A Better Approach to Asset Allocation for NDTs

Transcription

1 2014 NUCLEAR DECOMMISSIONING TRUST FUND STUDY GROUP May 18-21, 2014 A Better Approach to Asset Allocation for NDTs Consistent modeling and portfolio construction Guy Coughlan Bruce Jurin

2 Disclaimer This presentation is provided for informational purposes only. Do not use this presentation as a primary basis for investment decisions or for decisions pertaining to funding, accounting, or related regulatory requirements. In preparing this presentation, Pacific Global Advisors may have relied upon and assumed, without independent verification, the accuracy and completeness of information provided by various third parties such as investment managers. Pacific Global Advisors is not able to independently verify the accuracy and completeness of such information and makes no representation as to the information s accuracy or completeness. This presentation may contain projections, forecasts or estimates. Pacific Global Advisors makes various assumptions in connection with such forward looking information. Actual events or conditions may differ from those assumed and not all relevant events or conditions may have been considered in developing the assumptions. Changes to the assumptions could have a material impact on the information presented herein. Any forward-looking information contained in this presentation (such as illustrative cash flow, yields or returns) is based upon certain assumptions about future events or conditions and is intended only to illustrate hypothetical results under those assumptions (not all of which are specified herein). No representation is made that the performance presented herein will be achieved. Nothing contained herein should be construed as legal, actuarial or accounting advice. You must retain an actuarial firm which is independent of Pacific Global Advisors to provide all actuarial services with respect to statutory and regulatory requirements. NOTHING IN THIS PRESENTATION CONSTITUTES AN OFFER OR SOLICITATION FOR THE PURCHASE OR SALE OF ANY FINANCIAL INSTRUMENT OR A COMMITMENT BY PACIFIC GLOBAL ADVISORS AND ITS AFFILIATES TO ENTER INTO OR FACILITATE ANY TRANSACTION. Therefore, when considering whether to purchase any financial instrument, or otherwise participate in any transaction, no reliance should be placed on the information in this presentation. Such information is general, partial, preliminary and subject to change. The information contained in this presentation is not intended to be and is not warranted to be complete in all respects and Pacific Global Advisors expressly disclaims the completeness and accuracy of such information. Nothing in this presentation should be construed as a recommendation to purchase any financial instrument or participate in any transaction, or as legal, tax, regulatory or accounting advice. Actual events or conditions are unlikely to be consistent with, and may differ materially from, those assumed. Accordingly, actual results will vary and the variation may be material. Information about the past performance of issuers, financial instruments and markets should not be viewed as indicative of future results. THESE MATERIALS CONTAIN HYPOTHETICAL PERFORMANCE RESULTS. HYPOTHETICAL PERFORMANCE RESULTS HAVE MANY INHERENT LIMITATIONS, SOME OF WHICH ARE DESCRIBED BELOW. NO REPRESENTATION IS BEING MADE THAT ANY ACCOUNT WILL OR IS LIKELY TO ACHIEVE PROFITS OR LOSSES SIMILAR TO THOSE SHOWN. IN FACT, THERE ARE FREQUENTLY SHARP DIFFERENCES BETWEEN HYPOTHETICAL PERFORMANCE RESULTS AND THE ACTUAL RESULTS SUBSEQUENTLY ACHIEVED BY ANY PARTICULAR TRADING PROGRAM. ONE OF THE LIMITATIONS OF HYPOTHETICAL PERFORMANCE RESULTS IS THAT THEY ARE GENERALLY PREPARED WITH THE BENEFIT OF HINDSIGHT. IN ADDITION, HYPOTHETICAL TRADING DOES NOT INVOLVE FINANCIAL RISK, AND NO HYPOTHETICAL TRADING RECORD CAN COMPLETELY ACCOUNT FOR THE IMPACT OF FINANCIAL RISK IN ACTUAL TRADING. FOR EXAMPLE, THE ABILITY TO WITHSTAND LOSSES OR ADHERE TO A PARTICULAR TRADING PROGRAM IN SPITE OF TRADING LOSSES ARE MATERIAL POINTS WHICH CAN ALSO ADVERSELY AFFECT ACTUAL TRADING RESULTS. THERE ARE NUMEROUS OTHER FACTORS RELATED TO THE MARKETS IN GENERAL OR TO THE IMPLEMENTATION OF ANY SPECIFIC TRADING PROGRAM WHICH CANNOT BE FULLY ACCOUNTED FOR IN THE PREPARATION OF HYPOTHETICAL PERFORMANCE RESULTS AND ALL OF WHICH CAN ADVERSELY AFFECT ACTUAL TRADING RESULTS. Pacific Global Advisors and its affiliates do not provide tax advice. Accordingly, any discussion of U.S. tax matters included herein (including any attachments) is not intended or written to be used, and cannot be used, (a) in connection with the promotion, marketing or recommendation of any of the matters addressed herein to another person or (b) for the purpose of avoiding U.S. tax-related penalties. 1

3 Executive summary Asset allocation for NDTs involves long-term modeling of assets and liabilities The long-term is fraught with uncertainties and complexities Many asset allocation models face challenges Detailed and complex models of the liabilities, based on an engineering perspective, that don t necessarily reflect consistent economic relationships over the long-term Models of asset performance that are not completely consistent with the models used for the liabilities We recommend a simple approach anchored in economics Consistent macro-economic relationships reflected in assets and liabilities, and preserved in stochastic risk modeling Capture impact of regulatory constraints 2

4 Agenda A consistent framework for asset allocation 3 NDT costs: Long-term forecasts 9 Assets & liabilities: Risk drivers and sensitivities 14 Conclusions 24 Appendix 26 3

5 The asset allocation process involves choosing an efficient asset portfolio to meet the NDT liabilities Assets Liabilities Macroeconomic consistency Long-term scenario generator Forecasts Risks Contributions Consistent Scenarios Portfolio optimizer Selected portfolio Penalty functions Rebalancing rules 4

6 Implementation of realistic asset allocation models involves models for the long-term evolution of key variables Asset allocation requires determining: Long-term forecasts For the decommissioning liabilities, as well as asset returns Long-term risks For the liabilities, asset and ultimately the funding level Need to ensure that: Long-term forecasts are sensible Consistent with long-term evolution of macro-economic variables Long-term risks are realistic Consistent long-term relationships between the key variables Focus on simple models with the right macro-economic relationships between the key variables 5

7 Careful modeling of the liabilities is extremely important for asset allocation For NDTs (just as for DB pension plans) the liability impacts: Contributions into the trust The realizable value of any surplus assets Economic impact for the sponsor Seeking returns above what is needed to meet liabilities Often has little value Adds additional risk to the sponsor Not risk-reward efficient A high-return, high-risk investment strategy has an asymmetric payoff: Often limited upside Large downside Destroys value Liabilities become ever more important as decommissioning approaches 6

8 PGA s asset allocation framework takes account of regulatory considerations Need to reflect the rules, requirements and constraints of regulators: State Utilities Commissions Nuclear Regulatory Commission These vary by regulator, but generally there are common features Surplus funds Cannot be accessed without NRC permission, which is unlikely May be trapped in the trust for decades Volatility in funded status Large swings in funded status may make it difficult to get regulatory recovery Stability of funded status and contributions is valued by many regulators Consistent modeling of the impact of real interest rates and inflation Provides scope for evaluating hedging strategies to reduce the risk of Surplus funds that become trapped Wild fluctuations in contributions 7

9 Other factors to be considered in the asset allocation model The impact of tax adjustments to the asset returns Potential use of escrow accounts as a more liquid, but less tax efficient, investment vehicle Ability to take advantage of the illiquidity premium of asset classes, given the long-term nature of NDT liabilities Active rebalancing strategies as the decommissioning date approaches Potential to separate fuel storage from other decommissioning investments 8

10 Agenda A consistent framework for asset allocation 3 NDT costs: Long-term forecasts 9 Assets & liabilities: Risk drivers and sensitivities 14 Conclusions 24 Appendix 26 9

11 Forecasting long-term costs is challenging Forecasting costs and other economic variables over 30, 60, 100 or 200 years is challenging The structure of the economy may change significantly in terms of: Labor force make up (demographics and skills) Productivity Relationship between capital and labor Access to capital Technology may completely change the nature of decommissioning Past technological advances have led to: Products becoming obsolete Improved products, which are more effective, cheaper, and/or more functional Completely new, unimagined products, which Meet existing demand in better ways Create entirely new demand (e.g., the internet, smart phones, computers) Given the uncertainties, it doesn t make sense to overcomplicate the forecasts 10

12 US health care expenditure provides a useful example of the challenges of long-term forecasts Healthcare costs as a percentage of GDP have been an increasing since the 1940s Long term extrapolation of the trend without context can lead to unrealistic results Linear extrapolation of health care cost leads to expenditure forecasts of: 31% of GDP by % by After some point, health care cost cannot grow faster than GDP US health care expenditure as % of GDP 35% 30% 25% 20% 15% 10% Trend Extrapolation 5% 0% Source: Bureau of Economic Analysis 11

13 Traditional modeling of NDT cost escalation is often trend based Example: Energy cost component Engineer forecast of long-term growth of 1.79% Leads to energy costs in 2067 that are 2.81x the 2006 costs Energy cost growth forecast (%) But is this consistent with inflation? Energy costs fall by 38% in real terms (assuming inflation of 2.5%) Is this realistic? Energy cost projection (2006 = 100) Inflation (2.5%) Total energy cost Source: Docket No. E002/M~08-, Schedule B.1 12

14 Other cost components are typically modeled in a similar way and combined using fixed weights Cost components growth forecasts (%) Labor Equipment & Materials Energy Burial Other Composite Cost Component Fixed Weight Labor 59% Equipment & materials 17% Energy 3% Burial 7% Other 14% Total 100% Source: Docket No. E002/M~08-, Schedule B.1 13

15 Agenda A consistent framework for asset allocation 3 NDT costs: Long-term forecasts 9 Assets & liabilities: Risk drivers and sensitivities 14 Conclusions 24 Appendix 26 14

16 NDT cost escalation, asset returns and expected inflation are tied together via macroeconomic relationships NDT assets: Asset risk premium Return on assets = Expected CPI inflation Real interest rate NDT liabilities: Liability cash flows grow with CPI inflation Non-inflation escalation Liability discount rate = Return on assets 15

17 NDT liabilities depend principally on real interest rates and noninflation escalation Change in... Impact on value of liabilities Reason Real interest rate Real interest rates up Down Discount rate up CPI inflation up Expected CPI inflation No change Discount rate up and Cash flows up Non-inflation escalation Up Cash flows up Note: This ignores tax effects 16

18 Key variables for the NDT asset allocation Asset risk premium Real interest rate Driven by macroeconomic factors Expected CPI inflation Non-inflation escalation Tax rate Part driven by macroeconomic factors Part driven by other factors, e.g., regulation Because of taxes, impact of CPI inflation is not totally neutral 17

19 Macroeconomic factors have been stable over long history Real GDP growth of 2~2.5% is a reasonable long term forecast assumption Historically the breakdown of GDP between labor and capital has been fairly stable: Labor earns 52~58% of GDP Capital earns 42~48% of GDP Compensation of employees / GDP (%) 60% 58% 56% 54% 52% 50% 48% Source: Bureau of Economic Analysis 18

20 The stock market and GDP are also tied together over the long run The stability of the GDP growth and the stability of the capital share of GDP imply stable longterm stock market returns The return on many forms of capital is usually greater than GDP growth because it is levered E.g., equity capital, real estate GDP growth and average leverage in the economy determines long-term stock market returns GDP vs stock market (1929 = 100) 1,000, ,000 10,000 1,000 GDP Norminal SP 500 Total Return Source: Bureau of Economic Analysis; NYU Stern School of Business 19

21 Non-inflation escalation is specific to Nuclear Decommissioning Labor component Labor productivity = output per unit of labor The labor share of GDP growth reflects the increase in labor productivity This means that the cost of labor across the whole economy will rise in real terms with GDP, but so will productivity But for the decommissioning industry, taken in isolation, this may not be the case To be conservative, PGA assumes real labor costs increase with GDP, while productivity remains the same Non-labor component Across the economy, non-inflation escalation must total to zero But for the decommissioning industry, taken in isolation, this may not be the case In fact, the increase in cost from additional regulations has been substantial 20

22 Decommissioning costs have been difficult to estimate because of non-inflation escalation The most reliable estimate of the cost of decommissioning is percent of the construction cost MIT nuclear engineering professor David Rose, November Plant Size (MWe) License expiration Decommission cost (mm) Build cost (mm) Fort St. Vrain $195 $240 Shippingport $98 $79 Yankee Rowe $608 + $8 p.a. 2 $39 Maine Yankee $635 $231 Connecticut Yankee $820 3 Zion 1 & Zion 2 2 x , 1997 $1,100 4 ~$580 1 November 1985 issue of The Bulletin of the Atomic Scientists, p Decommissioning of Yankee Rowe took 15 years (5 times as long as it took to build it) and cost 608 mm (15 times the original build price of $39 mm) 3 Connecticut Yankee's draft updated cost estimate includes an increase of approximately $270 million over the cost estimate reported in November 2002 and $410 million over the April 2000 FERC rate case settlement 4 Zion 1 and 2 sold by Exelon to Energy Solutions in 2007 Sources: See Appendix 21

23 Sensitivities (durations) of nuclear decommissioning liabilities with respect to the key variables Example based on decommissioning in 2030s using DECON 60 year forecast Asset risk premium Duration = -20 years* Real interest rate Duration = -20 years* Expected CPI inflation Duration = +10 years* Non-inflation escalation Duration = +30 years * Tax adjusted Without the tax effect, the durations of the real interest rate and the asset risk premium would equal -30 years, the same as non-inflation escalation The greater duration of the liabilities comes from the fact that tax impacts asset returns only The duration of expected inflation comes only because the tax impact of inflation. Without the differential tax effect, the impact cancels 22

24 Improving the process Typical process Engineering consultant estimates decommissioning costs, using extrapolations in line with government publications Asset consultant performs a bottom-up analysis of expected returns, based on methodologies used for pensions Both analyses may be of high quality, but they use: Different methodologies Different assumptions for expected inflation Better process PGA performs analysis based on consistent assumptions and methodologies for both assets and liabilities Leads to a very different, but more realistic, result Regulator has made positive comments about this modeling approach 23

25 Agenda A consistent framework for asset allocation 3 NDT costs: Long-term forecasts 9 Assets & liabilities: Risk drivers and sensitivities 14 Conclusions 24 Appendix 26 24

26 The framework for asset allocation needs to ensure consistency across forecasts and risks Consistent macro-economic relationships Based on assumptions for real GDP growth, labor productivity Consistent long-term forecast of expected level of key variables Real interest rate, asset risk premium, CPI inflation, non-inflation escalation Consistent volatility and correlations relationships Ensure key relationships are preserved in risk modeling Ensure the same model is used for each of the key variables throughout In particular, use the same model for expected CPI inflation for both: Asset returns Liability cash flow growth Keep it simple and keep it anchored in the economics 25

27 Agenda A consistent framework for asset allocation 3 NDT costs: Long-term forecasts 9 Assets & liabilities: Risk drivers and sensitivities 14 Conclusions 24 Appendix 26 26

28 Sources for decommissioning examples Fort St. Vrain: Shippingport: Yankee Rowe: Maine Yankee: Decommissioning experience report. See: Connecticut Yankee: SEC 8-K filing for Uil Holdings Corp 12/05/03. See: &Find=connecticut+yankee&Page=1&List=Docs&Show=Each Zion 1 and 2: