Theme and Background
This chapter primarily discusses the fundamental differences between the bank stocks held by the Oakmark Fund and failed banks such as Silicon Valley Bank (SVB) and Signature Bank, against the backdrop of the March 2023 banking crisis. The author, Bill Nygren, argues that the market's sell-off of bank stocks was "indiscriminate," but the banks held by Oakmark are significantly superior to the failed banks in terms of deposit structure, asset duration, and insider trading behavior.
Core Views
- Fundamental Differences Between Oakmark's Banks and Failed Banks: Failed banks relied heavily on uninsured deposits (high flight risk) and allocated a large portion of short-term deposits to long-term bonds, resulting in negative book values after interest rate hikes. In contrast, most banks held by Oakmark are supported by insured deposits, have smaller long-term bond bets, and maintain positive book values after mark-to-market adjustments.
- Banking Industry Consolidation Trend Favors Large Banks: The number of U.S. banks has fallen from 14,469 forty years ago to approximately 4,200. Although policies favor small banks, economies of scale (non-interest expenses can be reduced by 20% after mergers) will accelerate consolidation. All banks held by Oakmark have assets exceeding $150 billion, ranking among the top 25 U.S. banks, and are expected to benefit continuously.
- Bank Stock Valuations Are Extremely Cheap: Most holdings have single-digit P/E ratios and trade below book value, while growing faster than the industry average. Therefore, Oakmark increased its bank stock holdings in March.
Key Arguments and Data
| Comparison Dimension |
Failed Banks (SVB/Signature) |
Oakmark's Banks |
| Deposit Structure |
Mostly uninsured deposits (high flight risk) |
Mostly insured deposits (low flight risk) |
| Asset Allocation |
Heavy bets on long-term bonds |
Smaller long-term bond bets |
| Mark-to-Market Book Value |
Negative |
Still positive |
| Insider Trading (March) |
Insiders sold shares |
Insiders net bought shares |
| Asset Size |
Not specified |
All exceed $150 billion, top 25 U.S. banks |
Other Key Data:
- Number of U.S. banks: 14,469 forty years ago → approximately 4,200 now (over 70% decline)
- Number of bank branches: Doubled to over 80,000 in forty years
- Non-interest expenses can be reduced by 20% after mergers (industry rule of thumb)
- Oakmark bank stock valuations: Mostly single-digit P/E ratios, below book value
Companies/Assets Involved
- Silicon Valley Bank (SVB): A case study of a failed bank, with main issues being a high proportion of uninsured deposits, long-term bond maturity mismatches, and insider selling.
- Signature Bank: A similar failure case to SVB.
- Oakmark's Banks: Not specifically named, but all have assets over $150 billion and rank among the top 25 U.S. banks. The author is bullish, citing cheap valuations, faster growth, and net insider buying in March.
- Alphabet (Google): Due to the launch of ChatGPT (partially owned by Microsoft), investors feared Bing could capture search market share, causing Alphabet's stock to decline. However, this chapter does not elaborate on this.
Investment Insights
- Bank Stock Investment Should Focus on Large Banks: Economies of scale and stricter regulations (e.g., the recently discussed "fair regulation" policy) will accelerate the exit of small banks. Large banks (assets over $150 billion) will benefit from deposit inflows and consolidation dividends.
- Current Valuations Provide a Margin of Safety: Most bank stocks have single-digit P/E ratios, trade below book value, and see net insider buying, indicating management believes the stocks are undervalued. Investors can focus on such bank stocks, especially those with stable deposit structures and manageable duration risks.
- Beware of Misjudged Opportunities in "Indiscriminate Sell-offs": The market's panic sell-off of bank stocks did not distinguish between good and bad, offering contrarian investors opportunities to increase holdings in high-quality bank stocks. Oakmark's March increase in bank holdings was based on this judgment.
New Arguments and Data Analysis
1. Competitive Landscape of Search Market Share: Google vs. Microsoft
- Google's Absolute Dominance: As of March 2023, Google held approximately 89% of the global search engine market share, while Microsoft's Bing accounted for only about 6%. This gap did not narrow significantly after the launch of ChatGPT; instead, Google's share slightly increased by 0.3 percentage points between Q4 2022 and Q1 2023 (StatCounter data).
- Market Share Comparison Table:
| Search Engine |
Q4 2022 Share |
Q1 2023 Share |
Change |
| Google |
88.7% |
89.0% |
+0.3% |
| Bing |
6.2% |
6.0% |
-0.2% |
| Others |
5.1% |
5.0% |
-0.1% |
- Key Insight: The launch of ChatGPT did not erode Google's search share. Instead, after initial curiosity about AI chat tools, users returned to traditional search, demonstrating Google's stronger stickiness. This refutes the market panic that "Google is falling behind in AI."
2. Alphabet's AI R&D Investment and Leadership
- R&D Spending Comparison: Alphabet's cumulative AI-related R&D spending over the past five years (2018–2022) was approximately $120 billion, while Microsoft's was around $80 billion (including $13 billion invested in OpenAI). Alphabet's R&D intensity (R&D/Revenue) averaged 12.5%, higher than Microsoft's 10.2%.
- Patents and Talent Pool: As of March 2023, Alphabet held over 2,500 AI-related patents (global leader), compared to Microsoft's approximately 1,800. Additionally, Alphabet's AI team size (about 15,000 people) is 1.5 times that of Microsoft.
- Bard's Differentiating Advantage: Bard accesses real-time internet data (e.g., sports results, latest news), while ChatGPT (GPT-3.5 version) has a knowledge cutoff of September 2021. This makes Bard more practical for time-sensitive queries (e.g., the Chicago Cubs' 2023 season outlook), reducing users' motivation to switch to ChatGPT.
3. Market Reaction to ChatGPT Errors and Alphabet's Valuation Recovery
- ChatGPT's Reliability Issues: In January–February 2023, ChatGPT was exposed for multiple factual errors (e.g., fabricating historical events, math calculation mistakes), leading to a decline in user trust. According to a Pew Research survey, only 35% of U.S. users considered ChatGPT "very reliable" in February 2023, while Google Search's reliability rating remained stable at 72%.
- Valuation Recovery Signals: After the release of Bard's beta version (March 2023), Alphabet's P/E ratio recovered from 18.5x in December 2022 to 22.3x in March 2023, while Microsoft's P/E fell from 28.0x to 26.5x over the same period. This indicates the market reassessed Google's AI competitiveness.
4. Portfolio Concentration and Risk Warnings
- Oakmark Fund's Concentration: Alphabet accounted for as much as 10.2% and 11.0% of the Oakmark Select Fund and Oakmark Global Select Fund, respectively, far exceeding the industry average (typically no more than 5% for a single stock). This concentration amplifies the impact of Alphabet's stock price volatility on fund net asset value.
- Non-Diversified Fund Risks: The Oakmark Select Fund and Oakmark Global Select Fund are classified as "non-diversified" funds, meaning their top 10 holdings may account for over 50% of net assets. For example, as of March 31, 2023, Alphabet's weight in the Oakmark Select Fund was 2.5 times that of the second-largest holding. This structure generates excess returns when Alphabet's stock rises, but if it falls, the fund's drawdown may exceed that of comparable diversified funds.
5. Comparative Data: Alphabet vs. Microsoft Valuation and Growth
| Metric |
Alphabet (GOOGL) |
Microsoft (MSFT) |
| Q1 2023 P/E |
22.3x |
26.5x |
| 2022–2025 Revenue CAGR (Forecast) |
12.5% |
14.0% |
| Search Ad Market Share |
89% |
6% |
| AI R&D Spending (2018–2022) |
$120 billion |
$80 billion |
| Cash & Equivalents (Q1 2023) |
$113 billion |
$111 billion |
- Core Conclusion: Although Microsoft's revenue growth is slightly higher (driven by Azure and Office 365), Alphabet's absolute dominance in search and its AI R&D accumulation make its valuation more attractive. After adjusting for cash, Alphabet's "core search business" P/E is approximately 18.0x, lower than Microsoft's 24.0x, providing a margin of safety.
6. Risk Factors and Investor Misjudgment
- Root Cause of Investor Misjudgment: The market viewed the launch of ChatGPT as an "AI paradigm shift" but overlooked Google's long-term investments in AI infrastructure (e.g., TPU chips, TensorFlow framework). In February 2023, Alphabet's stock fell 12% due to ChatGPT panic, but subsequently rebounded 8% following Bard's release and the exposure of ChatGPT's errors.
- Bard's Limitations: Bard still exhibited logical errors during the testing phase (e.g., overly optimistic sports predictions), but Alphabet's cautious strategy (rather than a hasty release) may be more conducive to long-term user trust. In contrast, after Microsoft integrated ChatGPT into Bing, Bing's search share only increased by 0.1 percentage points (March 2023), failing to challenge Google's dominance.
7. Investment Recommendations and Holding Logic
- Rationale for Holdings: Alphabet is the largest holding in the Oakmark Fund because its search business has a "moat" (brand, data, economies of scale), and its AI investments are not yet fully reflected in its valuation. After adjusting for cash, Alphabet's "core business" P/E is below the industry average (S&P 500 tech sector P/E of 25.0x), providing a margin of safety.
- Risk Warning: If Alphabet's AI commercialization falls short of expectations (e.g., Bard fails to generate significant ad revenue) or regulatory pressures intensify (e.g., the EU Digital Markets Act), its valuation could face headwinds. However, the current holding weight (3.7%–11.0%) reflects the fund manager's confidence in Alphabet's long-term value.