Switching options and the impact on business strategy and risk management.

Johnson, Larry A.

INTRODUCTION

There were 4,560 Independent Power Producers (IPP's) in the United States in 2006 and account for a substantial amount of the 17.9 percent of all electricity generated by natural gas (The Energy Information Administration, 2006). A common business strategy for IPP's is to make a forward sale of electricity to a utility and purchase the natural gas to generate the electricity on a short-term basis. Operators often prefer to delay the purchase of natural gas because the natural gas generating units have real optionality. The plant operator has the option of delivering to the utility electric power generated from the plant or purchasing electricity from the daily electric power market. The operator compares the heat rate efficiency of the plant (number of million British Thermal Units of natural gas to produce one megawatt of power) and the wholesale price of electric power and delivers whichever is less expensive at the time. This real switching option (See Culp, pp. 312-314) can add value to the generator, but does impact the firm's risk management strategy and hedging decision. The "generate versus purchase" option can lead to over or under hedged positions.

OBJECTIVE

The objective of this study is to explore the impacts of having the option to switch from delivering electricity generated by natural gas at the plant versus purchasing electricity from the daily electric power market on the effectiveness of the natural gas hedge.

METHODOLOGY

Optimal Hedging Through the Use of Minimum Variance Hedge Rations

The theory of hedging as a means of reducing price risk has a long history. (see Johnson (1960), Franckle (1980), Ederington, and Culp and Millar(1999)). The effectiveness of a hedge from an economic perspective depends on the amount of risk reduction obtained through the off-setting hedged position. Since the cash flows of the hedge position rarely completely off-set the cash flows of the hedged item due to the basis risk, the optimal off-setting position may be more or less than the hedged item.

A common approach to matching the cash flows of the hedged position to the hedged item is to determine the minimum variance hedge ratio or more commonly known as the optimum hedge. (see Culp (2001), p.p 53-55) The variance minimizing hedge ratio for equal quantities of the hedged item and the futures contract can be determined by regressing cash price changes to the changes in the futures price and observing the beta coefficient.

Hedging Given Both Price and Volume Risk

Most minimum variance hedge studies assumed volume was known and constant. Development of the portfolio model of hedging with applications to include both price and volume variability were developed by McKinnon (1967) who states that the greater the volume variability relative to price variability, the smaller will be the optimal forward sale; and, the more negatively correlated the price and volume the smaller will be the forward sale. Other work involving hedging with both volume and price risk were Rolfo (1980), Conroy and Rendleman (1983), and Miller and Kahl (1986).

Volume uncertainty in this study is addressed through the inclusion of a switching option. Switching options are a variation of real options where firm management has flexibility of altering decisions to add value to the firm. Culp ((2001) p. 312) defines a switching option where the management can switch either inputs or outputs during the production process. This type of analysis is common in electric power generation since integrated utilities have a number of options for power generation including nuclear, hydroelectric, natural gas, and fossil fuels. Most of these assets are valued as real options and managed accordingly.

Methodology for the Natural Gas Price Risk Hedge Given Volume Uncertainty

This study extends the methodology used by Johnson, (2008 pp. 23-24) to include the value of the switching option.

"The simulation incorporated the hypothetical case where an Independent Power Producers (IPP) enters into a forward sales agreement with an electric utility where the IPP will deliver electricity from the Combined Cycle Electric Generating Station on a daily basis. The IPP and electric utility enters into the agreement where the electric utility would purchase 100Megawatts (MW's) for the on-peak hours. (6a.m. to 10:00p.m. Central Pacific Time (C.P.T.)). The plant requires 8,000 million British Thermal Units (MBTU's) of natural gas per megawatt hour (MWh) of electricity. The IPP would deliver electricity from the plant on the days when the generation cost of the plant is less than the wholesale electric market or purchase electricity from the market, whichever is less. Assume the IPP would purchase electricity from the Into Entergy electric market. Also assume the natural gas can be delivered to the plant at a price equal to the Henry Hub natural gas price. Power price is for the on-peak period 5 days per week (Monday through Friday) for 16 hours per day (6 a.m. to 11:00 p.m., CPT).

Assume the IPP routinely forward sells the electricity and subsequently hedges the natural gas price exposure three months prior to delivery using NYMEX natural gas futures. Prior to the delivery month the IPP rolls the natural gas hedge into the prompt month on the last trading day before delivery. Natural gas hedged positions are lifted each day during the month of delivery. The hedge using natural gas futures is assumed to commence at the closing price of the natural gas futures contract on the first trading day of the month for a contract three months out. The quantity of futures contracts will be the equivalent amount of the forecasted natural gas needs. The futures contract settlement price will be the closing price on the last day of trading and then the closing price of the prompt contract for each day during the month of delivery.

A simplified version of the futures hedge will be:

[[PI].sub.f(t+3)] = ([P.sub.f(t+3)] [V.sub.(t+3)])--([P.sub.ft] [V.sub.(t+3)])

where:

[[PI].sub.f(t+3)] = Return to the future's position;

[P.sub.f(t+3)] = Futures price at contract expiration;

[V.sub.(t+3)] = Forecast volume; and

[P.sub.ft] = Initial futures price.

The cash position will compare the next day Into Entergy on-peak power price to the generation cost of the natural gas generator based upon the heat rate and the next day price for natural gas delivered at the Henry Hub. If the price of electricity is lower than the generation cost, the electricity would be purchased and the futures hedge lifted.

The return to the cash position will be the difference between projected cash position at the initiation of the hedge with the actual cost of natural gas purchases during the delivery month. The expected futures cash position will be the natural gas futures price and the forecast volume. Specifically, the return to the cash position:

[PI.sub.c(t+3)] = ([Pc.sub.(t+3)] [V.sub.(t+3))] - [(Pa Va).sub.(t+3)]

where:

[[PI]c.sub.(t+3)] = Return to the cash position;

[Pc.sub.(t+3)] = Cash price at when hedge is placed;

[V.sub.(t+3)] = Forecast volume;

Pa = Cash Price of natural gas delivered; and

Va = Volume of natural gas purchased."

Optimal Hedge Ratios for the Natural Gas Hedge

A simple regression was performed with the simulated profits from the cash position regressed against the returns of the futures position. The resulting beta coefficient of the futures position was used to determine the optimum hedge ratio (minimum variance) designated as [N.sub.f], where:

[N.sub.f] = - Cov ([[PI].sub.c(t+3)], [[PI].sub.f(t+3)])/ Var [[PI].sub.f(t+3)]

Model Specification for the Switching Option

The actual cost of natural gas generation [(PaVa).sub.(t+3)] is impacted by the option to switch to the daily power market. Therefore, the simulation included the option of switching from natural gas generation to the purchase of electricity on each day of delivery. Specification for the switching option is:

Min [[(PaVa).sub.t+3)], [PC).sub.(t+3)]] = If [[(PaVa).sub.(t+3)] < [PC.sub.(t+3)], [(PaVa).sub.(t+3)], [PC.sub.(t+3],]

where,

[(PaVa).sub.(t+3)] = Natural gas generation cost per MWh (Heat rate time the natural gas price in MMBTU divided by 1000); and

[PC.sub.(t+3)] = Purchase price of electricity in MWh

Model Specification Incorporating Optimal Hedges with the Switching Option

The simulation includes a feedback loop where the predetermined optimal hedges ([N.sub.f]) were incorporated into the model with the switching option:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where,

[[PI].sub.(t+3)] = Total return of the optimal hedge including the switching option.

DATA AND ANALYSIS

The simulation included primary seasonal months for the period July 2001 to January 2006. Natural gas hedges were the same as the seasonal power months. All power forward prices, daily power cash prices, and daily cash natural gas prices were obtained from trading day summaries provided by the Intercontinental Exchange (ICE). The forward power contracts used were Into Entergy Summer (July/August), Winter (January/February), and Spring (March/April) with daily power cash prices being Next Day Into Entergy. Natural gas cash prices were the next day Henry Hub. Futures prices used in the analysis were NYMEX closing prices and obtained from the Federal Energy Regulatory Commission's website data base of natural gas futures prices.

Natural gas hedges were entered into on the first trading day three months prior to the month of delivery. The natural gas hedges continued until the last trading day before the day of delivery and then rolled into the prompt month. During the month of delivery, the natural gas hedges were lifted daily until the last day before delivery and the remaining hedged positions rolled to the next prompt month.

RESULTS

Optimal hedge ratios were determined across all months, all seasons, individual months, and by individual season. (see Johnson, 2008) "The hedge analysis across all months gave an optimal hedge ratio of-1.38 (Table 1), meaning that the optimal hedge was 1.38 times expected natural gas usage. The hedge analysis across all months had an Adjusted [R.sup.2] of 0.46 and significant t statistic (-4.85, P = .00006) for the beta coefficient of -1.38".

Table 1 summarized the cash flows from the hedged position relative to changes in the expected physical gas cash flows by month. Table 1 also shows the cash flows from purchased power switch option for each month and the average price of natural gas during that month. During the months in 2001 and 2002 natural gas prices were generally lower (less than $5.00/MMBTU) limiting the use of the power switch option. Hedges in 2003-2006 months were of generally higher natural gas price months ($4.97/MMBTU-$9.45/MMBTU). During these months natural gas purchases were less than expected and were largely offset by power purchases.

Table 1 also presents a summary of the value of the switching option both in dollars per MW per month and in average dollars per MWh. Across all months, the switching option was $1.99 per MWh with the highest being $23.20 per MWH in April of2000. The switching option had very little value during the peak summer months when power prices were at their seasonal peak and natural gas prices usually exhibit their seasonal trough. Winter is somewhat different in that even though power prices are high, natural gas prices are higher relatively due to heating demand. Spring had the most value for the swing option due to higher natural gas prices and seasonally low power prices.

The optimal hedge ratio for all months and the optimal hedge ratios by season were simulated to determine the extent of over or under hedging (Table 2) (see Johnson, 2008, p. 29). Over-hedged positions cash flows were as high as 498.64% of expected natural gas purchases in April of 2002 to as low 7.49% of expected natural gas purchases in March 2002. Over-hedging occurred in six months and under-hedging occurred during 12 months. Nine months resulted in risk increasing hedges meaning the cash flows from the physical positions and the hedged items were positive.

CONCLUSION

The volume uncertainty associated with the switch option of purchased electricity and the volatility of natural gas cash prices during delivery relative to the futures price of the prompt month led to the ineffectiveness of most of the natural gas hedged positions. The results suggest that while hedging natural gas price exposure is intuitively appealing, the volume risk of the switch option can lead to over and under-hedged positions, may preclude one from using hedge accounting, adversely impacts cash flow management, and can even be risk increasing. However, the switch option has value which may offset these inefficiencies and this value should be taken into consideration when making hedging decisions.

While these results are specific to the hedging activities of an IPP, they have broad implications for hedging where volume risk is present.

SUGGESTIONS FOR FUTURE RESEARCH

The study is limited in scope but does provide insight into the use of real options and hedge effectiveness. The study could be extended to include other risk management strategies that might be more effective in matching the cash flows of the physical position and the hedge such as the use of daily call options rather than futures contracts for the hedge of natural gas.

Independent Power Producers sell into competitive electric power markets and are not subject to state or federal rate setting agencies. Regulated electric utilities that use hedging strategies for their own natural gas generation could have an impact on rate structures since these activities do impact their cost to serve. This type of study could address the needs of rate making authorities.

REFERENCES

Financial Accounting Standards Board. Accounting for Derivative Instruments and Hedging Activities: FAS B Statement as Amended and Interpreted, February 2004.

Conroy, R. and R. Rendlemen, Jr. "Pricing commodities when both price and output are uncertain," The Journal of Futures Markets. Summer 1983.

Culp, C.L. The Risk Management Process: Business Strategy and Tactics. John Wiley and Sons.

Culp, C.L. and Millar, M. H. Corporate Hedging in Theory and Practice: Lessons From Metallgesellschaft. London: Risk Books.

Ederington, L. "The hedging performance of the new futures markets," Journal of Finance, 34, 157-170, (March 1979).

Energy Information Administration. "Existing generating capacity," EIA Homepage, Electric Power, http://www.eia.doe.gov/oiaf/aeo/index.htm 2006

Federal Energy Regulatory Commission's. "Natural gas prices" FERC Homepage, http://www.ferc.gov/marketoversight/mktgas/southeast/archives.asp 2006

Franckle, C.T. "The hedging performance of the new futures markets: Comment," Journal of Finance, 35, 1273-1279, (December 1980).

InterContinental Exchange. Summary of price settlements, 2000-2006.

Johnson, L.A. "The impact of volume risk on hedge effectiveness: The case of an independent power producer." The Journal of Energy Markets, Vol. 1, No. 1, Spring 2008. 21-30.

Johnson, L. "The theory of hedging and speculation in commodity futures," Review of Economic Studies, 27:139-151 (1960).

Miller, S.E., and K.H. Kahl. "Forward pricing when yields are uncertain," Review of Futures Markets, 6:21-39 (1986)

McKinnon, R.I. "Futures markets, buffer stocks, and income stability for primary producers," Journal of Political Economy 75: 844-861 (1967)

Rolfo, J. "Optimal hedging under price and quantity uncertainty: The case of cocoa producers," Journal of Political Economy, 88:100-116 (1980)

Larry A. Johnson, Dalton State College

Table 1. Summary of Cash Flows From The Natural Gas Hedge, Physical
Natural Gas Purchases Versus Expectations, and Power Purchases,
2001-2006.

 Purchased Purchased
 Physical Purchased Power Power
 Month Hedge Nat Gas Power Gains Gains

 Dollars per MW per Month $/MWh

July 01 -6012 5531 0 0 0.00
Aug 01 -6205 6518 0 0 0.00
Jan 02 -1107 2996 1461 47 0.13
Feb 02 -1742 1392 0 0 0.00
Mar 02 37 -677 0 0 0.00
Apr 02 2141 -593 1309 23 23.20
July 02 -1866 1694 0 0 0.00
Aug 02 -463 1458 0 0 0.00
Jan 03 2804 7645 9433 1253 3.40
Feb 03 5577 3812 11011 1063 3.32
Mar 03 5577 3812 11011 901 2.68
Apr 03 3699 3157 5163 295 0.84
July 03 -170 352 0 0 0.00
Aug 03 -544 1882 1270 42 0.13
Jan 04 4271 9039 9398 2274 6.46
Feb 04 1056 6850 6693 290 0.91
Mar 04 1483 11197 11927 456 1.24
Apr 04 1357 12278 13800 827 2.35
July 04 644 600 448 308 0.87
Aug 04 -1002 4186 2027 138 0.39
Jan 05 -4949 18864 12244 1822 5.42
Feb 05 -4476 18588 13056 966 3.02
Mar 05 -1881 13023 11264 405 1.10
Apr 05 1119 10771 11467 712 2.12
July 05 -1284 4981 3921 42 0.12
Aug 05 2708 -1876 0 0 0.00
Jan 06 -16626 41936 19211 5468 15.53

 Nat Gas Power
 Month Price Price

 $/MMBTU $/MWh

July 01 5.19 41.85
Aug 01 5.22 39.17
Jan 02 2.32 19.45
Feb 02 2.28 20.82
Mar 02 3.02 26.63
Apr 02 3.42 31.75
July 02 3.00 33.87
Aug 02 3.08 28.38
Jan 03 5.36 40.58
Feb 03 7.38 56.88
Mar 03 6.30 48.52
Apr 03 5.30 42.87
July 03 5.06 44.44
Aug 03 4.97 43.43
Jan 04 6.08 43.10
Feb 04 5.39 43.18
Mar 04 5.38 41.94
Apr 04 5.71 43.38
July 04 5.93 50.36
Aug 04 5.43 45.91
Jan 05 6.16 44.63
Feb 05 6.09 45.75
Mar 05 6.93 55.34
Apr 05 7.20 56.11
July 05 7.57 67.39
Aug 05 9.45 85.67
Jan 06 8.76 54.58

Table 2. Summary of Percentage Change in Cash Flows from Hedges
versus Changes in Cash Flows from the Expected Natural Gas Purchases
for Optimal Hedge Ratio Simulations Across All Months, 2001-2006.

Month July 01 Aug 01 Jan 02 Feb 02

Hedge/Expected Physical -150.00% -131.39% -50.99% -172.71%

Month Aug 02 Jan 03 Feb 03 Mar 03

Hedge/Expected Physical -43.83% 50.61% 201.91% 161.71%

Month Jan 04 Feb 04 Mar 04 Apr 04

Hedge/Expected Physical 21.28% 18.28% 15.25% 147.94%

Month Feb 05 Mar 05 Apr 05 July 05

Hedge/Expected Physical -19.93% 14.34% -35.56% -199.12%

Month Mar 02 Apr 02 July 02

Hedge/Expected Physical -7.49% -498.64% -152.07%

Month Apr 03 July 03 Aug 03

Hedge/Expected Physical -66.74% -39.89% 65.21%

Month July 04 Aug 04 Jan 05

Hedge/Expected Physical -33.03% -36.20% -33.23%

Month Aug 05 Jan 06

Hedge/Expected Physical -54.71% -15.61%