants going abroad - compare to Schedule data
  • look at associated organization data to try and filter out transfers

• longitudinal DQ shift data

  • determine how to identify movers
  • characterize baseline
  • quantify increase

• Find way of vizing multiple distributions other than transparency

• data quality issues

  • double reporting of lines 5000 and 5050
  • cross-check gifts data off T1236
  • run down cause of 2024 activities issue
  • some sort of estimate of the fraction of grants value should actually be gifts

Introduction

The key issue to explore is whether qualifying disbursements have increased with the shift to a higher disbursement quota, starting in 2023. Our primary focus is on the contributions of organizations that have historically tended to disburse less than required by the revised DQ (i.e., less than 5% on non-charitable assets valued at $1 million or more). Organizations that have historically contributed above this level are not the target of this shift. While understanding their behaviour is interesting, it is not central to the policy question.

Secondary issue is understanding the impact of any increases in qualifying disbursements on foundation assets. Likely going to be very difficult to disentangle from market events; also note that there will be significant lag effects given that the basis of value is 24 months prior (need to understand market dynamics 36 months ago),

Background

It is important to understand that this policy change has been enacted in an environment of tremendous growth in foundations and foundation assets, with the further significant complication of a global pandemic and its ongoing aftermath.

Between 2013 and 2023, the number of private foundations increased by over a quarter (26.4%), while the number of public foundations has contracted by about 7.1%.

The pandemic does not seem to have significantly affected growth in the number and assets of foundations, but it does seem to have affected granting patterns. The percentages of both public and private foundations making gifts to qualified donees decreased somewhat during the pandemic (potentially even somewhat prior, starting in 2018). While rates of making gifts have increased since the depths of the pandemic, they have not yet reached pre-pandemic norms, at least among private foundations.² The new ability to make grants to non-qualified donees is clearly not the driver of the pandemic-period decline as this only came into effect in 2022, though this channel looks to be materially increasing.

Growth in foundations and pandemic-related shifts

Note that 2024 filings are not complete, which has an impact on most of these visualizations.

Number of foundations

Understanding the types of foundations that are driving these patterns is going to be significant for future research.

Asset value

Need to understand the extent to which DAFs are driving public foundation growth. Private foundation growth dominated by MasterCard foundation, but do need to understand how other foundations are faring. Although our primary focus is on determining whether gifts are increasing, we also need to understand any shifts in assets - likely too early for this, but need to put down some sort of marker to fill the vacuum.

% making gifts

I have a sneaking suspicion that at least some of this is business life-cycle dynamics: public foundations gifting at lower levels before de-registering and newly registered private foundations not yet gifting. For context, the number of foundations required for the observed 2017 - 2019 shifts would be on the order of 80 public foundations and 75 private foundations.

Value of gifts

Looking at the step function shift in private foundation support from 2020, MasterCard and Temerty together account for about $1.1 billion of the 2021 shift; 2022 is a little more broadly based, with MasterCard accounting for about half and transformative gifts to HUCs from various foundations driving a big chunk of the rest; for 2023 things are even more broadly based - MasterCard accounts for about a third, with much of the rest from private foundations making substantial gifts (again, generally to HUCs). For 2023, need to run down the tails of the distribution; this may actually be a real policy-induced shift. Key is going to be separating out the contributions from foundations with non-charitable assets vs. others - look for other analyses to not do this.

% making grants

Value of grants

Paralleling the increase in the number and assets of foundations, the fraction of foundations reporting unused assets on lines 5900 and 5910 is increasing.³ Below the surface, there is an increase in both the number of private foundations and the percentage of public and private foundations (both new and previously existing) reporting assets not being used in their charitable programs. Not surprisingly, the absolute value of these assets is increasing significantly, both in absolute terms and as a percentage of total foundation assets.

Non-charitable assets summary

Again, keep in mind that 2024 filings are not complete. Clearly there has been a significant increase in the reporting of non-charitable assets. Will have to see whether this is due entirely to the new form, or filers paying more attention given the increased CRA focus on the issue. Note also the impact on non-charitable assets when excluding MasterCard Foundation.

It is particularly important to keep in mind that if the new filers tend to have smaller fractions of their assets in non-charitable form (this seems a reasonable guess), it could potentially juice naive cross-sectional assessments (i.e., if we just restrict the analytical universe to foundations with non-charitable assets in any period, we’ll potentially end up including new, generally better performing foundations in the post-DQ measures and excluding them from prior measures).

% Foundations

Amazing what happens if one gives someone a form to fill out rather than vague, non-enforced guidance.⁴ To really nail the case down, need to divide 2023 into pre- and post-DQ change components.

% of total assets - wi. MC

Note that these line items are naive - they are not adjusted for permission to accumulate property, nor do they account for permission to reduce disbursements.

% of total assets - excl. MC

Once we exclude MasterCard Foundation we can see that for private foundations a materially higher fraction of total assets (about 10 percentage points 2022 / 2023) are being reported as non-charitable.

Line 5900 summary

Line 5910 summary

DQ shifts

The new DQ requirements took effect for taxation years beginning on or after January 1, 2023. By extension, this means that only foundations with a fiscal year end of December 31 were required to meet the new requirements for 2023. For this this reason, 2023 filings are divided into pre- and post-DQ change groups. From 2024 onwards, all foundations are subject to the new requirements.

Although most public foundations have earlier year ends, nearly three quarters of private foundations have December 31 year ends. Significantly, private foundations with December 31 year ends account for roughly 85% of private foundation assets not used in charitable programs.

Cross-sectional data

Have qualifying disbursements increased?

Looking at the aggregate data (total qualifying disbursements against varying measures of total assets), there is little indication that disbursements have increased significantly as a result the new DQ requirements. The observed 2023/2024 fluctuations have more to do with the types of foundations that have December 31 year ends vs. earlier year ends and/or file early in the 6-month period after year end (and dominate current 2024 filings) than with any observed changes in aggregate disbursements.

It is important to understand that we shouldn’t necessarily expect to see significant shifts in the aggregate measure. Bottom line, this measure is just too blunt to show anything meaningful given: the rapid increase in asset value, the 24 month lag between assets and the calculation of the DQ requirement, and the extent to which the increase in asset value is being dominated by a single very large foundation (MasterCard Foundation) all reduce the acuity of this measure.

Total qualifying disbursements

Note that these figures are restricted to foundations reporting non-charitable assets. This is different from the universe above, which was all foundations.

% of Line 5900

% of Line 5910

% of Line 815

% of total assets

Median data

Looking at median values⁵ for disbursements as a percentage of non-charitable assets, there is clear evidence for an increase in qualifying disbursements by private foundations. The evidence for public foundations is less clear, given the 2023 pre-DQ change spike (i.e., year-end effects could be predominating here).

Median qualifying disbursements

% of Line 5900

% of Line 5910

% of Line 815

% of total assets

What are foundations disbursing for?

If the new DQ measures are effective, we should see qualifying disbursements accounting for increased percentages of expenditures. If qualifying expenditures do not account for a larger proportion of total expenditures, this would indicate a policy failure if not by the letter of the law, certainly by its spirit.

Complicating our assessment, we are seeing some significant problems with the reporting of qualifying expenditures - principally double reporting of lines 5000 and 5050; need to both control for this and understand the impact on reporting.

In the sections below, the tabs labelled total and average show the shifts in qualifying disbursement patterns the others evaluate the potential impact reporting issues:

• filtered: aggregate with problematic records filtered out
• problem rate: percentage of records that are problematic
• % asset value: percentage of total foundation assets captured by problematic records
• % L815: percentage of L815 (non-charitable assets, adjusted for permission to accumulate) captured by problematic records
• % disbursement / gift / activity / grant value: percentage of total expenditure type captured by problematic records

Qualifying disbursements as a fraction of total expenditures

Bottom line, it does look like we are seeing increases in the percentage of expenditures allocated to qualifying disbursements. These can be seen in the aggregate private foundation estimates (both raw and filtering out problematic records) and in the average data for both public and private foundations.

Overall, the percentage of clearly problematic records (qualifying disbursements > total expenditures) though reasonably high does not appear to involve enough assets or disbursements to call these conclusions into doubt.

Total

Filtered

Average

Excludes records where qualifying disbursements are greater than total expenditures (logically impossible).

Problem rate

% asset value

% L815

% disbursement value

Gifts as a fraction of total expenditures

Incorporating the new data, looks like there isn’t huge movement in the aggregates, though the incompleteness of the 2024 data could well be a factor.

Whether one believes that there has been an increase in gift levels very much depends on the timespan one is looking at. There is some indication of near term changes (i.e., since 2020/2021), but post-DQ change levels are actually quite in line with pre-pandemic norms.

Problematic records (gifts > total expenditures) don’t look to be having a massive impact here, though should probably take a look at qualifying disbursements > total expenditures filter as well.

Total

Filtered

Average

Excludes records where gifts are greater than total expenditures (logically impossible).

Problem rate

% asset value

% L815

% gift value

Charitable activities as fraction of total expenditures

Significant 2024 increase in aggregate among public foundations, but this looks to be a reporting error. Private foundation data volatile enough that it is difficult to be certain. Again, need to be cautious because of the incompleteness of 2024 data.

Looking at the averages, there appear to be some near-term increases for private foundations, reversing the long-term trend. Public foundations look much more in-line with long term trends (which increased after 2018). Key question here will be whether we are seeing substitution of charitable activities for gifting or if this is an artifact of increased reporting of non-charitable assets.

Problematic records (activities > total expenditures) are relatively rare but do appear to have a material impact on the aggregates for public foundations. Average will be substantially unaffected by this, given the nature of the measure (one charity, one equally weighted data point).

Total

Filtered

Average

Excludes records where activities are greater than total expenditures (logically impossible).

Problem rate

% asset value

% L815

% activity value

Grants as fraction of total expenditures

Role of grants is clearly increasing and the amounts involved are significant enough that they account for a material component of the shift. Bigger impact on public foundations than private, likely driven by the strong international component.

Error rate is non-existent here, due to the relatively marginal character of the support (i.e., few organizations are going to give enough grants that there is a chance that grants > total expenditures). Given this, need to also see what the impact of a qualifying disbursements > total expenditures filter is.

Parenthetically, two issues here: 1) MasterCard has a major impact here, accounting for a significant fraction of total granting, 2) there are issues with an indeterminate fraction of grants (i.e., gifts and operating expenses being mis-identified as grants); unsure what the monetary impact of this is.

Total

Filtered

Average

This data presentation excludes records where grants are greater than total expenditures (logically impossible).

Problem rate

% asset value

% L815

% grants value

Changes in the distributions of disbursement percentages

In these vizes, “Baseline” refers to the filing period prior to the DQ change, “Post-DQ changes” refers to the filing period immediately after the DQ change, and “Sustaining” refers to the second period following the DQ change, if the foundation has filed a T3010 for the period.

These vizes restrict observations to foundations with qualifying disbursements equivalent to 10% or less of non-charitable assets, so as to give us an indication of what is going on in the part of the distribution that is the target of the policy shift.

In interpreting these vizes, note that it is completely normal to see significant levels of disbursement below the mandated thresholds. Because organizations can pull prior over-contributions forward for up to five years and following over-contributions for one year to make up annual shortfalls, it is quite common for in-compliance foundations to be slightly or even significantly below the DQ threshold in multiple years and counter this by being above it in a few (or even one) year(s).

Year over year change

Year over year changes look at changes from one year to the next, without accounting for the fact that DQ compliance is calculated over a 7 year period (charities are able to pull forward excess qualifying expenditures up to five years and pull them back up to one year to compensate for under-disbursement in a particular year).

Foundations with L815 >= $1e6

Scope is foundations with non-charitable assets of $1 million or more.

Foundations with L815 < $1e6

Scope is foundations with non-charitable assets less than $1 million.

In-range foundations with L815 > $1e6

Scope is foundations that had qualifying expenditures ranging between 3.45% and 5% of non-charitable assets prior to the DQ shifts.⁶ The in-scope range is computed accounting for the ability to pull forward excess qualifying expenditures to compensate for under-disbursement.

Rolling changes

Rolling changes attempt to account for the fact that DQ compliance is calculated over a 7 year period. The data show what we would expect - there has been a material increase in qualifying disbursements, but the scale of the increase is small, given that organizations have only had one or two periods at the higher contribution rate, and shows an increase in succeeding years.

Foundations with L815 >= $1e6

Scope is foundations with non-charitable assets of $1 million or more.

Foundations with L815 < $1e6

Scope is foundations with non-charitable assets less than $1 million.

In-range foundations with L815 > $1e6

Scope is foundations that had qualifying expenditures ranging between 3.45% and 5% of non-charitable assets prior to the DQ shifts. This is computed accounting for the ability to pull forward excess qualifying expenditures to compensate for under-disbursement.

Longitudinal data

While the above analyses are consistent with an increase in disbursements, they are cross-sectional (rather than longitudinal) and do not focus on foundations that have historically tended to disburse less than 5% of non-charitable assets. Beyond the relatively limited analytic potential, I include them primarily because most other analyses will take these types of approaches. To really evaluate whether disbursements are increasing to meet the new requirements, we need to look at the data longitudinally (i.e., tracking changes in individual foundation behaviour) and focus on the foundations that are actually the object of the policy change.

Broadly, it looks like there has been a significant increase in qualifying disbursements. Depending on how one defines the increase (i.e., against baseline pre-DQ shift organizational norms or against an arbitrary 3.5%) the scale of the increase appears to be somewhere in the range of $660 million to $1 billion.

While this top line figure is certainly striking, it is notable that MasterCard Foundation accounts for two thirds of the increase, regardless of how one defines baseline.

To progress towards a publication, first we need to explain the boundaries of the target population - how many foundations report non-charitable assets of $1 million or more, disbursing less than 5% of the value of non-charitable assets annually.⁷ Then we need to characterize their role, in terms of the fraction of assets (both total and non-charitable) they control, as well as the proportion of qualifying disbursements they account for.

One could try to argue that the missing non-charitable assets data is causing an under-estimate of any DQ shifts, but this shouldn’t actually be viewed as a major deal for exploring this issue. Collectively, these foundations account for only about a tenth of total assets, but about a quarter of total expenditures and about 45% of charitable program expenses - in sum, they don’t have a huge financial role, have pretty high flow through of revenues, and aren’t primarily focused on grantmaking.

Appendix A - Aggregates Excluding MasterCard Foundation

Growth in foundations and pandemic-related shifts

Note that 2024 filings are not complete, which has an impact on most of these visualizations.

Asset value

Value of gifts

Value of grants

Non-charitable assets summary

% of total assets

Note that these line items are naive - they are not adjusted for permission to accumulate property, nor do they account for permission to reduce disbursements.

Line 5900 summary

Line 5910 summary

Total qualifying disbursements

Note that these figures are restricted to foundations reporting non-charitable assets. This is different from the universe above, which was all foundations.

% of Line 5900

% of Line 5910

% of Line 815

% of total assets

Appendix B - Methodology

Data completeness

Line 815 missingness

Baseline

        range >= $1 million < $1 million L815 missing
  Above range         15.0%        10.8%         3.6%
     In range          7.5%         2.0%         0.2%
  Below range          3.4%         2.4%         0.1%
 Missing L815          0.0%         2.2%        52.9%

1st post-DQ

        range >= $1 million < $1 million L815 missing
  Above range         17.5%        13.2%         3.1%
     In range          8.0%         2.0%         0.1%
  Below range          3.0%         2.4%         0.2%
 Missing L815          0.0%         2.3%        48.2%

2nd post-DQ

        range >= $1 million < $1 million L815 missing
  Above range         17.7%        13.6%         2.4%
     In range         10.0%         2.3%         0.0%
  Below range          2.4%         1.9%         0.2%
 Missing L815          0.0%         2.5%        47.0%

Evaluating impact of Expenditures reporting issues

Filtered qualifying disbursements as a fraction of total expenditures

These analyses exclude records where the difference between activities (Lines 5000, 5010, 5020, 5040, 5045, and 5050) and total expenditures exceed 2% of expenditures and $500. Broadly, the story seems to be the same as with the unfiltered data.

Total

Filtered

Average

Excludes records where qualifying disbursements are greater than total expenditures (logically impossible).

Problem rate

% asset value

% L815

% disbursement value

Filtered gifts as a fraction of total expenditures

Total

Filtered

Average

Excludes records where gifts are greater than total expenditures (logically impossible).

Problem rate

% asset value

% L815

% gift value

Filtered charitable activities as fraction of total expenditures

Total

Filtered

Average

Excludes records where activities are greater than total expenditures (logically impossible).

Problem rate

% asset value

% L815

% activity value

Filtered grants as fraction of total expenditures

Total

Filtered

Average

This data presentation excludes records where grants are greater than total expenditures (logically impossible).

Problem rate

% asset value

% L815

% grants value

---

1. don’t forget that increased reporting is going to have its own effects – will this systematically over- / under-state impact?↩︎

2. Figures for 2024 reflect only foundations that have filed as of the end of September, 2025 and should not be viewed as reflecting the population as a whole.↩︎

3. Again, 2024 figures are from partial filings and should not be taken as a reliable indicator of change.↩︎

4. Snarky, but true.↩︎

5. We focus on median values because mean values are too skewed by outliers.↩︎

6. We use this range because such a high fraction of charities appear to be slightly under the threshold, likely due to computation error. This doesn’t appear to be intentional.↩︎

7. One could argue that foundations disbursing less than 3.5% are out of scope, but the net effect of these foundations is so small that it doesn’t materially affect the analysis and it might look like we are trying to low ball any shifts if we exclude them.↩︎