{"id":"93521292-5c9a-4c01-a007-0478f429c2a3","arxiv_id":"2505.24159","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A Lagrangian-dual settlement with a security charge achieves revenue adequacy and neutrality for energy, reserves, and transmission prices in LP market-clearing models.","lead":"This paper introduces a security charge for contingency-constrained electricity markets that makes generator revenues exactly match consumer payments while allocating reserve costs to the generators whose outages create them. It also derives separate prices for up and down reserves and a congestion payment for transmission lines from the same dual framework.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Security charges depend on which optimal dual solution is selected; with dual degeneracy, the causation-based allocation is non-unique and the central cost-causation claim is underdetermined.","rationale":"The reader's weakest assumption was the LP/nonconvexity scope, a genuine boundary on the strong-duality identity. My concern is complementary and, I think, more directly aimed at the paper's headline 'causation-based' claim: even inside the LP scope, the dual solution used for settlement is not proved unique. Since the paper's examples are solved with CPLEX and only one optimal dual is reported, the security charges and prices could be artifacts of solver basis selection. Revenue adequacy and neutrality are robust to dual non-uniqueness, so the mathematical core of the paper stands. However, the cost-allocation story—the paper's main novelty—is underdetermined unless a selection criterion or uniqueness condition is supplied. This reinforces the reader's conditional verdict rather than changing it: the framework is mathematically coherent as a construction, but its interpretation as a unique causation-based allocation needs additional support. I set verdict_should_be to UNCHANGED because the conditional verdict already captures the need for such caveats, and my concern strengthens that conclusion without overturning it.","tokens_in":22369,"tokens_out":11399,"duration_ms":149193,"concrete_test":"Construct a degenerate variant of the two-bus example, e.g., add a generator at bus 1 with the same offer and capacity as generator 1 so the primal has alternate optimal bases. Solve (15)-(31) with two different LP solvers or basis orientations and compare security charges and transmission revenue p^f_l F_l. If the charges differ across optimal dual solutions, the causation attribution is non-unique; if they coincide, the concern is settled. An analytical check is also possible: derive the KKT conditions and verify whether the optimal dual solution is a singleton.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central cost-causation claim presupposes that the Lagrange multipliers defining security charges and prices are well-defined attributes of the market outcome. In the LP models (1)-(6) and (15)-(31), the optimal dual solution is not proved unique, and in general it is not. Whenever the dual optimal face has multiple extreme points (e.g., due to primal degeneracy or redundant binding constraints), different optimal dual solutions produce different security charges C^g_i, different up-/down-reserve prices, and different transmission revenues, while preserving revenue adequacy and neutrality. The simplex solutions reported in Sections 2 and 3 represent one arbitrary basis selection. The paper offers no selection rule or uniqueness argument, so the 'causation-based' attribution is underdetermined: the same physical dispatch can support multiple settlements, each defensible as dual-based. This does not invalidate Theorems 1 and 2, which hold for any optimal dual, but it undermines the claim that charges 'reflect the different contributions of generators' and the attendant reliability-incentive narrative.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proposes a pricing and cost-allocation framework for contingency-constrained, co-optimized energy and reserve markets. It introduces a security charge that is subtracted from each generator's revenue, defined from the Lagrangian dual of an LP market-clearing model, and it derives separate nodal prices for up-spinning reserves, down-spinning reserves, and transmission services. In a single-bus model and a two-bus network model, the paper compares its settlement with that of Arroyo and Galiana (2005), shows that the proposed scheme removes the revenue imbalance, and states two theorems: revenue adequacy (nonnegative profits for generators and consumers) and revenue neutrality (consumer payments equal generation plus transmission revenue). Proofs are given in Appendices B-E, and two numerical examples illustrate exact settlement balance.","tokens_in":22507,"tokens_out":17801,"duration_ms":168814,"significance":"If the results stand, the framework offers a clean, dual-based method to repair the revenue mismatch of a well-known pricing scheme while preserving nonnegative profits. The paper's strengths are its explicit, checkable price definitions, its complete proofs of revenue adequacy and neutrality for the stated LP models, and its exact numerical reconciliation in both examples. The separation of up- and down-reserve prices and the explicit transmission remuneration are useful contributions. However, the central 'causation-based' interpretation is not fully supported unless the optimal dual solution is unique or a selection rule is provided; the same physical dispatch can in principle support multiple dual-based settlements. The paper is also silent on the scope limitation that the guarantees rely on strong duality and therefore do not extend to nonconvex market-clearing models.","major_comments":[{"comment":"The security charge C_i^g, the up/down reserve prices, and the transmission price are all defined as functions of one optimal dual solution Pi*. The paper does not prove that the optimal dual solution of (1)-(6) or (15)-(31) is unique, and in LP degeneracy it generally is not. Different optimal dual extreme points can yield different security charges and different reserve/transmission prices while leaving the primal dispatch and the total settlement balance unchanged. The numerical results were obtained with CPLEX's simplex algorithm, which selects one basis without economic justification. Because the paper's central claim is that these charges 'reflect the different contributions of generators' and provide reliability incentives, the absence of a uniqueness result or a selection rule leaves the cost-causation attribution underdetermined. Theorems 1 and 2 remain correct for any optimal dual, but the causation narrative needs either a uniqueness argument, an explicit selection rule, or a reformulation of the result as a family of admissible revenue-neutral settlements.","section":"Sections 2.3 and 3.3 (Eq. (46)-(47), Tables 22-27)"},{"comment":"The abstract and conclusion state revenue neutrality and adequacy without qualification, but the proofs in Appendices B-E rely essentially on strong duality for the LP models (1)-(6) and (15)-(31). If the market-clearing model includes binary unit-commitment variables, startup costs, or other nonconvexities, the strong-duality identity can fail, and the neutrality and adequacy guarantees may fail as well. The paper does not mention this limitation in the conclusion or in the discussion of future work. Please add an explicit scope statement clarifying that the proposed settlement is guaranteed only for the continuous linear-programming formulation, and temper the unqualified 'no missing money' claim in the abstract.","section":"Abstract and Appendix E (Eq. (E.1))"}],"minor_comments":[{"comment":"The sign convention for the line-flow multipliers is nonstandard: pi^{f-} is described as nonpositive even though it multiplies (F - H\\theta). Since pi^{f-} is apparently the negative of the usual upper-bound multiplier, please state this explicitly to avoid confusion.","section":"Section 3.3, Eq. (32)-(34)"},{"comment":"The term 'missing money' is used to describe a situation where generation revenue exceeds consumer payments by $5,200. That is a revenue surplus, not the usual meaning of 'missing money' (insufficient revenue). Please use a different term, such as 'settlement surplus' or 'revenue imbalance'.","section":"Section 2.2, Table 7"},{"comment":"The future-work paragraph lists multi-period settings and renewable uncertainty but does not mention nonconvex market clearing (binary unit commitment, startup costs). Adding that direction would be a natural and relevant extension, given the LP-only scope of the current proofs.","section":"Conclusion, Section 4"}],"recommendation":"major_revision","confidential_remarks":"The paper is a good fit for EJOR and the core LP-based revenue-neutrality construction is sound. The main obstacle is the dual-degeneracy issue: because the security charge and prices depend on an arbitrary optimal dual solution, the 'causation-based' contribution is underdetermined unless a selection rule is supplied. I also recommend that the authors add an explicit nonconvexity limitation statement, since the unqualified abstract claim could mislead readers. With these revisions, the paper would be a solid contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know up front. The actual contribution is a settlement rule — a security charge, separated up- and down-reserve prices, and a transmission rent term — all derived from the Lagrangian dual of a contingency-constrained LP. And the two headline theorems, revenue adequacy and neutrality, are correct but close to strong duality restated as a bookkeeping identity. They validate the construction; the interesting load is carried by the causal story, which is argued by example rather than proven.\n\nIt does several things well. The security charge is clean: generator i pays the dual price of the contingency in which it is out, applied to its pre-contingency output plus reserves. In the single-bus example it deletes the $5,200 missing-money gap in the Arroyo-Galiana settlement while leaving every profit non-negative. Separating the contingency multipliers into positive and negative parts to price up- and down-reserve distinctly is a nice touch, and the transmission term fixes a real gap in the 2005 scheme. The two cases reconcile numerically, and the appendix proofs are correct within the declared LP scope.\n\nThe soft spots, in proportion. First, the abstract and conclusion claim efficient operations and reliability incentives. Nothing in the paper derives those; what is proven is non-negative profit and an accounting identity. Second, the dual-degeneracy concern lands: when the dual optimal face has more than one point, different optimal multipliers give different security charges, all revenue-neutral. The paper gives no selection rule, so the causal attribution is underdetermined in exactly those cases. I would not call this fatal — the same is true of LMP schemes generally — but the cost-causation language should be softened or a tie-break rule provided. Third, the LP assumption: add unit-commitment binaries and strong duality fails, taking neutrality with it. The paper is honest that the models are linear, but the practical claims outrun that scope. Fourth, the Shapley/nucleolus literature is cited as motivation and never compared against. The paper flags multi-period and renewable extensions as future work, but it never notes the nonconvexity or degeneracy issues, which are the ones a referee would want addressed.\n\nThis is for market-design people working on joint energy-reserve clearing who want a concrete, provably revenue-neutral alternative to ad hoc uplift. It deserves a serious referee. I would send it out, with a request for a dual-selection rule or a scaling-back of the causal claims.","headline":"A clean and correct dual-based extension of the Arroyo-Galiana pricing scheme; the security charge is new and the accounting is right, but the causal story needs a dual-selection rule and the efficiency claims outrun the proofs.","tokens_in":23082,"tokens_out":7453,"would_cite":true,"duration_ms":84683,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["90C05","90C46","91B24"],"pacs":[],"model":"deepseek-v4-flash","headline":"A security charge levied on the generators whose outages drive reserve needs makes consumer payments exactly match generation and transmission revenues in contingency-constrained electricity markets.","keywords":["causation-based pricing","security charges","operating reserves","contingency-constrained scheduling","revenue adequacy","revenue neutrality","transmission pricing","Lagrangian duality"],"falsifier":"Take the paper's two-bus example and tighten the line so the line-outage contingency also binds, making a second flow multiplier nonzero at the same time as the generator-1 outage, then recompute the proposed settlement and check whether consumer payments still equal generation plus transmission revenue; alternatively, add a single binary unit-commitment variable with a startup cost and observe whether the balance fails, which would show the LP strong-duality premise is doing the work.","tokens_in":22117,"feed_emoji":"⚡","tokens_out":12782,"duration_ms":108296,"temperature":0.7,"pith_summary":"This paper proposes a pricing scheme for electricity markets that co-optimize energy, operating reserves, and transmission under contingency constraints. The central claim is that the Lagrange multipliers of the post-contingency power-balance constraints can be re-attributed so that each generator pays a security charge for the contingency states in which that generator's own outage creates the need for reserves; the generator whose possible failure actually binds the system pays, and the others pay nothing. With that charge in place, the settlement becomes revenue-adequate (no participant is forced to a loss) and revenue-neutral (total consumer payment equals total generation revenue, or generation plus transmission revenue in the network model), eliminating the missing-money imbalance shown by the earlier Arroyo and Galiana pricing. The scheme also prices up- and down-spinning reserves separately and pays transmission lines for the capacity dedicated to delivering reserves in every contingency state. If correct, the framework gives markets a way to allocate reliability costs to the agents that cause them without ad hoc out-of-market compensation.","feed_headline":"New security charge removes missing money in power markets","feed_subtitle":"Generators whose outages force reserve purchases pay for that risk; market accounts balance exactly.","key_machinery":"The load-bearing object is the Lagrangian dual function of the clearing LP. The authors build the dual by relaxing the power-balance constraints (and, in the network model, the line-flow limits) and then read each term as a settlement account: consumer payment, generator profit, consumer profit, and, for the network model, transmission congestion rent. The decisive algebraic step is rewriting the contingency-state revenue $\\psi^g_{ik}(\\pi_k, x_i) = \\pi_k a^g_{ik}(g_{i0} + r_i^{up})$ as $\\pi_k(g_{i0} + r_i^{up})$ over all states minus $\\pi_k(g_{i0} + r_i^{up})$ over the states $k \\in K^{OFF}_i$ in which the generator is unavailable; that subtracted term is the security charge. Strong duality then turns the equality between primal cost and dual value into the revenue-neutrality identity, and the fact that each agent's term in the dual is a linear program over non-negative variables gives non-negative profits. Splitting each contingency price as $\\pi_{bk} = \\pi^+_{bk} - \\pi^-_{bk}$ yields separate up- and down-reserve prices, and the absolute values of the line-flow bound multipliers form the transmission price.","core_discovery":"The paper's central claim is that the dual of a contingency-constrained energy-and-reserve market-clearing LP contains a complete, self-balancing settlement, provided the revenue that each generator would collect in contingency states is first reduced by a security charge for the states in which that generator is out of service. In the single-bus model the charge is $C_i^s = \\sum_{k \\in K^{OFF}_i} \\pi_k^*(g_{i0} + r_i^{up})$, the product of the contingency power-balance prices and the generator's pre-contingency output plus up-reserve. Because $\\pi_k^*$ is positive only for the contingencies that actually constrain the system, the charge falls on the generator whose potential outage drives the reserve requirement, exactly implementing cost causation. The same rearrangement applied to the network model splits the contingency prices into positive and negative parts to price up- and down-spinning reserves separately, and adds a transmission price $p_l^f = |\\pi^{f*}_{l0}| + \\sum_k |\\pi^{f*}_{lk}|$ per MW of line capacity. Theorems 1 and 2, with network extensions in Appendices D and E, prove via strong duality that generator and consumer profits stay non-negative and that consumer payments exactly equal generation revenue, or generation plus transmission revenue in the network case. The two worked examples show the previously reported scheme leaving shortfalls of $5,200 and $6,900 that the new settlement closes.","pith_inferences":["Because the neutrality identity is a bookkeeping consequence of strong duality, it should hold in every LP instance of the paper's class; a cheap stress test is to generate many random cost, capacity, and contingency instances and verify that consumer payment minus generation and transmission revenue is identically zero in each.","The same dual-rearrangement logic should carry over to probabilistic security criteria, where each contingency's price would be weighted by its probability when computing the security charge, yielding causation-based charges under stochastic reserve requirements.","The paper notes that over-frequency events are the demand-side mirror of generator outages; a symmetric construction would levy security charges on loads whose sudden disconnection forces downward-reserve purchases, completing the causation story on both sides of the market.","If implemented, the security charge would give the paying generator a direct financial incentive to lower its forced-outage rate, turning what looks like a penalty into a reliability investment signal, a behavior the paper motivates but does not model."],"forward_implications":["In the single-bus model the proposed settlement makes total generator revenue exactly equal the consumer payment, closing the $5,200 gap left by the Arroyo-Galiana pricing on the paper's example.","In the network model consumer payments exactly match the sum of generation revenue and transmission revenue, closing the $6,900 gap and removing the need for out-of-market compensation.","Generators whose outages bind the contingency constraints pay security charges while others pay nothing, so reserve costs follow the cost-causation principle while all profits remain non-negative.","Up- and down-spinning reserves receive distinct nodal prices, so a unit that deploys downward reserve in a contingency with a negative price is paid according to that separate price rather than a blended security price.","Transmission lines are remunerated for the capacity they dedicate to reserve deliverability across all contingency states, giving line owners an economic incentive tied to reserve provision."],"supporting_citations":[{"why":"Supplies the energy and security price definitions this paper keeps, and the missing-money imbalance the security charge is designed to remove.","marker":"(Arroyo and Galiana, 2005)"},{"why":"Provides the Lagrangian dual function construction the authors use to read the settlement off the dual.","marker":"(Conejo et al., 1999)"},{"why":"Supplies the strong duality and complementary slackness theorems on which the revenue adequacy and neutrality proofs rest.","marker":"(Bertsimas and Tsitsiklis, 1997)"},{"why":"Establishes the marginal spot-pricing viewpoint that justifies using Lagrange multipliers as prices.","marker":"(Schweppe et al., 1988)"},{"why":"Empirical study showing large generator outages dominate reserve costs, the cost-causation premise the security charge implements.","marker":"(Matamala et al., 2024)"},{"why":"Argues reserve costs should be borne by the units responsible for them, the principle this paper operationalizes as a security charge.","marker":"(Badesa et al., 2025)"},{"why":"Early proposal to allocate reserve costs by generator capacity and unavailability, the lineage the security charge continues.","marker":"(Strbac and Kirschen, 2000)"},{"why":"Standard reference for the deterministic N-1 security criterion that defines the paper's contingency set.","marker":"(Wood et al., 2014)"}],"fun_headline_variants":["Causation-based pricing ends power market shortfalls","Security charge balances contingency-constrained markets","Contingency-aware pricing closes missing money gap"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire construction assumes the market-clearing problem is a linear program with continuous variables, so the primal and dual optima coincide; the revenue adequacy and neutrality guarantees can break down once binary unit-commitment decisions, startup costs, or other nonconvexities enter the model.","fun_headline_variants_meta":{"raw":{"variants":["Causation-based pricing ends power market shortfalls","Security charge balances contingency-constrained markets","Contingency-aware pricing closes missing money gap"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000669,"raw_usage":{"total_tokens":3101,"prompt_tokens":1044,"completion_tokens":2057,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":660,"completion_tokens_details":{"reasoning_tokens":2021}},"tokens_in":660,"tokens_out":2057,"duration_ms":15921,"temperature":1.0,"reasoning_tokens":2021,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:33:04.699832+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the paper's two-bus example and tighten the line so the line-outage contingency also binds, making a second flow multiplier nonzero at the same time as the generator-1 outage, then recompute the proposed settlement and check whether consumer payments still equal generation plus transmission revenue; alternatively, add a single binary unit-commitment variable with a startup cost and observe whether the balance fails, which would show the LP strong-duality premise is doing the work.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the energy and security price definitions this paper keeps, and the missing-money imbalance the security charge is designed to remove."},{"cited_title":"J., Arroyo, J","cited_arxiv_id":null,"evidence_quote":"Provides the Lagrangian dual function construction the authors use to read the settlement off the dual."},{"cited_title":"and Tsitsiklis, J","cited_arxiv_id":null,"evidence_quote":"Supplies the strong duality and complementary slackness theorems on which the revenue adequacy and neutrality proofs rest."},{"cited_title":"C., Caramanis, M","cited_arxiv_id":null,"evidence_quote":"Establishes the marginal spot-pricing viewpoint that justifies using Lagrange multipliers as prices."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Empirical study showing large generator outages dominate reserve costs, the cost-causation premise the security charge implements."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Argues reserve costs should be borne by the units responsible for them, the principle this paper operationalizes as a security charge."},{"cited_title":"and Kirschen, D","cited_arxiv_id":null,"evidence_quote":"Early proposal to allocate reserve costs by generator capacity and unavailability, the lineage the security charge continues."},{"cited_title":"J., Wollenberg, B","cited_arxiv_id":null,"evidence_quote":"Standard reference for the deterministic N-1 security criterion that defines the paper's contingency set."}],"review_version":1}