Giverny Capital Asset Management Q1 2026 Letter

To our clients & friends:

For the first quarter ended March 31, 2026, the Giverny Capital Asset Management (“GCAM”) model portfolio performed as follows: 1

GCAM Performance	Quarter ended	One-year ended	Three-years ended	Five-years ended	*Annualized Since Inception
3/31/2026	3/31/2026	03/31/2026	03/31/2026	3/31/2026
Portfolio Return -Net	-6.88%	8.52%	15.60%	8.65%	15.81%
S&P 500 TR	-4.33%	17.80%	18.32%	12.06%	18.46%
Excess Return -Net	-2.54%	-9.29%	-2.72%	-3.42%	-2.65%
* Inception Date 04/01/2020
Returns longer than one year are annualized

An investor friend sighed recently that it feels like we are fitting entire years’ worth of activity into each quarter and wouldn’t it be nice to have the opposite sometime? Alas, we are not living in those times. In January, when it appeared the Federal Reserve might vote for several interest rate cuts this year, our portfolio was up about 3% for the month. In February, we gave up some of the gains and in March, the US bombing campaign against Iran caused oil prices to spike, inflation to tick up and the stock market to teeter. Our portfolio was down 8% during the month of March and nearly 7% for the quarter.

The S&P 500 Index returns include the reinvestment of dividends and other earnings. The Index is an unmanaged, capitalization-weighted Index of common stocks of 500 major US companies. The Index does not incur expenses and is not available for investment.

In recent days, news of a temporary ceasefire to the hostilities in Iran prompted a market rally: as of the market close on April 13th , the S&P 500 Index was slightly up for the year-to-date, as are we. No doubt we will have another month's worth of activity next week.

What to write about such a volatile time? I do not have an opinion on how long the conflict in Iran will last or how it might end, what will happen with oil prices or whether the US economy will sputter in coming months. I do know that our portfolio companies as a group have strong earnings, healthy balance sheets and deep competitive moats. That was true when we were down in March and it is true in April.

I don't think it serves much purpose to focus on short-term market fluctuations, except perhaps to repeat that markets prefer predictability, which we do not have. As a result, volatility may remain the norm. It might be more useful to cover company-specific news that has affected our performance. In particular, a primary factor in our lagging results for the first quarter and the 12 months ended March 31st was the steep decline in the price of Constellation Software.

At the end of the first quarter one year ago, Constellation was our largest position and a 7.9% overall portfolio weight. The Toronto company's stock price in US dollars was $3,168. On March 31st , Constellation closed at $1,755 and represented a 3.9% weight despite us not selling a share. That amounts to a -44.5% return over one year.

Software businesses have been marked down dramatically over the past six months or so, as the market believes AI-coded software will be cheap to produce and capable of replacing many existing applications. However, I am skeptical that Constellation will be displaced quickly or easily.

To test this thesis, over the past few months we spoke to former Constellation employees, some of whom now work at competitors. Their feedback was consistent: the vast majority of Constellation's assets are data repositories or so-called Enterprise Resource Planning (ERP) systems that would be expensive and disruptive to displace, even if the act of coding got much cheaper. One of the most helpful insights from this round of conversations was the importance of service. When a small business owner needs help, they want to call someone they trust and who understands the intricacies of their business. Support is as important as the code itself.

In addition to reaching out to industry experts, we had the eye-opening experience of moving our own performance reporting software from one vendor, Morningstar (MORN), to another, Advyzon, over the past few months.

We used Morningstar since the inception of our firm to provide us with analytics on our portfolio, and as a service for clients to do the same for their own accounts. It was not integrated into my daily workflow, meaning I had to connect to Morningstar to run reports. This is the easiest kind of software to replace: a specific tool that is segregated from our operating system.

In 2025, Morningstar announced it would exit the performance reporting business, forcing us to find a new provider. Between conducting due diligence on three vendors, training on the new provider's system, managing the transfer of six years of historical data on several hundred separately managed accounts into a new system, and then troubleshooting small-but-persistent issues with historical data being different in the two vendors' systems, our operations chief Al Munro spent more than 100 hours on this transition and I, perhaps 15.

Happily, we like the new system and we're going to save a few thousand dollars a year in fees. The truth, however, is only a fool would replace a "good enough" software package to achieve marginal savings, especially when it means moving client account data around. The time Al and I spent on this switch amounted to years' worth of future cost savings, plus a few gray hairs.

Turning then to Constellation, it typically functions as the ERP system for its customers, and it might hold decades worth of personnel records, customer data and accounting history. Often, that operating data is precious and, for regulatory or legal reasons, should not be compromised. One thing switching software vendors does every time is compromise data. It was frustrating for us to replace a tool. It would be a nightmare to interrupt the daily operations of our organization. I don't foresee America's small businesses or municipalities replacing their ERP systems unless they are badly broken or egregiously expensive.

Meanwhile, Constellation churns out cash. Over the past five years, Constellation's free cash flow has grown 18% annually, from about $1.2 billion in 2020 to $2.7 billion in 2025. The stock over the five years ended March 31st compounded at a modest 4.8% annualized rate. This disconnect between business results and stock price tells us the market believes the cash flow trajectory is unsustainable. For sure, the cost of creating new software is collapsing. That does not mean the cost of replacing software is also collapsing.

I don't believe there is a good economic argument for replacing operating software that works and often constitutes less than 2% of an organization's operating budget. This is especially true for the kind of small enterprises that lack internal expertise to troubleshoot a conversion. Such firms are the foundation of Constellation's business.

That said, the market is probably right that we will see slower growth rates for software firms in the future as clients steer innovation budgets to AI applications rather than new software products. For now, Constellation is growing: Wall Street thinks free cash flow will step up from $2.7 billion in 2025 to at least $3.0 billion this year, or more than $140 per share. If so, the stock trades at an 8% free cash flow yield.

Two caveats to this thesis. I spoke recently to a friend who works at one of America's largest companies. They are focused internally on using AI applications to eliminate tens of thousands of jobs. Investment is flowing to AI, not software, and to the extent companies license software by headcount, that model could be endangered.

Second, throughout the software industry we do not see companies responding to the implosion of their share prices by buying back stock. Constellation has always said acquisitions provide higher returns than share repurchase, but senior executives are not rushing to buy either, so far as I am aware. So, I have my reasons for believing Constellation's business will prove durable, but I would be more inclined to add to our position if insiders were, as Warren Buffett likes to say, greedy when others are fearful.

During the quarter, we exited one position and established one new position.

We exited Ametek in February at $232, after trimming the position in the fourth quarter. I am an admirer of Ametek and am appreciative of the fine return we earned as shareholders over our six years of ownership. However, Ametek's rising PE multiple of ~27x felt too high relative to its long-term earnings growth rate of about 10% and its return on equity of 15%. The growth and return on equity figures are solid, but do not justify a growth stock multiple.

We used most of our Ametek proceeds to establish a new position in American Express (AXP) in March, at a price of $294. Probably many of you hold at least one American Express card. It's one of the premier status brands in the world, with a customer base of prime borrowers who often pay hundreds of dollars a year for the privilege of earning lucrative rewards. Those same cardholders generally do not revolve loan balances, meaning Amex earns much more money from transaction fees and annual cardholder dues than it does in interest on monthly balances. It makes money because people transact with the card to earn rewards, not because they need to borrow money to make ends meet.

It is on my mind that we may be living in a time of peak affluence. I read recently that the United States now has more than 430,000 households with a net worth above $30 million. Simultaneously, the federal government is running irresponsible budget deficits, many college-educated young people can't get career-track jobs or afford housing, AI may threaten the future of white-collar work, and income inequality mainly seems to worsen. Could higher taxes, lower federal spending, an AI-led white-collar recession, a push for redistributive economic policies or perhaps some combination of all of them bode ill for the kind of folks who hold American Express cards?

Maybe. A simpler question would be: do cardholders like the product and want to use it? The answer there is a clear yes. Spending on the card has been rising faster than overall consumer spending in recent years and lately has been accelerating. Paraphrasing an American Express executive at a recent investor meeting, “We keep raising the card fee to match the value, and the demand keeps rising.” Cardholders seem keenly aware of the value they’re getting back on their spending.

Similarly, merchants very much want the patronage of American Express’ wealthy cardholders, to the point that they fund a sizable portion of the rewards programs. This creates a virtuous cycle of American Express driving high spenders to participating merchants, who rebate a portion of their spending through rewards. Restaurants and shops want the customers, the customers want the rewards, and Amex plays matchmaker.

We bought our shares after a recent 25% drop in the stock price amid outstanding earnings performance and conservative accounting for potential future credit losses. Amex continues to add premium cardholders at remarkable rates: it added 5.8 million new consumer card accounts in 2025 at an average annual fee above $210. Further, it is growing disproportionately among Gen-Z and Millennials: the company says premium cardholder growth with younger adults is 15% in the US and 20% internationally. Our purchase price represents a PE multiple of about 15 times the 2027 earnings estimate of ~$20 per share, a fair price for a business growing earnings 11%-13% per year and generating a return on equity above 30%.

In addition to these two changes, we had some small trims and additions to existing positions. We trimmed Installed Building Products during the quarter after CEO Jeff Edwards announced the sale of a portion of his holdings after a big run-up in the stock price. We did the same. IBP remains a large holding and we remain fans of management.

We added to positions in Kinsale, TWFG and Hawkins during the quarter. I have written at length about all three in recent letters and will spare you the repetition. We ended the quarter with 6% cash. We do not lack ideas, but we feel comfortable holding some dry powder during chaotic times.

Drilling down a bit on our performance, below are the five most positive contributors to our return over quarter, followed by the five most negative contributors.

Top 5 Contributors for Quarter Ended 03/31/2026	Ending Weight %	Total Return %	Contribution to Return %
Taiwan Semiconductor (TSM)	5.1%	11.8%	0.4%
AMETEK, Inc.	0.0%	13.5%	0.4%
Watsco, Inc.	4.3%	8.8%	0.3%
Hawkins, Inc.	2.8%	8.3%	0.2%
Installed Building Products, Inc.	4.4%	3.0%	0.2%
	16.6%		1.4%

I don't believe quarterly results tell us a lot, but it's worth noting Taiwan Semiconductor performed very well when its leading customers such as Nvidia (NVDA), Broadcom and Apple saw their valuations cut during the quarter. TSMC has long traded at a PE multiple discount to its customers, but the gap is narrowing.

Bottom 5 Contributors for Quarter Ended 03/31/2026	Ending Weight %	Total Return %	Contribution to Return %
Constellation Software Inc.	4.0%	-27.0%	-1.2%
Alphabet Inc. Class A (GOOGL)	10.7%	-8.1%	-0.9%
Meta Platforms Inc Class A (META)	5.9%	-13.3%	-0.9%
HEICO Corporation Class A (HEI.A)	4.4%	-16.3%	-0.7%
Arista Networks, Inc.	9.6%	-6.3%	-0.7%
	34.5%		-4.5%

We have owned all five of these since our inception, and all but Constellation have generated excellent performance over our six years in business. The portfolio is not a museum, and I will make changes as needed, but this is a "bottom five" that boasts strong balance sheets, hefty operating margins, high returns on capital and, with the possible exception of Constellation, deep competitive moats.

I am giving a talk in Omaha on Friday morning, May 1st , at Robert Miles' Value Investor Conference, held every year before the Berkshire Hathaway (BRK.A, BRK.B) annual meeting. If you're traveling to the meeting this year, it would be nice to see you there.

With every good wish,

David M. Poppe

---

PS – you can learn more about us at Giverny Capital Asset Management and you can contact us at gcaminfo@givernycapital.com.

---

References

1. GCAM’s model portfolio is a Poppe family account (“portfolio”). The portfolio does not pay an advisory fee, but the returns presented herein assume the deduction of an annual advisory fee of 1% to show what a client account’s performance would have been if it had been invested the same as the portfolio. The returns reflect reinvestment of dividends and other earnings. Past performance is not necessarily indicative of future results.

---

---

Original Post

Editor's Note: The summary bullets for this article were chosen by Seeking Alpha editors.