Two-Stage Electricity Markets with Renewable Energy Integration: Market Mechanisms and Equilibrium Analysis

Nathan Dahlin; Rahul Jain

REVIEW 2 major objections 6 minor 20 references

Two-Stage Electricity Markets with Renewable Energy Integration: Market Mechanisms and Equilibrium Analysis

T0 review · 2 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A two-stage electricity market with renewable uncertainty can clear both stages efficiently through a sequential competitive equilibrium, whose prices are the dual variables of a single stochastic dispatch program.

desk verdict The existence construction and the DED game results are solid, but Theorem 2(i) is overclaimed because Definition 1 clears markets only in aggregate, not node by node. read the letter →

arxiv 1909.00508 v2 pith:KXKURMI3 submitted 2019-09-02 cs.GT cs.SYecon.GNeess.SYq-fin.EC

Nathan Dahlin , Rahul Jain This is my paper · ORCID

classification cs.GTcs.SYecon.GNeess.SYq-fin.EC MSC 91B2691A1090C15

keywords two-stageelectricitymarketsequentialcompetitiveequilibriumrenewableenergyintegrationdemandresponseDCpowerfloweconomicdispatchNashdesign

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that a two-stage electricity market with renewable generation can be made to clear efficiently: the market operator chooses a day-ahead dispatch and a per-scenario real-time adjustment, or recourse, and the paper proves that a sequential competitive equilibrium exists under strictly convex costs. That equilibrium is constructed from the solution of a single two-stage stochastic program, with optimal dual variables serving as the stage-one and stage-two nodal prices. The paper then establishes analogues of the two welfare theorems: every sequential competitive equilibrium allocation is efficient, and every efficient allocation can be supported by such an equilibrium. It also designs a market mechanism in which participants submit linear bids, and shows that under a congestion-free or a monopoly-free condition an efficient Nash equilibrium exists. The upshot is a formal case that coupled two-stage markets, rather than independently run forward and real-time markets, could integrate high-renewable uncertainty without sacrificing economic efficiency.

What carries the argument

The load-bearing object is the sequential competitive equilibrium (Definition 1), a two-stage analogue of Walrasian equilibrium in which, given stage-one prices and stage-two price schedules, every generator and load-serving entity maximizes profit or utility in each stage and markets clear in both stages in every renewable scenario. The argument is carried by the two-stage stochastic program with recourse, (SPP-P), whose KKT conditions coincide with the individual optimality conditions of all agents once the dual variables are announced as prices; the same dual variables double as locational marginal prices in the bidding game. An equivalence result recasts the two-stage planner problem as a single convex program, and the welfare proofs use the decomposition of (SPP-P) into an ISO redispatch problem together with the individual agent problems.

What would settle it

Construct the two-bus, one-congested-line example of the paper and check whether an allocation that meets the aggregate clearing condition (39) and every agent's optimality conditions, but violates the nodal balance equation at the exporting bus, can be admitted as a sequential competitive equilibrium; if the definition admits it, the first welfare theorem is false.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is Theorem 1: under strictly convex, increasing, differentiable cost functions for generation, demand response, and blackouts, a sequential competitive equilibrium exists and is given by the primal solution to the two-stage stochastic program (SPP-P) together with optimal dual variables as prices. Theorem 2 states that every such equilibrium supports an efficient sequential allocation and, conversely, every efficient allocation can be supported by such an equilibrium. In the strategic setting, when participants submit linear bids and the ISO solves a dynamic economic dispatch problem, the efficient bid profile—each participant bidding the relevant locational marginal price—is a Nash equilibrium provided either no transmission constraint binds or each bus hosts at least two generators and two load-serving entities. Without the reformulation in terms of LSE utility functions, the paper exhibits a counterexample in which bidding marginal demand-response costs destroys the existence of an efficient Nash equilibrium.

Load-bearing premise

The central claim rests on the premise that aggregate market clearance in each stage, meaning total generation equal to total purchases summed over the whole network, guarantees that the equilibrium allocation obeys the per-bus power balance and line-capacity constraints of the DC network model; if that premise fails, efficiency of every sequential competitive equilibrium is not established.

Editorial extensions

If this is right

A two-stage market with recourse can clear both stages simultaneously from a single optimization, eliminating the need to run day-ahead and real-time settlements as independent markets.
The dual variables of the dispatch program provide the prices the ISO should announce, so efficient dispatch is implementable as a market mechanism.
With linear bids, the efficient bid profile is a Nash equilibrium under either no congestion or at least two competitors at every node, so strategic behavior need not destroy efficiency in those cases.
If load-serving entities bid marginal demand-response costs instead of electricity valuation functions, the efficient equilibrium can fail, so bid format matters as much as network topology.
Demand response and managed blackouts can be priced into the real-time market as recourse options without compromising the welfare theorems.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The existence and welfare results likely extend to more than two stages, since the argument only requires stage-by-stage convexity and the same dual-variable-as-price construction, but the paper does not prove this.
Using an AC power flow model with losses would probably break the exact welfare theorems, because aggregate clearing would no longer coincide with feasible physical flows; testing this would bound the role of the DC approximation.
One could calibrate the model to real renewable forecast-error distributions and compare expected welfare against the current multi-settlement design; the paper identifies the comparison but does not quantify it.
The congestion-free and monopoly-free conditions suggest a practical heuristic: market design should either keep transmission uncongested or ensure two or more competitors at every node.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The existence construction is probably right, and the two-stage market design is a real extension of the earlier CDC model—but Theorem 2(i) is stated too broadly. Definition 1 clears markets only in aggregate, so the first welfare theorem does not follow from the proof as written. The good news: the fix is small, and most of the paper survives.

What is genuinely new: the network-constrained SCEq existence result built from KKT dual variables of the social planner problem; the efficient linear bid profile; and the two sufficient conditions (congestion-free and monopoly-free) for an efficient Nash equilibrium. The counterexample in Example 1, showing that LSEs bidding marginal demand-response costs can profitably deviate even under the monopoly-free condition, is the most interesting piece. It justifies the reformulation in terms of LSE utility functions. Adding DC flow constraints and the DED game changes the analysis in a non-cosmetic way.

The soft spots are real but concentrated. The stress-test counterexample is valid: one generator at bus 1 with marginal cost 10, one LSE at bus 2 with demand 10, price 10 at bus 1 and 0 at bus 2, line limit 1. Both agents optimize, aggregate supply equals aggregate purchase, stage-2 quantities are zero, but 10 units would have to flow over a 1-unit line. So an SCEq under Definition 1 need not be node-feasible. The proof of Theorem 2(i) just asserts that market clearing is equivalent to the ISO problem; it isn't. Strengthening Definition 1 to require per-node balance and feasible angles within line limits fixes the statement, and the constructive equilibrium of Theorem 1 already satisfies those constraints. The existence theorem and the mechanism section survive.

Minor issues: Lemma 1 is deferred to [9] rather than proved; acceptable, but annoying in a paper that builds on it. There are also typos in the displayed welfare objective (25), the KKT labeling in the proof of Theorem 1, and a duplicate index in (115). None undermine the main construction.

Who should read this: people working on two-stage or stochastic electricity market design, and anyone who cares about the gap between aggregate market clearing and network feasibility in competitive-equilibrium definitions. It deserves a serious referee. The right recommendation to the editor is: send it, but make the authors tighten Definition 1 and re-state Theorem 2(i).

Referee Report

2 major / 6 minor

Summary. The paper proposes a two-stage electricity market with a forward stage and a real-time stage with recourse, including renewable generation, demand response, blackout costs, and DC network constraints. It defines a sequential competitive equilibrium (SCEq) with aggregate market clearing, proves existence by taking primal and dual solutions of a two-stage social planner problem, states first and second welfare theorems for SCEq, describes a market mechanism, and then studies a bidding game in which generators and LSEs submit linear bids. It claims an efficient Nash equilibrium exists under either a congestion-free or a monopoly-free condition, and provides a counterexample showing that bidding demand-response costs directly can destroy that guarantee.

Significance. The constructive existence proof is a genuine contribution: Theorem 1 uses the KKT system of the social planner's problem to produce prices at which each agent's individual problem is solved, and the inclusion of DC network constraints and blackout costs makes the model more realistic than much of the existing sequential-equilibrium literature. The paper also gives a clean decomposition argument (ISO plus agent problems) that is reusable. The significance is limited, however, by the fact that the main welfare theorem is false under the stated definition of equilibrium and by a gap in the Nash-equilibrium claims.

major comments (2)

[Section III, Definition 1 and Eq. (39); Theorem 2(i)] The SCEq market-clearing condition is aggregate over the whole network. One scalar equality per stage cannot imply the N nodal balance equations (26)-(27) or the line-flow limits (28)-(29), so the statement in the proof that aggregate clearing is 'equivalent' to the ISO problem (74)-(80) is not correct. As a concrete counterexample, take two buses with a generator at bus 1 having c1(y)=y^2/2, an LSE at bus 2 with D=10 and blackout cost cbo(z)=z^2, a single line with fmax=1, and W=0 almost surely. With P1,1=10 and P1,2=0 the generator produces 10 and the LSE buys 10, so Eq. (39) holds, but the allocation requires 10 units of flow on a 1-unit line and is not feasible for (SPP-P). Thus Theorem 2(i) is false as stated. The repair is to add to Definition 1 the requirement that there exist theta_1 and theta_2(w) such that the equilibrium quantities satisfy (26)-(29); with that strengthened definition the first welfare theorem can be proved from individual optimality plus feasibility.
[Section VI-A, Theorems 3 and 4, proof around Eqs. (126)-(130)] The Nash proofs assume that after any unilateral deviation the ISO problem (DED) still has a finite solution. This is not guaranteed under the linear-bid format. On a single bus with at least one generator and one LSE, if all equilibrium bids equal lambda*, an LSE that deviates upward to bL > lambda* makes the DED objective contain (bL - lambda*) yL with yL = yG = t; increasing t does not change the nodal balance, so the objective goes to +infinity and no optimal dispatch exists. Hence the payoff for that deviation is undefined, and the strategy space in Definition 2 is not well-defined as stated. This configuration is not excluded by either sufficient condition: a single bus with two generators and one LSE is congestion-free (and satisfies Assumption 2 if either generator alone can serve the load), and a bus with two generators and two LSEs satisfies the monopoly-free condition. The theorem needs either a bid cap or price cap, a demand bound, or an explicit treatment of unbounded DED instances, together with a tie-breaking rule because when all bids equal lambda* the DED has multiple optimal dispatches and the proof selects the efficient one.

minor comments (6)

[Section V-A, after Eq. (90)] The statement that ui,k is 'convex' should read 'concave': ui,k is the negative of a convex value function. The KKT conditions stated later are the correct ones for a concave utility, so this appears to be a typo, but it should be corrected because concavity is what makes (SPP-U) a concave maximization.
[Eq. (14)] The expression for pi_SPP(w) appears to drop the c2,k terms and contains a misplaced parenthesis; the sum should be over c1,k(yG1,k) + c2,k(yG2,k(w)), not c1,k(yG1,k + yG2,k(w)).
[Definition 1] The tuple defining an SCEq omits zL*2(·), although zL*2 is part of the LSE's decision and is used in the proof of Theorem 2; it should be included for clarity.
[Lemma 1] The proof is deferred by citing the same method as Lemma 1 in [9]; for a journal version, a self-contained proof or a precise statement of the equivalence should be included.
[Eq. (25)] There are missing parentheses in the second term of the objective; it should read sum_w c2,k(yG2,k(w)) p_w.
[Theorem 3 and Theorem 4] Assumption 2 is stated as necessary for the Nash results, but it is not used explicitly in the displayed proofs; the authors should explain where it enters.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the equilibrium and efficiency results are proved by KKT verification against explicitly defined agent problems; self-citations are to elementary reformulations and do not smuggle in the target results.

full rationale

None of the paper's derivation steps reduces to its own inputs by construction. Theorem 1 constructs prices as optimal dual variables of (SPP-P) and then proves individual optimality by matching the KKT systems of (SPP-P) with those of (GEN-P) and (LSE-P), with market clearing following from summation of the nodal balance constraints (26)-(27). This is a verification, not an assumption of the result. Theorem 2(ii) is the same constructive argument. Theorem 2(i) contains an apparent proof gap: Definition 1 imposes only aggregate market clearing in Eq. (39), while the ISO problem (74)-(80) requires nodal balance and line-flow constraints, so the sentence in the proof that aggregate clearing 'is equivalent to posing the following ISO problem' is not justified by the paper's equations. That is a correctness concern, not circularity: the SCEq is not defined as 'solves SPP-P', and no fitted parameter is renamed as a prediction. Lemma 1 is cited from the authors' prior work [9], but it is an elementary two-stage/single-stage decomposition whose stated assumptions (strict convexity, finite scenarios) do not include the equilibrium existence or efficiency conclusions; it is independent support rather than a load-bearing self-citation that assumes the target result. Similarly, the efficient bid profile in Proposition 1 is constructed from the dual solution and then shown to be a Nash equilibrium by deviation bounds; this is the normal fixed-point content of an equilibrium proof, not a circular input. There are no empirical predictions, no fitted parameters, and no uniqueness theorem imported from the authors' prior work to forbid alternatives. The only flagged issue, the Theorem 2(i) proof gap, is an omitted justification and should be weighed as a correctness risk, not as circularity.

Assumptions & free parameters 0 free parameters · 8 assumptions · 0 invented entities

No fitted parameters or invented physical entities. The paper's assumptions are structural, such as convex costs, finite scenarios, DC flow, and inelastic demand, plus design choices for the bidding game. The main unproved imported result is Lemma 1 from [9].

assumptions (8)

domain assumption Assumption 1: c1,k, c2,k, cdr,k, cbo,k are strictly convex, increasing, differentiable, and nonnegative over R+.
Used throughout for KKT sufficiency and uniqueness of SPP-P solutions; restricts admissible cost structures.
domain assumption Renewable generation W has finite support in [0,wk] with common-knowledge probability mass p.
Enables the finite-scenario stochastic program and linear bid schedules indexed by w; excludes continuous distributions and private information.
domain assumption DC power flow model with susceptance matrix B and line limits fmax, with losses ignored.
Network representation used in SPP-P and the mechanism; standard approximation, not exact AC physics.
domain assumption Aggregate demand Dk is inelastic and not price sensitive.
Used in constraint (3) and the LSE problems; no demand elasticity in the model.
standard math Lemma 1 from [9]: two-stage SPP1-SPP2 is equivalent to single-stage SPP-P.
Invoked in Lemma 1 and proved only by reference to the authors' CDC 2019 paper; needed for Theorems 1 and 2.
domain assumption Assumption 2: SPP-U remains feasible when any one generator is removed in either stage.
Precludes monopoly price effects in Theorems 3 and 4; not stated until Section VI.
domain assumption Congestion-free or monopoly-free condition holds for Nash existence.
Theorems 3 and 4 are conditional on one of these; neither is guaranteed by the market model.
ad hoc to paper Linear bid format and locational marginal pricing administered by the ISO.
The DED game restricts strategies to scalar linear bids and assumes LMP settlement; this is a design choice, not derived from the underlying market.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Two-Stage Electricity Markets with Renewable Energy Integration: Market Mechanisms and Equilibrium Analysis." pith.science (2026). https://pith.science/paper/KXKURMI3

@misc{pith2026190900508,
  author       = {Pith},
  title        = {Pith review of: Two-Stage Electricity Markets with Renewable Energy Integration: Market Mechanisms and Equilibrium Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KXKURMI3}},
  note         = {Machine review of arXiv:1909.00508}
}

read the original abstract

We consider a two-stage market mechanism for trading electricity including renewable generation as an alternative to the widely used multi-settlement market structure. The two-stage market structure allows for recourse decisions by the market operator, which are not possible in today's markets. We allow for different conventional generation cost curves in the forward and the real-time stages. We have considered costs of demand response programs and black outs, and adopt a DC power flow model to account for network constraints. Our first result is to show existence (by construction) of a sequential competitive equilibrium (SCEq) in such a two-stage market. We argue social welfare properties of such an SCEq, and then design a market mechanism that achieves social welfare maximization when the market participants are non-strategic. We also show that under either a congestion-free or a monopoly-free condition, an efficient Nash equilibrium exists.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 19 canonical work pages

[9]

A two-stage market mechanism for electricity with renewable generation,

N. Dahlin and R. Jain, “A two-stage market mechanism for electricity with renewable generation,” in 2019 IEEE 58th IEEE conference on decision and control (CDC) . IEEE, 2019

work page 2019
[1]

The use of marginal energy costs in the design of U.S. capacity markets,

R. Moye and S. P. Meyn, “The use of marginal energy costs in the design of U.S. capacity markets,” in 51st Hawaii International Conference on System Sciences, HICSS 2018, Hilton Waikoloa Village, Hawaii, USA, January 3-6, 2018 , 2018. [Online]. Available: http://aisel.aisnet.org/hicss-51/es/markets/5

work page 2018
[2]

Electricity grid operators forecast load shapes to plan electricity supply,

S. Hoff, “Electricity grid operators forecast load shapes to plan electricity supply,” July 2016, [Online; posted 22-July-2016]. [Online]. Available: https://www.eia.gov/todayinenergy/detail.php?id=27192

work page 2016
[3]

Renewable energy,

C. E. Commission, “Renewable energy,” January 2019

work page 2019
[4]

California to rely on 100 percent clean electricity by 2045 under bill signed by gov. jerry brown,

L. Dillon, “California to rely on 100 percent clean electricity by 2045 under bill signed by gov. jerry brown,” September 2018, [Online; posted 10-September-2018]. [Online]. Available: http://www.latimes.com/politics/la-pol-ca-renewable-energy-law-signed-20180910-story. html

work page 2018
[5]

Renewables cover about 100% of german power use for ﬁrst time ever,

S. Amelang, “Renewables cover about 100% of german power use for ﬁrst time ever,” January 2018, [Online; posted 05-January-2018]. [Online]. Available: https://www.cleanenergywire.org/news/renewables-cover-about-100-german-power-use-ﬁrst-time-ever 22

work page 2018
[6]

The trouble with electricity markets: understanding California’s restructuring disaster,

S. Borenstein, “The trouble with electricity markets: understanding California’s restructuring disaster,” Journal of economic perspectives , vol. 16, no. 1, pp. 191–211, 2002

work page 2002
[7]

4 Utilities Betting Billions on Renewable Energy,

T. Hoium, “4 Utilities Betting Billions on Renewable Energy,” September 2017, [Online; posted 28-September-2017]. [Online]. Available: https://www.fool.com/investing/2017/09/28/4-utilities-betting-billions-on-renewable-energy.aspx

work page 2017

Show all 20 references

[8]

An overview of demand response: Key-elements and international experience,

N. G. Paterakis, O. Erdinc ¸, and J. P. Catal ˜ao, “An overview of demand response: Key-elements and international experience,” Renewable and Sustainable Energy Reviews , vol. 69, pp. 871–891, 2017

2017
[10]

Dynamic competitive equilibria in electricity markets,

G. Wang, M. Negrete-Pincetic, A. Kowli, E. Shaﬁeepoorfard, S. Meyn, and U. Shanbhag, “Dynamic competitive equilibria in electricity markets,” in Control and Optimization Theory for Electric Smart Grids , A. Chakrabortty and M. Illic, Eds. Springer, 2011

2011
[11]

A control theorist’s perspective on dynamic competitive equilibria in electricity markets,

G. Wang, A. Kowli, M. Negrete-Pincetic, E. Shaﬁeepoorfard, and S. Meyn, “A control theorist’s perspective on dynamic competitive equilibria in electricity markets,” IFAC Proceedings Volumes, vol. 44, no. 1, pp. 4933–4938, 2011

2011
[12]

Equilibria in two-stage electricity markets,

A. Gupta, R. Jain, K. Poolla, and P. Varaiya, “Equilibria in two-stage electricity markets,” in 54th IEEE Conference on Decision and Control, CDC 2015, Osaka, Japan, December 15-18, 2015 , 2015, pp. 5833–5838. [Online]. Available: https://doi.org/10.1109/CDC.2015.7403136

2015
[13]

Dynamic economic dispatch game: The value of storage,

W. Tang and R. Jain, “Dynamic economic dispatch game: The value of storage,” IEEE Transactions on Smart Grid , vol. 7, no. 5, pp. 2350–2358, 2016

2016
[14]

Mas-Colell, M

A. Mas-Colell, M. D. Whinston, J. R. Green et al., Microeconomic theory. Oxford university press New York, 1995, vol. 1

1995
[15]

The role of a market maker in networked cournot competition,

S. Bose, D. W. Cai, S. Low, and A. Wierman, “The role of a market maker in networked cournot competition,” in 53rd IEEE Conference on Decision and Control . IEEE, 2014, pp. 4479–4484

2014
[16]

Dc power ﬂow revisited,

B. Stott, J. Jardim, and O. Alsac ¸, “Dc power ﬂow revisited,” IEEE Transactions on Power Systems , vol. 24, no. 3, pp. 1290–1300, 2009

2009
[17]

Dynamic competitive equilibria in electricity markets,

G. Wang, M. Negrete-Pincetic, A. Kowli, E. Shaﬁeepoorfard, S. Meyn, and U. V . Shanbhag, “Dynamic competitive equilibria in electricity markets,” in Control and optimization methods for electric smart grids . Springer, 2012, pp. 35–62

2012
[18]

Business practice manual for market instruments,

C. ISO, “Business practice manual for market instruments,” March 2021, [Online; posted 30-Mar-2021]. [Online]. Available: https://bpmcm.caiso.com/BPM Document Library/Market Instruments/BPM for Market 20Instruments V64 redline.pdf

2021
[19]

Boyd and L

S. Boyd and L. Vandenberghe, Convex optimization. Cambridge university press, 2004

2004
[20]

J. M. Morales, A. J. Conejo, H. Madsen, P. Pinson, and M. Zugno, Integrating renewables in electricity markets: operational problems . Springer Science & Business Media, 2013, vol. 205

2013

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [9]

A two-stage market mechanism for electricity with renewable generation,

N. Dahlin and R. Jain, “A two-stage market mechanism for electricity with renewable generation,” in 2019 IEEE 58th IEEE conference on decision and control (CDC) . IEEE, 2019

work page 2019

[2] [1]

The use of marginal energy costs in the design of U.S. capacity markets,

R. Moye and S. P. Meyn, “The use of marginal energy costs in the design of U.S. capacity markets,” in 51st Hawaii International Conference on System Sciences, HICSS 2018, Hilton Waikoloa Village, Hawaii, USA, January 3-6, 2018 , 2018. [Online]. Available: http://aisel.aisnet.org/hicss-51/es/markets/5

work page 2018

[3] [2]

Electricity grid operators forecast load shapes to plan electricity supply,

S. Hoff, “Electricity grid operators forecast load shapes to plan electricity supply,” July 2016, [Online; posted 22-July-2016]. [Online]. Available: https://www.eia.gov/todayinenergy/detail.php?id=27192

work page 2016

[4] [3]

Renewable energy,

C. E. Commission, “Renewable energy,” January 2019

work page 2019

[5] [4]

California to rely on 100 percent clean electricity by 2045 under bill signed by gov. jerry brown,

L. Dillon, “California to rely on 100 percent clean electricity by 2045 under bill signed by gov. jerry brown,” September 2018, [Online; posted 10-September-2018]. [Online]. Available: http://www.latimes.com/politics/la-pol-ca-renewable-energy-law-signed-20180910-story. html

work page 2018

[6] [5]

Renewables cover about 100% of german power use for ﬁrst time ever,

S. Amelang, “Renewables cover about 100% of german power use for ﬁrst time ever,” January 2018, [Online; posted 05-January-2018]. [Online]. Available: https://www.cleanenergywire.org/news/renewables-cover-about-100-german-power-use-ﬁrst-time-ever 22

work page 2018

[7] [6]

The trouble with electricity markets: understanding California’s restructuring disaster,

S. Borenstein, “The trouble with electricity markets: understanding California’s restructuring disaster,” Journal of economic perspectives , vol. 16, no. 1, pp. 191–211, 2002

work page 2002

[8] [7]

4 Utilities Betting Billions on Renewable Energy,

T. Hoium, “4 Utilities Betting Billions on Renewable Energy,” September 2017, [Online; posted 28-September-2017]. [Online]. Available: https://www.fool.com/investing/2017/09/28/4-utilities-betting-billions-on-renewable-energy.aspx

work page 2017