California’s Largest Charitable Remainder Trusts: An IRS Data Analysis

Klaus Gottlieb, Esq.

Aug 29

9 min read

Open the full PDF article

By Klaus Gottlieb, Esq.

Jurisdiction

Federal, with California fiduciary-law overlay

Primary Statutes

IRC §§ 6034, 664, 4941, 4946, 4947(a)(2); Cal. Prob. Code §§ 16002, 16003, 16045–16054

Key Authorities

Treas. Reg. §§ 1.664-1(a)(7), 53.4947-1(c); IRS Form 5227, Split-Interest Trust Information Return; IRS Statistics of Income Form 5227 microdata, filing years 2016, 2017, 2020, 2021, and 2022

Last Reviewed

August 2026

Category

Charitable Planning — Research / Data

At a Glance

What this is, in one sentence. I analyzed five years of IRS Form 5227 microdata to see what California’s very largest charitable remainder trusts actually look like: what type of CRT they use, where they are located, and, most importantly, who serves as trustee.

The population. The data identify 4,241 unique California-addressed CRTs reporting assets of $50 million or more across the five filing years studied. Approximately 83% are charitable remainder unitrusts (CRUTs) and 17% are charitable remainder annuity trusts (CRATs).

The surprise. Only 15.6% were classified as having institutional trustees. Roughly 67% were managed by private individual trustees, with another 17.8% unclassified. Even if every unclassified trustee were assumed to be institutional, private trustees would still outnumber institutional trustees by at least two to one.

The practical lesson. Institutional trusteeship is plainly not the only model used for very large CRTs. But private trusteeship works only if the trust has the necessary infrastructure: professional tax administration, clear self-dealing rules, proper valuation procedures for unmarketable assets, and a serious successor-trustee plan.

The warning label. IRS administrative data are messy. Asset values are self-reported, the IRS documentation appears to reverse the CRUT and CRAT classification codes, the early and later filing years are not directly comparable, and the California field reflects the filer’s mailing address rather than necessarily the trust’s legal situs. The numbers need to be read accordingly.

1. What Can IRS Data Tell Us About Large CRTs?

Charitable remainder trusts are discussed constantly in estate-planning, charitable-giving, and investment circles. Empirical information about how they are actually structured and administered is much harder to find.

Form 5227 provides an unusual window into that world. A CRT files the Split-Interest Trust Information Return annually, and the IRS Statistics of Income division releases public microdata for selected filing years. Five available years — 2016, 2017, 2020, 2021, and 2022 — contain name-level data that include trust names, trustee names, addresses, trust classification, and financial information.

I filtered those files for California filers reporting at least $50 million in assets and then removed apparent charitable lead trusts that had been miscoded as remainder trusts. The result is an unusual look at the highest end of the California CRT market.

The numbers matter less because most donors will ever establish a $50 million CRT than because very large trusts reveal how sophisticated donors solve the same structural questions faced by a $500,000 or $2 million CRT.

And one answer stands out.

2. Finding One: The CRUT Dominates

Of the 4,241 unique CRTs in the dataset, 3,531 — approximately 83% — are CRUTs. Only 700, or about 17%, are CRATs.

That roughly five-to-one ratio makes structural sense.

A CRAT pays a fixed dollar amount and cannot accept additional contributions after initial funding. A CRUT pays a percentage of annually revalued assets and can be designed to receive additional contributions. For founders, executives, and other donors whose wealth may become liquid over several transactions or several years, that flexibility can matter.

CRATs can also have greater difficulty satisfying the § 664(d)(1)(D) 10% remainder requirement when interest rates are low, particularly for younger beneficiaries or long payout periods. And because the annuity payment remains fixed while assets can decline, a poorly performing CRAT can progressively consume its own principal. A CRUT automatically reduces its dollar distribution as the trust value falls.

The preference for CRUTs at the very high end is therefore not especially surprising.

The trustee data are.

3. Finding Two: Private Trustees, Not Banks, Are the Dominant Model

The common mental picture of a $50 million or $100 million charitable trust involves a bank or trust company sitting in the trustee’s chair.

The IRS data tell a different story.

Trustee classification	Trusts	Share
Private individual	2,825	~67%
Institutional	660	15.6%
Unclassified	756	17.8%

Among trusts that could be classified, there are approximately 4.3 private trustees for every institutional trustee. And the result does not disappear under aggressive assumptions. If all 756 unclassified trustees were institutional, private trustees would still outnumber institutional trustees by more than two to one.

Nor does trust size explain the difference. Among trusts over the $50 million reporting threshold, reported median values were broadly similar between the private and institutional groups. The data do not show institutional trustees becoming markedly more common as reported trust size rises.

That is probably the most important finding in the study.

4. Why Would Someone With a Very Large CRT Use an Individual Trustee?

The IRS data can tell us what people do. They cannot establish why they do it. But several explanations are plausible.

Many large CRTs begin with concentrated stock, private-company interests, real estate, or another asset about which the donor has unusually deep knowledge. The donor may not want to hand control over the timing of diversification, investment-manager selection, or portfolio strategy to a bank.

At the largest asset levels, donors may already have family-office personnel, investment advisers, accountants, custodians, and lawyers performing much of the work an institutional trustee would otherwise supply. Adding a corporate trustee can therefore duplicate infrastructure already in place while adding another layer of fees.

More importantly, the trustee function can be unbundled.

One person or entity can exercise fiduciary and investment authority. A brokerage firm can provide custody. A specialist CRT administrator can handle § 664(b) tier accounting, Form 5227, K-1s, annual calculations, and distributions. The donor does not necessarily have to purchase all three services from the same institution.

That model is not confined to $50 million trusts.

It can be even more relevant at $500,000 or $1 million, where a conventional institutional trustee fee consumes a larger percentage of the trust’s expected return.

But private trusteeship comes with conditions.

5. The Three Risks a Private Trustee Cannot Ignore

Self-dealing

A CRT is a split-interest trust subject to the private-foundation self-dealing rules through IRC § 4947(a)(2). The donor who created the trust is ordinarily a disqualified person, and a donor serving as trustee has an additional fiduciary role.

IRC § 4941 imposes an initial excise tax of 10% of the amount involved on a disqualified person who engages in self-dealing. A foundation manager who knowingly participates can face a separate 5% tax, and an uncorrected transaction can ultimately generate a 200% second-tier tax on the self-dealer.

Ordinary payments of the required annuity or unitrust amount are specifically accommodated by the CRT rules. That does not turn the trust into the donor’s personal account.

Related-party transactions, use of trust property, and decisions designed to confer an impermissible personal benefit require particular care. The administrative file should reflect that distinction from the beginning.

Unmarketable assets

Real estate, closely held stock, restricted securities, and similar property create a second problem.

Under Treas. Reg. § 1.664-1(a)(7), a CRT holding an unmarketable asset must have that asset valued either exclusively by an independent trustee or through a current qualified appraisal by a qualified appraiser. A grantor who is also the noncharitable beneficiary does not satisfy the regulation’s definition of an independent trustee.

A self-trusteed CRUT holding public securities after diversification is therefore very different from a self-trusteed CRUT still holding the donor’s apartment building or closely held company stock.

Trustee succession

This may be the easiest risk to overlook.

A lifetime CRT created by someone in his or her fifties can remain in existence for decades. The original trustee may eventually become unwilling, incapacitated, or deceased.

Every privately trusteed CRT therefore needs a succession plan that works without improvisation: a named successor, a mechanism for appointing later successors, or an institutional trustee available as a backstop. Otherwise an incapacity can leave an irrevocable split-interest trust needing court intervention merely to keep operating.

6. California Fiduciary Law Still Applies

Federal tax compliance is only half of the trustee’s job.

An individual trustee administering a California CRT is also subject to the state’s fiduciary rules, including the Uniform Prudent Investor Act, the duty of loyalty, and the duty of impartiality among beneficiaries.

That last duty is particularly important in a CRT. The trustee is administering one pool of property for beneficiaries whose economic interests point in different directions: the current recipient benefits from distributions during the trust term, while the charitable beneficiary receives what remains.

California Probate Code § 16003 requires the trustee to deal impartially with multiple beneficiaries and to take their differing interests into account when investing and managing the trust.

Serving as trustee therefore preserves control. It does not eliminate fiduciary responsibility.

7. Geography: Where the Largest California-Addressed CRTs Appear

The dataset is heavily concentrated in the state’s principal wealth centers.

The five highest-count cities were:

Redwood City — 311
San Francisco — 264
Los Angeles — 254
San Diego — 128
Palo Alto — 121

By broader region, the Bay Area accounted for 36.0% of the trusts, Los Angeles 14.8%, San Diego 6.6%, and Orange County 5.3%. But a substantial 37.3% fell elsewhere in California.

The institutional list is also revealing. Universities are unusually prominent. Stanford, Caltech, UC Berkeley, Pomona, UCLA, USC, Santa Clara, and Occidental all appear among the leading institutional trustees, alongside financial institutions such as BNY Mellon and State Street.

One important qualification: the IRS state field is a mailing-address field, usually reflecting the trustee’s address. It does not conclusively establish the trust’s legal situs or the donor’s residence.

8. A Warning About the Numbers

The dataset is useful. It is not clean.

Several limitations are important enough that I would not publish the findings without them.

First, Form 5227 asset amounts are self-reported and are not routinely audited. Some entries are plainly impossible: the largest reported value in the dataset is approximately $586 billion. For that reason, I do not treat the reported aggregate asset value as a reliable measure of the actual wealth held in these trusts.

Second, the IRS documentation for the trust-type field appears to reverse CRATs and CRUTs. The published documentation says code 1 means CRAT and code 2 means CRUT. Applying that mapping produces a result inconsistent both with IRS national statistics and with the trust names in the data. Cross-validation shows that the empirical mapping is the reverse: code 1 corresponds to CRUT and code 2 to CRAT.

Third, there is a dramatic break between the 2016–2017 files and the 2020–2022 files. The later years contain thousands of qualifying records where the earlier years contain only dozens. That almost certainly reflects an IRS processing change rather than an explosion in the number of giant CRTs.

Finally, classification of trustees as private or institutional necessarily involved pattern matching of trustee names. About 18% could not be classified, and some misclassification is unavoidable. But the private-trustee finding survives every plausible allocation of the unknown group.

The exact percentages should therefore be treated as estimates.

The direction of the result is much harder to dismiss.

9. What the Largest CRTs Tell Us About Ordinary CRTs

A $50 million charitable remainder trust is unusual. The problems it must solve are not.

The largest trusts provide a useful demonstration that trustee, custodian, investment manager, and tax administrator do not have to be the same person or institution. Private trusteeship is not an eccentric structure reserved for small trusts. It is, at least in this dataset, the predominant structure among some of California’s largest CRTs.

That does not mean every donor should serve as trustee.

It means trustee selection should be treated as an engineering problem rather than a reflexive choice between “do it yourself” and “hire a bank.”

For a private trustee, the architecture needs four pieces:

competent custody and investment execution;
professional § 664 accounting and tax administration;
procedures for self-dealing and unmarketable assets; and
a successor-trustee structure that can survive the donor.

The IRS data cannot tell us whether any particular CRT is well administered. What they can tell us is that private trusteeship is already commonplace at the very highest asset levels.

The practitioner’s job is not simply to decide who occupies the trustee box on the day the CRT is signed. It is to make sure the structure can function safely for the next twenty or thirty years.

The Bottom Line

The headline finding from this study is not how large these trusts are.

It is who runs them.

Roughly two-thirds of the California CRTs classified in this $50-million-plus dataset are managed by private individuals rather than banks, trust companies, universities, or other institutions. That result should change the way practitioners think about trustee selection.

Self-trusteeship can preserve investment control and eliminate an institutional fee layer. Modern custodians and specialist CRT administrators make it possible to outsource much of the operational work without surrendering the trustee role. But that model succeeds only when the tax and fiduciary hazards are dealt with deliberately — particularly § 4941 self-dealing, valuation of unmarketable assets, professional § 664(b) administration, and trustee succession.

Those problems do not become less important because a trust is smaller.

In several respects, they become more important.

Read the complete published analysis, including the methodology, tables, institutional-trustee rankings, data limitations, and citations.