Mirant Corp. filed for chapter 11 protection on July 14, 2003. With claims in excess of $10 billion, it was the largest bankruptcy filing of the year. In April 2005, a hearing for emergence from chapter 11 began in Fort Worth, Texas. The focus of the hearing was the valuation of Mirant's enterprise value to determine the new ownership structure at emergence.

The valuation hearing was a bitterly fought battle of experts representing the debtors, creditors, equity-holders and certain convertible security-holders. At least 11 different experts submitted expert reports, and many of these experts testified at the hearing. There were numerous points of contention about the valuation methods employed by the various experts. In fact, the valuation experts differed on nearly every aspect of their valuation methodology. Some differences were minor; however, others could change the value by hundreds of millions of dollars, even billions. Generally, the experts employed two methods to determine Mirant's enterprise value: a comparable company multiple valuation and a discounted cash-flow valuation (DCF).

Comparable Company Multiple Valuation

In performing a comparable company valuation, all of the experts used Earnings Before Interest Taxes Depreciation and Amortization (EBITDA) multiples, as EBITDA is the most widely followed metric in Mirant's industry. The calculation of EBITDA multiples was hotly contested, as experts disagreed on how to calculate the two main components of the multiple: the comparable companies' enterprise value (the numerator) and their EBITDA (the denominator). With Mirant having EBITDA of approximately $1 billion, the multiple used to value its operating assets had a huge impact on value.

Comparable Companies' Enterprise Value. One of the comparable companies, NRG Energy, had a note receivable asset worth over $800 million. One valuation expert believed that the notes were nonoperating assets and reduced enterprise value by this amount, resulting in a lower multiple. Other experts argued that the notes were operating assets and should not be excluded from the enterprise value. This adjustment affected the median multiple used to value Mirant. Specifically, reducing NRG's enterprise value by the notes receivable reduced Mirant's enterprise value by over $500 million.

---

Because of the broad range of assumptions used in the methodologies for valuing Mirant's operating assets as well as its nonoperating assets, the range in values calculated for Mirant was extremely wide.

---

Comparable Companies' EBITDA. Two of the issues that affected the calculation of the comparable companies' EBITDA had substantial impacts on Mirant's value: mark-to-market gains and losses, and income from unconsolidated investments. While all of the comparable companies comply with FAS 133 (Accounting for Derivative Instruments and Hedging Activities), only one comparable company detailed its gains and losses from marking-to-market its derivative contracts on gas and energy. Its EBITDA, using the numbers in its income statement, was reduced by more than $250 million because of the mark-to-market losses. Certain experts reversed those losses, claiming that they were noncash losses, thereby reducing the multiple. Other experts disagreed, arguing that the remaining comparable companies did not detail similar gains/losses, and therefore the experts were inconsistent in their adjustments.

Another hotly-disputed issue was the inclusion of income from minority investments. It is common in the energy industry for companies to take minority stakes in power-producing assets or to develop joint ventures, neither of which is consolidated in the operating income. Instead, income attributable to those investments is included in the income statement but below the operating income line as a separate item.

Certain experts insisted that income from unconsolidated investments needed to be included in EBITDA because it was income from normal operations and it was also the way the comparables reported their EBITDA to analysts and investors. However, these experts failed to include the unconsolidated debt associated with this income in the comparable companies' enterprise value. This inconsistent approach results in a downward biased multiple, thereby reducing Mirant's enterprise value.

Timeliness of Information. The first round of the experts' reports were submitted in February. Rebuttal reports followed in March; the hearing started in mid-April and concluded in late June. During this time period, Mirant continued to operate and forecasts were revised as actual information became available. The comparable companies sold divisions of their businesses, operating agreements expired, Form 10-Ks for the fiscal year ended December 2004 were issued, and then Form 10-Qs for the first quarter ended March were issued. In addition, the comparables and the analysts revised their projections for the 2005 fiscal year and then issued their projections for the 2006 fiscal year. All this information had to be considered in the valuation calculations. As a result, throughout this period, the experts had to continuously revise their valuation reports. Some of the experts believed that forward multiples were the only correct multiples to use in valuing Mirant because the market is only concerned with future performance as an indicator of value. They used one- and two-year forward multiples to calculate a range of enterprise values for Mirant. Other experts advocated the use of both forward and historical multiples, arguing that both types are widely accepted in valuation literature and that the market considers both past and future performance when determining value.

Discounted Cash-Flow Valuation

The second method used to value Mirant was a discounted cash-flow valuation. Mirant provided a business plan with detailed projections covering a 10-year period, which mostly eliminated dispute among the experts as to the expected cash flows. However, the other assumptions required to complete the discounted cash-flow valuation varied greatly between valuation experts.

Discount Rate. Cash flows in a DCF model are usually discounted at the company's weighted average cost of capital (WACC). As implied by its name, the WACC is derived as the weighted average of the company's cost of equity and cost of debt based on the company's capital structure. Mirant's after-tax cost of debt was generally agreed upon by the experts to be around 5.4 percent based on Mirant's business plan.

The cost of equity was calculated by experts using a two-step process. The first step was to calculate Mirant's beta, based on its expected correlation with the stock market. Beta is an element of the cost of equity when using the Capital Asset Pricing Model (CAPM). Mirant's expected beta was calculated using the comparable companies. The comparable companies' unlevered (not influenced by capital structure) betas were calculated, and the median was applied to Mirant. However, the betas varied depending on the methodology being used. In one instance, an expert used a longer time horizon that required him to exclude one of the comps. By doing so, the beta was significantly lower than that of some of the other experts.

Once the beta was calculated, it was applied to the CAPM formula that considers beta, the market-risk premium and the risk-free rate to determine a company's cost of equity. While the experts agreed on the risk-free rate, there was considerable disagreement about the market-risk premium. Certain experts used the long-term equity risk premium of 7.2 percent based on Ibbotson's "Stocks, Bonds, Bills and Inflation" study that covers the period 1926-2004. One expert cited recent studies that indicate a declining risk premium. Rather than 7.2 percent, which is based on data since 1926, the risk premium was estimated to be in the range of 2.55-4.32 percent when looking at much more recent data.

One caveat to the CAPM formula is that whenever warranted, experts add a specific risk premium. The specific risk premium is sometimes calculated using Ibbotson's size premiums, according to the company's size. In this case, Mirant's size warranted a size premium of 0.5 percent. However, some experts felt that an additional premium in excess of 0.5 percent was warranted. One expert applied an additional premium of 1 percent, another expert used 2 percent, and a third expert used 4.2 percent. The additional premiums of 1 and 2 percent were presented without substantiation. The additional premium of 4.2 percent was based on a study of distressed companies. However, following its emergence from bankruptcy, Mirant was not expected to be distressed.

Because Mirant had substantial operations in the Caribbean and Philippines, experts disagreed on the correct application of the WACC for its international cash flows. Some experts felt that the WACC should only be applied to the cash flows from domestic operations. Those experts split the DCF analysis into operations by geographic segment, then used different discount rates for the foreign cash flows based on the country's specific risk and comparable companies in those same areas.

Other experts disagreed with that method. They argued that most of the comparable companies used to calculate Mirant's WACC also have international operations, and therefore Mirant's WACC already reflects foreign operations. Moreover, one expert argued that studies have shown that foreign operations actually decrease rather than increase the risk, therefore the discount rates used for the foreign operations should not be higher than Mirant's WACC. Certain experts discounted all of Mirant's cash flows at the same discount rate, using its WACC.

Terminal Value. The final hotly contested and very significant factor of the DCF valuation was the determination of a terminal value. Because Mirant's own cash-flow projections covered 10 years and its business is expected to operate in perpetuity, an estimate of the value of the business at the end of the 10-year period was a necessary input for the DCF model. The experts all used two methods to calculate a terminal value: an EBTIDA multiple at the end of the 10-year period and an expected perpetual growth rate for Mirant's cash flows.

The terminal EBITDA multiples used by the experts were basically the same as the multiples calculated for the comparable company ("compco") valuation. Most experts took the range of multiples they had previously calculated and applied it to Mirant's expected EBITDA in 10 years. However, one expert who had determined a range of multiples for the compco of 8.0-9.0x used in his DCF analysis terminal multiples of 6.5x-7.5x. No substantiation was provided for lowering the multiple other than a feeling that multiples would be lower in the future. The effect of using the low range was to decrease Mirant's valuation substantially.

The terminal growth rates applied to Mirant's year-10 cash flow also varied among experts. One expert used a perpetual growth rate of 2 percent. Another expert used a range of 1-3 percent. A third expert opined that 3 percent was the absolute minimum growth rate that could be used for the terminal growth. His justification was that based on Mirant's own projections, its free cash flows were projected to grow at an annual compounded rate of 6.11 percent from 2006-14. In addition, he argued that the long-term inflation rate as calculated by Ibbotson was 3 percent. The expert's argument was that based solely on inflation, cash flows should be growing at 3 percent, even before accounting for real GDP growth.

Nonoperating Assets

Once the value of Mirant's operating assets was determined via the compco and DCF valuations, the experts added the value of Mirant's nonoperating assets that were not captured by either of those two methods. The most basic nonoperating asset was cash. Most experts added the value of the cash Mirant expected to have as of the date of emergence from bankruptcy. This methodology was consistent with that used to calculate the compco multiples where total cash was subtracted from the comparable companies' enterprise values. However, one expert added only the "excess cash" he calculated for Mirant as of the date of emergence. Excess cash was calculated by subtracting restricted cash and expected cash payments under the reorganization plan from total cash at emergence. By doing so, he decreased cash and Mirant's value by $800 million.

Another contested issue was the value of Mirant's net operating loss (NOL) that it could use to reduce future tax payments. Generally, experts agreed that Mirant was expected to emerge with large NOLs. However, experts were unsure as to the country of incorporation at emergence and its effect on the NOL. Experts also disagreed on the appropriate discount rate to use for calculating the present value of the expected tax savings resulting from the NOL's tax shield. Some experts felt that the applicable discount rate was the cost of equity because tax payments are made after debt is serviced and therefore available only to shareholders. Another expert advocated the use of the after-tax cost of debt because the NOL tax shield provided safe, reliable cash flows whose realization was nearly assured in the case of Mirant.

One of the largest differences in experts' opinions was the valuation of excess working capital. Some experts either did not perform any analysis of working capital or did not find any excess. One expert looked solely at accounts receivable to determine whether Mirant had excess accounts receivable outstanding. Based on historical-days sales outstanding, he determined that Mirant's expected excess accounts receivable is $200 million. Another expert looked at Mirant's working capital ratio (current assets/current liabilities). He compared Mirant's working capital ratio to that of the comparable companies and found that Mirant's was considerably higher than those of the comps. Based on the difference between Mirant and its peer group, he determined that $3.1 billion of liquidity was being withheld from creditors in the form of Mirant's excess of current assets.

Because of the broad range of assumptions used in the methodologies for valuing Mirant's operating assets as well as its nonoperating assets, the range in values calculated for Mirant was extremely wide. The lowest value concluded by an expert was $7.4 billion and the highest was $13.5 billion, with a difference of $6.1 billion. While the Mirant valuation hearing illustrates how complex and challenging those hearings can get, it also shows how important it is to have a strong expert to present complex issues in plain English as well as, whenever justified, to credibly impeach the other side's expert.

---

Footnotes

¹ Cowritten by Brad Orelowitz and Mark Marcus. The authors are, respectively, managing director, senior expert, senior manager and analyst at the Michel-Shaked Group in Boston. Allen Michel and Israel Shaked are also professors at Boston University's School of Management. MSG is a firm providing corporate finance, consulting and expert-witness services. Return to article