Freddie Mac 2012 Annual Report Download - page 80

management judgment. Our models may not perform as well in situations for which there are few or no recent historical
precedents. We have adjusted our models in response to recent events, but there remains considerable uncertainty about
model results.
Models are inherently imperfect predictors of actual results. Our models rely on various assumptions that may be
incorrect, including that historical experience can be used to predict future results. It has been more difficult to predict the
behaviors of the housing and credit capital markets and market participants over the past several years, due to, among other
factors: (a) the uncertainty concerning trends in home prices; (b) the lack of historical evidence about the behavior of deeply
underwater borrowers, the effect of an extended period of extremely low interest rates on prepayments, and the impact of
widespread loan refinancing and modification programs (such as HARP and HAMP), including the potential for the
extensive use of principal reductions; and (c) the impact of the concerns about deficiencies in foreclosure documentation
practices and related delays in the foreclosure process.
We face the risk that we could fail to implement, operate, or adjust or use our models properly. For example, the
assumptions underlying a model could be invalid, or we could apply a model to events or products outside the model’s
intended use. We may fail to code a model correctly or we could use incorrect data. The complexity and interconnectivity of
our models create additional risk regarding the accuracy of model output. While we have processes and controls in place
designed to mitigate these risks, there can be no assurances that such processes and controls will be successful. This risk may
be elevated to the extent that we have difficulty attracting and retaining employees with the necessary experience and skills.
We have increased our use of third-party models. This may expose us to additional risk, as third-parties typically do not
provide us with proprietary information regarding their models. As a result, we may not fully understand the risks associated
with the use of such models.
Management often needs to exercise judgment to interpret or adjust modeled results to take into account new
information or changes in conditions. The dramatic changes in the housing and credit capital markets in recent years have
required frequent adjustments to our models and the application of greater management judgment in the interpretation and
adjustment of the results produced by our models. This further increases both the uncertainty about model results and the risk
of errors in the implementation, operation, or use of the models.
We face the risk that the valuations, risk metrics, amortization results, loan loss reserve estimations, and security
impairment charges produced by our models may be different from actual results, which could adversely affect our business
results, cash flows, fair value of net assets, business prospects, and future financial results. For example, our models may
under-predict the losses we will suffer in various aspects of our business. Changes in, or replacements of, any of our models
or in any of the assumptions, judgments, or estimates used in the models may cause the results generated by the model to be
materially different from those generated by the prior model. The different results could cause a revision of previously
reported financial condition or results of operations, depending on when the change to the model, assumption, judgment, or
estimate is implemented. Any such changes may also cause difficulties in comparisons of the financial condition or results of
operations of prior or future periods.
Due to increased uncertainty about model results, we also face increased risk that we could make poor business
decisions in areas where model results are an important factor, including loan purchases, management and guarantee fee
pricing, asset and liability management, market risk management, and quality-control sampling strategies for loans in our
single-family credit guarantee portfolio. Furthermore, any strategies we employ to attempt to manage the risks associated
with our use of models may not be effective. See “MD&A — CRITICAL ACCOUNTING POLICIES AND ESTIMATES”
and “QUANTITATIVE AND QUALITATIVE DISCLOSURES ABOUT MARKET RISK — Interest-Rate Risk and Other
Market Risks” for more information on our use of models.
Changes in our accounting policies, as well as estimates we make, could materially affect how we report our financial
condition or results of operations.
Our accounting policies are fundamental to understanding our financial condition and results of operations. Certain of
our accounting policies, as well as estimates we make, are “critical,” as they are both important to the presentation of our
financial condition and results of operations and they require management to make particularly difficult, complex or
subjective judgments and estimates, often regarding matters that are inherently uncertain. Actual results could differ from our
estimates and the use of different judgments and assumptions related to these policies and estimates could have a material
75 Freddie Mac