In accordance with the valuation policies set forth on the company's website, the values of TFG's CLO equity investments are determined using a third-party cash flow modeling tool. The model contains certain assumption inputs that are reviewed and adjusted as appropriate to factor in how historic, current and potential market developments (examined through, for example, forward-looking observable data) might potentially impact the performance of TFG's CLO equity investments. Since this involves modeling, among other things, forward projections over multiple years, this is not an exercise in recalibrating future assumptions to the latest quarter's historical data.
Subject to the foregoing, when determining the U.S. GAAP-compliant fair value of TFG's portfolio, the company seeks to derive a value at which market participants could transact in an orderly market and also seeks to benchmark the model inputs and resulting outputs to observable market data when available and appropriate. Please refer to the 2011 Annual Report for a more detailed description of the cash flow projection and discounting process.
2.1 Forward-looking CLO Equity Cash Flow Modeling Assumptions Recalibrated in Q4 2012:
The Investment Manager reviews and, when appropriate, adjusts in consultation with TFG's audit committee, the CLO equity investment portfolio's modeling assumptions as described above. At the end of Q4 2012, certain key assumptions relating to defaults were recalibrated. Those relating to recoveries, prepayments and reinvestment prices were unchanged from the previous quarter.U.S. CLOs - default assumptions recalibrated For the U.S. deals, near-term default assumptions were unchanged but medium-term default multiples were reduced to reflect, among other things, the perceived decline in concern over the so-called "maturity wall". These changes, which are detailed in the table below, had a positive impact on the undiscounted future projected cash flows of the U.S. deals.
Variable Year Current Assumptions Prior Assumptions CADR 1.0x WARF-implied 1.0x WARF-implied 2013 default rate (2.2%) default rate (2.2%) 1.0x WARF-implied 1.5x WARF-implied 2014 default rate (2.2%) default rate (3.3%) 1.25x WARF-implied 1.5x WARF-implied 2015-2016 default rate (2.7%) default rate (3.3%) 1.25x WARF-implied 1.0x WARF-implied 2017 default rate (2.7%) default rate (2.2%) 1.0x WARF-implied 1.0x WARF-implied Thereafter default rate (2.2%) default rate (2.2%) Recovery Rate Until deal maturity 73% 73% Prepayment Rate Until deal 20.0% p.a. on loans; 20.0% p.a. on loans; maturity 0.0% on bonds 0.0% on bonds Reinvestment Price Until deal maturity 100% 100%European CLOs - default assumptions recalibrated For the European deals, an elevated default multiple was maintained in the near term, but the medium term multiple was recalibrated higher bringing it in line with the U.S. deals, and reflecting some of the enhanced risks during that period, including the percentage of loans maturing. For European deals, this change resulted in a reduction in future undiscounted projected cash flows.
Variable Year Current Assumptions Prior Assumptions CADR 1.5x WARF-implied 1.5x WARF-implied 2013-2014 default rate (3.1%) default rate (3.1%) 1.25x WARF-implied 1.0x WARF-implied 2015-2017 default rate (2.6%) default rate (2.1%) 1.0x WARF-implied 1.0x WARF-implied Thereafter default rate (2.1%) default rate (2.1%) Recovery Rate Until deal maturity 68% 68% Prepayment Rate Until deal 20.0% p.a. on loans; 20.0% p.a. on loans; maturity 0.0% on bonds 0.0% on bonds Reinvestment Price Until deal maturity 100% 100%These key average assumption variables include the modeling assumptions disclosed as a weighted average (by U.S. dollar amount) of the individual deal assumptions, aggregated by geography ( i.e. U.S. and European). Such weighted averages may change from month to month due to movements in the amortized costs of the deals, even without changes to the underlying assumptions. Each individual deal's assumptions may differ from this geographical average and vary across the portfolio. The reinvestment price, assumptions about reinvestment spread and reinvestment life are also input into the model to generate an effective spread over LIBOR. Newer vintage CLOs may have a higher weighted-average reinvestment spread over LIBOR or shorter reinvestment life assumptions than older deals. Across the entire CLO portfolio, the reinvestment price assumption of 100% for U.S. deals and European deals with their respective assumed weighted-average reinvestment spreads, generates an effective spread over LIBOR of approximately 284 bps on broadly syndicated U.S. loans, 272 bps on European loans, and 328 bps on middle market loans.