Exelon 2014 Annual Report Download - page 61

to Exelon’s current best estimate as to the timing of ceasing generation operations at the Oyster Creek unit in 2019. Generation has
successfully secured 20-year operating license renewal extensions for seventeen of its nuclear units (including the two Salem units
co-owned by Generation, but operated by PSEG), and none of Generation’s applications for an operating license extension has
been denied. For its remaining seven operating units, Generation is in various stages of the process of pursuing similar extensions
and has filed license renewal applications for six operating nuclear units and has until 2021 to seek license renewal for one operating
nuclear unit. Generation’s assumption regarding license extension for ARO determination purposes is based in part on the good
current physical condition and high performance of these nuclear units, the favorable status of the ongoing license renewal
proceedings with the NRC, and the successful renewals for seventeen units to date. Generation estimates that the failure to obtain
license renewals at any of these nuclear units (assuming all other assumptions remain constant) would increase its ARO on average
approximately $300 million per unit as of December 31, 2014. The size of the increase to the ARO for a particular nuclear unit is
dependent upon the current stage in its original license term and its specific decommissioning cost estimates. If Generation does not
receive license renewal on a particular unit, the increase to the ARO may be mitigated by Generation’s ability to delay ultimate
decommissioning activities under a SAFSTOR method of decommissioning.
Discount Rates. The probability-weighted estimated future cash flows using these various scenarios are discounted using credit-
adjusted, risk-free rates (CARFR) applicable to the various businesses in which each of the nuclear units originally operated. The
accounting guidance required Generation to establish an ARO at fair value at the time of the initial adoption of the current accounting
standard. Subsequent to the initial adoption, the ARO is adjusted for changes to estimated costs, timing of future cash flows and
modifications to decommissioning assumptions, as described above.
Under the current accounting framework, the ARO is not required or permitted to be re-measured for changes in the CARFR that
occur in isolation. This differs from the accounting requirements for other long-dated obligations, such as pension and other post-
employment benefits that are required to be re-measured as and when corresponding discount rates change. If Generation’s future
nominal cash flows associated with the ARO were to be discounted at current prevailing CARFRs, the obligation would increase
from approximately $7.0 billion to approximately $8.6 billion. The ultimate decommissioning obligation will be funded by the NDTs.
The NDTs are recorded on Exelon’s Consolidated Balance Sheets at December 31, 2014 at fair value of approximately $10.5 billion
and have an estimated targeted annual pre-tax return of 6.0% to 6.3%.
To illustrate the significant impact that changes in the CARFR, when combined with changes in projected amounts and expected
timing of cash flows, can have on the valuation of the ARO: i) had Generation used the 2013 CARFRs rather than the 2014 CARFRs
in performing its third quarter 2014 ARO update, Generation would have reduced the ARO by approximately $190 million as
compared to the actual decrease to the ARO of $125 million; and ii) if the CARFR used in performing the third quarter 2014 ARO
update (which also reflected increases in the amounts and changes to the timing of projected cash flows) was increased or
decreased by 100 basis points, the ARO would have decreased by $230 million and increased $40 million, respectively, as
compared to the actual decrease of $125 million.
ARO Sensitivities. Changes in the assumptions underlying the foregoing items could materially affect the decommissioning
obligation. The impact to the ARO of a change in any one of these assumptions is highly dependent on how the other assumptions
will change as well.
57