Discussion: Modeling the Budgetary Costs of FHA’s Single Family Mortgage Insurance Laurie Goodman, Center Director Housing Finance Policy Center Urban Institute MIT Center for Finance and Policy Inaugural Conference September 12-13, 2014 Cambridge, MA 1
13
Embed
Discussion: Modeling the Budgetary Costs of FHA’s Single Family Mortgage Insurance
Discussion: Modeling the Budgetary Costs of FHA’s Single Family Mortgage Insurance. Laurie Goodman, Center Director Housing Finance Policy Center Urban Institute MIT Center for Finance and Policy Inaugural Conference September 12-13, 2014 Cambridge, MA. Discussion. - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Discussion: Modeling the Budgetary Costs of FHA’s Single Family Mortgage Insurance
Laurie Goodman, Center DirectorHousing Finance Policy CenterUrban Institute
MIT Center for Finance and PolicyInaugural ConferenceSeptember 12-13, 2014Cambridge, MA
2
DiscussionModeling the Budgetary Costs of FHA’s Single Family Mortgage Insurance
• Forecast cash flows on FHA mortgages to calculate subsidy rates on loan guarantees– model prepayment and defaults– model loss given default– simulate future results using 1000 paths; variables include home prices,
interest rates and unemployment
• Estimate appropriate subsidy rate on a fair value calculation– use OAS to capture discount rate on securities with government guarantee– credit risk premium is derived from partial insurance coverage provided by
Quibbles• Greater transparency on actual and estimated cash flows
– Would allow researchers to experiment with different assumptions about discounting, other changes. At minimum, providing a duration measure for remaining balance by vintage year would be helpful.
• Loss-given-default results were not transparent and may not be indicative of current period– 85% of liquidated loans had mark-to-market LTVs of 70-100, reflecting
data ending in 2009. MSA level mark-to-market was used, with no accounting for price tier. No summary statistics were presented.
• More transparency on calculations of fair value subsidy rates
4
Larger Comments
• Models used to estimate prepayments, defaults, and loss given default were estimated over long period in which parameters are unstable– All errors accumulate in one direction; loans made in
later years will have longer average durations than loans made in earlier years
• Fair value subsidy rate calculation raises concerns– OASs are much higher than those calculated by Wall
Street firms; an explanation would be in order
5
Huge structural changes in past two decadesResult: Shifts in prepayment and default curves all point to mortgage loans remaining outstanding for longer periods of time• Homeowner mobility is down• The prepayment S-curve is flatter, as the costs of origination are up sharply• Foreclosure timelines have extended, modifications are a permanent part of the new loss
mitigation landscape• FHA mortgages are assumable, which will prove important in a rising rate environment
Additional analysis: The careful estimation of the prepayment/default models give a false sense of precision. It is important to recognize structural changes in the market, which could be done within the confines of their analysis.
Policy implications: These structural changes have important implications for FHA pricing, as the profitability of new FHA guarantees would rise sharply.
6
Homeowner mobility has been declining for decades
19891990
19911992
19931994
19951996
19971998
19992000
20012002
20032004
20052006
20072008
20092010
20112012
20130
2
4
6
8
10
5.1
Total owner-occupied mobility (percent of households)
Source: US Census Bureau, CPS Historical Geographic Mobility Tables.
Originators processing fewer loans per month, resulting in higher cost of originationApplications per retail underwriter per month
Source: Mortgage Bankers Association, Peer Group Program conducted by MBA and STRATMOR Group.
Prepayment curve by incentive for 30-year fixed rate GNMA product
Sources: Mizuho Securities and Urban Institute.
Refinancing incentive (bps)
Prep
aym
ent s
peed
(per
cent
per
ann
um)
AS-MMM0814
Foreclosure timelines have extended (PLS only)
12/1/1
0
4/1/1
1
8/1/1
1
12/1/1
1
4/1/1
2
8/1/1
2
12/1/1
2
4/1/1
3
8/1/1
3
12/1/1
3
4/1/1
4
8/1/1
40%
10%
20%
30%
40%
50%
60%
70%
80%
90%
% Delinquent Greater Than 18 Months
12/1/1
0
4/1/1
1
8/1/1
1
12/1/1
1
4/1/1
2
8/1/1
2
12/1/1
2
4/1/1
3
8/1/1
3
12/1/1
3
4/1/1
4
8/1/1
40
5
10
15
20
25
30
35
40
45
50
Liquidation lag (months to liquidate) % of Liquidations Greater than 18/24 Months DQ (% of Balance)
Source: Amherst Securities analysis of CoreLogic and 1010Data.
2007 (Q3-Q4)
2008 2009 2010 2011 2012 2013 2014 YTD0
1,000,000
2,000,000
3,000,000
4,000,000
5,000,000
6,000,000
7,000,000
8,000,000HAMP mods Proprietary mods Liquidations
Sources: Hope Now Reports and Urban Institute.Note: Liquidations includes both foreclosure sales and short sales.
Cumulative modifications and liquidations (millions of loans)
10
Modifications now a permanent part of loss mitigation landscape
11
12
Thoughts on fair value subsidy rate
• Calculated as present value (PV) of the guaranteed loan less present value of the non-guaranteed loan– PV of the guaranteed loan is calculated as MBS cash flows, discounted
by Treasury notes of comparable maturity + the option adjusted spread.
– PV of the non-guaranteed loan is calculated as cash flows of a portfolio of non-guaranteed loans discounted at the Treasury note rate + the OAS + the market risk premium.
• The OAS in this paper is far higher than is calculated by any Wall Street Firm, and merits an explanation or at least an acknowledgement