Conoco s Value and IPO: Real Options Analysis 1

Transcription

1 FIN 673 Professor Robert B.H. Hauswald Mergers and Acquisitions Kogod School of Business, AU Conoco s Value and IPO: Real Options Analysis 1 As you might recall a standard DCF analysis of Conoco s free cash flows would value the company at billions below the market s implied valuation. To reiterate our findings: Valuation of Conoco Downstream $ 7,77,997, Upstream Developed $ 7,255,11,18.78 Upstream Undeveloped $ 6,36,29, Upstream Unproven $ 929,,. Overhead $ (5,483,224,4.87) Enterprise Value $ 16,84,813, Less: Debt Assumed $ 4,953,,. Equity Value $ 11,131,813, IPO Percentage 3.5% Value Sold $ 3,395,23, Shares Sold 191,45, Value Per Share $ Unaccounted for/share $ I am grateful to Richard Shockley for these solutions. 1

2 which compares unfavorably to the IPO price of $23 per share. I ll now demonstrate that the option value in the proven undeveloped reserves makes up the majority of this underestimation. With the 1998 price of oil so low, the naïve NPV analysis valued the undeveloped reserves at only $6,393 million. This implicitly reflected the assumption that Conoco was committing $2,886 million to development no matter what. Of course, oil prices could continue to drop and if they did (and indeed real oil priced fell below $1 per barrel within a year), Conoco would delay or even perhaps abandon development of the reserves. Conoco s plans were to develop the field to replace the production at the alreadydeveloped fields to provide input for the downstream operations. However, if the price of oil fell very low Conoco might not develop at all, and if the price of oil surged after 1998 Conoco might find it optimal to go ahead and develop the field and sell the oil on the spot market. So we need to consider an American option analysis to value the undeveloped but proven Lobo tract. First, let s think about the basics. The source of uncertainty is the world price of crude oil, and management s flexibility is to delay (perhaps indefinitely) the development of the undeveloped reserves. For this analysis, the underlying asset will be oil out of the ground. The simplest way to approach this is to build a binomial tree of oil prices and then figure out the value of the American option on the field. The 1/98 average spot value of crude (weighted by the % of each type in Conoco s reserves) was $12.62/BOE, and the annual standard deviation in returns on oil prices was estimated to be 15%. (How do you calculate this? Write down the price of oil at the end of each year (,t-1, t, t+1,..). Then take the ratio of year t s price over year t-1 s price, to get the price relative for each year. Then take the natural logarithm of this to get the continuously-compounded return for year t. The standard deviation of this series of log price relatives is the standard deviation of return, which we use in the analysis). The leases on the undeveloped properties have 12-year lives. The real risk-free return per year (for a 12-year horizon) was estimated to be 3%. The strike price is the cost of development, or $2.886 billion. The fixed life of the lease is important, because it imposes a shortfall cost of not developing the tract (or a cash benefit or convenience yield of holding a developed tract over an option to develop). The Texas Railroad Commission has the authority to regulate the quantity of oil extracted from all oil and gas properties in Texas, and they administer this authority in a way that forces producers to extract at the optimum extraction rate. Conoco would only be allowed to extract 29mm BOE of hydrocarbons from the Lobo tract, so the tract could be fully exploited in seven years. Hence, if Conoco waits until there are only six years left on the lease, the effective size of the undeveloped reserves shrinks. That is all we need to do the analysis. We just need to be careful in how we proceed. First, let s do the simple part: build a binomial tree of oil prices over 12 years, with one step per year (for demonstration purposes). Since our standard deviation of the underly- 2

3 ing is 15% and we have one step per year, u = exp(sigma*(t)^.5) = exp(.15*(1/1)^.5) = and d = 1/u =.967. So our lattice of oil prices looks as such: Binomial Lattice of Oil Prices I believe you remember how to do that. What is the next step? Well, we need to make sure we frame our problem correctly. Conoco has an American option on the field they can open it any time they want. But they can t take all of the oil out at one time. The maximum that may be taken out in any one year is 29 mm BOE. So once again, we need to first build a tree of developed field values: if we had the field open and developed at each date in time, what would it be worth? Here there will be no erosion in the value of the field until the constraint of 29mm BOE per year becomes binding. So there is no erosion in value in years 1999 through 24. But if the field is not developed by the end of 24 (or the beginning of fiscal 25), then the value erodes at a rate of 29mmBOE per year. Got it? So start in the last period. The most that could be extracted from an open field in 211 is 29mmBOE, so we plug that amount times the net revenue from production (market price minus extraction costs), after taxes, into the last period of the open field tree. That is the price someone would pay us for one year s production in

4 Value of Developed Field (Assuming Value Erosion Starts in 25) =[29*( )]*.65 =[29*( )]*.65 Now we work backwards (just like in the precious metal mine from the last lecture). In 21, the value of an open field is the value of current production plus the 21 expected value of year 211 production. We calculate the discounted expected value of the field in 211 using our risk-neutral probabilities q = [exp(r*t)-d]/[u-d] =.5637 and 1-q =.4362 along with our risk-free discount rate exp(r*t) = exp(.3*1/1) = ,83 7,141 5,15 3,675 2,582 1,772 1, Value of Developed Field (Assuming Value Erosion Starts in 25) ,786 = [29*( )]*.65 +[(.5637*9,83)+(.4362*7141)]/ = [29*( )]*.65 +[(.5637*-16)+(.4362*)]/ ,786 9,83 12,158 7,141 8,73 5,15 6,191 3,675 4,39 2,582 2,916 1,772 1,883 1,172 1, Repeat this step backwards in time through 25, the first year in which value erosion will take place. What you will have are the 25 values of the field if undeveloped at that time. How much would someone be willing to pay us for the field in each state at 25? 4

5 Value of Developed Field (Assuming Value Erosion Starts in 25) ,44 = [29*( )]*.65 +[(.5637*26,374)+(.4362*18,755)]/ ,45 = [29*( )]*.65 +[(.5637*1,838)+(.4362*578)]/ ,44 26,374 25,96 24,379 21,472 16,786 9,83 18,393 18,755 18,529 17,523 15,498 12,158 7,141 12,726 13,111 13,65 12,444 11,72 8,73 5,15 8,527 8,93 9,16 8,681 7,793 6,191 3,675 5,417 5,833 6,17 5,894 5,364 4,39 2,582 3,112 3,538 3,796 3,829 3,565 2,916 1,772 1,45 1,838 2,15 2,299 2,232 1,883 1, ,165 1,244 1, At this point, there is no more value erosion working backwards. At this point I will argue to you that the development will never actually take place before 25 since there is no value erosion before then (i.e. the option is non-dividend paying before 25, so it is never exercised early). So our values of a developed field will just be the discounted expected value of the future year s field values. Just roll back using q, 1-q and the risk-free discount rate. Value of Developed Field (Assuming Value Erosion Starts in 25) ,35 = [(.5637*26,44) +(.4362*18,393)]/ ,298 = [(.5637*3,112) +(.4362*1,45)]/ ,35 26,44 26,374 25,96 24,379 21,472 16,786 9,83 15,45 18,393 18,755 18,529 17,523 15,498 12,158 7,141 1,572 12,726 13,111 13,65 12,444 11,72 8,73 5,15 6,958 8,527 8,93 9,16 8,681 7,793 6,191 3,675 4,281 5,417 5,833 6,17 5,894 5,364 4,39 2,582 2,298 3,112 3,538 3,796 3,829 3,565 2,916 1,772 1,45 1,838 2,15 2,299 2,232 1,883 1, ,165 1,244 1, Continue rolling back this way through year (the beginning of fiscal 1999, or the date of the IPO). 5

6 Value of Developed Field (Assuming Value Erosion Starts in 25) ,99 1,953 13,119 15,646 18,596 22,35 26,44 26,374 25,96 24,379 21,472 16,786 9,83 7,339 8,92 1,768 12,928 15,45 18,393 18,755 18,529 17,523 15,498 12,158 7,141 5,89 7,154 8,729 1,572 12,726 13,111 13,65 12,444 11,72 8,73 5,15 4,477 5,619 6,958 8,527 8,93 9,16 8,681 7,793 6,191 3,675 3,315 4,281 5,417 5,833 6,17 5,894 5,364 4,39 2,582 2,298 3,112 3,538 3,796 3,829 3,565 2,916 1,772 1,45 1,838 2,15 2,299 2,232 1,883 1, ,165 1,244 1, Now we have our tree of developed field values. This is what we get if, at any point, we exercise the option by investing the $2.886 billion in development of the field (the strike price). So how do we value the option? Again, start at the end and ask the following: If we have not developed by the beginning of 211, would we then? Obviously we will only develop in 211 if the value of a developed field with one year s production left is worth more than the strike. So write in the max of the value of the developed field minus the strike (develop), or zero (don t develop). Value of American Option on Developed Field, and Optimal Exercise Boundary ,944 6,944 = max(9,83-2,886; ) 4,255 2, = max(-2,886; ) Notice something else here I put a shaded box on the states of nature in which I do exercise the option. This is important to me I need to know the optimal exercise strategy as well as the value. Now we roll backwards, with the extra calculation we talked about in the derivatives pricing review of American options. In 21 we will ask the question: If I have not exercised by now, is it more valuable to develop the field immediately (exercise the option) or wait (keep the option alive)? The value of exercising is what you get minus the cost: the value of a developed field in 21 (with two years production left) minus the 6

7 strike. The value of waiting is the expected present value of the option to make the decision in 211 just calculate this using q, 1-q and the risk-free discount rate. Here s what it looks like. Value of American Option on Developed Field, and Optimal Exercise Boundary ,9 = max[16,786-2,886 (strike); (.5637*6, *,255)/1.35 (wait) =13,9 (strike) = max[-41-2,886 (strike); (.5637*+.4362*)/1.35 (wait) = (wait) 13,9 6,944 9,272 4,255 5,844 2,264 3, ,423 3 Notice again the blue blocks where I optimally develop if I have not done so already. Notice something important here: There is no way to get to a shaded state in 211 without passing through a shaded state in 21. Thus, there will be no development in 211 it will occur earlier (if ever at all). Repeat this step exactly the same way for every period back to This gives us the value of the option: Value of American Option on Developed Field, and Optimal Exercise Boundary ,73 8,471 1,559 13,9 15,878 19,234 23,158 23,488 23,2 21,493 18,586 13,9 6,944 4,887 6,364 8,131 1,21 12,649 15,57 15,869 15,643 14,637 12,612 9,272 4,255 3,319 4,527 6,11 7,771 9,84 1,225 1,179 9,558 8,186 5,844 2,264 1,99 2,924 4,157 5,641 6,44 6,13 5,795 4,97 3, ,535 2,531 2,947 3,131 3,8 2,478 1, We notice two things: the value of the American option on the undeveloped reserves is $6.73 billion (as opposed to only $3.655 billion in the naïve analysis). Second, Conoco s optimal policy is to wait as long as possible: no development takes place before 25, and if prices are very low in 25 development may wait until 26 (or may never occur). More on that later. Let s go back to the valuation of the Conoco IPO. What do we have now? 7

8 Valuation of Conoco Downstream (mm) $1,742 Upstream: Proven Developed (mm) $5,448 Proven Undeveloped (mm) $6,73 Unproven (mm) $929 Corporate Overhead (mm) ($5,245) Total Enterprise Value (mm) $18,577 Less Debt Assumed (mm) $4,953 Implied Equity Value (mm) $13,624 Value Per Share $21.7 So I m still $1.3 per share too low. Did the market overpay? Not necessarily. If you recall, I valued the unproven reserves at Conoco s acquisition price, implicitly assuming zero NPV. Conoco may have purchased them at an attractive price (due to geophysical expertise within the company, etc.); moreover, there may have been good news about the tract since the purchase was made. Furthermore, Conoco s geological expertise itself generates options on future unproven reserves. The market s valuation of the remaining option-generating capability of Conoco was $19.39 billion - $ billion = $813 million. This is not a bad estimate of the value of Conoco s geophysical expertise but I have a hunch it is too low. Do you think I have overvalued Conoco s proven reserves, or do you think the market is undervaluing the company s more hidden options? Finally, let s go back and make sure we understand the exercise boundary. Below I have replicated the original lattice of oil prices. I have superimposed the optimal exercise boundary, and I have removed the unnecessary information. Take a look. 8

9 Binomial Lattice of Oil Prices with Optimal Exercise Strategy = Develop = Wait and See = Abandon = Irrelevant (Decision Already Made) Conoco has a date with destiny in 25 it will most likely exercise its development option then. On the other hand, if oil prices are very low ($5.13/BOE), the company will walk away from the development option altogether. However, if oil prices are just somewhat low ($6.93/bbl), Conoco will optimally give up a year s production to see how prices move until 26. If prices go up, development will occur; if prices go down, the development option enters a range in which it will never pay to exercise. It is no coincidence that Conoco s date with destiny arrives at exactly the time the company s developed reserves become completely depleted. This illustrates how strategy and option analysis go hand in hand Conoco s managers needed to structure their leases on the undeveloped properties so that they maximized the value created in the context of the business. Figuring out the optimal exercise times before entering the leases was critical. 9