Proposer-Builder Separation (PBS)

• PBS is an architectural paradigm in blockchain that decouples block proposing and block building, reducing centralization risks and mitigating MEV extraction.
• Mathematical and agent-based models show that while PBS equalizes proposer rewards through competitive builder auctions, builder centralization persists due to skill and latency advantages.
• Emerging solutions like on-chain auctions, MEV burn mechanisms, and protected order flow protocols are proposed to improve network security, reduce censorship, and enhance reward fairness.

Proposer-Builder Separation (PBS) is an architectural paradigm in blockchain (most prominently, Ethereum) that decouples the roles of block proposing and block building. Instead of validators (proposers) assembling and proposing blocks themselves, specialized builders compete to construct blocks, submitting bids to proposers in an auction mechanism. PBS aims to mitigate centralization risks, address MEV (Maximal Extractable Value) extraction, and promote decentralization and reward equity at the consensus layer by outsourcing block construction to a competitive builder market.

1. Centralization Dynamics Without and With PBS

In traditional block production, block producers with superior extraction skills (algorithmic, informational, or access-based) accrue higher rewards and can rapidly concentrate stake. The equilibrium in such a system is mathematically characterized by:

$$\max \sum_{i} \mu_i \frac{r}{c}\, x_i \left(1 - \frac{x_i}{2}\right) \quad \text{subject to} \quad \sum_i x_i = 1,\quad x_i \ge 0,$$

where $x_i$ is the stake share, $\mu_i$ the skill multiplier, $r$ the base reward, and $c$ the staking cost. Even modest $\mu_i$ heterogeneity leads to high centralization—e.g., one skilled block producer can hold up to ~67% of stake under realistic parameterizations (Bahrani et al., 22 Jan 2024).

PBS intervenes by separating the propositional privilege (validator/consensus role) from the construction/ordering privilege (builder/auction role). Under ideal PBS, the proposer’s reward is

$R_i = \max\{b^*, r_i\},$

with $b^*$ the highest builder bid and $r_i$ any private block built by proposer $i$ (typically suboptimal under FOSD assumptions). The key insight is that as the builder market grows, reward variance among proposers shrinks as

$$\frac{R_i}{R_j} \le 1 + O\left(\frac{1}{\log k}\right),$$

for all $i, j$, where $k$ is the number of competing builders (Bahrani et al., 22 Jan 2024). Thus, economic forces driving stake toward skilled proposers are neutralized, preserving consensus-layer decentralization.

2. Game-Theoretic and Stochastic Models of PBS

Mathematical and agent-based models provide a rigorous understanding of PBS outcomes.

Builder Centralization: PBS auctions among heterogeneous builders lead to centralization on the builder side. Order Flow Auctions (OFAs), wherein users’ transactions are sold to builders via competitive bidding, exhibit Nash equilibria where the most skilled builder bids less per unit MEV and captures disproportionate rewards, with equilibrium bids derived by solving polynomial equations (up to quartic in two-builder games) (Ma et al., 17 Feb 2025). These models show that builder centralization is an inevitable outcome when there are persistent skill or access differentials.

Proposer Decentralization: Proposer stakes evolve as a martingale (Ma et al., 17 Feb 2025):

$$\mathbf{E}\bigl[\omega_{j,t+1}\mid \mathcal{F}_t\bigr] = \omega_{j,t}$$

for stake share $\omega_{j,t}$, implying long-run preservation of stake distribution and hence decentralization in proposer space, even as builders centralize.

Role-Selection Equilibria: Co-evolutionary agent-based models for role selection (searcher vs. builder) show dynamic equilibria depend non-monotonically on bundle conflict probability. With low conflict, equilibria are searcher-dominated; with higher conflict, agents rationally shift toward building (Li et al., 19 Mar 2025). Adaptive strategies emerge through reinforcement learning and genetic algorithms, with system equilibrium frequencies computed via α‑Rank analysis.

3. PBS Auctions, Private Order Flow, and Monopoly Risks

In MEV-Boost and MEV auctions, block builders submit bids to proposers, who select the block yielding maximal revenue. When all information is public, builder bids approach proposers’ full MEV value, capturing negligible margin. However, private order flows dramatically disrupt these equilibria:

• Builders with privileged access to private flows (searcher-preferred channels, private RPCs) have idiosyncratic block value distributions $v_t^k$, leading to asymmetric Nash bidding strategies (Wang et al., 16 Oct 2024).
• The equilibrium strategy for a “strong” builder is to bid a lower fraction of its block value (lower bid/value ratio), capturing more profit and increasing dominance.
• Empirical analysis finds that although private order flows constitute only ~12% of transactions, they account for >54% of block rewards. The top builders capture 95%+ of winning auctions and retain >27% profit margins. A positive feedback loop emerges in which private order flow attracts more wins, which further increases future private order flow, leading to almost sure monopoly in the builder market (Wang et al., 16 Oct 2024). Robust process models confirm that the fraction of private order flow converges to either 0 or 1 for each builder in the long run.

PBS thus shifts centralization pressure from proposers (validators/stakers) to builders, particularly to those with preferred access and low-latency infrastructure (Gupta et al., 2023, Wu et al., 24 Dec 2024).

4. Implications for Network Security, Censorship, and User Welfare

Receiver centralization at the builder layer raises several systemic concerns:

• Censorship: Empirical studies show that after PBS and MEV-Boost on Ethereum, nearly 46% of blocks are produced by actors explicitly enforcing OFAC-based or other censorship policies (Wahrstätter et al., 2023). Average inclusion delays for censored transactions nearly doubled (from 15.8s to 29.3s), increasing vulnerability to sandwich attacks and double-spend risks.
• Trust and Relay Dependence: Relays—critical intermediaries—may not honor full value or censorship pledges, with empirical gaps in value delivery and inadvertent inclusion of “censored” transactions (Heimbach et al., 2023).
• Order Flow Auctions and User Rebates: OFAs are increasingly employed, returning a fraction $\mu h_i$ of builder bids directly to users and the remainder to validators $(1-\mu) h_i$ (Ma et al., 17 Feb 2025, Janicot et al., 26 May 2025). While this recycles MEV to originators, builder centralization persists, and inclusion quality becomes contingent on the chosen OFA, its fee structure, and competitiveness.
• Latency Games and Liveness: Validators and relays strategically introduce artificial latency to capture higher MEV bids, with annualized APR increases of ~1.58% achievable for optimally-configured operators (Natale et al., 2023). However, this raises network inefficiency, gas burn, and further centralization risk by penalizing less sophisticated validators.

5. Future Directions: Enshrinement, Mitigation, and Open Issues

Enshrined PBS (ePBS): There is a strong movement toward bringing PBS into the protocol layer (see EIP-7732), replacing off-chain MEV-Boost relays with on-chain, trustless auctions (Koegler, 22 Jun 2025). Features under consideration include:

• On-chain Builder Auctions with public bids attached to block headers, eliminating external relay trust assumptions.
• Builder Staking and Accountability, penalizing non-delivery or slow payloads.
• Payload Timeliness Committees to enhance liveness.
• Native MEV Burn/Burn Auctions: Demand builders burn a fraction of their bid, reducing MEV windfalls and instilling deflationary pressure.
• MEV Smoothing: Pooling MEV across validator committees to reduce variance and reward inequity.

The Free Option Risk: ePBS introduces a new liveness fragility; builders withhold payloads in high-volatility periods, effectively exercising a risk-free short-dated option (Mazorra et al., 29 Sep 2025). Theoretical models and empirical block data demonstrate that the value and exercise probability of this option increase with market volatility, option window length, and reliance on external signals (e.g., CEX–DEX arbitrage). On volatile days, up to 6% of blocks may be rendered empty. Mitigation strategies include shortening the option window or explicitly penalizing option exercise, with dynamic penalty tuning via online algorithms to target acceptable exercise rates.

Protected Order Flow Protocols (PROF, PROF-Share): Mechanisms like PROF enforce transaction order protection within builder proposals. PROF guarantees that user-specified transaction order is respected, with negligible latency penalty (~6ms baseline), and maintains profit-maximizing incentives for proposers (Babel et al., 5 Aug 2024).

6. Economic and Market Structure Analysis

• Execution Tickets: PBS enables non-builder entities—especially large investors with low capital costs—to dominate block proposal rights via Execution Ticket (ET) markets. Although PBS allows even unsophisticated ET buyers to sell the ticket to a competitive builder market, dominance cascades to entities with superior risk appetite and funding, re-introducing centralization risk at new loci (ticket ownership) (Burian et al., 21 Aug 2024).
• Market Efficiency and Oligopoly: Game-theoretic analyses in the presence of network latency and private order flow access confirm ongoing oligopoly: only low-latency or privileged-access builders can sustain aggressive but lower-bid strategies, depressing proposer MEV revenue and reducing auction efficiency (Wu et al., 24 Dec 2024).
• Role Selection and Co-Evolution: Market equilibrium between block building and bundle sharing (searchers) is sensitive to payoffs shaped by bundle conflict probability. A surge in agent role-switching occurs as payoffs evolve, with empirical role frequency analysis (α‑Rank) providing predictions for market configuration under varying conditions (Li et al., 19 Mar 2025).

7. Summary Table: PBS Effects by Stakeholder

Stakeholder	Effect of PBS	Comments
Proposers	Homogenized rewards; decentralization preserved	Provided builder market is highly competitive
Builders	Rewards/market share concentrate with skill/access	Private order flow and low latency amplify dominance
Users	Inclusion timing, MEV rebates depend on OFA design	Vulnerable unless protected protocols like PROF used
Validators	Lower centralization but subject to relay/builder	Trust assumptions persist unless enshrined

PBS, as deployed and envisioned for protocol-level enshrinement, addresses validator centralization and, with sufficient competitive builder diversity, secures reward equity for proposers. However, market realities—private order flow, asymmetric infrastructure, and OFA economics—enable persistent builder centralization, which, unless mitigated via robust protocol design, risks recasting old oligopolies in new technical roles. Emerging research proposes technical, economic, and cryptoeconomic responses—on-chain auctions, MEV burn, penalty mechanisms, and protected order flow—to contain unintended consequences and maintain the decentralization mandate at all layers (Bahrani et al., 22 Jan 2024, Wang et al., 16 Oct 2024, Koegler, 22 Jun 2025, Mazorra et al., 29 Sep 2025).