Hedging and Price Risk Management: A California Perspective Todd Strauss Senior Director, Energy Policy, Planning, and Analysis Pacific Gas and Electric Company NARUC – Staff Subcommittee on Accounting and Finance Spring 2008 Meeting 1 April 2008

Contents Introduction to energy procurement at PG&E How PG&E’s hedging programs work Learning and insights from PG&E experience in hedging

Pacific Gas and Electric Combined gas and electric utility in northern and central California Gas 4 million customer accounts $4 billion annual revenue 850 bcf (30% bundled) Electric 5 million customer accounts $9 billion annual revenue 86,000 GWh (92% bundled) 20,000 MW peak load

PG&E Energy Procurement Costs heavily tied to gas commodity prices Electric: 45% of energy tied to natural gas commodity prices Combined utility; separate portfolios Wholesale portfolios for bundled electric customers and core gas customers are managed separately Decoupling Balancing account treatment provides no incentive for utility to earn more dollars by selling more energy

PG&E Energy Procurement: Objectives Reliability Meet obligation to serve Environment Customer cost Reasonable level Stable Cost recovery Cost allocation Shareholder earnings (core gas incentive mechanism)

PG&E Energy Procurement: Regulatory Regimes Bundled electric portfolio Procurement plan review and approval Procurement plan includes products, processes, strategies Includes hedging plans Compliance review of activities Were utility’s actions consistent with plan? Procurement Review Group (PRG) Consumer advocates, California PUC staff, and other non-market participants representing stakeholder interests in electric procurement Advisory role to PG&E Core gas portfolio Incentive mechanism Based on basket of monthly indices Short-term oriented

Contents Introduction to energy procurement at PG&E How PG&E’s hedging programs work Policies and principles Bundled electric portfolio Core gas portfolio Learning and insights from PG&E experience in hedging

A. Bundled Electric Portfolio: PUC’s Risk Policy Risk policy established by California PUC in 2003 Cost variability measure To-expiration Value-at-Risk (TeVaR) Customer Risk Tolerance (CRT) 1 cent per kWh Risk management policy Compare TeVaR with CRT If TeVaR > 1.25  CRT, then meet and confer with Procurement Review Group

Portfolio Cost Variability Bundled electric portfolio consists of Load obligations Resources (supply-side and demand-side) Sources of portfolio cost variability Resource cost (natural gas price) Resource quantity (forced outages, hydro generation) Load uncertainty: heat rate of marginal resource in supply stack Probability distribution of portfolio cost Each source of variability can be described probabilistically Result is probability distribution of portfolio cost

1. Cost Variability Measure TeVaR is a measure of the width of the probability distribution of portfolio cost Cost distribution (and therefore TeVaR) is modeled, not observed Inputs include market data (forward curves, volatilities, and correlations) and portfolio resources/instruments See next slide for comparison between TeVaR and Value-at-Risk (VaR)

SIDEBAR Risk Measurement: VaR and TeVaR Value-at-Risk (VaR) Answers the question of how large the deviation could be between portfolio value in the future and portfolio value today This deviation is in the context of a particular time horizon (1 to 5 days in the future) and confidence interval (e.g., 95th percentile) Time horizon is typically short: 1 to 5 days To-Expiration Value-at-Risk (TeVaR) is VaR with liquidation horizon carried to delivery Load obligation cannot be unwound like instruments in a trading book

Reducing Cost Variability Activities that narrow the cost distribution: Adding fixed-price resources to the portfolio Hedging If it doesn’t narrow the cost distribution, it isn’t hedging, it’s speculation!

Hedging Strategy Changes Cost Distribution Candidate strategies differ by amounts and product mix Cost distribution is narrowed by Greater hedging quantities More swaps/forwards/futures and fewer options

Hedging and Cost Distributions: A Numerical Look Some analysts like tables of numbers, others like graphs Table has additional information (lines 18-27)

Hedging Strategies and Cost Distributions The question Which hedging strategy is best? is transformed into the question Which cost distribution do bundled electric customers prefer? Extensively discussed with PRG strategies differ by amounts and product mix

2. Customer Risk Tolerance How much of an increase in cost can customers tolerate? is operationalized as the question How wide should the probability distribution of portfolio cost be?

Customer Risk Tolerance (CRT) How wide should the probability distribution of portfolio cost be? This is a risk preference This is a policy issue Current policy set by California PUC is 1 cent per kWh For PG&E bundled electric portfolio, this corresponds to incremental portfolio cost of about $800 million Customer survey was/is intended to inform policymaking

3. California PUC’s Risk Management Policy Compare TeVaR with CRT In words: compare estimated width of probability distribution of portfolio cost with the stated 1 cent per kWh target for the width If TeVaR > 1.25  CRT, then meet and confer with Procurement Review Group Stakeholder discussion of the situation is required No particular portfolio action is required 1.25  CRT is referred to as the “notification level” This is very different from a trading limit

B. Core Gas Portfolio Regulatory regime is short-term incentive mechanism Based on basket of monthly indices Misalignment of interests: hedging for customers looks like speculating from shareholder perspective Longer-term hedges deviate from monthly spot price index Hedging narrows probability distribution of customer cost Hedging increases probability distribution of shareholder gain/loss Disconnect identified: significant hedging for bundled electric customers, and no hedging for core gas customers Bundled electric customers and core gas customers are largely the same residential households and small commercial customers, with the same risk preferences for gas service costs as for electric service costs

Core Gas Portfolio: History of Hedging In 2005 PG&E initiated PUC filing regarding hedging for core gas customers PG&E requested that all benefits and costs related to hedging go to customers, and be outside of incentive mechanism Hedging has been performed for past 3 winter seasons, with hedges outside of incentive mechanism Advisory group process analogous to bundled electric PRG has been established Customer risk tolerance survey is underway PUC about to begin a broad proceeding looking at hedging in context of existing incentive mechanism structure

Contents Introduction to energy procurement at PG&E How PG&E’s hedging programs work Learning and insights from PG&E experience in hedging Hedging vs. speculating Hedging: costs vs. cash flows Risk vs. regret

Hedging vs. Speculating: Behaviors Market view Hedging takes the market as is: the market (i.e., what is currently transactable) is not right or wrong, it just plain exists Speculating takes a view on where the market is headed and acts on that view Market timing Hedging executes transactions relatively evenly over time, to diversify timing risk, perhaps using dollar cost averaging Speculating uses event-driven trading to time the market, perhaps trading in and out of positions

Hedging vs. Speculating: Objectives Hedging objectives Manage TeVaR Protect against price blowout scenarios Flatten positions that arise from physical assets and obligation to serve Speculating objective Earn an outsized return on risk capital

Hedging: Costs vs. Cash Flows Q: What is the cost of hedging? A: Transaction costs. Broker fees Financing margin and collateral Bid-ask spreads A: Option premiums are cash outflows, not costs. Hedging has little impact on expected portfolio cost.

Cash Flows vs. Net Costs: Forwards Q: At the time a forward (swap/forward/future) contract is executed, how much money trades hands? A: Zero. Q: At settlement, does money trade hands? A: Yes. Q: At time of execution, how much money is expected to trade hands at settlement? Q: Therefore, at execution, expected net cost of forward is zero? Q: But what about cost at settlement? A: See “regret.”

Cash Flows vs. Costs: Options Q: At the time an option contract is executed, how much money trades hands? A: Buyer of option pays seller of option the option “premium.” Q: At settlement (option expiry), does money trade hands? A: If option is in the money, seller of option pays buyer of option the difference between market price and option strike price. Q: At time of execution, how much money is expected to trade hands at settlement? A: The option premium plus the interest associated with the time value of money. Q: Therefore, at execution, expected net cost of option is zero? A: Yes. Q: But what about cost at settlement? A: See “regret.”

Risk vs. Regret Risk is “potential negative impact that may arise from a future event” Regret is “distress of mind for what has been done or failed to be done” Hedging example: quantity hedged is always wrong in hindsight—too little or too much Hedging example: buying options Most of the time, these options won’t pay back the option premium, and will regret buying them When these options do pay out, will regret not having swaps instead Hedging example: whether to hedge or not to hedge Hedging seems to have more regret than not hedging Risk is prospective, regret is retrospective

Conclusion: Overcoming Regret Ask yourself: Is the hedging strategy designed to reduce risk or to avoid regret? Focus on the total portfolio—physical and financial—not just the hedge book Focus on the exposure ($) of a potential event separately from the probability of that event occurring Establish risk benchmarks and measure the portfolio against those benchmarks Include as a hedging objective: Manage option premium expenses