Pricing the Federal Guarantee of
Fannie Mae and Freddie Mac
Deborah Lucas
Congressional Budget Office
Any opinions expressed may not be those of the Congressional
Budget Office
2
Background
• Last year over 90% of residential mortgages originated
in the U.S. carried some type of federal guarantee
▫ Fannie & Freddie, FHA, VA
• Implicit federal guarantees thought to motivate excessive
risk-taking by GSEs
• Many proposals for the future of the secondary mortgage
market call for some form of explicit federal guarantees
▫ More or less limited, possibly only catastrophic
▫ Concern that a policy of “no guarantees” could be
interpreted as a return to implicit guarantees
• Risk-based pricing of federal guarantees provides an
incentive to limit risk-taking
3
Setting “risk-based prices”
• Policymakers, practitioners, and academics may have
quite different views of what this means in practice.
• Policymakers
▫ Many think fees must only cover expected guarantee
payouts
▫ No need to charge more for systematic (market) risk
▫ Government doesn’t need to “make a profit”
▫ And there are practical difficulties
 Even pricing to an average loss rate is difficult when tail risk is
the dominant risk
 There will be many years of zero cost (or a profit); hence cost
estimates based on recent historical data tend to be biased down
 Even when risk is recognized as a legitimate cost, there may be
no tools or infrastructure to quantify it.
4
Setting “risk-based prices”
• Practitioners (e.g., Fannie Mae)
▫ Recognize capital costs associated with guarantees
▫ But those capital costs don’t reflect the full cost of risk:
▫
Assets
Liabilities
Mortgages
Debt
Equity
……………
Fee Income
Gov’t Guarantee
…………
Guarantees
WACC = (D/V)rD +
(E/V)rE
▫ Capital is costly and guarantees increase capital needs
▫ Debt cost not affected so not priced in guarantee fees
• What is missing is the value of the federal guarantee
▫ Most of the systemic risk is absorbed by the guarantee – it is what
makes the debt safe and price-insensitive to risk
5
Setting “risk-based prices”
• A theory-based approach recognizes loan guarantees as put options
▫ Or equivalently, highly levered positions in the underlying mortgages.
• Systematic risk is concentrated in those guarantees
• That risk is expensive – fairly priced federally loan guarantees carry a
high risk premium
• Using risk-based (market) pricing in federal budgeting is important for
aligning the cost of policy alternatives with their economic effects
▫ e.g., Security sales by the GSEs at market prices should not generate budget
costs or budget savings
• What does an options-pricing approach suggest about the value of
federal mortgage guarantees?
▫ “Valuing Government Guarantees: Fannie and Freddie Revisited” (Deborah
Lucas and Robert McDonald), in Measuring and Managing Federal Financial
Risk, University of Chicago Press, 2010
▫ “An Options-Based Approach to Evaluating the Risk of Fannie Mae and
Freddie Mac” (Deborah Lucas and Robert McDonald), Journal of Monetary
Economics, 2006.
6
General conclusions of analyses
• The guarantee creates a wedge between the value of operating
assets and the market value of debt and equity equal to the present
value of the future stream of income generated by the guarantee.
▫ This affects the initial conditions for derivatives-based estimates.
▫ It also affects the government’s assessment of the value of its asset
holdings (e.g. Fannie, Freddie, AIG)
• The presence of an implicit guarantee does not fundamentally
change the relation between the volatility of levered equity and the
underlying assets (with some caveats).
▫ This leaves intact the standard equations underlying derivative-based
pricing approaches (KMV/Merton).
• Estimates based on bond spreads are upwardly biased when no
correction is made for the lower default rate for guaranteed firms.
▫ In comparison to a similar firm without an implicit guarantee,
guaranteed firms optimally default less often to preserve the value of
future lower borrowing costs.
7
Fair guarantee fees –- Options-based
estimates for Fannie Mae and Freddie Mac
Table 3: Base Case 2005 Guarantee Value Estimates
Fannie
Freddie
Horizon
10
10
Guarantee Cost
14.46
9.16
($ billions)
Premium Rate (bp) 20.53
16.46
Implied Equity
49.99
45.40
Value ($ billions)
Default Prob.
0.19
0.18
(risk-neutral)
Default Prob.
0.050
0.033
(actual)
Default trigger
1.08
1.07
(L/A)
Fannie
20
35.49
Freddie
20
29.50
27.01
55.78
22.91
48.73
0.34
0.34
0.084
0.059
1.13
1.11
8
Fair guarantee fees vary significantly with
asset value.
Table 7: 2005 Guarantee Value Estimates, Varying Initial Equity for
Fannie
Fannie
Fannie
Fannie
-5% assets
-10% assets
Horizon
20
20
20
Guarantee Cost 35.49
50.84
72.50
($ billions)
Premium Rate
27.01
44.64
80.81
(bp)
Implied Equity 55.78
36.02
23.84
Value
($ billions)
Default Prob.
.34
.43
.56
(risk-neutral)
Default Prob.
.084
.175
.394
(actual)