TL;DR
The paper introduces the Multi-BMBY Mechanism (MBM), a game-theoretic framework for restructuring ownership in private companies with many shareholders. It extends the classic two-party "Buy Me, Buy You" (BMBY) logic to multi-owner settings, ensuring that shareholders bid their true valuations while preserving the relative power dynamics among those who remain.
What Problem Does This Paper Address?
Privately held startups often end up with "messy" cap tables—a diverse mix of early investors, VCs, angels, and former employees. As a company matures or prepares for an exit, it often needs to recapitalize: reducing the number of owners to simplify decision-making.
However, doing this fairly is incredibly difficult because:
- No Market Price: Unlike public stocks, there is no objective ticker price to tell you what a share is worth.
- Information Asymmetry: Different owners value the company differently based on their private information or strategic interests.
- Strategic Gaming: Traditional negotiations are prone to manipulation, where owners might lie about their valuations to get a better deal.
- Power Dynamics: Simply selling to the highest bidder might destroy the proportional voting power that remaining owners wish to maintain.
The authors aim to create a mechanism that is strategyproof (no one benefits from lying), budget-balanced (no outside cash is needed), and proportionality-preserving (the relative ownership ratios of the remaining partners stay the same).
Methods
The authors propose the Multi-BMBY Mechanism (MBM). It works as follows:
- Bidding: Every shareholder $i$ submits a private bid $b_i$ representing what they think the entire company is worth.
- Ranking: The planner sets a target number of owners, $\bar{m}$. The agents are ranked by their bids.
- Randomization (The Secret Sauce): To ensure the mechanism is strategyproof, the final number of owners $m$ is randomly chosen to be either $\bar{m}$ or $\bar{m}-1$. The probabilities are carefully calculated based on the initial shares of the top bidders.
- Pricing: The price per share is set based on the bid of the $\bar{m}$-th ranked agent (the "threshold agent").
- Allocation: The top $m$ bidders remain owners. Their shares are increased proportionally to their original holdings. The bottom $n-m$ bidders are bought out at the calculated price.
This approach draws inspiration from McAfee’s double auction and trade-sacrifice mechanisms, but adapts them to a setting where every participant is simultaneously a potential buyer and a potential seller.
Datasets
Not discussed in the paper. As a theoretical contribution to mechanism design and game theory, the paper relies on mathematical proofs rather than empirical datasets. The authors do provide a welfare analysis using a theoretical model of uniform valuations and equal initial shares to benchmark the mechanism's performance.
Key Insights and Ideas
- Randomization for Truthfulness: The most striking insight is that by making the status of the "threshold agent" (the one on the bubble of staying or leaving) uncertain, the mechanism removes the incentive to shade bids. The threshold agent’s expected utility is zero, which anchors the entire system into a strategyproof equilibrium.
- Proportionality as a Constraint: Unlike standard auctions that just want the highest bidder to take everything, MBM recognizes that in a corporate setting, the ratio of ownership between remaining partners is often a legal or strategic requirement.
- Efficiency vs. Stability: The paper acknowledges the "Myerson-Satterthwaite" impossibility—you can't have perfect efficiency, budget balance, and individual rationality all at once. MBM chooses to prioritize budget balance and strategyproofness, accepting a "second-best" efficiency that is still a significant improvement over the status quo.
- Weak Group Strategyproofness: The mechanism is resistant to collusion. A group of shareholders cannot coordinate their bids to all strictly benefit at the same time.
Key Results
- Theoretical Guarantees: The authors prove that MBM is strategyproof, weakly group strategyproof, budget-balanced, individually rational, and proportionality-preserving ex-post efficient.
- Welfare Improvement: The mechanism is guaranteed to result in an expected social welfare that is at least as high as the initial allocation.
- The 50% Bound: In a scenario with many agents, equal initial shares, and uniform valuations, the MBM is shown to preserve at least 50% of the social welfare compared to the "first-best" (but unattainable) scenario where the highest valuer takes everything. As the target number of owners $\bar{m}$ decreases, the welfare increases.
Limitations and Future Work
- Linear Utilities: The current model assumes agents have linear utility (they value every percentage of the company the same way). In reality, owners might have non-linear preferences (e.g., valuing the jump from 49% to 51% much more due to control).
- Share Caps: The mechanism doesn't currently account for constraints like "no single owner can hold more than 50%."
- Fundraising: The authors suggest that future versions could exploit the "price gap" between what buyers are willing to pay and what sellers are willing to accept to help the firm raise new capital during the restructuring.
- Complexity of Implementation: While mathematically elegant, explaining a randomized threshold mechanism to a board of directors might be a hurdle for real-world adoption.
Summary updated