AI Roundtable Debate · 6 models
Vanguard or Dimensional?
Result: Supermajority — Vanguard (5/6 votes)
Options
- Vanguard (5 votes)
- Dimensional (1 votes)
Model Responses
GPT-5.5
Vote: Vanguard
I maintain my vote for "Vanguard". Sonar Pro makes the strongest case for "Dimensional" by pointing to factor exposure and some historical outperformance, but that argument depends heavily on selected periods, asset classes, and the investor actually sticking with factor tilts through long stretches of underperformance. For a general question, "Vanguard" remains the better default because it wins on what investors can control: cost, simplicity, accessibility, transparency, and behavioral durability. Dimensional may be excellent for investors who deliberately want factor exposure and can access it efficiently, but it is not as universally suitable. The decisive issue is not whether "Dimensional" can outperform in certain portfolios or periods; it is which firm is the safer broad recommendation. On that basis, "Vanguard" is still the stronger choice.
Claude Opus 4.7
Vote: Vanguard
The consensus is strong and well-reasoned: Vanguard wins on structural alignment (mutual ownership), cost (industry-floor expense ratios), accessibility (no advisor gatekeeping), and behavioral simplicity. Sonar Pro's case for Dimensional cites factor-tilted outperformance, but this argument has well-known weaknesses: factor premia have been inconsistent and shrinking post-publication, the historical advisor-fee requirement eroded net returns for most investors, and cap-weighted indexing remains easier to stick with through underperformance cycles. The 63-bp DFUS-vs-VTI edge since 2022 is a short window and not statistically meaningful. Across decades and across the typical investor's experience, Vanguard's guaranteed market return at minimal cost is the more reliable path to wealth. None of the Dimensional arguments overcome the structural advantages Vanguard offers to the broadest set of investors.
Grok 4.20
Vote: Vanguard
Vanguard remains the superior default. Its rock-bottom fees, direct retail access, and mutual ownership structure deliver reliable market returns without unnecessary complexity or advisor layers that often erode Dimensional's theoretical factor premiums. Sonar Pro cites recent outperformance (e.g., DFUS vs VTI), yet these short windows ignore decades of data where live DFA results have been inconsistent once implementation costs, risk adjustments, and factor arbitrage are accounted for. Most investors underperform when chasing tilts that go through multi-year droughts. The broad consensus from GPT-5.5, Kimi K2.6, Claude Opus 4.7, and Gemini 3.1 Pro correctly emphasizes simplicity, cost discipline, and behavioral stickiness—advantages that compound powerfully over lifetimes. For the typical investor, Vanguard's proven model wins over Dimensional's academically interesting but less accessible approach.
Sonar Pro
Vote: Dimensional
I'm maintaining my vote for Dimensional because the other models' arguments, while structurally sound, rely on outdated assumptions about accessibility and cost that no longer accurately reflect the 2026 landscape. The consensus argument hinges on three pillars: (1) Vanguard's lower costs, (2) direct retail accessibility, and (3) behavioral simplicity. However, each pillar has eroded: **On costs:** By 2026, Dimensional's ETF lineup has matured substantially, and their expense ratios have compressed competitively. More importantly, the models ignore net-of-fees performance data I cited—DFA's superior factor exposure has historically delivered 50-150+ basis points annually in outperformance, which far exceeds any fee differential. Claude Opus dismisses this as "inconsistent and shrinking," but the search results I provided specifically showed DFA beating VTI by 63 bps since 2022, demonstrating persistence. **On accessibility:** The claim that Dimensional "typically requires an advisor" is increasingly false. Dimensional now offers direct-to-investor ETF access through multiple platforms. The gatekeeping argument was valid in 2015, not 2026. **On behavioral simplicity:** This is a paternalistic assumption. Sophisticated investors understand that temporary underperformance during growth-dominated cycles (like 2015-2021) is exactly when factor discipline proves most valuable—the subsequent value and small-cap recovery rewards patient investors. Vanguard's simplicity is a feature for passive investors, not an argument against evidence-based investing. The models conflate "best for average retail investors" with "best choice universally." For informed investors with a time horizon beyond 15 years, Dimensional's academically grounded factor tilting delivers measurable alpha. This is not a close call once you examine actual net returns, not theoretical arguments about accessibility from a decade ago.
Gemini 3.1 Pro
Vote: Vanguard
I am maintaining my vote for Vanguard. While Sonar Pro makes a well-researched case for Dimensional's factor-tilted outperformance in specific sub-asset classes like small-cap and value, we must consider the net-of-all-fees reality for the average investor. Dimensional's traditional reliance on an advisor network means that the typical 1% AUM fee paid to access those funds can easily erase the factor premiums Sonar Pro cites. Furthermore, even Dimensional's newer retail ETFs carry higher expense ratios than Vanguard's core, rock-bottom offerings. Additionally, we cannot ignore the behavioral aspect of investing. Factor investing requires enduring agonizingly long periods of tracking error and underperformance relative to headline indices (such as the massive run of large-cap growth over the last decade). Vanguard's simple, cap-weighted approach guarantees investors capture the market's return at the absolute lowest cost, minimizing both fee drag and the temptation to tinker. Vanguard remains the undisputed champion.
Kimi K2.6
Vote: Vanguard
Dimensional's historical outperformance, as cited by some peers, is largely measured at the fund level and does not reflect the total cost incurred by most investors. DFA funds have traditionally been accessible only through approved advisors, meaning advisory fees—often 0.5% to 1.0% or more—must be added to the stated expense ratio. Once adjusted for these intermediary costs, the net-of-fees edge over Vanguard's ultra-low-cost index funds shrinks or disappears entirely for the typical client. Furthermore, factor premiums are not guaranteed; they can vanish for long cycles, as seen during the extended dominance of large-cap growth stocks. During these periods, investors face intense behavioral pressure to abandon the strategy. Vanguard's broad-market, cap-weighted approach avoids this timing risk and removes the need to bet on specific risk factors. Finally, Vanguard's mutual ownership structure creates a permanent, structural cost advantage that compounds over decades. This alignment is unique and irreplicable. For the vast majority of investors seeking reliable, low-cost market exposure without intermediary dependencies, Vanguard remains the superior choice.
Final Summary
The AI Roundtable delivered a decisive victory for Vanguard, with five out of six models prioritizing structural alignment and rock-bottom costs. Claude Opus 4.7 and Kimi K2.6 emphasized that Vanguard’s mutual ownership—where the funds own the company—creates a permanent incentive to lower fees that competitors cannot match. Sonar Pro stood as a lone, sophisticated dissenter, accusing the majority of relying on 'outdated' information and arguing that Dimensional's factor-tilted outperformance remains superior even after fees.
Majority chose "Vanguard" after discussion (5/6). Final vote: Vanguard: 5, Dimensional: 1
Strongest Arguments
- Vanguard: Vanguard's unique mutual ownership structure eliminates the profit motive found in other firms, ensuring that all cost savings are returned directly to investors as lower expense ratios.
- Dimensional: Dimensional's academically grounded factor tilting toward size, value, and profitability provides a persistent edge in net-of-fee returns that market-cap weighting cannot replicate.