[Senate Hearing 113-35]
[From the U.S. Government Publishing Office]
S. Hrg. 113-35
RETURNING PRIVATE CAPITAL TO MORTGAGE MARKETS: A FUNDAMENTAL FOR
HOUSING FINANCE REFORM
=======================================================================
HEARING
before the
SUBCOMMITTEE ON
SECURITIES, INSURANCE, AND INVESTMENT
of the
COMMITTEE ON
BANKING,HOUSING,AND URBAN AFFAIRS
UNITED STATES SENATE
ONE HUNDRED THIRTEENTH CONGRESS
FIRST SESSION
ON
EXAMINING THE IMPORTANCE OF BRINGING PRIVATE CAPITAL BACK TO MORTGAGE
MARKETS, CONSIDERING WHAT MECHANISMS WOULD ACHIEVE THIS GOAL, LIMIT
TAXPAYER RISK, AND FACILITATE A STABLE AND LIQUID MORTGAGE MARKET
__________
MAY 14, 2013
__________
Printed for the use of the Committee on Banking, Housing, and Urban
Affairs
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Available at: http: //www.fdsys.gov /
_____
U.S. GOVERNMENT PRINTING OFFICE
81-438 PDF WASHINGTON : 2013
-----------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Printing
Office Internet: bookstore.gpo.gov Phone: toll free (866) 512-1800; DC
area (202) 512-1800 Fax: (202) 512-2250 Mail: Stop SSOP, Washington, DC
20402-0001
COMMITTEE ON BANKING, HOUSING, AND URBAN AFFAIRS
TIM JOHNSON, South Dakota, Chairman
JACK REED, Rhode Island MIKE CRAPO, Idaho
CHARLES E. SCHUMER, New York RICHARD C. SHELBY, Alabama
ROBERT MENENDEZ, New Jersey BOB CORKER, Tennessee
SHERROD BROWN, Ohio DAVID VITTER, Louisiana
JON TESTER, Montana MIKE JOHANNS, Nebraska
MARK R. WARNER, Virginia PATRICK J. TOOMEY, Pennsylvania
JEFF MERKLEY, Oregon MARK KIRK, Illinois
KAY HAGAN, North Carolina JERRY MORAN, Kansas
JOE MANCHIN III, West Virginia TOM COBURN, Oklahoma
ELIZABETH WARREN, Massachusetts DEAN HELLER, Nevada
HEIDI HEITKAMP, North Dakota
Charles Yi, Staff Director
Gregg Richard, Republican Staff Director
Dawn Ratliff, Chief Clerk
Kelly Wismer, Hearing Clerk
Shelvin Simmons, IT Director
Jim Crowell, Editor
______
Subcommittee on Securities, Insurance, and Investment
JON TESTER, Montana, Chairman
MIKE JOHANNS, Nebraska, Ranking Republican Member
JACK REED, Rhode Island BOB CORKER, Tennessee
CHARLES E. SCHUMER, New York RICHARD C. SHELBY, Alabama
ROBERT MENENDEZ, New Jersey DAVID VITTER, Louisiana
MARK R. WARNER, Virginia PATRICK J. TOOMEY, Pennsylvania
KAY HAGAN, North Carolina MARK KIRK, Illinois
ELIZABETH WARREN, Massachusetts TOM COBURN, Oklahoma
HEIDI HEITKAMP, North Dakota
Kara Stein, Subcommittee Staff Director
Brian Werstler, Republican Subcommittee Staff Director
Alison O'Donnell, Senior Economic Advisor
(ii)
C O N T E N T S
----------
TUESDAY, MAY 14, 2013
Page
Opening statement of Chairman Tester............................. 1
Opening statements, comments, or prepared statements of:
Senator Johan................................................ 2
WITNESSES
Mark A. Willis, Ph.D., Resident Research Fellow, Furman Center
for Real Estate and Urban Policy, New York University.......... 3
Prepared statement........................................... 34
Andrew Davidson, President, Andrew Davidson & Co., Inc........... 6
Prepared statement........................................... 42
Phillip L. Swagel, Ph.D., Professor, International Economic
Policy, Maryland School of Public Policy, University of
Maryland....................................................... 7
Prepared statement........................................... 65
Robert Van Order, Ph.D., Chair, Department of Finance and
Professor of Finance and Economics, George Washington
University School of Business.................................. 9
Prepared statement........................................... 70
(iii)
RETURNING PRIVATE CAPITAL TO MORTGAGE MARKETS: A FUNDAMENTAL FOR
HOUSING FINANCE REFORM
----------
TUESDAY, MAY 14, 2013
U.S. Senate,
Committee on Banking, Housing, and Urban Affairs,
Subcommittee on Securities, Insurance, and Investment
Washington, DC.
The Subcommittee met at 3:19 p.m. in room SD-538, Dirksen
Senate Office Building, Senator Jon Tester, Chairman of the
Subcommittee, presiding.
OPENING STATEMENT OF CHAIRMAN JON TESTER
Senator Tester. I would like to call this hearing to order
of the Securities, Insurance, and Investment Subcommittee
titled Returning Private Capital to Mortgage Markets: A
Fundamental for Housing Finance Reform.
I want to welcome our witnesses, and I look forward to
hearing from them this afternoon about some of the benefits and
challenges associated with bringing private capital back to the
mortgage markets. And, more specifically, we will get at how
different risk-sharing mechanisms will achieve this goal while
limiting taxpayer risk and facilitating a stable and liquid
mortgage market. This is a crucial issue within the context of
Fannie and Freddie as they consider and engage in different
risk-sharing mechanisms later this year and also as we consider
what the future of housing finance reform might look like.
We know that private capital will not automatically return
to this market without an understanding of how the new system
will function and what the rules of the road will be. Part of
that is understanding more fully the mechanisms that are being
considered to facilitate credit risk-sharing and seeing how
investors respond to the enterprises' trial run later this
year.
The FHFA has directed Fannie and Freddie to complete $30
billion in risk-sharing transactions and has directed them to
consider expanded mortgage insurance, senior/subordinated
securities and credit-linked securities.
While these mechanisms provide opportunities for the
enterprises to better manage credit risk, we must also
understand how they will limit taxpayer risk, impact mortgage
affordability and accessibility for consumers in all
communities and all markets, facilitate market liquidity and
stability, including in times of economic stress, interact with
the TBA market and scale over time.
We need to understand what different mechanisms would mean
for different market participants and players in the mortgage
market, like small financial institutions, for example, and how
these mechanisms should be structured and what levers exist to
ensure that private capital actually returns to the mortgage
markets and that it does not leave taxpayers on the hook if it
disappears in times of stress.
The good news is the housing market is showing signs of
strength, property values are rising, and the enterprises are
turning a profit, which once seemed unimaginable. The
announcements of Fannie and Freddie last week are good signs,
but they cannot lull us into complacency with the status quo.
And I do not believe they should be used as an excuse to leave
the enterprises in perpetual conservatorship.
Now is the time for us to be working toward solutions that
will bring private capital back into our mortgage markets and
create a vibrant and competitive mortgage market that is built
to last and able to withstand the next crisis. A marketplace
that uses instruments like the 30-year mortgage will keep the
dream of home ownership within reach of hardworking middle-
class Americans.
From my conversations with Members of this Committee, I
think the opportunities to build consensus are there and the
time to get serious is right now. I look forward to working
with my colleagues, who will roll up their sleeves and get to
work.
We have some great witnesses with us here today, and I am
looking forward to hearing from all of them as we drill down on
this important topic.
And, with that, I will turn it over to the Ranking Member,
Senator Mike Johanns.
STATEMENT OF SENATOR MIKE JOHANNS
Senator Johanns. Mr. Chairman, let me start out and just
say thank you for holding today's hearing on this very
critical, timely and important topic.
And I would like to thank our witnesses for their insights
and their expertise. I am going to have a very brief opening
statement because I am most interested in hearing from you.
I would say at the outset of the hearing, though, that I
believe there is a bipartisan consensus that is emerging around
the idea that the status quo with Fannie Mae and Freddie Mac is
not an acceptable approach. A nationalized housing market with
an implicit Government guarantee, I would argue, is not good
government.
I am hopeful that risk-sharing provisions such as those
discussed today will begin to form the basis for private
capital to come back into the housing market. The level of
risk-sharing, the amount of private capital, the percentage of
the Government guarantee--well, those are complex details and
details we need to come to grips with before we can go forward.
Perhaps today's hearing can start to shed some light on the
appropriate path forward. So I look forward to hearing from the
witnesses, forging ahead on the important work of reforming a
broken housing finance system.
Thank you, Mr. Chairman.
Senator Tester. Thank you, Senator Johanns.
Does any other Member have an opening statement?
OK. I want to welcome our witnesses here today--four folks
who have spent quite a bit of time working on housing issues--
and I want to thank them for their willingness to take time out
of their busy schedule to be here this afternoon.
First, we have Dr. Mark Willis, who is a Resident Research
Fellow at the Furman Center, and he teaches housing and
community development policy jointly at New York University's
Law and Wagner Schools. Before joining the Furman Center, Mark
was a visiting scholar at the Ford Foundation and has also
worked in community development at JPMorgan Chase and for the
city of New York.
Welcome, Dr. Willis.
Next, we have Mr. Andrew Davidson. He is the President of
Andrew Davidson and Company, a New York firm which he founded
in 1992. His firm specializes in the application of analytical
tools for mortgage securities and other asset-backed
securities. Prior to founding Andrew Davidson and Company, Mr.
Davidson worked at Merrill Lynch where he was a managing
director, producing research reports and analytical tools to
evaluate mortgage-backed securities.
Welcome to you, Mr. Davidson.
Then we have Dr. Phil Swagel. Dr. Phil Swagel is a
Professor at the University of Maryland School of Public
Policy, where he teaches international economics, and is an
academic fellow at the Center for Financial Policy at the
university's Robert H. Smith School of Business. Previously,
Dr. Swagel served as an Assistant Secretary for economic policy
at the Treasury Department, acting as a member of the TARP
Investment Committee, advising Secretary Paulson on all aspects
of economic policy.
Welcome to you, Dr. Swagel.
And last, but certainly not least, Dr. Robert Van Order is
the Chair of the Department of Finance and a visiting professor
of real estate and finance at George Washington School of
Business. Prior to his teaching career, Dr. Van Order served as
a chief economist at Freddie Mac from 1987 until 2003,
pioneering the development of Freddie's models of mortgage
default, prepayment and pricing. Prior to joining Freddie Mac,
Dr. Van Order served as a director of the Housing Finance
Analysis Division, for HUD.
Welcome to you, Dr. Van Order.
Each of you will have 5 minutes for your oral statement.
Your complete written testimony will be made part of the
record.
And I think we will start with you, Dr. Willis.
STATEMENT OF MARK A. WILLIS, Ph.D., RESIDENT RESEARCH FELLOW,
FURMAN CENTER FOR REAL ESTATE AND URBAN POLICY, NEW YORK
UNIVERSITY
Mr. Willis. Thank you very much, Chairman Tester, Ranking
Member Johanns and Members of the Committee. I thank you for
the opportunity to testify today on the role for private
capital in reforming mortgage markets.
My name, as you now know, is Mark Willis, a resident
research fellow at the Furman Center for Real Estate and Urban
Policy at NYU. I am on the Mortgage Finance Working Group
convened by the Center for American Progress and have done
research for the Bipartisan Policy Working Group.
My comments today reflect my own views and should not be
attributed to any of the organizations to which I am
affiliated.
I want to make two major points in my testimony.
First, restoring private capital's historic role in the
financing of home mortgages--that is the financing of large
jumbo mortgages--should be a straightforward matter once
regulatory uncertainties are resolved.
Second, requiring the use of private credit risk-taking
capital in front of a Government guarantee is also possible,
but imposing such requirements should only be implemented after
we have tested their impact on access to, and affordability of,
mortgages that serve the vast bulk of the home buyer market.
By test-driving different approaches, we will be better
able to weigh the costs and benefits of having private capital
take more of the risk and avoid unnecessarily disrupting the
availability and affordability of new mortgages.
I assume we are here because of a desire to decrease the
risk of loss to taxpayers by having private capital absorb some
amount of that loss.
Some argue that the private sector is better able than the
public sector to absorb and price the risk. However, we should
note that the Government had to bail out purely private credit
risk-takers whose mispricing helped fuel the subprime boom and
bust.
While it is a challenge for Government, or anyone, to set
exactly the right fee for providing a wrap, Government requires
less of a return than private risk takers do to provide that
guarantee.
It is also argued that sharing risk with private investors
could add an extra set of eyes to assess credit standards and
underwriting criteria and monitor whether the loans are being
properly underwritten and serviced.
Additionally, it is hoped that the active involvement of
private sector actors will discourage, if not prevent, attempts
by Government officials to fiddle with underwriting and other
standards for political gains.
While private capital is now, and always has been, the
source of all the funding of home mortgages, it consists mainly
of what are called rate investors. Credit investors, on the
other hand, have generally only funded so-called jumbo loans--
loans traditionally above $417,000.
FHA can take a number of steps to move us down the road of
housing finance reform in a measured and informed way.
Now that the housing market seems to have stabilized, it is
time to let the jumbo market again stand on its own without a
Government guarantee. It accounts, after all, for some quarter
of the dollar volume of per year and over 8 percent of all
mortgages by unit count.
The best way to trim back is to raise the g-fee on all
loans over $417,000 until the private sector is able to capture
as much of the market as it is willing to finance. By not
formally pulling out until the private sector has moved in, the
Government can avoid inadvertently leaving big gaps in the
marketplace.
Once the securitization market for jumbo mortgages is
functioning at scale, it will be possible to see how willing a
private mortgage market is to offer mortgages that are
comparable to what is available in the conforming market,
including, for example, long-term, fixed-rate mortgages which
are well priced and available without regard to geography or
other factors that would limit access to those that now have
it.
If the loans are comparable, then the limit could be
reduced again in stages. At each stage, the same test for
comparability should be applied.
On a second path, FHFA should also continue its current
quest to determine the costs and constraints of bringing in
private credit risk-taking capital ahead of Fannie and Freddie.
Shared risk does offer the potential to reduce the burden that
could ultimately fall on taxpayers.
However, using such private capital has drawbacks as well--
higher cost to borrowers and potentially tighter underwriting
standards. Moreover, requiring private capital to take first
loss will limit the Government's ability in times of economic
stress to ensure the continued availability of mortgage
financing unless a way is found to dial back that requirement
when, as we saw not so long ago, private capital quickly
abandoned the mortgage market.
Of all the options discussed to share risk, only those
involving insurance seem compatible with the continued
availability of well priced, longer-term, fixed-rate mortgage
products with rate locks from 30 to 90 days. The two main types
that are often most mentioned are mono-line companies and
credit-linked notes, and these have differences based on
regulatory requirements plus the functioning of the capital
market versus an insurance market.
In addition to testing the cost and viability of different
options for sharing risk, it is important to be able to assess
the tradeoff between the cost to the borrower and the degree of
risk-sharing.
Another focus for reform is to ensure that the lower cost
of funds made possible by the Government wrap pass through to
borrowers.
As for the transition, once a determination is made as to
the degree of risk-sharing that is optimal, the provision of
the Government wrap can be moved to another entity such as the
Government National Mortgage Association known as Ginnie Mae.
The remaining functions in Fannie and Freddie could then
continue in a new legal entity.
New entrants should also be allowed, if not encouraged, to
compete with the successors to Fannie and Freddie in
securitizing mortgages that are eligible for the Government
wrap.
Let me conclude by saying that by running tests that
provide market-based information we will able to proceed in a
measured and informed way and so make sure that we do not
unnecessarily and unintentionally impair access to, and
affordability of, housing finance to the vast bulk of the
housing market.
Thank you.
Senator Tester. Thank you, Dr. Willis. We appreciate your
testimony.
Mr. Davidson, proceed.
STATEMENT OF ANDREW DAVIDSON, PRESIDENT, ANDREW DAVIDSON & CO.,
INC.
Mr. Davidson. Chairman Tester, Ranking Member Johanns and
the Members of the Subcommittee, today we are discussing the
financial structure of the mortgage market. In my written
statement, I discuss rates investors and credit investors and
how to bring them back into this market. However, it is
important to keep in mind that mortgages are more than just a
bundle of financial risks.
Mortgages are a set of legal documents that create
obligations for borrowers and lenders.
Mortgages are a complex payment system that transfers
capital from investors to borrowers and mortgage payments back
from borrowers to investors.
Mortgages are an instrument of Government economic and
social policy as well as monetary policy.
And mortgages are the hopes and dreams of millions of
Americans seeking a better life.
Because of the multiple dimensions of mortgages, housing
finance reform is both difficult and essential.
Today, there are trillions of dollars invested in GSE-
guaranteed MBS. This investment largely flows through the TBA
market, which is an essential component of our housing finance
system. Unfortunately, that market now relies almost completely
on the U.S. Government and taxpayers to bear the credit risk.
There is a substantial amount of private capital currently
bearing credit risk, just not in the GSE market. Due to the
poor performance of the underlying loans in the private-label
market, many securities that were purchased with the idea that
they were low-risk investments are now subject to credit risk.
There is about $600 billion of investment in credit-
sensitive bonds, much of it by firms that bought the bonds from
the original investors. The existence of this large market has
created tremendous analytical and investing expertise in
mortgage credit risk.
Thus, the issue is not so much how do we return capital to
the mortgage market but how to structure the market so this
capital and expertise is deployed for new GSE-guaranteed
mortgages and not just for legacy private-label mortgages.
Capital markets mechanisms, such as senior/subordinated
bonds and credit-linked notes, can serve this purpose. If
properly structured, these approaches will not impose an undue
burden on borrowers.
My firm analyzed the cost of private-market credit
enhancement for the Bipartisan Policy Center's Housing
Commission. While we found a wide range of costs, depending on
the risk characteristics of the loans, we also found that
credit costs using private markets would be about the same as
the current guarantee fees charged by Fannie Mae and Freddie
Mac for similar quality loans as they are originating now.
The choice between senior/subordinated bonds and credit-
linked notes, as well as other approaches, are primarily
regulatory as opposed to economic. All securitizations are a
particular pathway through a thicket of regulations that affect
disclosure, tax treatment, regulatory capital and other
operational requirements. Finding the right path has slowed the
process of Fannie Mae and Freddie Mac utilizing these
structures.
Capital markets solutions will generally be most effective
if they allow the broadest range of investors for both the
guaranteed securities and the credit-sensitive securities.
Protection from double taxation, exemption from securities
registration, allowing reinvestment from both an SEC and tax
perspectives, and simplifying risk retention rules are all
necessary. CFTC oversight and CFPB QM requirements also need to
be addressed. Clear and consistent rules would increase
liquidity and lower the cost to borrowers.
While the use of these instruments will lower the risk to
taxpayers from Government guarantees of MBS, the stability of
the mortgage finance system, and many other goals associated
with Government involvement in the mortgage market depend more
on the industrial organization of the mortgage market than the
form of credit enhancement. Thus, the success of any system of
housing finance will also depend critically on the path to get
from the current structure of the housing finance system to the
desired outcome.
While there is much to commend the idea of shutting Fannie
and Freddie down and starting again, I believe the best path
forward will be to transform the GSEs from what they are to
what we would like them to be. In this way, there are three
possible paths:
One, we capitalize the GSEs to shareholder-owned companies.
We tried this, and it did not work.
Two, nationalizing the companies and have them operate as
Government-owned corporations--I do not think this will work
either.
Or, three, transform Fannie Mae and Freddie Mac into
issuer-owned cooperatives. Freddie was originally owned by
mortgage lenders, and I believe we should return to that type
of structure.
Cooperatives offer several advantages. They can establish
standards and provide access to all originators. They can be
effective at monitoring loan quality. They can provide a
mechanism for risk retention and true sale treatment.
Cooperatives, using capital markets credit instruments, can
provide a stable source of financing through periods of stress.
And, most importantly, the TBA market can be maintained
throughout the transition from conservatorship to cooperative
ownership so that mortgages can continue to fulfill their many
roles.
Thank you for your interest in my comments. I look forward
to your questions.
Senator Tester. Well, thank you, Mr. Davidson. We
appreciate your comments.
Dr. Swagel.
STATEMENT OF PHILLIP L. SWAGEL, Ph.D., PROFESSOR IN
INTERNATIONAL ECONOMIC POLICY, MARYLAND SCHOOL OF PUBLIC
POLICY, UNIVERSITY OF MARYLAND
Mr. Swagel. Thank you. Thank you, Chairman Tester, Ranking
Member Johanns and Members of the Committee. Thank you for the
opportunity to testify on the vital topic of returning private
capital to mortgage markets.
The Government now stands behind more than 90 percent of
new mortgages, distorting the economy and putting taxpayers at
risk. Bringing in private capital is the essential first step
in housing finance reform.
This can take several forms. One would be the capital of
new firms that compete in conforming securitization. A second
form would be private-label securitization and balance sheet
lending; that is, mortgage origination with no Government
guarantee. And the third would be risk-sharing under which
firms would sell non-guaranteed tranches of guaranteed MBS or
use other forms of risk-sharing.
This can proceed without legislative action, and indeed,
under the FHFA strategic plan it is moving forward. Ultimately,
however, investors will require legislative action for this to
take off in scale. After all, private-market participants will
naturally hesitate to invest in 30-year mortgages unless they
understand the rules in the future.
The policy levers to bring in private capital include four
possible levers. One is raising the price of the Government
guarantee. Two is reducing the quantity of insurance offered by
the Government or otherwise narrowing the scope of the
mortgages eligible for the Government guarantee. Number three
would be opening the housing finance system to new competition
that, in turn, brings in private capital. And four would be
requiring the firms that securitize Government-insured MBS to
arrange for the first loss private capital to take losses
before the Government guarantee. My written testimony provides
details on several forms of this first loss capital.
Reducing or eliminating the Government role in housing
finance involves moving forward with all four policy levers.
As reform moves forward, at first, all conforming mortgages
will still be guaranteed by the Government but behind first
loss private capital at the MBS level.
As the price for the Government insurance increases and as
more private capital is required in front of mortgages to
receive a Government guarantee, eventually, not all conforming
mortgages will receive the Government guarantee. Some
conforming mortgages will choose to be originated without a
guarantee, and the share of the Government in housing finance
will go below 100 percent, or the nearly 100 percent it is
today.
Eventually, as the pricing of the guarantee fee becomes
high enough and as so much private capital is required,
eventually, no MBS securitizers will purchase the Government
guarantee. Eventually, that would be a fully private market.
So, in other words, to reach an outcome of a fully private
market, at first the housing financing system must transition
through the intermediate steps in which the Government
guarantee recedes. There is a sense in which it is useful to
formalize the Government guarantee so that it can recede.
Interest rates for mortgages will rise as reform proceeds,
reflecting the compensation demanded by private investors for
taking on housing credit risk. In a sense, this reflects the
fact that the previous system was undercapitalized. Whether it
is possible for reform to arrive a fully private system depends
on the social and political reaction to these higher mortgage
interest rates.
I worry that a housing finance system that is notionally
private will inadvertently recreate the implicit guarantee in
the previous system. In my view, it would be better for the
inevitable Government involvement in housing finance to be made
explicit.
Taxpayers should be compensated for taking on housing risk.
Rather than leaving it implicit, they should be paid for
providing a guarantee, and there should be considerable private
capital ahead of the secondary Government guarantee.
Housing finance reform that brings back private capital
should proceed immediately even without resolving the question
over the eventual role of the Government.
The policy levers required to move forward with reform are
the same for all of the options under consideration, including
a system that is fully private, or at least notionally fully
private.
Bringing in private capital is the essential first step in
reform.
Thank you very much.
Senator Tester. Thank you, Dr. Swagel.
Dr. Van Order.
STATEMENT OF ROBERT VAN ORDER, Ph.D., CHAIR, DEPARTMENT OF
FINANCE AND PROFESSOR OF FINANCE AND ECONOMICS, GEORGE
WASHINGTON UNIVERSITY
Mr. Van Order. Thank you, Senator Tester and Committee
Members. I am very pleased to be here.
This, obviously, is an important topic. I would like to
start out also by saying it is actually a very difficult topic,
intellectually as well as policy-wise. There are lots of
different structures that have been tried in the United States
for mortgages, and around the world. In some ways they are very
similar, and in some ways they are very different, and they all
have flaws.
The topic, it seems to me, at hand is particularly about
long-term fixed-rate mortgages. Adjustable-rate mortgages can,
and have been, done around the world very easily by banks and
funded with deposits.
The question mostly for us, I think, is funding them
through capital markets and how you do it.
There seems to be a consensus emerging of the Government as
a kind of backup, as putting it as the last guarantor and
trying to get as much capital in front as you can. And I think
that is right and appropriate.
It is very much analogous to the role of deposit insurance
and the way they have historically funded things like
adjustable-rate mortgages and other things, where the
Government backs up a Government insurance company in a bank
that has capital in front of it. These are very common
structures.
One of the things I want to do--I have three points.
The first one I want to talk about is what actually private
capital is because it is actually quite ambiguous. Right now,
something like 90 percent of the mortgages in the country are
done through companies that are more or less owned by the
Government. But, if you went back a few years, Fannie and
Freddie actually, you could argue, had private capital; that
is, they had private shareholders.
If you went back a longer time ago, you saw the savings and
loans that actually clearly had private capital. They all got
into trouble.
And it seems to me while private capital is an important
way of thinking of it, what is really important is where the
risk is and how the risk gets controlled. Both of those are
important.
And I think this structure of having the Government at the
end is an interesting one. It is also not a very new one, and I
think that is one of the things I first want to remind you of.
When I first got to Freddie Mac, I was curious about
guarantees and things because, you know, it was what my
employer had. And we took a look historically and asked the
question, what fraction of mortgages have benefited from
guarantees from the Government, either directly from FHA, from
deposit insurance or from GSEs?
The answer was it had always been about 90 percent. The
only difference had been the structure of it. And that is about
what it is now.
The only time it was actually less than that was from about
2003 until 2006, with the private-label market.
This is not an unusual state of affairs. It is a common
one. And the question is, what is the structure to manage it?
Second, I have included in my testimony an appendix which
is a little bit of data on credit risk and what actually has
happened over the last few years. And I think it is important
to understand, when we get into regulating things and talking
about changes, what was and was not the problem.
What was not the problem, particularly, was low downpayment
loans or low-income loans. They do have high default rates.
They are sometimes a problem. But, in terms of the lift, what
was it that really went south?
It was not particularly those. They did do badly, but the
increase in their costs and decline in the credit quality was
roughly the same as for lots of other products.
In a period like we had a few years ago, when property
values fall by anywhere from 25 to 50 percent, even loans with
big downpayments suddenly have negative equity, and even
borrowers that have good credit history and have high incomes
default more.
What did seem to matter was two things:
One was the year in which the loan was originated. That is
an easy one. If the loan was originated in 2003, that was
golden, not because the world was necessarily better--it was a
little bit--but because property values went up like crazy for
the next few years, and that covers a lot of mistakes. Loans
originated in 2006, with the same observable characteristics,
were way worse.
The second thing that mattered was the channel. There are
some data--neat data--that I, as an academic, can get and lots
of data now are proprietary, but some free data from FHFA on
defaults by GSE loans and by private-label loans. The private-
label channel was worse. Not only was it worse in the sense
that on average it had higher defaults, but the lift, the
increase in defaults, from 2003 to 2006 was worse.
The point here partly is in looking at this a lot of what
happened is stuff that is difficult to observe from the
outside. There were some changes in the downpayment structure,
some changes in the credit score of borrowers, but those were
not nearly enough to explain what happened to defaults. It was
a bunch of different things.
And what I want to do finally is to talk about setting up
incentive structures and push for contingent capital,
particularly from the standpoint of management having a stake
in the downside. So one of the things I want to leave you with
is a proposal for various types of management incentive
programs which look like holding bonds in a company, so that
while they get the upside if they hold shares in the company,
they also are forced to participate in the downside because
their bonds will be converted into capital.
Thank you.
Senator Tester. Thank you, Dr. Van Order. We appreciate it
very much.
Thank you all for your testimony.
I think we will do 5-minute rounds and 5-minute questions,
and we will do as many rounds as you want.
I am going to start, and I will not direct it to anybody. I
assume you all want to put your two bits in.
I want to start by drilling down a little bit to better
understand how credit risk-sharing mechanisms might impact
mortgage affordability--something that you have all addressed.
The very basic question, which is, how exactly would these
mechanisms be priced in the cost of a mortgage or mortgage
rate?
You can start, Dr. Willis.
Mr. Willis. So the cost to the borrower includes--let's say
there is a Government wrap and there is private capital ahead
of it. So it would include the sum of the cost of the
Government wrap.
Let us look at the example of Ginnie Mae, which charges a
fee for its wrap. And I think in the proposed idea here of a
Government wrap for a conforming market on MBS going forward,
there should be a charge based on that risk, and then the
private sector requires a return on their taking the risk.
There is a huge difference between what Government needs to
charge and what the private sector needs to charge. The private
sector requires a large reward for taking risk.
The Government, on the other hand, in economics jargon, is
not risk-adverse. It can and, under the Federal Credit Reform
Act, does charge for its risk at a much lower price. It looks
to get the full money back. It looks to get its return on an
expected basis based on the Government borrowing rate, but that
yields a much lower cost here.
And there is a debate about whether the Government should
be charging the same as the private sector, but as long as it
does not and is being properly compensated under the Federal
Credit Reform Act, any change in the mix between what is the
risk the Government wrap is taking versus the risk to the
private sector affects the cost. The more you increase the
share that is private sector, the more you are increasing the
cost, the more expensive the cost of insurance.
Senator Tester. Mr. Davidson.
Mr. Davidson. Yes, there are two different components we
need to think about the cost.
One is the cost of credit enhancements; let's say there is
a 3 percent capitalization requirement. Or, let's say a 4
percent subordination that is going to be from the credit
buyers. The other 96 percent is basically being covered by the
rates investors. And so the most important thing is protecting
that 96 percent, in terms of the cost to the borrowers, and
that is really why this Government guarantee becomes important.
And not in good times, but in bad times, the spread
between, let's say, treasuries and mortgage rates can go up a
lot if there is no confidence in investing in those loans.
On the other hand, if we just look at the cost of the
credit side--you know, let's say in good times--and, let's say,
3 to 4 percent is what needs to go into this credit enhancement
piece. And then the borrower--the investors require, let's say,
10 percent to 15 percent excess returns. That just translates
into a 30 basis points to 60 basis points cost of credit.
The GSEs are currently charging about 50 basis points.
So to that 30 basis points at the low end you have to add
operational costs and the costs for a wrap. That is at the 50
basis points.
And so it might raise guarantee fees 30 basis points more.
Or, if we want to broaden from where we are now, it might be a
little bit more.
But we are not talking about changing mortgage rates by 50
or 100 basis points. We are talking within the range of where
we are now.
Senator Tester. Good.
Dr. Swagel.
Mr. Swagel. Chair, I will just add two things.
I agree on the range of pricing. I mean, the Fed is going
to be raising interest rates by a lot more, and that is going
to be a lot more noticeable.
Senator Tester. That is true.
Mr. Swagel. And it is important to have this layering of
risk--the loan level capital, the real downpayments, private
mortgage insurance with real capital and then the MBS level
capital.
So, two small notes in addition. The guarantee fee pricing
could actually go down in the future? I think it is going to go
up. It should go up now. The Government is not being
compensated. But when there is enough private capital, a good
deal of private capital, then the insurance the Government is
providing is less costly; it is less risky.
So this can go in both directions.
You know, on the other hand, I do not think the Government
is all that great an investor, and I think it is important to
make sure that Government is properly compensated for taking on
this risk.
And one benefit of the way that the risk is calculated now
is the Congressional Budget Office, I think, does us all a
favor by calculating the risk on the same basis as the private
sector and really protects the taxpayers from selling the
insurance or counting a profit for selling underpriced
insurance.
Senator Tester. Dr. Van Order.
Mr. Van Order. Yes, I agree on the accounting dimension.
It seems to me that there is a lesson to be learned from
the private-label market in terms of structuring. The risk-
sharing in the private-label market and the risk-sharing in
corporate structures in general is not so much sharing as
prioritizing.
And I think the idea is to set up structures that look like
that, where you know exactly who takes the first loss, exactly
who takes the second loss, the third, the fourth, and you put
the Government at the end, saying, yes, there really is a
guarantee. This makes it much easier to do things like a TBA
market, but to share the risk by slicing it in different
directions and really understanding the priorities.
One of the things with the structuring was there really
were different clienteles that took the different pieces of the
deals. It blew up. There were some problems with it. But the
underlying idea that there were different clienteles who would
take different strips of the risk and having the Government at
the back is what we have been doing for the last 50 years in
one way or the other.
Senator Tester. OK, Senator Johanns.
Senator Johanns. Thank you, Mr. Chairman.
Let me focus on some specific items that are of interest to
me. I want to start out with the GSEs.
You know, we turn the clock back a few years, and that was
a mess, to say the least. Do you have a sense or, maybe beyond
a sense, do you have proof that you could offer to me that the
GSEs are slowly working their way out of the mess, becoming
profitable; there is a light at the end of the tunnel?
Anyone one of you.
Mr. Willis. The piece that I would suggest to think about
here is the g-fee, which is now more than twice of what it was
historically. The main losses in Fannie and Freddie were in the
Alt-A piece. They are not making Alt-A loans now.
Senator Johanns. Right.
Mr. Willis. So it appears that they have more--the fee that
they are charging relative to the risk they are taking seems to
be very favorable.
So, if you look at those ratios I talk about, it seems
pretty clear why they are making money on the additions to the
portfolio--the mortgages they are adding today.
Mr. Davidson. I do not think we can really talk about the
profitability of the GSEs without having a capital model in
place and some measure of what risks they are taking on and
whether or not they are properly insuring through enough
capital for those risks in a bad environment.
But even beyond that, my sense is at least the new book of
business is well capitalized, or is earning enough return that
would cover a reasonable amount of capital for a downside risk.
But it is hard to go beyond that.
Senator Johanns. It is hard to go beyond that.
Mr. Swagel. Yes.
Senator Johanns. And it is hard to have a crystal ball and
say, well, the real estate market is going to continue to
improve, and I mean if, if, if.
But what I am getting to here is it seems to me that there
is healing going on. There is healing in the market. There is
healing in the book of business they are doing. They have
changed the way they have done business.
So this kind of leads to my next question. If, in fact,
that is the case, is there a point then where we can say to the
GSEs, if you want to go out there and do business in the new
world, fine, but pull back on the Government guarantee? Is that
a doable model?
All of you have talked about Government guarantee and this
and that, but I am really getting down to brass tacks here. Is
it a doable model to move them off the Government guarantee
program--because I think that gets you in trouble--and have a
system that is capitalized through the private sector?
Mr. Van Order. Yes, I would say I think it is doable, but
there is always the question.
Fannie and Freddie never had an explicit guarantee. They
always wrote on their paper that this not guaranteed by the
Government, which almost certainly meant that it was because no
one else had to put that on their securities.
Senator Johanns. And 100 and some failed----
Mr. Van Order. So how do you do this; you see, I am going
to get you off the guarantee, but this time we mean it?
It seems to me the way to do it is in the context of this
structure of having the Government at the end of the queue,
explicitly there, and then having capital rules in terms of
minimum capital requirements.
But I also like stress tests. And I am going to come back
to my retention capital. I like the idea of having a part of
their capital, a part of their debt, be something that when
times get tough can be automatically converted into equity, to
give incentives to management and to have angry bond holders
who have seen suddenly seen their bonds converted into
worthless--not worthless, but shares.
So I think you can do it. I think you do not want to
pretend to not have the guarantee. I think you want to have it
there, keep it at the end and as small as you can, and work
your way back.
Mr. Swagel. Yes.
Senator Johanns. Go ahead.
Mr. Swagel. I was going to say very quickly I think it can
recede, that exactly what you said can happen. It is just going
to take place slowly.
The market is dominated by the Government. We are not going
to get to zero, or something close to zero, quickly. But
bringing in private capital is the first step.
And I think eventually then the guarantee needs to be
formalized. Rather than guaranteeing Fannie or Freddie as
firms, transfer the guarantee to the MBS themselves and then
let that guarantee shrink.
Senator Johanns. You know, I am running out of time here,
but here is the thought that goes through my mind. It seems to
me that the GSEs are out there. You are right; there is not
really an explicit guarantee. But guess what? Taxpayers stepped
in and bailed out the system. Without that, we do not know
where we were at.
The challenge I think we face in going forward is
politicians, of course, have a tendency to want to improve upon
every system, and so they come in and say let's do some more of
this or more of that. And all of a sudden there are some quirky
things going on, and all of a sudden there are some mortgages
that probably are not very good. Then we get down to a point
where we were a few years ago.
My philosophy is if we as politicians want to boost home
ownership, do a program to boost home ownership. Put money
into. Appropriate money for it. Buy down the interest rate. Do
whatever to try to boost home ownership.
And let's just be honest. That is what we are doing.
I think what we ended up with, to me, seemed like kind of a
hybrid thing. Well, we did not really guarantee. But guess
what? We did guarantee because we had to step in.
So that is what is going through my mind, and maybe there
will be some follow-up questions to try to see if that makes
any sense or not.
Thank you, Mr. Chairman.
Senator Tester. You bet.
Senator Reed.
Senator Reed. Thank you, Mr. Chairman.
And, gentlemen, thank you for your testimony.
Dr. Swagel, I thought in reading your testimony you make
the point that even for those who want a privatized system a
first step is some type of Government guarantee to begin the
process.
In fact, I think you made an excellent point that the
option the Treasury laid out can be viewed not just as options
but as actually three necessary steps to get to a final end
point, which is a privatized system with a standby residual
Government role.
Could you elaborate on that? Is that the concept you have?
Mr. Swagel. Absolutely. It feels like a long time ago--
February of 2011--when the Treasury put forth their paper, and
they put forth the three options with the receding Government
guarantee.
And what was striking to me was that they were presented as
three separate options, but to get to the first one you have to
go through the third and second. It really is as the guarantee
recedes, and that can recede by having the Government guarantee
fewer mortgages or by, as we have all said, having the amount
of capital in front of the guarantee go up.
But they are all in the same line of the guarantee
receding. You have to go through those stages.
Senator Reed. Right. Now one of the things I think you all
generally pointed out is that if you have MBS, for example, you
want the first loss to be borne by a private insurer and that
insurer has to be well capitalized.
So it raises several issues. One is, what do you mean by
well capitalized, and two, who is going to make sure they are
well capitalized?
So do you have any----
Mr. Swagel. Yes. I mean, I was at the Treasury during the
crisis, and obviously, AIG used to be AAA.
So, the private capital--there are lots of different ways
of doing it. We want to make sure it is real capital. We want
to know who exactly takes the losses----
Senator Reed. Right.
Mr. Swagel.----and to make sure the loss-bearing capacity
is there.
So, in a sense, I think that is a nice part about the kind
of so-called A-B structure, that the people who own the B
pieces--they are taking the losses and everyone understands
that.
Senator Reed. A final point too is that in your testimony
you point out that--in fact, I think--am I over time?
Senator Tester. You are fine.
Senator Reed. Well, fine. Excuse me.
You point out that waiting to do this and not moving
forward on GSE reform is a choice in and of itself. And I
presume that you think that is the wrong choice. Or, can you
comment upon that?
Mr. Swagel. I do. So I worry that, in some sense, waiting
until there is the perfect solution means that we do not do
anything, and doing nothing means the GSEs stay in Government
hands. You know, the Government share is 90 percent.
And while there are many excellent people at the two firms,
inevitably, they are going to leave as the status of the firms
is uncertain.
So I do worry about making that choice.
Senator Reed. Gentlemen, can I open it up--because I have
raised these questions. I wonder if Dr. Willis or Mr.
Davidson----
Mr. Willis. What I would like to add to these three
options, in that it provides a path, is that--what I would add
to that is that at each stage you ought to test and see whether
the change in additional costs or changes in affordability and
access are something that is a concern here.
So you go bring private capital back into the jumbo market.
Then you can see whether you want to lower that limit any
further because, as we know right now, credit is very tight, so
a little unfair. The jumbo market likes 60 percent loan-to-
value. Banks prefer adjustable mortgages to put in their
portfolios.
So, yes, that is the sequence, but I do not think it is
inevitable you want to march all the way down that path.
Senator Reed. Right.
Mr. Willis. And, when we get to the end, we do know that
banks cannot hold of these mortgages. So we are going to have
to watch as we go down here, what the impact is--extending
further down the road to reply purely on the private sector.
And then the question is, where we have a conforming
market, how and to what degree do we want to have private
capital ahead?
Senator Reed. I think there is another aspect too here--
that the markets go up and down. And there are times when the
private market, just because of the returns and the economic
climate, will cover all the needs. But there are those times
where we deliberately want to support the market, and we have
to have that option.
So a complete system where there is no way for the
Government to come back in--and one of the things Senator
Johanns suggested is that maybe it is through a different
mechanism--a housing trust fund or--but I think even in terms
of this guarantee and this privatization there has to be at
least a way to come back and move the market.
Mr. Willis. Well, to have a mortgage market if you are
relying on private capital to take first loss and it abandons
the market as it did----
Senator Reed. Right.
Mr. Willis.----you will not have any, no matter what
subsidy you provide.
So you do need to at least have FHA available and be ready
to scale up. And some people may think you need more than FHA
to be there, like we had the conforming market this last time
around.
Senator Reed. One of the ironies is that as the private
capital fled in 2008 and 2009, the only game in town was Fannie
and Freddie, but at least we were able to keep a housing market
until we started seeing other factors contribute to the re-
emergence.
Mr. Willis. I would just add even in the case in Fannie and
Freddie--they could not do over 80 percent loan-to-value
because the private money in the mortgage insurance industry
also basically left the market.
Senator Reed. Thank you very much, Mr. Chairman.
Senator Tester. Thank you, Senator Reed.
Senator Corker.
Senator Corker. Thank you, Mr. Chairman. You are doing a
great job.
I enjoy your hearings. Thank you. Thanks for having this
one.
And I appreciate all the witnesses being here.
To the last point--I was going to go in a different order,
but Dr. Willis, you made one of my points, and that is when we
had the crisis, because the mono-lines are all sort of part of
the system, when the crisis--when you have a systemic crisis
like that, basically they are gone. Is that correct?
Mr. Willis. They do not appear to have had enough capital.
That is correct.
Senator Corker. So, if you were going to have a system
where you had private sector risk in front, you could do it
numbers of ways. One would be a credit-linked note; I think you
all referred to that. You could have A piece or B piece. But,
if you really relied on the system to have private capital up
front solely through insurance--I think there should be
insurance on loans above 80 percent. I agree with that.
But if you look at that, that is really not capital when
you need the capital. Would that be correct?
In other words, if you have private sector--let's say you
had a 90-10 ratio and 10 percent was private sector, if you are
relying on insurance, the insurance would not be there at the
very time you were hoping for it to be there. Is that not
generally a true statement?
Mr. Willis. Again, that was the case--the idea, not to
disagree, but the idea was you pick up a lot of the capital.
That is right.
Senator Corker. History will show that I do not think
private mortgage insurance will be a very good first-loss
position because it would not be there when you need it.
Let me ask you a question, all of you. Is there any reason
we would not go ahead and be dropping loan limits?
I mean, you could be sort of getting to where Johanns was
heading a minute ago by taking us from where we are today at
725 down to where we were pre-crisis. I mean, the average home
in America today is much less than, or less than, where it was
back when this began. So is there any reason we would not begin
doing that today?
Mr. Willis. There are two ways to do it.
Absolutely, from my point of view, I agree we should go
down at least to $417,000 and then maybe test whether to go
lower.
Many proposals are to lower the loan limits in stages. And
my fear is that you are not opening the whole market at once.
So private investors may not feel justified to----
Senator Corker. So, adjust that down over a period of time.
Mr. Willis. My preferred way would be to increase the g-fee
above $417,000 until the private sector comes in so that you
always have the existing system that can provide mortgages to
the point where the private sector comes in. If you just
withdraw down, you might leave gaps in the market that you do
not need to.
So it is an issue of how you do it, not whether you should.
Senator Corker. Does anybody disagree with we ought to
figure out some way of lowering the limits? If you do, say yes.
Mr. Swagel. No, absolutely, I agree. We give a tax break to
$1.1 million in mortgages. So you can imagine scaling down, or
refocusing, those benefits on people who really need them to
buy a house.
Senator Corker. I assume everybody else generally agrees.
Mr. Van Order. Yes, the $417,000, which is sort of the
underlying limit right now, is based on when property values
were at their peak.
The way it is done now, you cannot lower it without
legislation. You have to wait until property values come up.
But it is certainly the case that even at $417,000, in most
parts of the country, that is a pretty big house.
Senator Corker. Right. Let me keep on moving through and
thank you for being the way you are.
The GSE's balance sheets today are pretty large, and I know
they have had to buy back some loans that have had some
difficulties. Do you know of any reason we would not be moving
quickly in this market, with rates where they are, to unload
that portfolio or at least take advantage of market conditions?
Mr. Davidson. My view, generally, is that we should not be
mixing securitization and portfolios. So anything we can do to
separate those activities--if you could just get Treasury and
the Fed to do their purchases from the GSEs----
Senator Corker. Fannie and Freddie, yes.
Mr. Davidson.----that would be great. It would definitely
be much better for those organizations.
Senator Corker. So does anybody disagree that we ought to
be taking advantage of market conditions and taking this
balance sheet down in a methodical thoughtful way?
Mr. Van Order. Well, yes, with an emphasis on methodical
and thoughtful.
The portfolio was not the problem for Fannie and Freddie.
It was credit losses on the regular business.
You can let it fade away. Whether you hold it or sell it
into the market, at this point, it does not affect anything
real.
Senator Corker. Well, it helps you, though, move to a
system more quickly, I think, where you have an explicit
Government guarantee, if that is where we are going to end up,
and it causes that process to move along in a much better way.
Mr. Van Order. The loans that they have in their portfolio
are basically two sorts.
Most of the time, until recently, they bought back their
own mortgage-backed securities. So you are not really changing
the credit guarantee by whether you hold them or whether you
sell them.
The other were the private-label securities, which are
amiss, and they are winding down. They turned out not to be as
bad as people thought.
So the private label--you could sell them, but they are a
big messy. The other stuff are already guaranteed anyhow.
Senator Corker. Go ahead, Phillip.
Mr. Swagel. I just wanted to add that I strongly agree with
what Andrew said about not mixing them, and in some sense the
portfolios were not the problem for Fannie and Freddie; they
were the problem for everyone else.
Senator Corker. Right.
Mr. Swagel. I mean, the systemic risk in the system came
from the funding required to fund the portfolios.
So I agree with what Andrew said.
Senator Corker. Very good. Is there a way--if we were to
set up an A and B piece, some ratio like I laid out earlier,
certainly not at 4 and 96 but at some number, would that
provide--is there mechanism that would provide the B-piece's
owners, the sub-piece, a way of actually doing underwriting in
a way that is better than just taking all comers?
Mr. Davidson. One of the issues is that the GSE portfolios
are gigantic, and so the traditional way you would underwrite a
sub-piece would be to go through the loans and do loan-level
analysis. I think if the GSEs are going to start selling off
their credit risk, people are going to have to take that risk a
little bit more on a generic basis and not on a loan-by-loan
basis and focus more on the processes that the GSEs have rather
than the loan-level data.
I mean, we love to look at loan-level data at our firm. We
have millions and millions of loans. But, when you have a
structure that has 100,000 or 200,000 or 300,000 loans in it,
that really makes it a lot more difficult to do loan-level due
diligence.
Senator Corker. If you were scaling up, though, a new
entity--let's say you had designed a way to basically move away
from Fannie and Freddie in a methodical way and you created an
entity that had an explicit guarantee and you were building up
the process there, where you had the A and B pieces or a
credit-linked note, either one that worked best, or maybe some
other mechanism.
Would there be a way, if you were building up and not
looking at the portfolio that is in place, but building up to
actually look at the B-piece's owners or the credit-linked note
pieces, to be able to look at the actual loans that were being
put forth and give greater underwriting ability to those people
who are actually buying the B piece? Is there a way to do that?
You still think it is too much scale?
Mr. Davidson. I do not think you will really have an
efficient system that is built on that, but there is no reason
why you should not have access to loans if you need it.
I think it would just be very difficult to say we have a
system with 25 million loans or 50 million loans and that the
investors are going to be focusing on 5 or 10 million of those
loans in that level of detail.
Senator Corker. Sorry, Mr. Chairman, for being over. You
are doing a great job leading. I am not doing a good job
following.
[Laughter.]
Senator Tester. Thank you, Senator Corker. Your apology
goes to Senator Warner.
Senator Warner. Yes, that is right. I am expecting to get
my extra 3 minutes added on as well, but to then say I am going
to build off my friend, Senator Corker's, comments.
Let's, again, keep looking at this A piece/B piece
structure with the assumption being that we are putting them--
you know, the current model, kind of out of its misery.
Wouldn't that B piece--I think you may have mentioned, Mr.
Davidson, if it had been at 4 percent--let's assume for
hypothetical's sake that we are at double that. Should we have
any concern that some have expressed, that if you double that,
you have got banks over here with a 4 percent tier-1 capital
requirement, that you may be--basically, too many of these
banks will simply hold these loans on their books since the
price of that private capital up-front will mean it is more
economical to keep these within the banking industry itself.
Mr. Davidson. The senior/sub A-B structures have an
advantage versus insurance in that if you require too much
capital--so let's say that I am right and we need 4 percent,
but to be careful we make it 8 percent, which might not be the
worst idea. We now have an extra 4 percent of capital, or 4
percent of credit enhancement, that is not going to take very
much risk.
If you can convince investors that it is not a very risky
instrument, they will price that not too much differently than
a senior bond. So the cost of having extra credit enhancement
is much less than having too much equity.
So, for example, if we told one bank, you have to have 4
percent equity, and another bank, you have to have 8 percent
equity, the 8 percent equity----
Senator Warner. Obviously----
Mr. Davidson. Right, they would be out of business.
But in the capital markets, too much credit enhancement
costs less. It is still costly but not----
Senator Warner. Would the panel concur with Mr. Davidson's
comments or disagree?
Mr. Swagel. Absolutely, and I would go even further--that
if it does encourage more balance sheet lending, as long as
banks have lots of capital and good regulation, well, that
would be fine.
Senator Warner. Any others want to weigh in on that?
[Pause.]
Senator Warner. I guess the other question too is that with
what we are hoping coming out of FHFA in terms of a greater
standardization of the securitization portal, with the
repurchasing agreements, with servicing agreement
standardization, should there be a particular--as we try to get
the FHA to kind of get the utility functions better done,
should there be any priority in any of those items that you
feel if you were in charge of FHFA right now, with getting one
of these right, right away?
Mr. Willis. I think they have already shown some priority
in terms of reps and warranties and with regard to servicing
guidelines here. And I think those are really important as well
as pooling and servicing agreements. You remember PSAs were a
huge mess because every PLS, for example, seemed to have a
different one.
So I think if part of what you are asking is are they
putting in place standardization that should help the market,
it seems to me pretty clearly that they are doing the right
piece here.
Others may want to comment on the common securitization
platform. It is always better, I guess, to improve that once
than improve it in both of the different divisions.
Senator Warner. And if they were getting it right, wouldn't
that, Mr. Davidson, get to the point in terms of the whatever
that private capital, first risk dollar is?
If there is this standardization that we all hope they are
getting toward and we all are encouraging them getting toward,
that should actually improve your ability to do that
underwriting.
Mr. Davidson. That is correct. So the more you are
operating under standards that are public, the less you need
the need for sort of the loan-level due diligence in
underwriting, provided that you have some process in place to
make sure that Fannie and Freddie are actually doing what they
said they are doing. And so that would help liquidity.
Senator Warner. Do you want to add?
Mr. Swagel. I was just going to add in addition to what
FHFA is doing I think a really hard thing is getting, in a
sense, what people broadly refer to as legal and regulatory
certainty--that if we want a restart of private-label
securitization there has to be some confidence about the
regulatory environment.
And it is a tough one because if a bank or lender does
something wrong, we should go after them. But they have to
understand the rules.
And since we want origination to take place within QM and
outside of QM, I think right now there is a lot of uncertainty
that is probably preventing some of that from happening.
Senator Warner. I want to get my last question in before my
time is expired, unlike certain Senators.
Let's assume for a moment that we were unwinding GSEs as we
know them now and we were creating these new issuers with this
private capital. One of the concerns as we try to think through
this--and we want to make sure we maintain this robust market--
is, how do we make sure that there is an ability for those
small community-based banks or credit unions or others to
access this market if they do not have access to the Fannie or
Freddie window?
Could we create a co-op? Is there some mutual? Is there
another way that we can make sure in one view of this new world
where they are going to still get equal treatment and fair
treatment?
And I would love to hear from each of you.
Mr. Van Order. I think that is actually tough, and I think
a part of the problem is there is this real conflict we have in
a lot of players who all can put bad things to you.
I mean, we had a similar problem a long time ago in the
savings and loan industry when there was a comparable collapse
and it was a lot of little institutions that took risks. And
they kind of all took the risks at the same time, and they
added up.
And one of the things I think you need to worry about as
you replace Fannie and Freddie is, are you replacing them with
something that has better incentives to control risks, or are
you replacing it with something like the S&L industry a long
time ago?
Senator Warner. I would love to hear everybody else, and
then I will give it to Senator Warren.
Mr. Swagel. I will go very quickly.
Yes, it is a really important issue in the sense that it
goes to the market structure of the industry, and I think what
we want in the future is competition. We want lots of firms
undertaking the role that Fannie and Freddie are doing now and
doing securitization with a guarantee.
And we see the negative effects of not enough competition
in the spread between mortgage-backed security yields and
mortgage interest rates. There is not enough competition in
origination, and so interest rates for home buyers are higher.
So home buyers are harmed by not enough competition in this
industry.
And so we want the competition of the small banks. I think
a mutual is a good way to do it.
I would say if there is competition and other firms
competing with Fannie and Freddie, some of those should serve
the smaller banks. The smaller banks should not have to go
through the big banks to do it.
Mr. Davidson. My concern with any of the multiple issuer/
multiple guarantor models is either a race to the bottom or
difficulty in regulating them. And that is why I favor
transforming Freddie and Fannie into issuer-owned cooperatives
with requirements of access to small issuers.
It has worked for the home loan banks. They give advances
to all their members, regardless of size, as long as they can
put up the appropriate amount to capital.
Mr. Willis. You just made part of the comment I was going
to.
I think it is a really important issue, and as I said, we
should do this one step at a time and see what is happening.
And if small originators are being eliminated, we need to think
carefully about the system we have created.
I would just add to the Federal home loan banks--they have
a voting system that allows the small banks to have a larger
share of the vote than they would have based on their capital,
and that might fit with the cooperative model that Andy is
mentioning.
Senator Tester. Congratulations, Senator Warner. You outdid
Corker.
Senator Warren.
Senator Warren. Thank you very much, Mr. Chairman. Thank
you, Ranking Member. I am glad you're having this hearing.
And thank you all for being here.
I just want to shift the question just a little bit. You
have been talking some about the transition and mixing within
that where we are trying to ultimately end up.
The question I want to ask is, what is the center of the
bull's-eye?
However we get there by transition, what exactly are we
aiming for?
And I read your testimony. I listened to all of you. And I
just want to see if I have this right across the board--that
you all favor layering the risk between the private market and
what the Government has to do, publicly; that you all favor
making explicit the guarantee that is there, that Fannie and
Freddie--or whoever would call whatever names we give them--are
making and then make people pay for that guarantee that they
are getting; and that the two variables, the sort of hard nuts
to solve in this, are how you ever get the pricing right when
the Government is doing the pricing and how you get the
layering details right; that is, who steps in at what point.
And the reason for that is that it bifurcates into two
points in time. One is the worry about the lending incentives;
that is, that the private market is out there with the right
set of incentives. The Government is, obviously, rather
different on that. And the second is whether or not you have
got adequate loss bearing capacity if there is collapse.
Is that a fair description--because if it is, we need to
know that--that that is really the center of what all four of
you at least are anchoring in on. And to the extent it is not,
I just want to hear from you.
So that is why I put it out there.
And, Dr. Van Order, maybe you would be the right place to
start.
Mr. Van Order. Well, I particularly am interested in the
incentives because any time you have structures like this,
which are good, there is always the potential for moral hazard.
We used to call this unbundling. The traditional bank or
savings and loan took all the risk. The securitization process
unbundled it. That is neat. It goes back to our founding
father, Adam Smith, and the division of labor.
The problem with it is as you do that, as you get along the
food chain, you have got the moral hazard problems. People can
abuse one another.
I think these are the right structures, but I have not seen
anything in them that really makes sure that that thing does
not happen.
And it happened like crazy a few years ago. It happened in
the private-label market; it happened in all sorts of
situations, where people did not feel they had a stake in what
they were selling to other people or they would not get caught.
So I like capital, but I really think you need to worry
about the incentives at the level of management and the
companies to really perceive themselves as being on the risk
all the way down, not just for a little bit of it.
Senator Warren. I understand your point, but I just want to
make sure that I am locked in on this, Dr. Van Order, and that
is while you would put the emphasis on the incentive question
and others might put it in other places, the basic structure as
I described it is where you would end up? I just want to be
sure on that.
Mr. Van Order. Mm-hmm.
Senator Warren. OK. Good. Let the record reflect you are
nodding yes.
Mr. Van Order. OK.
Senator Warren. OK. Good.
Dr. Swagel.
Mr. Davidson. Sorry.
Senator Warren. I was just going to come up the line here.
Mr. Swagel. I agree. I agree, absolutely.
The emphasis I would add, since we are going to agree on
that emphasis, is on the shrinking of the guarantee. So make it
explicit because I think it will be there.
If you do not make it explicit, it is implicit, but then
shrink it.
Protect the taxpayers. Provide better incentives. And then
I think that also helps address the price----
Senator Warren. But, Dr. Swagel, let me just push back.
Mr. Swagel. Please.
Senator Warren. How do you shrink it if at the end of the
day we talk about a Fannie and Freddie that had no explicit
guarantee and yet when it all falls apart, when the market
falls apart, the answer is the United States taxpayer is called
to come in and backstop?
Mr. Swagel. Yes.
Senator Warren. So I do not even understand what it means
to say that part of the target is to shrink it.
Mr. Swagel. So I would use the tools that I mentioned--have
more first-loss private capital----
Senator Warren. No, no, no. I am not asking you about tools
to shrink it. I am asking you, how the end can be to shrink it
in a world where you have got the guarantee and if the world
falls apart the U.S. taxpayer will be called on to come in?
Every single player in the marketplace after 2008 now knows
that.
And so I do not see how it is anything different from we
price for it and we try to figure out how to layer in a way to
deal with the incentives.
I just do not understand how you can talk about shrinking
it.
Mr. Swagel. Well, it might be a semantic difference. In my
mind, shrinking the guarantee is when the Government covers
less than every conforming mortgage and when there is first-
loss private capital before the guarantee.
Senator Warren. Well, this is layering question. That was
the second half we were doing. So maybe we are in the same
place then.
Mr. Swagel. Yes. So it is really just a matter of how much
layering and then what is the market share of the Government.
Senator Warren. All right. Fair enough. Fair enough.
Mr. Davidson.
Mr. Davidson. So I guess my main adjustment to that thought
is that I think there are different segments of the mortgage
market that are going to have different structures.
So, for example, there is the FHA portion where the
Government is providing the credit piece and the wrap on the
MBS. Then in this other layer we have been talking about what
is currently the GSE market. That is the area where we have the
multiple layering of private capital and Government guarantee.
Senator Warren. But let me ask, Mr. Davidson, is there
really--do you have any doubt that the Government implicitly
backstops the entire market?
Mr. Davidson. So, you know, the Government did not backstop
the AAA private-label bonds, which are now trading at 50. So it
did not backstop the entire market.
What the Government did was step in and say we will make
sure there are new mortgages that we will stand behind. But it
did not stand behind the private-label MBS. It did not stand
behind the CDO market.
Senator Warren. Well, one could argue, respectfully, that
when you step into a market that has collapsed and offer
funding into it, you have backstopped that market.
Yes, we imposed some losses. And this is what I was talking
about. That is what the layering does. It imposes losses, but
it does not change the fact that the Government is the one that
still backstops the entire market.
Mr. Davidson. I guess what I am saying is I want to segment
backstopping loans that already existed and backstopping new
loans. Clearly, every new loan is now guaranteed fully by the
U.S. Government, but there are loans that were originated that
are not being backstopped by the United States in the legacy
book.
Senator Warren. Oh, fair enough. Fair enough.
Dr. Willis.
Mr. Willis. I agree that we should keep a guarantee on the
MBS.
I also agree that the Government has got to come in if
there is a systemic failure, and so we might as well charge now
and buildup a reserve against that contingency to protect the
taxpayers.
I think there is a role for private risk ahead of that
Government wrap. I think we should test for that to see what
effect it has.
Whatever your view of MI going forward (I recognize that
Senator Corker is not in favor of the MI), I do not think
anybody wants to defend the way it was.
But there are a lot of changes perhaps in the rescission
rules, capital, all sorts of things, with regard to insurance.
And you could talk about them even taking a deeper loss than
above 80-20. They actually take 35 percent. There also were
issues with reps and warranties.
Senator Warren. Fair enough.
Mr. Willis. So that is a way to do risk.
So there are a number of ways to do that.
As for getting the pricing right, everybody wants to point
out the Government did not do it well. Quite honestly, FHA--
that loss is relatively small here, particularly compared to
what the private sector lost here. Ginnie Mae is still doing
very fine here.
I do not suggest that the Government not charge fully based
on the Fair Credit Reporting Act accounting for the risk, but I
will point out that if worse comes to worst and they make a
mistake, we have ways of--you know, the Government is not going
to go out of business, and it can survive.
And people disagree about this. It then can raise the rates
and pay itself back in a sense by penalizing future borrowers,
but that may not be a fair way to do it.
Senator Warren. Thank you, Mr. Willis.
I appreciate the difference in focus for each of you, but I
take it we at least understand the structure that all four of
you are putting forward.
Thank you. I did my best to go over.
Senator Tester. You did well.
Senator Warren. I am trying to follow the lead of my
seniors.
Senator Tester. Yes, it is a poor example that they set,
though, Senator Warren.
[Laughter.]
Senator Tester. No, it is just fine.
And we will have another round. We have got plenty of time.
I mean, this is a very important issue.
And thank you, guys, for your straightforward testimony. I
very much appreciate that.
As you know, the FHFA has directed the enterprises to
engage in about $30 billion of risk-sharing transactions this
year and directed them to consider things like expanded
mortgage insurance, senior/subordinated securities, credit-
linked securities--things you guys have talked about. And I am
sure that this is going to be an effort to pave the way for
further risk-sharing in the future.
Are there specific things that we, as policymakers, should
be looking at as these transactions unfold and market
participants respond?
Mr. Willis. What I am concerned about in the direction they
are going now in is in order to do the $30 billion they seem to
think that they have to raise g-fees even higher--g-fees that
are now twice as high as they were before, credit quality on
new mortgages probably better than it has ever been. And that
is solely in order to make sure that they can run this
experiment without losing money.
And I think they are foreclosing policy options for you all
by doing that.
They should run these experiments, see in fact what they
cost, and then you can have an open discussion about what the
tradeoffs are here in terms of bringing private capital, in
terms of what kind of products, as I have said, what kind of
underwriting, et cetera.
Right now, they are marching ahead, as I read it, to
increase the g-fee just so they can run this experiment. And,
as I said, I think that forecloses or makes it harder for you
to choose some of the options that you may want to choose once
we see the end of this experiment.
Mr. Davidson. I would say in the experiment, you know, one
of the biggest problems why we did not have a deal a year ago
is the regulatory impediments to doing the kinds of
transactions that make economic sense. So I would hope that you
would sit down with FHFA, Fannie Mae and Freddie Mac and ask
them what hoops were they trying to jump through that prevented
them from doing the transaction that they thought would be most
economic.
Senator Tester. OK. Excellent.
Mr. Willis. I think we should recognize that Freddie
actually has done both of these, right?
They do K Series, which is a senior/subordinated on the
multi-family, and they did what is called moderns--others here
can comment more about them--which were using credit-linked
notes. So they have tried this in the past.
Senator Tester. OK.
Mr. Swagel. Right. I agree with what Andrew said.
You know, the $30 billion seems like a small amount to me,
but I understand that they want to test it out.
You know, I think g-fees can go up and down. As I said,
they will go down as the amount of private capital comes back.
But I think the key is for taxpayers to be protected. We
should not be underpricing this insurance. We should all
understand rates will go up.
In a sense, at the beginning, when the market is illiquid
and there is only a little bit of this risk-sharing, there will
be a premium. Investors will demand a premium. So rates will be
too high, but hopefully, this market will develop, and that
will smooth that over time.
Senator Tester. So you anticipate the rates will be--when
the offerings are made, they will be higher than they probably
will be later on.
Mr. Swagel. Eventually, just because it will be such a
small security class, that there will be a liquidity premium.
Senator Tester. Gotcha.
Mr. Van Order. Yes, I think this is fine.
My only question with experiments like this is to learn
something you need to see them through a few cycles. One of the
things that happened a few years ago with subprime loans was
people began pricing them by just looking at one cycle. So this
is something that will require a lot of patience.
I do not care that much about the pricing, but if you are
going to do this, you have got to watch it for a while because
I think things will look good for a while. And we want to be
careful if they look good, projecting it too far.
Senator Tester. Gotcha.
Look, the structures that are under consideration are not
new, but the effort, I think, with Fannie and Freddie is
largely experimental, where they are going.
What specific issues would you be considering with respect
to scalability of these mechanisms, both as enterprises
increase risk-sharing transactions and as we consider the
transition out of conservatorship?
Go ahead, Mr. Davidson.
Mr. Davidson. So the credit-linked note structures are
extremely scalable if you can get investors because they do not
alter any of the operations of Fannie Mae and Freddie Mac.
Unfortunately, they are not the most investable right now, and
they create some issues for REITs because they are not REMIC
transactions.
The senior/subordinated transactions, which are the
cleanest and what has been done on the multi-family side, are
great transactions for your less liquid products, but they are
not going to be as scalable.
So it is sort of a tradeoff. You have a clean structure
that is not scalable and a scalable structure that is not, you
know, clean from a regulatory standpoint.
Mr. Willis. I did not come here to defend the insurance
side of this, but I will point out the insurance is probably
quite scalable here particularly because there is a reinsurance
market out there as well. So all of these are just relative.
I agree with everything else that was said here but just
wanted to add that.
Senator Tester. OK, Senator Corker.
Senator Corker. Thank you again for your testimony, and I
want to follow up on Senator Warren's comments about the
shrinking.
I do think there was a semantics issue, and I just want to
point out that let's just say you had a 90-10 model, just for
talking purposes. You would have automatically shrunk the
public side by 10 percent, immediately on the front end.
And I will say that if Fannie and Freddie in this last go-
round had just had a 5 percent capital buffer there would have
been zero taxpayer losses.
So it is not a small thing to talk about a 90-10, and that
is an immediate shrinking. So I do think you all are talking
past each other just a little bit on that issue.
On the mortgage piece that you just brought up, Dr. Willis,
it is true that insurance is very scalable. But you do not
really have much when you are done, right?
I mean it is not the same as hard cash, which is what you
would have with a credit-linked note or with A or B piece. It
is hard cash. It is there. It is gone. And if something
happens----
Mr. Willis. Well, again, I am not an expert in insurance.
There is a concept called risk-to-capital.
Senator Corker. You have a highly leveraged product in
front of you. Is that----
Mr. Willis. Right, right. But if you leveraged it at--let's
say you think you want 10 percent. You could leverage it at 10
to 1.
I mean, you could require that level of capital for them to
have, and you would have the same protection there.
Senator Corker. You would have 1 percent of the 10. You
only have 1.
See, with the credit-linked note or with the A and B piece,
you have 100 percent cash.
Mr. Willis. For that portion that----
Senator Corker. For that portion.
Mr. Willis. Absolutely right.
Senator Corker. So if you had a 10 to 1 leverage, you would
have 1 percent cash in front.
Mr. Willis. Right.
Senator Corker. I guess we have had numbers of discussions.
So it is a very different model.
And I am all for private insurance on the portion above 80
percent. I just think as an equity piece that is not really
what you are getting. Does that make sense?
Mr. Willis. I have to defer to you on the insurance, in the
leverage, but all I am pointing out is you could require them
to hold more capital. You could require them to hold more
reserves equivalent to whatever you wanted elsewhere.
The insurance model is a different model. It builds up
capital over time. There are other--there are differences that
you are referring to.
But, as a matter of practice, you could make them hold; say
it is only acceptable capital in front of the Government
guarantee if you hold the certain ratio here.
Senator Corker. It would still be a leverage ratio.
Mr. Willis. OK.
Senator Corker. I think everybody agrees to that.
Dr. Van Order, the incentive piece--I am interested in what
you are saying there.
I mean, if you--let's just say you had an A and a B piece,
the not scalable but superb model. And I assume that with the B
piece, whoever held that would want to be able to determine who
the servicer is because they would have all the risk. Is that
correct?
Mr. Van Order. Yes.
Senator Corker. So talk to me a little bit about the kind
of things that would be disincentives, or take them in the
wrong direction, because it seems to me they would have every
incentive to make sure that what they were doing was not going
to fail. Tell me how that could go awry.
Mr. Van Order. Everybody in the deal has limited liability.
And the question is, at what point do you run out of yours?
But, yes, if you have got someone that is on the hook from
beginning to end and they are in a position to handle the
information properly from the servicer, then, yes, it should
work pretty well. It should mimic.
What we would like to do is take the--we need the guarantee
at the end. What we would like to do is take it off the table
in terms of behavior. Right. So set up something so that they
behave as if they did not have the guarantee and still have the
guarantee.
Let me give you an example. You were talking about raising
capital ratios to 5 percent. The problem with raising capital
ratios to 5 percent is you might change the behavior of the
people because at 5 percent a lot of safe assets are not
profitable anymore and you might move them into riskier assets.
And that is----
Senator Corker. But on the B piece, they would not be doing
that. That is not what they would be doing.
Mr. Van Order. That is a different question. You are right.
Senator Corker. That is just your equity.
So, I guess, tell me how we might go awry.
Mr. Van Order. Well, because when you have an equity
position, you have a limited liability.
So what you would want to do if you are in an equity
position is you might want to set up something where 90 percent
of the time everything was fine and 10 percent of the time it
went wrong and it went wrong in a huge way because your
liability is limited.
Maybe I have got it turned around.
Senator Corker. Let me go--thank you very much.
Let me go to another point.
So we have a superb non-scalable model which is the A and B
piece, or we have a credit-linked note which is scalable but
not quite as superb. How do we figure out a way to solve the
problem?
Mr. Davidson. My view is it has really only to do with the
current regulatory structure. Right.
So it really has nothing to do with the economics because
the economics of senior/sub and the economics of credit-linked
note are virtually identical.
You have some amount of cash that is sitting on the
sidelines. It is earning some return. And, if there are losses,
it gets taken away from the investor, which is opposite of the
insurance model where you keep your cash and then when
something goes wrong you hand over your cash.
Senator Corker. If you have any.
Mr. Davidson. If you have any left.
So, really, if these experiments work and Congress wants to
pursue this, it is not going to be that hard for you to change
some of the rules so that senior/sub or credit-linked notes
work both extremely well.
Senator Corker. So the impediment is not due to the
structure; it is due to us.
Mr. Davidson. Yes.
Senator Corker. And if we were to craft a piece of
legislation where we wanted an A and B piece or credit-linked
note, we could do so in a manner that both were very scalable
and superb. Is that correct?
Mr. Davidson. That is correct.
Mr. Willis. I will just add; you also have to deal with the
compatibility with the TBA market, right, which today it is
not, as I think you pointed out in your testimony.
Mr. Davidson. That was my scalable point--was interfering
with the TBA market.
Senator Corker. This has been an outstanding hearing, and I
thank all of you.
Mr. Swagel. Right. It just seems like that you would have--
the A and B really would be quite different. Right?
The A would be liquid. The TBA market would be concentrated
there. And then the B would just be very different. And
different types of investors would have different interests.
Mr. Willis. But you would not be able to use the TBA market
now to provide--easily provide--a rate lock, a forward
committed rate, because you would not be able to do that with a
B piece without legislative change.
Mr. Swagel. Right.
Senator Tester. Thank you, Senator Corker.
We have got a vote at 5.
You each have 7 minutes because my close is very short. Go
ahead.
Senator Warner. Well, mine is going to be even quicker than
that--one quick question.
Assuming we follow up on what Senator Corker is saying, we
have got this A and B piece. I actually do think there is a way
to structure--I do not have quite the concerns Dr. Van Order
does about that if we have got that 10 percent risk capital up
front, and I think we can take care of most extraordinary risk.
What I worry about, though, is trying to make sure on the
servicer piece. I try to understand the servicer piece more and
more and the 25 to 50 basis points that they get paid. Usually,
they get paid additional for back fees and other things through
the operations. That is really a nice revenue stream there.
How do we make sure that if we have got this private
capital up front, or even a Government backstop later, that if
the servicer is not aligned the right way--back to your point,
Dr. Van Order, in terms of incentives--we can get that servicer
out of that role and get somebody that is better aligned, in,
when they have got such a flush revenue stream there that may
or may not be rewarding actual performance?
Mr. Van Order. Actually, the Ginnie Mae market is a neat
example of this because they have, for years and years,
required very high servicing fees, and yet the originators and
sellers would rather not receive them because they would rather
sell the securities at higher prices.
What Ginnie Mae uses them for is the servicers are
responsible for making the timely payment. Ginnie Mae is a
backup to that, but the servicers are responsible for making
it.
And it is profitable. The right to service--to pool the
mortgages--trades at a positive price, and the reason is it is
valuable.
And what Ginnie Mae has done--and Fannie and Freddie did a
little bit, but Ginnie Mae did in a really neat way--was if you
did not do this they could take all of your servicing, just
take it away from you, and sell it to someone else.
So, actually, that was an example of--this was like a
performance bond which they could take away from you, and I
think that is a neat way of setting it up.
Senator Warner. I would love for you to get me some more
information about that.
Does anybody else want to have a quick comment?
Senator Warren. Good. Thank you.
Senator Tester. Senator Warren.
Senator Warren. Thank you.
And I am going to have to be quick too because I have to go
preside.
So let me ask the one question, something we have not
talked about, and that is we were talking about pricing--how
the Government prices for risk and how the private-market
prices for risk.
And I think it was you, Mr. Davidson, who talked about with
millions of mortgages in Fannie's or Freddie's--or whatever its
successor is. The question is how do you price out of that to
try to sell off some of that and how difficult that is when we
are dealing with so many mortgages.
So the question I want to focus on for just a second is
about data tagging, whether or not you have any familiarity
with the idea behind this. The notion that we now have better
capacity than ever in the past to tag even something like a
mortgage with lots of information--information about the
characteristics of the borrower, information about the
characteristics of the originator--and to keep that information
with that mortgage we can now do, or at least we can
theoretically do we could do, as it is sliced and diced and put
into different pools and keep its performance information
always pegging in so that, over time, you have this incredibly
robust data base about performance.
Now you do not have cycles yet until it plays out over
longer and longer periods, but you do have incredibly detailed
information about how it is working.
Does that potential change just in the technology of what
we can do or what we can do, I hope, in the near future that we
were not able to do up through the present--does that give us a
better capacity either for the Government to price risk or for
the private market to be pricing the risk as it buys from the
Government and, therefore, solve some of our pricing problem?
Any thoughts around that? Anybody?
Mr. Swagel?
Mr. Swagel. Sure. I think, yes, that is exactly right. So I
very strongly agree.
In some sense, the previous model was you do not need to
know any of this because Fannie and Freddie are there.
Senator Warren. Yes.
Mr. Swagel. So this is just--it is sunlight.
Senator Warren. This is like inverting that approach.
Mr. Davidson. So, I mean, Fannie and Freddie just released
a tremendous amount of delinquency data that the market never
had before.
Senator Warren. Yes.
Mr. Davidson. Which is, you know, crucial to price these
instruments.
But I guess I would caution that more data does not make it
more correct data. And so----
Senator Warren. What does that mean?
Mr. Davidson. We still have the issue that someone is
originating a loan. They are writing down what their income is
on that document, and maybe it is verified in some fashion. And
there is an appraisal. If the income and the appraisal are
false because there was fraud, then all of the rest of the
analysis you do after that is not valuable.
Some of that is----
Senator Warren. Fair enough, but let me just break in
there, Mr. Davidson, just because I am going to run out of
time.
It is also the case that you can start picking up the
originators that are not the ones who are following the rules
because now you can isolate by originator for whose loans are
going sour fast.
Mr. Davidson. Yes. So I think a good representation and
warranties and good data about that will improve this whole
process.
Senator Warren. Good.
Mr. Willis. It is another hearing, but do not forget second
liens in your database.
Senator Warren. Good point. Good point, Dr. Willis.
Mr. Van Order. And a neat thing that is going on in
developing these mega databases is linking it up with the
borrower and the information from credit repositories. I think
one of the neat things going forward will be the early warning
systems in mid-stream so you see this guy's credit score has
deteriorated. How predictive is that?
And splicing these together is not easy, and there are
serious issues, but I think that is pretty neat.
Senator Warren. So I hear a glimmer of optimism. Good.
Thank you very much.
Thank you so much. My apologies and excuse me.
Senator Tester. Perfect. Thank you, Senator Warren.
I just want to thank the witnesses again. I especially want
to thank the fact that if you guys had something to say you
were not bashful about saying it, and I very much appreciate
that. That is what hearings should be about--is getting good
information so that we can make good decisions.
And I think this hearing has underscored the importance of
getting private capital back into the marketplace, where we are
protecting taxpayers and how we transition Fannie and Freddie
out of conservatorship.
I certainly look forward to working with the folks here
today and a whole lot of others that I know are very concerned
about the housing finance system and concerned about us really
doing the right thing and building a housing finance system
that is going to last well into the future.
Just some housekeeping, this record will remain open for 7
days, and any additional comments and any questions can be
submitted for the record at that point in time.
With that, once again thanking the witnesses, and this
hearing is adjourned.
[Whereupon, at 4:50 p.m., the hearing was adjourned.]
[Prepared statements supplied for the record follow:]
PREPARED STATEMENT OF MARK A. WILLIS
Resident Research Fellow
Furman Center for Real Estate and Urban Policy
New York University
May 14, 2013
Chairman Tester, Ranking Member Johanns, and Members of the
Committee, I thank you for the opportunity to testify today on the role
for private capital in reforming mortgage markets. I am Mark A. Willis,
a Resident Research Fellow at the Furman Center for Real Estate and
Urban Policy at New York University and an adjunct professor at the
Wagner School of Public Service, also at NYU. Previous to that, I was
the Executive Vice President for community development at JPMorgan
Chase, having started many years earlier at one of its predecessor
institutions, Chase Manhattan Bank, as president and founder of its
Community Development Corporation. Before that, I served as an
economist at the Federal Reserve Bank of New York and worked in New
York City government.
Since 2008, I have focused much of my research work on housing
finance reform and have written articles, consulted for a number of
organizations, lectured, and participated on numerous panels on this
topic. I am also a member of the Mortgage Finance Working Group
convened by the Center for American Progress and have conducted
research for the Housing Commission of the Bipartisan Policy Center.
The views contained in this testimony are mine and should not be
attributed to any of the organizations to which I am affiliated.
I want to make two major points in my testimony today: First,
restoring private capital's historic role in the financing of home
mortgages--financing jumbo mortgages--is a straightforward matter once
regulatory uncertainties are resolved. Second, requiring the use of
private, credit-risk-taking capital for the much larger remainder of
the housing finance market is also possible, but such requirements
should only be implemented after we have tested their impact on access
to and affordability of mortgages for the vast majority of the home
buyer market. By test-driving different approaches, we will be better
able to weigh the costs and benefits of having private capital take
more of the risk and avoid unnecessarily disrupting the availability of
new mortgages.
Background
Before discussing the details of increasing and deepening the role
private capital plays in mortgage markets, it is useful to clarify the
different roles it has played in housing finance in the past. Private
capital is today, and always has been, the source of all the funding of
home mortgages. What we are discussing now is to what extent this
private capital is insured or guaranteed by the Government.
It is also useful to understand that there are two types of
investors bringing private capital. First, there are those investors
who are interested in taking only interest rate risk and prepayment
risk (these are called ``rate'' investors). These investors purchase
mortgage-backed securities (MBS) where there is essentially no credit
risk, e.g., when the Government insures against the risk of borrower
default. Then, there are those who take that credit risk, either alone
or in addition to interest rate and prepayment risk (these are called
``credit'' investors). It is generally understood that the pool of
capital available from these rate investors far exceeds that for the
credit investors.\1\
---------------------------------------------------------------------------
\1\ For example, rate investors in Fannie Mae and Freddie Mac debt
and mortgage-backed securities have included sovereign funds which see
the U.S. Government as providing a guarantee against credit loss.
---------------------------------------------------------------------------
In recent decades, much of the funding for mortgages has come
through the secondary market, as opposed to through financial
institutions that make loans and hold them on their own books.\2\ Many
investors purchased MBS issued by Fannie Mae and Freddie Mac, which are
backed by mortgages below what has been called the conforming loan
limit (pre-crisis was set at $417,000), believing that there was an
implicit Government guarantee of these securities. Direct investors in
Fannie and Freddie also appear to have felt shielded from credit risk,
despite the technical fact that these two agencies had been privatized
decades ago, and it turns out that, at least for the debt investors,
they were right--the Government has fully stood behind those
securities. Even the equity investors in the Agencies appear not to
have paid enough attention to the riskiness of the mortgages they
backed or owned. In the end though, these investors did bear the cost
of the financial failure of these two firms. Accordingly, most of the
investors in this part of the mortgage market have in the past only had
to concern themselves with interest rate and prepayment risk since the
Government has guaranteed timely payment of principal and interest on
all securities.
---------------------------------------------------------------------------
\2\ Banks do hold some mortgages in portfolio but 1) are limited in
their appetite for long duration instruments such as 30-year fixed-rate
mortgages, 2) look to diversify their assets to guard against sharp
losses in any one sector of the economy, and 3) have only limited
portfolio capacity in any case. While bank assets have grown in recent
years, they still barely exceed the total value of mortgages
outstanding and bank deposits fall well short of this total (See
Federal Reserve Flow of Funds reports).
---------------------------------------------------------------------------
Only larger loans (the so-called jumbo loans) and loans that were
subprime or labeled Alt-A--the financing and regulating of these latter
types of loans I take as beyond the scope of this hearing--have
traditionally been financed with private, credit-risk taking capital
mainly by banks \3\ and by investors in private-label mortgage-backed
securities (PLS).\4\ Before the crisis, loans larger than $417,000 were
not eligible for purchase by Fannie Mae and Freddie Mac (``Fannie'' and
``Freddie''-collectively the ``Agencies''). Thus, investors in these
securities and banks had to cover default risks.
---------------------------------------------------------------------------
\3\ Banks also buy MBSs to hold in portfolio. These securities may
or may not involve credit risk depending on whether the MBS is covered
by a Government guarantee.
\4\ If subprime and/or Alt-A re-emerges as an asset class, then it
seems likely that a secondary market to fund it will also be able to
re-emerge once a healthy jumbo PLS market has been reestablished.
---------------------------------------------------------------------------
Following the 2007 onset of the nationwide decline in housing
prices and the great recession, the Government expanded the range of
loans that were eligible for purchase by Fannie and Freddie by raising
the size limits for eligible loans. The Economic Stimulus Act of 2008
temporarily raised the loan limit in some parts of the country with
high housing prices as high as $729,750 for them as well as for the
Federal Housing Administration's (FHA) insurance program. While the
limit for the Agencies has now fallen back somewhat to $625,000, that
limit still encompasses over 97 percent of the mortgages and almost 90
percent of the dollar volume originated annually for the purchase of
homes.\5\ So it is not surprising that some 90 percent of the mortgages
for the finance of home purchases rely on the Government guarantee,
which means that the taxpayers remain entirely on the hook if defaults
should exceed the financial capacity of Fannie and Freddie to absorb
any resulting losses.\6\
---------------------------------------------------------------------------
\5\ These percentages are based on an average of the annual HMDA
data for home purchase loans by owner-occupants for the years 2004
through 2011.
\6\ The U.S. Treasury has inserted $187.5 billion in capital into
Fannie and Freddie in the form of senior preferred stock and has
received back $121 billion in dividends. As of August 2012 all of the
earnings of Fannie and Freddie are being swept back into the Federal
budget. Based on the amount and rate of recent payments and sweeps from
Fannie and Freddie the Government appears well on its way to recovering
the full amount of the capital it invested.
---------------------------------------------------------------------------
Calls for using private, credit-risk-taking capital to decrease the
risk of loss to taxpayers are made on several grounds: First, there is
the simple desire to have private capital absorb some amount of the
loss. Second, some argue that the private sector is better able than
the public sector to price the risk,\7\ although the latter is
something of a specious argument given that the Government had to bail
out purely private credit risk takers whose mispricing helped fuel the
subprime boom and bust. While it is a challenge for Government (or
anyone) to set exactly the right fee for providing a ``wrap,''
Government does have one advantage: it can cover losses out of tax
revenues and even recoup those losses by raising the premium it charges
going forward for providing the guarantee (as it has done through the
addition of loan level price adjustments and increases in the so-call
g-fees charged by Fannie and Freddie and as FDIC has done with regard
to deposit insurance). A third potential benefit of having private
investors take credit risk alongside the Government could be an extra
set of eyes to assess credit standards/underwriting criteria and
monitor whether loans are being properly underwritten and serviced.
Additionally, it is hoped that the active involvement of private sector
actors will discourage, if not prevent, attempts by Government
officials to fiddle with underwriting and other standards for political
gains.
---------------------------------------------------------------------------
\7\ It is worth noting that the premiums charged by both FHA and
GNMA (the Government National Mortgage Association guarantees MBS
backed by FHA-insured mortgages) have been sufficient up to now to
cover any losses out of the reserves they have built up over time. It
is also likely that, given the rules governing FHA, a transfer of less
than a $1 billion may be triggered during the next fiscal year even
though FHA has enough money to cover foreseeable losses for the next 7-
10 years and can expect higher net earnings going forward since it has
raised premiums while the credit quality of new loans has risen
significantly.
---------------------------------------------------------------------------
In the end, it is important that America have a housing finance
system that can provide mortgage products that are well-priced and
accessible and safe for all borrowers who can sustain home ownership,
while at the same time minimizing systemic risk to the economy as a
whole.
Taking Steps that Make Sense
The FHFA can take a number of steps to move us down the road of
housing finance reform in a measured and informed way:
Turn the jumbo mortgage market back to the private sector
Now that the housing market seems to have stabilized, it is time to
let the jumbo market stand on its own without a Government guarantee.
It accounts, after all, for some quarter of the dollar volume of
mortgages per year and over 8 percent of all mortgages by unit
count.\8\ While it may in the end make sense to trim back further the
share of the market eligible for the Government wrap (see discussion
below), opening up the market above $417,000 should provide a very
significant opportunity to attract private, credit-risk-taking capital.
---------------------------------------------------------------------------
\8\ These percentages are based on an average of the annual HMDA
data for home purchase loans by owner-occupants for the years 2004
through 2011.
---------------------------------------------------------------------------
The best way to trim back is to raise the g-fee on all loans over
$417,000 until the private sector is able to capture as much of the
market as it wants to. An alternative approach, which some have
suggested, is to lower loan limits one step at a time to $417,000, but
this could leave parts of the market underserved or even unserved,
especially if the private sector is reluctant to gear up to serve this
market until it is of sufficient scale to justify the costs of setting
up and running the necessary market infrastructure and to offer
investors sufficient liquidity. While Project Restart \9\ and other
efforts are underway to re-think the workings of a PLS market, some
remaining regulatory uncertainties, especially with regard to risk
retention under Dodd-Frank, probably need to be resolved if the private
sector players (e.g., investors, originators, servicers, etc.) are
going to be sufficiently motivated to take the lead to reach agreement
among themselves on the rules of the road necessary to be able to come
back at scale. Otherwise, it may take a g-fee increase that is
significantly above what pricing should be needed in the long run for
the private sector to compete for the jumbo business.
---------------------------------------------------------------------------
\9\ Project Restart has been a project of the American
Securitization Forum.
---------------------------------------------------------------------------
Then evaluate whether to lower the loan limit below $417,000
Once the jumbo private securitization market is functioning at
scale, it will be possible to evaluate the impact of any further
lowering of the limits for the private sector to begin to serve the
vast bulk of the mortgage market which lies below $417,000. This part
of the market (74 percent in dollar terms and 92 percent of the units)
has historically benefited from a Government guarantee and not relied
on credit investors.
If there is no significant difference in what private, credit-risk-
taking capital proves willing to finance in the jumbo market compared
to the existing offerings in the conforming market, then further
testing of the right level for the loan limit should be undertaken. But
any expansion of the non-conforming market should only be done in
stages. At each stage, the goal should be to ensure that the additional
market segment will continue to offer a comparable range of mortgages
including, for example, long-term, fixed-rate mortgages, which are
well-priced and available without regard to geography or other factors
that would limit access to those that now have it.\10\ Similarly, if
the jumbo market requires higher downpayments and/or higher FICO
scores, then extending it into the heart of the mortgage market would
risk excluding many potential home buyers, particularly first-time and
lower-wealth home buyers.\11\
---------------------------------------------------------------------------
\10\ Most observers agree that broad access to a long-term (15 or
30 years), fixed-rate mortgage, which have been at the core of the U.S.
housing market, is critical for a healthy housing market. Many are
concerned that the availability of this type of product might be put in
jeopardy without the Government guarantee.
\11\ See, for example, Roberto G. Quercia, Lei Ding, and Carolina
Reid, ``Mortgage Underwriting and Access to Credit'', University of
North Carolina, Center for Community Capital, October 6, 2011, which
reports that some 40 percent of mortgages made between 2004 and 2008
were made with a downpayment of less than 20 percent. http://
www.ccc.unc.edu/documents/Mtge.Under.
Access.Credit.CFPB.10.6.11.No.2.pdf.
---------------------------------------------------------------------------
An additional option is to consider varying the loan limits based
on variations across metropolitan areas in median home prices. For
example, a loan limit of $417,000 preserves a Government guarantee for
only 14 percent of the mortgage market in San Francisco (27 percent of
the units) while the comparable percentages for Dallas and New York are
83 percent and 47 percent (95 percent and 68 percent) respectively.\12\
---------------------------------------------------------------------------
\12\ These percentages are based on an average of the annual HMDA
data for home purchase loans by owner-occupants for the years 2004
through 2011.
---------------------------------------------------------------------------
Weigh pros and cons of requiring private, credit-risk-taking capital
ahead of a Government guarantee on ``conforming'' MBS
Since these tests are likely to reveal the value of having a
conforming market supported by a Government guarantee on MBSs backed by
qualified mortgages, FHFA should continue its current quest to
determine the costs and constraints of bringing in private, credit-
risk-taking capital ahead of Fannie and Freddie in that conforming
market. Sharing risk does offer the potential to reduce the burden that
could ultimately fall on taxpayers.
However, using such private capital has drawbacks as well. Private
sector investors need to be rewarded for taking risk, and they may
require tighter underwriting standards (called credit overlays) than
the Government is willing to insure. It is simply naive to expect
private investors to adjust their expectations of an acceptable return
in order to make home ownership more accessible and affordable or to
put capital at risk during market downturns. Their presence may also
make it harder for smaller originators to have access to the Government
wrap. Finally, it is important to consider any systemic risk posed by
involving private capital across the board. Even if the investors in
the MBS are protected by the Government wrap, it is important to ensure
that large losses by private capital in this position do not result in
the need to once again bail out the private sector.
Shifting risk onto the private sector is also likely to raise the
cost of mortgages. Government is able to provide its guarantee at lower
cost because, unlike private investors, it is does not have to be
rewarded with a high rate of return for taking risk. Government has the
ability to recover from losses by tapping other sources of revenue and
so will not be put out of business if, by some unexpected set of
circumstances, losses exceed the existing reserves built up by charging
for the guarantee. Private capital, on the other hand, needs a high
return to take on such risk and so its use will push up the rate that
borrowers will have to pay. While some economists argue that the
Government should charge the same as the private sector to take on risk
(so-called ``fair value'' pricing), the accounting spelled out in the
Federal Credit Reform Act is designed to ensure that the Government is
appropriately compensated for the risk it takes, based on the
Government's borrowing rate.\13\
---------------------------------------------------------------------------
\13\ Of course, every sector of the economy would like to have
access to money at lower cost based on U.S. Treasuries plus a charge
for risk that does reflects the Government's borrowing costs and not
those of risk-averse investors. Favoring the housing sector is
consistent with the belief that housing provides important social
benefits. For a discussion of the FCRA versus fair value, see John
Griffith, ``An Unfair Value for Taxpayers,'' Center for American
Progress, February 9, 2012. http://www.americanprogress.org/issues/
budget/report/2012/02/09/11094/an-unfair-value-for-taxpayers/.
---------------------------------------------------------------------------
Moreover, reliance on private capital to take first loss will limit
Government's ability in times of economic stress to ensure the
continued availability of mortgage financing through the conforming
market, which can moderate the impact on housing prices and
consequently on household wealth in the event of an economic
downturn.\14\ As we saw most recently with the housing bust and great
recession, private capital can move quickly to withdraw from the
mortgage market. New PLS originations disappeared and private mortgage
insurance became close to unavailable. When that happened, Fannie and
Freddie (along with FHA) stepped in to provide that essential
countercyclical liquidity. Therefore, a requirement for private capital
to be ahead of a Government guarantee will make the availability of
mortgages backed by the Government guarantee (other than FHA) highly
pro-cyclical unless a way can be found to dial back that requirement
quickly.\15\ Alternatively, the Government can decide to rely just on
FHA to ensure continuation of a housing market that functions well
enough for both buyers as well as sellers.\16\
---------------------------------------------------------------------------
\14\ As Mark Zandi wrote, ``the FHA shows how Government action
during the Great Recession forestalled a much worse economic fate. If
FHA lending had not expanded after private mortgage lending collapsed,
the housing market would have cratered, taking the economy with it.''
See Mark Zandi, ``FHA role may be bloated, but we'd be much worse off
without it'' The Washington Post, December 15, 2011, available athttp:/
/articles.washingtonpost.com/2011-12-15/news/35285815_1_mortgage-loans-
private-mortgage-mortgage-securities.
\15\ If there is to be such a ``dial'', it will be necessary to
determine who would make the decision, according to what criteria
(e.g., would there be automatic triggers or would the decision be
delegated to a Government entity like Treasury, the Fed, or HUD), and
what changes in procedures would need to be instituted to replace the
functions that private capital was expected to perform. A similar issue
exists with regard to any lifting of the loan limits for the FHA, but
in this case FHA is already structured to work without relying on
private capital.
\16\ FHA was able to filled part of the gap opened up by the
withdrawal of private capital by allowing its share of the home
purchase to rise from less than 5 percent in 2006 to more than 30
percent in 2009. See U.S. Department of Housing and Urban Development,
``Annual Report to Congress Regarding the Financial Status of the FHA
Mutual Mortgage Insurance Fund Fiscal Year 2012'' (2012), available at
http://portal.hud.gov/hudportal/documents/huddoc?id=
F12MMIFundRepCong111612.pdf.
---------------------------------------------------------------------------
Another aspect of the current housing market that needs to be
preserved is the To-Be-Announced (TBA) market, where Fannie and Freddie
``pass-through'' MBS \17\ are traded today.\18\ Most observers agree
that the loss of this very deep and liquid market, which benefits from
its appeal to rate investors, would likely raise the cost of mortgages
and jeopardize the continued availability of well-priced, longer term,
fixed-rate mortgage products with rate locks from 30 to 90 days.\19\
Unless we are prepared to do without such a market, we should consider
bringing in private, credit-risk-taking capital only if it is
compatible with a well-functioning TBA market.
---------------------------------------------------------------------------
\17\ The term ``pass-through'' refers to the fact that the payments
made on the mortgages that back the security are simply passed through
to the holders of the MBS in proportion to their investment.
\18\ A TBA market also exists for GNMA securities but it serves
only mortgages that qualify under the FHA, VA, or RHS programs.
\19\ For example, see James Vickery and Joshua Wright, ``TBA
Trading and Liquidity in the Agency MBS Market,'' New York: Federal
Reserve Bank of New York, 2013, available at http://www.newyorkfed.org/
research/epr/2013/exesum_vick.html.
---------------------------------------------------------------------------
For risk-sharing with a Government wrap, focus on insurance options
because they are compatible with a TBA market
Unfortunately, one of the private, credit-risk-sharing vehicles
commonly discussed is incompatible with the TBA (To-Be-Announced)
market and particularly with its ability to allow for rate locks. This
option looks to structure an MBS into at least two tranches, one senior
and one subordinate, also called A and B pieces.\20\ The senior (``A'')
piece would retain the Government guarantee of timely payment of
interest and principal while the ``B'' piece would be sold off to
private investors who would stand to lose all of their investment
before the GSE would take any losses.\21\ Mortgage payments are first
distributed to the investors in the ``A'' piece, who are thereby
shielded from losses that are less than or equal to the payments owed
on the ``B'' piece. In other words, shortfalls in payment from
borrowers are absorbed first by ``B'' investors, and only if losses are
in excess of what the ``B'' piece can absorb will the ``A'' piece
suffer losses. As a result of being willing to take the first loss,
investors in the ``B'' piece look to be paid more than those holding
the ``A'' piece, yielding an interest-rate charge to borrowers which is
a weighted average of the two interest rates plus other charges such as
a servicing fee, etc.
---------------------------------------------------------------------------
\20\ No judgment is being made here as to the relative merits of a
senior subordinate structure for the PLS market where there is no
Government guarantee and no TBA market.
\21\ This structure is even now being used by Freddie Mac for some
of its multifamily MBS in its K-series, but it should be noted that
multifamily MBS does not trade in a TBA market.
---------------------------------------------------------------------------
While the ``A'' piece with its Government wrap would be able to
trade in TBA market, the ``B'' piece would not. Trading the ``B'' piece
in the TBA market would violate SEC rules that prohibit the selling of
securities where the underlying mortgages have not been identified in
advance. The MBSs that Fannie and Freddie guarantee are specifically
exempted from this requirement as are those guaranteed by GNMA.\22\
This means that it will not be possible to use the TBA market to price
the ``B'' piece in advance, making it a lot harder and presumably more
expensive for loan originators to offer borrowers a rate lock. Also,
the ability to raise capital using this structure is highly dependent
on credit rating agencies, which will have to assess the risk inherent
in the ``A'' piece if it is to trade without a Government wrap and yet
whose role in the crisis was significant and has yet to undergo reform.
---------------------------------------------------------------------------
\22\ The Government National Mortgage Association (GNMA) guarantees
securities with mortgages backed by FHA or other mortgages backed by a
Government agency.
---------------------------------------------------------------------------
Instead, FHFA should focus on insurance type options for that extra
layer of protection for taxpayers. Insurance can work with traditional,
pass-through MBSs that, with a Government wrap, should be able to trade
the same as Fannie and Freddie Mac MBS now do in a TBA market.
There are two main types of insurance that are most often mentioned
for this purpose. One type is provided through mono-line insurance
companies that are in business solely to insure mortgage risk. A
variant of this approach is the private mortgage insurance (PMI)
business that was a response to the statutory requirement that Fannie
and Freddie obtain third-party coverage on loans with a LTV ratio above
80 percent. On a number of accounts, the PMI model evidenced major
shortcomings when hit by the latest housing bubble and bust. However,
with changes in rescission rules, tighter capital-to-risk rules, and
enhanced regulation and supervision at the state and Federal levels
\23\ it may be possible to come up with a design that would be
acceptable to the customers of their product, e.g., Fannie and Freddie,
and to the regulators and credit rating agencies.\24\
---------------------------------------------------------------------------
\23\ In general, insurance companies are subject to state
regulation with no Federal oversight bodies comparable to those that
exist for banks.
\24\ Even with risk-sharing ahead of the Government wrap at the MBS
level, requiring PMI might still make sense at the loan level for loans
with LTV greater than 80 percent. Alternatively, a risk-sharing system
could be built on a strengthened PMI model but with first-loss coverage
much deeper than 30-35 percent.
---------------------------------------------------------------------------
An alternative way to insure first loss would be to use credit-
linked notes (CLNs)-a type of security which can be bought and sold in
the public credit markets.\25\ With CLNs, private investors put their
capital at risk by purchasing the notes. These CLN investors advance
the full amount of the note, and these funds are held in trust (thus
this is described as a ``funded'' insurance model).\26\ If a loss
occurs, the funds go to cover the losses; if no loss occurs, the funds
are returned to the CLN investor. In the meantime, the CLN investors
receive regular payments which provide them with a return on their
capital. Compared to the Senior-Sub or the traditional mortgage
insurance model, this structure is more flexible as to what risks can
be covered. Rather than covering an individual loan, or a single
security, a draw on the CLN can be triggered by performance of a so-
called ``reference pool.'' This reference pool can be as simple as the
specific mortgages backing that MBS, or a broader group of mortgages,
or a cross section of a GSE's entire book of business, or other even
broader economic indicators such as the unemployment rate or house
price index.
---------------------------------------------------------------------------
\25\ CLNs have been used previously by Freddie Mac under a program
called Freddie Moderns.
\26\ It is a separate matter if the investors themselves can absorb
the losses without becoming insolvent or potentially creating systemic
risk across a broader segment of the financial system.
---------------------------------------------------------------------------
A key factor in the choice among these alternative forms of
insurance comes down to the overall cost imposed on borrowers for a
given amount of protection. While in theory, the cost of the private
capital to cover a given amount of risk should be the same regardless
of the institutional form, these two types of insurance are subject to
very different regulatory regimes which can affect the relative costs
of providing the coverage. Moreover, there are other differences that
should also be taken into account such as impact on small originators,
on the widespread availability of mortgages across geographies and all
segments of our society, scalability, ability to modify and restructure
loans, etc.
Investors in CLNs, for example, may be more restrictive in the
types of loans they are willing to insure and in dealing with smaller
originators and originators that work in only a limited number of
geographies. Investors in CLNs may prefer to work with originators that
have been rated for the quality of their origination and servicing
systems or have large, diversified pools of mortgages while insurance
companies may find that working with as many originators as possible
over time may help them diversify their risk, rather than seek to
diversify one MBS at a time. At least in theory, though, a large enough
``reference pool'' could accomplish the same thing for CLNs. Insurance
companies may also be more motivated and capable than the principals in
the CLN to provide a second set of eyes to monitor the origination and
servicing systems to minimize loss and be more flexible in allowing for
loan modifications and refinancing.
As for scalability, both approaches would seem to be able to scale
up, assuming that there is sufficient private capital willing to invest
in the stock of the insurance company or buy the CLNs. The insurance
companies also have access to re-insurers which can add to their
capacity to take on risk. For CLNs, a critical element for them to be
able to compete effectively may be sufficient scale to provide
liquidity for the trading of these securities.
Test for the optimal allocation of risk-sharing versus cost to the
borrower
In addition to testing the cost and viability of different options,
it is important to keep in mind that any incremental costs will have to
be borne by borrowers. Since, as noted earlier, the Government does not
need to charge as high a premium for taking on a given amount of risk
as private capital requires, the higher the degree of risk-sharing, the
higher the likely cost to borrowers.
In order to sort out the tradeoff between the cost to the borrower
and the degree of risk sharing, the Government needs to be explicit in
how much to charge to buildup an appropriately sized reserve to protect
itself and thus the taxpayer from having to call on tax dollars. The
amount it needs to charge (and the size of the reserve it needs to
buildup) depends on how much risk it is taking. The amount of risk, in
turn, depends on how much it lays off on the private sector as well as
the underwriting and servicing standards it sets.
At one extreme, Government can require enough private capital to be
able to absorb all expected loss with a high degree of certainty. In
this case, the Government would only need to impose a very small charge
to cover the de minimis probability of the tail risk that it retains.
Alternatively, the amount of first loss placed on the private capital
can be limited or none at all, leaving the Government with more of the
risk, which it can cover at lower cost than the private sector would
likely be willing to do.
In evaluating its risk, the Government will also need to take into
account counterparty risk, that is, whether the insuring entity will be
able to come up with the money it has promised. For CLNs, this issue
may not arise if the notes are fully funded. However, once the full
amount of the notes has been paid out, the Government must make up any
difference. In contrast, insurance companies may be able to pay out
more but they are regulated based on risk-to-capital which means that,
at any point in time, there is a limit to the losses that they can
cover. However, if desired, there is a way to structure an insurance
contract similar to the protection provided by a CLN and that would be
to set a cap on total payout, i.e., include a stop-loss provision.
Insurance companies also could have more discretion to allow loans
to be modified or refinanced if the buyers of the CLNs require highly
prescriptive rules for the treatment of any mortgages that are in the
``reference pool.'' Similarly, insurance companies would seem to have
more ability to rescind coverage in the case of defects in the
origination process, but this is technically an issue of the language
in the insurance contract requiring payment contrasted with the
language in a CLN as to when it also must pay out.
To determine the tradeoff between more risk-taking by the private
sector and the cost of mortgages, it makes sense to test some different
structures with varying amounts of risk being laid off on the private
sector. The challenge in designing these tests will be to choose which
levels of protection to test in order to get a good idea of the
parameters of the tradeoff. To do this, it is essential for the
Government to determine the appropriate premiums it needs to charge for
different levels of residual risk.
Ensure cost savings are passed on to borrowers
Regardless of the final structure chosen, it is critical to ensure
that the borrower benefits from the lower cost of funds made possible
by the Government wrap. Many have argued that the duopoly of Fannie and
Freddie allowed them to capture for their shareholders and senior
management excess profits that limited the benefit of the implicit
Government guarantee from flowing to borrowers. Going forward, this
means that all the players--from the originators to the servicers to
the securitizers to the private, credit-risk-sharers (if there are
any)--provide their services based on a competitive price. If any of
these markets lack sufficient competition, then it will be essential
for Government to intercede in some way to prevent monopoly like
profits at the expense of the borrower. One proposal that has been put
forth to deal with this possibility is set up a cooperative to
securitize the MBSs and retain first-lost risk with the originators
putting up the capital and being themselves members.\27\ I leave it to
others to identify other alternatives that would also help ensure the
savings are passed on to borrowers.
---------------------------------------------------------------------------
\27\ See Toni Dechario, Patricia Mosser, Joseph Tracy, James
Vickery, and Joshua Wright, ``A Private Lender Cooperative Model for
Residential Mortgage Finance'' Staff Report No. 466 (Federal Reserve
Bank of New York, 2010), available at http://www.newyorkfed.org/
research/staff_reports/sr466.pdf. A similar proposal has been put forth
by Andy Davidson.
---------------------------------------------------------------------------
Transition
Separate out the provision of the Government wrap from Fannie and
Freddie for ``conforming'' MBS and re-launch Fannie and Freddie
without any Government guarantee, either implicit or explicit
Once a determination is made as to the degree of risk-sharing that
it considers optimal (that share could be zero), the provision of the
Government wrap can be moved to another entity such as the Government
National Mortgage Association (GNMA/Ginnie Mae).\28\ The remaining
functions in Fannie and Freddie could then continue in a new legal
entity or entities. New entrants should also be allowed, if not
encouraged, to compete with the successors to Fannie and Freddie in
securitizing mortgages.
---------------------------------------------------------------------------
\28\ The Housing Commission of the Bipartisan Policy Center has
proposed that the role of ``public guarantor) be performed by GNMA or
by a newly created Government entity.
---------------------------------------------------------------------------
Coordinate any changes with attention to the role of FHA and the single
securitization platform being developed by FHFA
While this hearing does not directly concern what role the Federal
Housing Administration (FHA) should play in a reformed housing finance
system, it is worth noting why its continuation is important and, in
particular, what changes should be made now to enhance its ability to
protect both taxpayers and future borrowers from being exposed to
unnecessary risk. FHA has three roles to play: First, FHA needs to
ensure that all those who can sustain home ownership have access to
reasonably priced long-term, fixed-rate mortgages. Second, FHA needs to
prevent the collapse of local, regional, or national housing markets
when the private sector pulls back from offering mortgages. Third, FHA
needs to promote innovation by piloting new products and underwriting
and servicing practices.
All three roles are important, but the provision of countercyclical
support to the new mortgage market is probably most relevant to this
hearing. If it is concluded that private capital should be brought in
ahead of the Government wrap and if there is no mechanism devised to
dial it back in the face of a withdrawal of private capital has
withdrawn, then it is essential to preserve FHA's ability to scale up
even more than it did during the most recent fall in housing prices and
the great recession. Even with Fannie and Freddie still originating
loans (although with limited support from the PMI industry to do loans
with LTV's in excess of 80 percent), FHA alone provided as much as 40
percent of mortgages for home purchase with over 70 percent of these
loans going to first time home buyers.\29\
---------------------------------------------------------------------------
\29\ U.S. Department of Housing and Urban Development, ``Annual
Report to Congress Regarding the Financial Status of the FHA Mutual
Mortgage Insurance Fund Fiscal Year 2012'' (2012), available at http://
portal.hud.gov/hudportal/documents/huddoc?id=F12MMIFundRepCong
111612.pdf.
---------------------------------------------------------------------------
It is worth noting that FHA also needs to pull back from the jumbo
market. As part of the response to the fall in house prices and the
financial crisis, the FHA was permitted to dramatically increase the
size of loans that it could offer. With the stabilization of the
housing market and the desire to crowd in private capital to the jumbo
market, the loan limit for FHA should be lowered to $417,000 if not to
the lower levels that prevailed earlier.
Impact of a single securitization platform
FHFA has announced plans to develop a single securitization
platform to replace the back office functions of the Agencies. While
its creation will not necessarily inhibit the ability to implement the
steps outlined above, it might be just as easy, if not easier, to
modify existing systems to accommodate the necessary changes. As
originally announced, the plans for this platform were very ambitious,
especially given the intention to design it with sufficient flexibility
to accommodate the wide variety of originators and originating
platforms beyond those of Fannie/Freddie. However, regardless of how
flexible and all inclusive the final product, it needs to incorporate
the possibility of providing first-loss protection either directly or
through third-party entities.
Conclusion
In exploring how to bring more private capital into the housing
finance system, there are a number of steps that FHFA should undertake.
One path is to restore private capital's historic role in the financing
of the mortgages bigger than $417,000. This should be able to be done
in a straightforward matter, once regulatory uncertainties are
resolved, by raising the g-fee until the private sector takes over that
part of the market. With actual information on the cost of and product
mix of loans being offered in the jumbo market, we will be better able
to evaluate the benefits of having a Government guarantee supporting
MBSs in a conforming market. Many housing market experts worry about
loss of a TBA market and of a well-price, fixed-rate mortgage with 15
and 30 year maturities. By taking one step at a time, we will be able
to see for ourselves if lowering loan limits further will limit access
and affordability of mortgages. By doing it in stages, it will be
possible to prevent unnecessarily harming the bulk of the housing
market.
Given the skepticism that a purely private mortgage market would
work well for the vast bulk of the housing market (save for the portion
served by FHA), FHFA should also continue to pursue its exploration of
the cost and structure for requiring private capital to take first loss
ahead of a Government guarantee. By test-driving different approaches,
we will be better able to weigh the costs and benefits of having
private capital take more of the risk and avoid unnecessarily
disrupting the availability and affordability of new mortgages.
______
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
______
PREPARED STATEMENT OF PHILLIP L. SWAGEL
Professor, International Economic Policy
Maryland School of Public Policy, University of Maryland
May 14, 2013
Chairman Tester, Ranking Member Johanns, and Members of the
Committee, thank you for the opportunity to testify on the vital topic
of returning private capital to mortgage markets. I am a professor at
the University of Maryland's School of Public Policy and a faculty
affiliate of the Center for Financial Policy at the Robert H. Smith
School of Business at the University of Maryland. I am also a senior
fellow with the Milken Institute's Center for Financial Markets and a
visiting scholar at the American Enterprise Institute. I was previously
Assistant Secretary for Economic Policy at the Treasury Department from
December 2006 to January 2009.
Bringing private capital back to fund mortgages and take on credit
risk is an essential element of housing finance reform, particularly
with respect to reform of the Government-Sponsored Enterprises (GSEs)
of Fannie Mae and Freddie Mac. Housing finance reform should ensure
that mortgages are available across economic conditions, while
shielding taxpayers from taking on uncompensated risk and protecting
the broader economy from the systemic risks that arose in the previous
system. Bringing about increased private capital as part of housing
finance reform will help protect taxpayers and improve incentives for
prudent mortgage origination by lenders and investors with their own
resources at risk.
The situation in housing finance today is that taxpayers fund or
guarantee more than 90 percent of new mortgages through the GSEs and
through Government agencies such as the Federal Housing administration
(FHA). Fannie Mae and Freddie Mac stand behind virtually all new
conforming mortgages through the two firms' guarantees on the mortgage-
backed securities (MBS) into which the two firms bundle the home loans
they purchase from originators. There is loan-level capital to absorb
losses in the form of homeowner downpayments and private mortgage
insurance (PMI), but no private capital at the level of the mortgage-
backed security (MBS) ahead of the financial resources of Fannie and
Freddie. With the U.S. Treasury committed to ensuring that Fannie and
Freddie remain solvent, the U.S. Government effectively backstops
conforming loans, leaving taxpayers exposed to considerable losses in
the event of another housing downturn-and this risk remains even while
the two firms are now profitable. Taxpayers further take on credit risk
in housing through the Government backstop on the Federal Home Loan
Bank (FHLB) system, and through guaranteed mortgages supported by the
Federal Housing Administration (FHA) and other Federal agencies. I have
previously testified on reforms to the FHA that would better protect
taxpayers while focusing the agency on its mission to expand access to
mortgage financing for low- and moderate-income families who have the
financial wherewithal to become homeowners.\1\ I thus focus here on GSE
reform.
---------------------------------------------------------------------------
\1\ February 28, 2013, Senate Banking Committee hearing on
``Addressing FHA's Financial Condition and Program Challenges, Part
II.'' http://www.banking.senate.gov/public/index.cfm?Fuse
Action=Files.View&FileStore_id=6283a07f-b4c3448a-82e0-d62cfb06bf61.
---------------------------------------------------------------------------
Bringing back private capital into housing finance would mean that
private investors would absorb losses as some mortgage loans inevitably
are not repaid. In some instances, this could involve mortgage loans
with no Government guarantee, while in others there could be a
secondary Government guarantee that kicks in only after private capital
absorbs losses (or the guarantee could be alongside private capital,
with losses shared). Private investors would be compensated for taking
on housing credit risk, so that it should be expected that mortgage
interest rates will increase as housing finance reform proceeds. This
interest rate impact reflects the facts that the previous system was
undercapitalized and provided inadequate protection for taxpayers.
It would be useful for reform to allow for a diversity of sources
of funding for housing, and for private capital to come in a number of
forms and through a variety of mechanisms. This will help make the
future housing finance system more resilient to economic and market
events that affect particular parts of financial markets and thus
impinge on the availability of funds for housing.
At the level of the individual loan, capital for conforming
mortgages will continue to be present from a combination of homeowner
downpayments, private mortgage insurance, and the capital of
originators that carry out balance sheet lending. The recent housing
bubble and foreclosure crisis highlighted the importance of homeowner
equity as a factor in avoiding foreclosures, as foreclosure rates were
especially elevated for underwater borrowers--those who owed more on
their mortgages than the value of their home. As reform proceeds, it is
vital to ensure that meaningful downpayments remain a central aspect of
underwriting and a requirement for mortgages to qualify for inclusion
in MBS that benefit from a Government guarantee. Similarly, regulators
must ensure that private mortgage insurers have adequate levels of
their own high-quality capital to participate in mortgages that receive
a Government guarantee.
The larger changes involved with the return of private capital to
mortgage origination will come at the level of the mortgage-backed
security. With nearly all securitization of conforming mortgages going
through the GSEs, there is essentially no capital at the MBS level. The
so-called profit sweep agreement between the Treasury Department and
the two GSEs prevents Fannie and Freddie from building up the capital
that would be the norm for an insurer. Housing finance reform should
involve changes on all of these dimensions so that private capital is
present at the MBs-level. These changes are discussed next.
Fannie and Freddie are setting up risk-sharing mechanisms to allow
private investors to invest in securities that will take losses ahead
of the firms' guarantee (that is, ahead of the taxpayer guarantee).
There is still little securitization of mortgages taking place without
a guarantee (private-label securitization of non-conforming loans), and
firms other than Fannie and Freddie are not allowed to compete in the
business of securitization of conforming mortgages with a Government
guarantee.
Risk-sharing by Fannie and Freddie on guaranteed single-family MBS
Risk-sharing could be implemented by having the two firms sell non-
guaranteed tranches of MBS that take losses either before or at the
same time as MBS tranches that receive the guarantee. This could be
seen as selling off subordinated tranches of guaranteed MBS. This would
be an incremental approach for bringing in private capital that could
proceed ahead of legislative action; indeed, work on this is under way
at both Fannie and Freddie as part of the FHFA strategic plan. Fannie
and Freddie both already share risk in different ways on their MBS for
multi-family properties so there are extant examples of such a
mechanism.
Risk-sharing would translate into higher mortgage interest rates.
The yields on these non-guaranteed tranches would be higher than on
securities with a guarantee--after all, investors will demand to be
compensated for taking on housing credit risk. Even so, these
securities would still be protected from losses by post-crisis
underwriting standards (which some would say are too careful), and by
homeowner downpayments plus any PMI. The interest rates on mortgages
facing homeowners would reflect a blend of the yields on the guaranteed
and non-guaranteed MBS, along with costs such as the fee (g-fee) paid
to the Government for taxpayer backing.
An important consideration as risk-sharing proceeds is that the
initial volume of non-guaranteed MBS likely would be modest. Yields on
the non-guaranteed tranches could thus be elevated by a liquidity
premium (that is, by investors' demands to be compensated for the lack
of liquidity in these new securities). It would be useful to spread any
interest rate impact across mortgages that are bundled into all
conforming securities until the risk-sharing program has proceeded
enough to provide a liquid market for the non-guaranteed MBS tranches--
or more likely, until all guaranteed MBS are protected by first-loss
tranches.
As envisioned in the FHFA strategic plan, selling subordinated
tranches of guaranteed MBS would allow for a return of private capital
to conforming MBS even before housing finance reform clarifies the
long-term status of the GSEs. A larger role for the private sector and
a receding Government guarantee could be brought about by increasing
the size of the subordinated tranches and thus providing more first-
loss protection ahead of the firms (and thus ahead of the need for the
Government to make good on its contractual obligation to keep the firms
solvent). Note as well that the appropriate guarantee fee to charge on
the senior MBS would eventually decrease as more private capital takes
losses ahead of the Government.
Capital brought in by firms that compete in conforming securitization
A fruitful avenue for housing finance reform would be to allow
other firms to compete with Fannie and Freddie in the securitization of
conforming MBS. Firms undertaking such securitization would be required
to maintain appropriate levels of capital, both their own and that of
other investors, to take losses ahead of the Government. All firms
would then pay for the Government guarantee that is secondary to
considerable private capital.
Allowing for such competition would be beneficial to ensure that
any inadvertent (but likely unavoidable) underpricing of the Government
guarantee is pushed through to homeowners in the form of lower interest
rates rather than allowing MBS securitizers to profit from an elevated
spread between (low) interest rates on MBS and (high) interest rates on
mortgages. Indeed, Scharfstein and Sunderam (2013) document that a lack
of competition results in just such an elevated interest rate spread,
to the detriment of potential borrowers.\2\
---------------------------------------------------------------------------
\2\ See David Scharfstein and Adi Sunderam, ``Concentration in
Mortgage Lending, Refinancing Activity, and Mortgage Rates,'' April
2013. http://www.hbs.edu/faculty/Publication%20Files/
Concentration_in_Mortgage_Lending_20130407_adfb023e-3c76-42df-9ede-
312925dae538.pdf
---------------------------------------------------------------------------
Fostering competition would further help address the problem that
Fannie and Freddie are too important to be allowed to fail. If enough
additional firms enter in the business of mortgage securitization, then
any such securitizer could fail while others continue to undertake
securitization. Entry and competition as part of housing finance reform
could thus help to avoid a situation in which mortgage financing is not
available to American homeowners, with potentially serious negative
impacts on the U.S. economy.
Two steps are vital to allow for entry and competition. The first
is the completion of the common securitization platform now being
developed jointly by Fannie and Freddie as part of the FHFA strategic
plan. A common securitization platform would unify the markets for MBS
packaged by the two GSEs--both are effectively guaranteed, but they
trade separately to the disadvantage of the less liquid Freddie Mac
securities. A common securitization platform would facilitate entry by
new firms that securitize guaranteed MBS in competition with Fannie and
Freddie, since the MBS of new entrants could trade in the same market
as MBS issued by Fannie and Freddie rather than trading separately and
facing a considerable liquidity disadvantage. In developing the common
securitization platform, it will be important to maintain the TBA (``To
Be Arranged'') market that facilitates desirable features of such as
the ability of homeowners to lock in interest rates.
The second step would be for the Government guarantee that now
backstops Fannie and Freddie as firms to switch instead to a guarantee
on qualifying MBS (rather than on the firms themselves). This step
requires Congressional action, since it would formalize the Government
guarantee on housing that is now merely a bilateral contract between
the Treasury and each GSE. The Government guarantee on housing would be
formalized, but only so that the guarantee could shrink by requiring
increased first-loss private capital before the guarantee. In other
words, the guarantee would be made explicit so that it could recede.
Housing finance reform must ensure that smaller financial
institutions have access to the housing finance system on terms equal
to those for the larger firms that dominate mortgage origination. The
reform discussed here meets this essential criterion in two ways. The
first is that the use of a common securitization platform would allow
regulators to enforce non-discrimination provisions that require firms
that obtain the secondary Government guarantee for their MBS to
purchase qualifying mortgage loans on equal terms from qualifying
lenders. That is, regulators would ensure that the system is open to
all conforming loans. At the same time, it would be natural for smaller
institutions to join together to form a securitizer on a mutual basis.
As an observation, the securitization and guaranty businesses of Fannie
Mae and Freddie Mac are generating substantial profits, reportedly on
the order of $20 billion per year combined between the two firms.
Forming a mutual securitization company would thus give smaller
institutions a share of these profits while ensuring that they do not
need to rely on larger firms for access to the housing finance system.
Firms competing in securitization of conforming MBS could have
several forms of private capital ahead of the secondary Government
guarantee, including both their own equity and capital arranged with
other private entities. For example, securitizing firms might purchase
MBS-level insurance from other private firms, much as individual
homeowners purchase private mortgage insurance. As with any such
insurance product, a key consideration is to ensure that the firms
providing MBS insurance maintain appropriate amounts of high-quality
capital.
An alternative to MBS insurance would be for MBS securitizers to
issue credit-linked securities in which private investors provide funds
to the securitizer in return for a yield (as usual with a fixed-income
security), with provisions that specify the losses to be apportioned to
the outside investors in the event of housing credit losses. Such
credit-linked securities would bring in private capital in a similar
fashion to the subordinated tranches of MBS discussed above.
The market for conforming MBS would thus include securities with
and without a Government guarantee. The common securitization platform
would again be important to ensure that the guaranteed securities trade
together in a liquid market for all issuers. The non-guaranteed MBS
tranches could then trade separately for each securitizer. Indeed,
investors willing to take on first-loss housing credit risk would be
expected to demand considerable information on the characteristics of
the mortgages in the MBS. A useful feature of the structure discussed
here is that the amount and high quality of the private capital is
clear--the non-guaranteed securities take losses up to the amount of
capital at risk.
Private label securitization
An increase in mortgage lending without a Government guarantee
would constitute a direct return of private capital to housing finance.
Housing finance reform along the lines of the process discussed above
would gradually increase the incentive for some mortgages that could
qualify for a Government guarantee to choose to go without one. The
increased incentive to avoid the Government guarantee would reflect the
costs that correspond to a requirement for an increasing amount of
first-loss private capital (risk-sharing), along with a higher fee
charged by the Government for the secondary guarantee on conforming
MBS. As an increasing amount of first-loss capital is required ahead of
the Government guarantee and as the g-fee insurance premium rises, so
too will the incentives rise for a larger-scale restart of private-
label securitization. At some point, if enough private capital is
required and the g-fee pricing is set high enough, some conforming
loans that qualify for the guarantee will choose not to purchase it and
prefer instead to arrange for non-guaranteed financing. This could
include securitization of non-guaranteed (private label) conforming
MBS.
If the Government no longer provides a guarantee for every
conforming mortgage, then an auction mechanism could be used to set the
price of the Government insurance. This would help to address the
difficult challenge of setting the price for the guarantee. One way to
achieve this outcome in which not all conforming mortgages are covered
by a guarantee would be to gradually reduce the amount of insurance
capacity offered by the Government. A safety valve mechanism could be
put in place under which the Government would offer additional
insurance capacity at a higher guarantee premium that market
participants would find unattractive in normal times and thus prefer to
arrange for private-label securitization but remain available in the
event of a future crisis in which funding for non-guaranteed
securitization dries up (as has been the case since the collapse of the
housing bubble in 2006).
Steps that make guaranteed MBS less attractive would similarly
boost the incentives for increased usage of private-label
securitization of non-conforming loans--mortgages that do not qualify
for the Government guarantee. This is because as costs for (conforming)
guaranteed loans increase, some borrowers who might have taken out a
conforming loan will instead turn to mortgages with non-conforming
features such as a principal amount above the conforming loan limit.
Even so, a broad restart of non-guaranteed securitization likely
requires further progress in reducing the uncertainties regarding the
regulatory environment and legal liability for loans that do not
qualify for the safe harbor in the CFPB's qualified mortgage (QM)
standard. Private-label MBS issuance was $4.2 billion in 2012 according
to data collected by SIFMA (the Securities Industry and Financial
Markets Association)--compared to more than $1 trillion in MBS issuance
covered by a Government guarantee.
Policy Levers to Foster a Return of Private Capital into Housing
Finance
The various channels through which private capital could return to
the housing finance system involve four main policy levers: 1) raising
the price of the Government guarantee; 2) reducing the quantity of
insurance offered by the Government or otherwise narrowing the scope of
mortgages eligible for the Government insurance; 3) opening the housing
finance system to new competition that brings in private capital; and
4) requiring firms that securitize Government-insured MBS to arrange
for first-loss private capital to take losses before the Government
guarantee.
Reducing or eliminating the Government role in housing finance
involves going further with these four policy levers. The jumping-off
point for reform is the current system in which there is no first-loss
private capital and taxpayers stand behind essentially all conforming
loans. It is instructive to consider the steps to move to a private
system in which there is no Government guarantee on conforming
mortgages (leaving aside the FHA and other smaller programs). To reach
a private outcome, the housing finance system will first transit
through the alternative in which there is a secondary Government
guarantee behind first-loss private capital at the MBS level but all
conforming mortgages continue to be insured by the Government (which
now provides a secondary guarantee). This first alternative is
precisely option three from the February 2011 Treasury-HUD White Paper
on ``Reforming America's Housing Finance Market.'' That is, Treasury-
HUD option three is a necessary first step in the move toward a private
housing finance system.
Moving further toward a private system from Treasury-HUD option
three involves additional increases in guarantee fees and a requirement
for yet greater first-loss private capital ahead of the secondary
Government guarantee. As these policy levers are utilized, eventually
only a modest share of mortgages will be included in MBS that receive
the secondary Government guarantee. Instead, most mortgages will be
funded privately, at least in normal times. In times of credit market
stress, a greater share of mortgages would avail themselves of the
Government guarantee, even at the cost of the higher g-fees and
increased private capital. This outcome is precisely option two from
the 2011 Treasury-HUD White Paper. Again, the second option in the
Treasury-HUD white paper is a necessary stage on the transition to a
private system.
Eventually as the policy levers are fully utilized, the pricing of
the guarantee fee will be so high that no MBS securitizers will
purchase the Government guarantee (or more simply, the amount of first-
loss private capital required in front of the guarantee is set at 100
percent, eliminating the guarantee). This outcome is option one in the
2011 Treasury-HUD White Paper.
In other words, ending up at a housing finance system that is fully
private involves a transition through intermediate steps in which there
is first private capital in front of a secondary Government guarantee
(Treasury-HUD option three) and then a stage in which the share of
guaranteed MBS declines and the share of private-label securitization
and non-guaranteed balance sheet lending increases (Treasury-HUD option
two). Rather than seeing the three options in the Treasury-HUD White
Paper as separate proposals, it is useful to note that they differ by
the settings of the policy levers of the price and quantity of the
Government backstop, the scope of conforming mortgages, and the amount
of required private capital. These levers in turn determine the share
of conforming mortgages that will be covered by the Government
insurance and thus the choice between the three Treasury options. In
other words, the seemingly distinct policy options often considered in
the debate over housing finance reform are better seen as points on a
spectrum that differ by the share of credit risk taken on by the
Government and by private investors. This approach is depicted in the
figure below.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Moving forward with Housing Finance Reform that brings back private
capital
The key question in housing finance reform remains the degree of
Government involvement, and especially whether there should be some
form of a Government guarantee on some housing credit risk, even if one
that takes effect only after private investors take losses first. I
have written previously that I see it as a political and social reality
that future U.S. Governments will intervene if potential home buyers
cannot obtain mortgage financing such as during a financial crisis.\3\
An implication is that a housing finance system that is notionally
fully private will inadvertently recreate the implicit guarantee in the
previous system that failed so badly and that left taxpayers with a
costly bailout. It would be better in my view for the inevitable
Government involvement to be made explicit. Taxpayers would be
compensated for taking on housing risk, with considerable private
capital ahead of the secondary Government guarantee.
---------------------------------------------------------------------------
\3\ Phillip Swagel, 2012. ``The Future of Housing Finance Reform.''
The B.E. Journal of Macroeconomics, volume 12 issue 3, article 11.
---------------------------------------------------------------------------
Housing finance reform that brings back private capital can proceed
without resolving the question over the eventual role of the housing
finance system. This is because the policy levers required to move
forward with reform are the same ones to reach any system with a
smaller role for the Government than today, including the system with a
secondary Government guarantee and the alternative in which there is no
role for the Government (at least no explicit role). Indeed, as noted
above, to reach the system with no Government guarantee, a partial
guarantee will be in place during a transition.
Whether it is possible for housing finance reform to arrive at a
system that is fully private (at least notionally) will depend on the
societal and political reaction to the higher mortgage interest rates
and reduced availability of credit that correspond to the increased
protection for taxpayers from a system with a greater role for private
capital. It is unclear whether a private housing finance system is
politically and socially feasible. But the way to find out is to start
by adjusting the policy levers that bring in private capital.
This implies that (the sometimes passionate) disagreements about
the role of the Government at the core of the policy debate over U.S.
housing finance reform are misplaced. The next steps are the same for
all plans now under serious consideration; namely, that the price the
Government charges to insure mortgages should rise, the volume and
scope of mortgages that the Government offers to insure should decline,
and the amount of private capital should increase.
The disagreement is over how far to turn the policy levers
affecting the price and quantity of the Government insurance, and how
that in turn will affect the interest rates and types of mortgage
products faced by American home buyers. How far to go toward a private
system will ultimately reflect a societal and political judgment about
the role of home ownership and the degree to which Americans support
public efforts to foster home ownership.
The alternative is to wait for reform until there is agreement over
the end point. Waiting to start with housing finance reform is a choice
in itself--to keep Fannie Mae and Freddie Mac in Government control and
to have little role for private capital. The longer that
conservatorship continues, the more likely it is that it becomes
permanent, with Fannie Mae and Freddie Mac in Government hands forever.
This would mean a long-run housing finance system that most acutely
puts taxpayers at risk while missing out on the possibilities for
innovation that are most likely to occur with a system driven by
private sector involvement and incentives. Such a nationalized housing
finance system is a default outcome if no reform is undertaken.
______
PREPARED STATEMENT OF ROBERT VAN ORDER, Ph.D.
Chair, Department of Finance and
Professor of Finance and Economics, George Washington University
May 14, 2013
Getting private capital back into the mortgage market is clearly an
important goal. Right now almost all mortgage lending is done via
Fannie Mae and Freddie Mac, which are under Government control via
conservatorship, and FHA and Ginnie, Mae, which are Government owned.
It wasn't always that way. Forty years ago the industry was dominated
by Savings and Loans, and more recently by Fannie and Freddie as
privately owned corporations. Beyond that, in the years after 2000 the
market in which mortgages were securitized became increasingly
dominated by ``private label'' securities. All of these institutions
have, to varying degrees, collapsed.
Appearances can be deceiving, and what is and is not private
capital can be difficult to determine. Indeed, whether capital is
private or not is not the most important question. What is most
important is who ultimately bears the risk and how it can be
controlled. In the cases of both the Savings and Loans and Fannie and
Freddie the Government provided (explicitly in the first case,
implicitly in the second) guarantees to shareholder-owned institutions,
and these guarantees subsequently required very large cash outlays. In
the case of private-label securities collapse in value caused a
financial panic, which provoked other bailouts and was the impetus to
the Great Recession. Making mortgage markets work again will require an
understanding of who is taking the risk.
It is very likely that any system that we end up with will have a
role for the Government as guarantor at the end of the process, and
that what we mean by having private capital in the market means having
private capital taking risk ahead of the Government. This requires
decisions regarding both the quantity of capital ahead of the
Government (e.g., capital ratios) and the types of incentives used to
keep risk under control. Discussions regarding risk-taking in the
residential mortgage market often focus on the risks presented by
specific mortgages or the risk inherent to the institutions that
originate or fund mortgages. As discussed below, this focus is
misplaced, as it is not obvious what specific properties make one
mortgage more risky than another and institutional form (or name)
matters less than specifics about the capital they hold.
A central point is that all this is very difficult. Many of the
things associated with the huge increase in defaults in the Great
Recession were close to unpredictable and certainly not easy for
regulators to control. As a result we need policies that provide
automatic solutions and incentives for those closest to the operations
of financial institutions, their management, to control risk-taking.
After reviewing some of the lessons learned I will focus on work done
with Rose Neng Lai at the University of Macau on the use of contingent
capital, both as a source of new capital in tough times and as a way of
providing incentives to the mangers of financial institutions to take
on less risk.
In the next section I review some of the issues involved in
guarantees. This is followed by a discussion of what data so far tell
us about what is important, followed by ways, including contingent
capital, of improving capital standards.
Market Structure and Guarantees
For decades almost all American mortgages have benefited from some
sort of Government guarantee, e.g., directly via FHA insurance, or
indirectly from deposit insurance for banks and Savings and Loans or
guarantees for Government-Sponsored Enterprises (GSEs) like Fannie Mae
and Freddie Mac. If financial markets were perfect, or close to it, and
transfer payments were easy to make, there would be little economic
justification for the Government to have a role in financing housing
and certainly no need to provide guarantees to get people into good
housing. Anything that needed to be done could be done with housing
vouchers or direct provision of housing services, letting the financing
take care of itself.
Guarantees can make sense outside of housing policy--deposit
insurance and GSE guarantees, for example, as a way of stabilizing
financial markets--and they can be justified in a ``second best'' sense
as a way of promoting housing and home ownership when transfer payments
are hard to make or there are inefficiencies in financing housing. But
guarantees also have important incentive effects.
Basics of Guarantees
Guarantees have two principle effects:
If not fully priced and regulated they lower the cost of
housing and alter resource allocation, redirecting investment
into housing and away from other uses. When targeted they
promote housing for particular classes of households. This is
``good'' to the extent that housing is under produced, which is
a hard case to make, or when targeting is important, for
instance to encourage home ownership.
They help prevent financial panics, by removing the
motivation for ``bank runs.'' However, if they are not well
regulated, they lower the cost of risk-taking and promote
excessive risk-taking.
The first effect is most closely associated with housing goals; the
second is indirectly associated with it but also has broad macro
effects. Both have costs, in terms of misallocated resources and
``bailout'' costs when institutions getting the guarantees fail.
The two costs are related; the bailout costs typically go along
with misallocated resources, but even without misallocation bailout
costs are disruptive and unpopular. In the United States a bailout of
the Savings and Loans insurance fund ultimately cost taxpayers around
$150 billion. For Fannie Mae and Freddie Mac cost is not clear because
they appear to be making money again and may pay back most of what the
Government injected, but still there was a bailout.
Guarantees have many of the characteristics of financial options in
that the owners of guarantees get the upside from risk-taking but have
limited liability on the downside. If a guarantee is not priced or
regulated properly, then recipients get downside protection at below
cost, essentially an underpriced insurance policy. This provides
incentives to take on risk to maximize upside returns without having to
worry about downside losses. Indeed, absent other factors, like
reputation or franchise value, maximizing wealth will tend to involve
maximizing the value of the guarantee, which in turn means maximizing
risk. As a result the subsidy that comes with guarantees changes
incentives. Because risk-taking is hard to observe and control, the
subsidy is hard to control, as are bailout costs once the guarantee is
in place.
Effects of guarantees and bailouts have been mixed. For instance,
while they have received considerable support, neither banks nor Fannie
Mae and Freddie Mac were a source of systemic risk, not because they
didn't take risk, but because their guarantees kept the values of their
deposits or debt from falling. That is the paradox of guarantees. They
make it easier to take on risk, but they also limit systemic risk and
bank runs. It's hard to have one without the other. Sometimes you can't
live with; sometimes you can't live without them.
Probably more important than bailout costs, however, are the
economic costs that come with recessions and Great Recessions. In the
Great Recession systemic risk happened mainly in the private ``shadow
banking'' system, which was not guaranteed (and because it was
perceived as not guaranteed), but which still took on excessive risk
and saw something akin to bank runs as investors lost confidence in the
ability of the system's assets to cover its liabilities (See Gorton
(2009) and FCIC (2011)).
Recent History
I have attached as an appendix a summary of some work on mortgage
default done with a colleague at George Washington, Jason Thomas. It
summarizes some of the data for the performance of loans (both those
securitized by Fannie and Freddie and those securitized through the
private-label channel) originated in 2003 and 2006, along with a simple
analysis of the risk of requiring low-income lending. I am putting it
there because I think a few pictures can summarize some important
trends in defaults, and because some of what has been thought to be
true about the surge in defaults is not true (or incomplete).
Major points are:
The usual suspects matter. Looking at 2003 and 2006 vintage
default rates, lower downpayment meant higher defaults if
credit scores are held constant, and vice versa for credit
score with downpayment constant.
There are tradeoffs. A low downpayment can be offset with a
higher credit score. What does seem to matter is low
downpayment combined with low credit score. This is an example
of risk layering.
Economic conditions were very important. Loans originated
in 2006 had much higher defaults than those in 2003 for all
categories (of credit score and downpayment) and for both
Fannie/Freddie and private-label mortgages.
The Channel is very important. Private-label securities had
much higher default rates, even controlling for credit score
and LTV, than did Fannie/Freddie mortgages.
What is Risky?
The above describes things that were the case all the time. A more
important issue is what things were risky in the sense of causing
bigger changes in defaults from the good years (e.g., 2003) to the bad
years (e.g., 2006). Main results are:
Low downpayment, by itself was not especially risky. This
was especially true for loans that were not risk-layered. In
particular, there is no clear relation between downpayment and
increase in default rate, holding credit score constant.\1\
Furthermore most of the loss for low downpayment loans sold to
Fannie and Freddie was taken by private insurance companies.
---------------------------------------------------------------------------
\1\ This is probably because when dealing with prices falling by 40
percent in some regions, even downpayments of 20 percent provide much
less protection than might be thought at loan origination.
Low credit scores did matter, as did risk layering. This
---------------------------------------------------------------------------
was true for both channels.
Loans with LTV from 75 percent to 85 percent had the
biggest increase for every level of FICO. This might be because
loans involving moral hazard were more likely to have
downpayments right at 20 percent, and these loans were more
sensitive to declines in property values. This ``hump'' in the
risk profile is entirely from the 2006 vintage (see Table 2 in
the appendix); there was no such hump in the profile of loans
originated in 2003.
The channel mattered; Private-label loans had much bigger
increases across loan characteristics, by roughly twice.
The housing goals added little to the risk of the GSEs.
Size is ambiguous. The biggest intuitions (Fannie and
Freddie) had the lowest default rates and the lowest increase
during the recession. The private-label market, which was
served by a wide range of institutions, was much worse. On the
other hand Fannie and Freddie were very big, and the market was
clearly sensitive to their behavior. A hard to quantify
dimension of size is that it can generate ``franchise value''
(aka monopoly power), which has a tendency to produce risk
aversion to protect the franchise.
Institutional performance
The above focused on defaults by loan product, channel, etc. An
interesting question going forward is what type of institution
structure do we need? I am inclined to think that while this is
important, it is not crucial and that the key questions are incentives.
A question behind all of this is the role of fixed-rate mortgages in
our economy. They tend to have lower default rates that do adjustable-
rate loans, but leave many intuitions subject to interest rate risk
because their value fluctuates with interest rate changes. The GSEs and
private-label markets both provided access for fixed-rate loans to bond
market investors. This was less the case with private label because it
securitized a considerable amount of adjustable-rate mortgages. In any
event, because the overall size of the mortgage market (around $10
trillion in outstanding balance) is about the size of all the assets in
the banking system it is likely that some sort of securitization
structure will be needed. This can be done in a lot of different ways.
Here I outline some things we have learned lately:
The best source for private capital still might be Fannie
and Freddie (or their clones-bond market institutions with the
Government at the back end). They are currently profitable, as
evidenced by combined net income of $13 billion per quarter
(before accounting for the change in deferred tax assets). In
any event it long run net outlays by Treasury may well be close
to zero; that is it is possible that the residual value of
Treasury's stake is at least as big as the amount of money it
has put in.
Private-label security issuance was hugely dependent on
Collateralized Debt Obligations (CDOs) to buy the riskiest
parts of their deals pieces. There are important information
asymmetries in this market, which were behind the huge losses
in it. Investors would need huge coupons to be willing to buy
such ``information-intensive'' pieces. These costs would flow
through directly to borrowers.
It may be more efficient if the loss-bearing private
capital layer is an equity claim to a mortgage insurer or GSE.
Raising capital on a deal-by-deal basis, as in the private-
label market, through subordinated tranches is less efficient
because of information costs.
Gross business volumes in 2012 between Fannie and Freddie
were $1.4 trillion. It is unlikely that the private-label
market could replace these volumes at current mortgage interest
rates. Mortgages are complex instruments with multiple embedded
options. Fannie and Freddie absorbed most of the mortgage
credit risk and reduced the interest rate risk through retained
portfolios. Maybe this didn't work perfectly, but we don't have
any historical evidence that banks and capital markets can
manage these risks better.
Comment
A problem with all of this is that a lot of went wrong was very
hard to predict, and some proposals, for instance limiting low
downpayment loans and low-income lending, are not likely to help much.
The structure going forward is probably going to be something like
GSEs, maybe more of them, maybe as co-ops or specially chartered
mortgage banks, but with as much private capital as possible and with
Government stepping at the back end.
This is because having the Government as final risk-taker is going
to be hard to avoid, and probably shouldn't be avoided. That role can
provide stability, but it leaves open a lot of questions about the
details of risk and capital.
Capital and Contingent Capital
Whatever we do, we'll need better capital rules. Capital provides a
cushion that protects debt holders and guarantors, and it provides
incentives to control risk because more investor money is at stake.
Before the crisis, Fannie and Freddie had two capital rules applied to
them: stress tests that simulated company performance under stressful
conditions and required that enough capital be held to survive them and
a minimum capital requirement that applied even if they passed the
stress tests. Clearly they did not have enough capital to withstand the
Great Recession.
There are limits, however, to how far we can get by relying on
capital ratios and shutting down insolvent institutions. It is very
difficult to know whether or not institutions are really insolvent.
This is in part because that is a difficult problem, but especially
because accounting measures of capital are not up to the task. They
tend both to overestimate and underestimate net worth, and they tend to
be pro-cyclical, requiring institutions to raise capital at exactly the
times when this is most difficult. This leaves us with stress tests and
manipulating incentives. I shall leave stress tests, which I believe
are an excellent way of improving on required capital ratios, for
another time and focus on incentives in the form of contingent capital.
Incentives: Contingent Capital
You don't have to agree with the recent bailouts to understand the
difficult choices facing the Fed and Treasury when institutions like
AIG, or Fannie and Freddie or a slew of banks get into trouble and
threaten the rest of the financial system. In a well-ordered world
there would be clear rules for resolution via bankruptcy: rules for
settling claims would be clear and acted on quickly; bondholders would
take over, and there would be no need for panic. This is something that
happens relatively easily in structured securitization deals, but not
for actual corporations. Bankruptcy is costly and time consuming.
Uncertainty can breed panic and bank runs, and leave us with a choice
between a bailout and a meltdown.
The current debate about financial regulation is largely about
avoiding that choice in the future. It has focused on making insolvency
less likely by making financial institutions hold bigger capital
cushions and on making insolvency less costly by setting up resolution
systems. These are daunting tasks. However, there is a relatively easy
way of starting to address the problem: We can require banks, and other
financial institutions, to issue contingent capital, for instance in
the form of Conditional Convertible (or ``CoCo'') bonds. This can be
done by requiring issuance of bonds that look like regular bonds most
of the time, but which are automatically converted into common stock
when capital levels are low.\2\
---------------------------------------------------------------------------
\2\ A relatively early example of CoCo bonds is the first issue of
the ``Enhanced Capital Notes'' by the Lloyds Banking Group PLC in
November 2009. These are subordinated debt that will be converted into
equity if the core capital falls to 5 percent of its risk-weighted
assets. Other examples of CoCo bond issues are from Rabobank in May
2010, Credit Suisse in February 2011, and the more recent ones from
Barclays Bank in April 2013. By classifying into Tier 1 capital, the
mandatory leverage ratio required in the Basel Accord can be met
easier.
---------------------------------------------------------------------------
The automatic conversion gives CoCo bond investors a strong
interest in risk management (they can't assume a bailout). If mandated
as a part of pay, the bonds give management incentives to control risk,
and movements in the market price of the bonds will provide daily
evaluation of banks' risk. CoCo Bonds also limit concern about
institutions being ``too big to fail;'' because conversion avoids
bankruptcy it mitigates concerns about disruptions.
Why do this instead of just requiring more equity and less debt?
One reason is that debt has advantages. It is easier to evaluate than
equity, so it attracts a wider range of investors. Meeting debt
payments imposes discipline on management, and debt has tax advantages.
A second reason for liking CoCo bonds is that they address a
problem that higher capital ratios cannot easily solve: the problem of
banks that are still solvent but with low capital ratios. Suppose that
the minimum capital ratio for banks is 5 percent of assets. Banks will
keep a cushion above the required level, but not by much because equity
is more costly than debt. If a bank's capital ratio falls to, say, 3
percent, it will either have to raise capital or lower assets (lend
less) to get back to 5 percent. It will be solvent but in trouble.
During a period when large numbers of banks are missing their
ratios and there is a great deal of uncertainty, raising capital is
difficult, putting banks in the position of having to cut back assets.
In the extreme, getting back to 5 percent by shrinking the balance
sheet would mean a 40 percent cut in assets. This problem will exist
even with higher capital ratios as long as banks keep their ratios just
above the minimum. With suitable triggers the decline in stock price
that accompanies banks' declining asset values will convert CoCo bonds
into equity, providing an automatic and countercyclical cushion.
CoCo bonds are not entirely appealing to holders of the bonds, who
will want a higher interest payment on their bonds and will worry about
premature exercise. A way of handling some problems is to make the
shares convertible back into bonds if the company subsequently
recovers.
It is important to emphasize CoCo bonds as a tool of monitoring and
management. Traded bonds will provide a market read on the state of the
banks, which will not be clouded by questions of bailouts. Including
them in management's compensation, can provide some disincentives for
risk-taking. To the extent management is compensated with stock and
stock options (or close substitutes) it has the incentive to take on
risk in much the same way as shareholders. Imposing CoCo bonds as a
part of their package forces them to take on some downside.
Furthermore, these do not have to be traded, which mitigates some of
the criticism around the trading of CoCo bonds. They can be designed in
very specific ways (for instance by tying conversion to the bank's
asset value), which can unravel most of the disincentives that come
from the asymmetric of outcomes to owning shares.
Comments
There are lots of structures that can work in principle. Given
failure of the private-label market to provide market stability, having
Government ultimately be a guarantor is probably necessary and not as
scary as it might sound; it can enhance affordability and liquidity in
the market where mortgages are traded-making a TBA market readily
available. But there needs to be capital and incentives to limit risk-
taking. While there are several ways of doing it I think that
contingent capital can move incentives in the right direction.
References
FCIC, FCIC, ``Report of Financial Crisis Inquiry Commission,'' Jan 2011
http://fcic.law.stanford.edu/.
Gorton, Gary. 2009. ``Slapped in the Face by the Invisible Hand:
Banking and the Panic of 2007.'' Working Paper. New York: National
Bureau of Economic Research.
Lai, Rose, Neng and Robert Van Order, 2013, Remuneration Packages and
CoCo Bonds: Providing Disincentives For Risk-Taking.
Stanton, Thomas, 1991, A State of Risk: Will Government-Sponsored
Enterprises Be the Next Financial Crisis? Harper Collins.
Thomas, Jason and Robert Van Order, 2011, ``Housing Policy, Subprime
Markets and Fannie Mae and Freddie Mac: What We Know, What We Think
We Know and What We Don't Know.''
Van Order, Robert. 2000. ``The U.S. Mortgage Market: A Model of Dueling
Charters.'' Journal of Housing Research 11, 2: 233-55.
Van Order, Robert and Anthony Yezer (2011), FHA Assessment Report,
Third Edition. http://business.gwu.edu/creua/research-papers/files/
FHA2011Q3.pdf.
______
APPENDIX
What Is Credit Risk and Where Does It Come From?\3\
---------------------------------------------------------------------------
\3\ Prepared for Conference on ``The Future of Housing Finance,''
cosponsored by George Washington University and PWC, May 14, 2013.
---------------------------------------------------------------------------
Robert Van Order
Oliver Carr Chair in Real Estate
Professor of Finance and Economics
George Washington University
May 2013
The last decade has provided us with some great data on credit
risk-across product types, origination channels and risk
characteristics, because the market has experienced both good times
and, especially, bad. Most of the data we have are for loans that have
been securitized, by the Federal Agencies (Fannie Mae, Freddie Mac,
often referred to as Government-Sponsored Enterprises or GSEs) or in
the ``Private Label'' securities (PLS) market. Much of the data has
been proprietary. The following sets of pictures from available data
sets provide some summary information on where the risks have been.
While obviously just a snapshot of a much wider set of data, they
capture some important stylized facts.
A key point is that risk is not about the level of defaults; rather
it is about dispersion.\4\ We all know that some types of loans default
more than others, but those differences can be priced, and they may not
be especially important. If, for instance, pools of low downpayment
mortgages always have the same very high default rates, we could
readily price the loans to at least cover losses, in which case the
pools would not have any risk and would have a fixed, risk-free return.
---------------------------------------------------------------------------
\4\ More important, but not possible to cover here, is the notion
that risk applies to a whole portfolio, not just individual asset
types, which means diversification should count too.
---------------------------------------------------------------------------
That does not happen often, of course, but the example is
illustrative. What matters regarding risk is how far default rates
(more broadly, default costs) vary from what is expected when
conditions change. From this perspective, from the data presented,
there are some surprises: Low downpayment loans, by themselves, were
not especially risky; nor were ``affordable'' loans that Fannie Mae and
Freddie Mac were mandated to buy.
Defaults and Risk
The following three pictures, set up as tables, summarize some of
the data supplied by the Federal Housing Finance Agency (FHFA).\5\ Here
the focus is on fixed-rate mortgages. (Adjustable-rate loans have
similar properties but worse experience in general.) The data set
covers loans bought by Fannie Mae and Freddie Mac and those put into
PLS. The tables present matrices that show performance of the loans for
different origination years, controlling for two important measures of
credit risk: loan-to-value ratio (LTV) and borrower credit score
(measured by the Fair Isaac (FICO) statistical credit score).
---------------------------------------------------------------------------
\5\ See http://www.fhfa.gov/Default.aspx?Page=313.
---------------------------------------------------------------------------
The four LTV classes include:
75 percent and below, the safest category
75 percent-85 percent, the most common category, which
clusters around 80 percent
85 percent-95 percent, high LTV loans, which cluster around
90 percent
95 percent, which contain 95 percent and higher
Credit scores range from low 500s to 800; they are put into
discrete buckets. While there is not a clear definition of subprime, a
reasonable definition for our purposes is that subprime covers anything
with a credit score below 640 and anything from 640-680 with a loan-to-
value ratio over 85 percent.
Performance is measured by the share of loans of that year's
originations that were ever 90 days delinquent from the time of
origination through 2009. Table 1 depicts defaults on loans originated
in 2003, a good year because property values rose rapidly in the
following 3 years. For instance, the table says for loans with LTV less
than or equal to 75 percent and FICO score below 640 6.9 percent of the
loans originated in 2003 ever had at least one spell where they were 90
days delinquent. Table 2. looks at the same measure for loans
originated in 2006, a bad year with sharply declining housing prices.
Table 3. presents the differences between Tables 2. and 1.
Table 1. Default Rates: 2003 Vintage (Ever Seriously Delinquent)
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Table 2. Default Rates: 2006 Vintage (Ever Seriously Delinquent)
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
Table 3. Differences between Table 2 and Table 1
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
As can be seen by looking at the tables, default rates have varied
greatly by product, vintage and mortgage characteristic. Major points
with respect to the first two tables are:
The usual suspects matter. Looking at 2003 and 2006 vintage
default rates, higher LTV meant higher defaults if FICO scores
were held constant, and vice versa for FICO with LTV constant.
There are tradeoffs. For instance, in Table 1, for Fannie/
Freddie data, 95 percent or greater LTV loans with credit
scores in the (680-720) range had about the same default rates
as those loans below 75 percent LTV loans with low credit
scores (6.5 percent vs. 6.9 percent rates). What looks to be
worst is not simply high LTV or low FICO, but high LTV combined
with low FICO score. This is an example of risk layering.
Economic conditions were very important. The 2006 vintage
had much worse defaults than the 2003 vintage for all
categories and for both Fannie/Freddie and private-label
mortgages. The story is worse than the tables suggest because
the 2006 loans had only 3 years of exposure until 2009; whereas
the 2003 loans had six.\6\
---------------------------------------------------------------------------
\6\ This data set only tracks defaults through 2009, so it is not
possible to have comparable 3-year periods of exposure.
The Channel is very important. Private-label securities had
much higher default rates, even controlling for credit score
and LTV, than did Fannie/Freddie mortgages.
What is Risky?
But what about risk? Risk of default is not the same as expected
level of default. As discussed above, we know that high LTV loans have
high default rates, but to be riskier they must have more volatile
losses, rather than simply higher losses. If losses on loans (more
broadly on portfolios of loans) are more volatile, then the risk of
insolvency is higher even if the loans are correctly priced.
The data sample depicted above is too narrow for complicated
measures of volatility or dispersion. But it does depict a very severe
sort of ``one-shot'' volatility from the extreme differences between
2003 and 2006. This measure of risk is akin to analysis from a stress
test. If two products both have their losses increase by the same
amount in the face of stress, then even if their losses are quite
different on average, they are equally risky (and have the same
implications for insolvency under that particular stress).
Consider Table 3. It depicts differences between the first two
tables. It shows sensitivity to the very poor economic conditions after
2006, relative to the good conditions following 2003. It is a natural
stress test.
Main results are:
Low downpayment loans were not especially risky.\7\ This is
especially true in the middle of the matrices; for most
elements of both the GSE and PLS matrices there is no clear
relation between LTV and increase in default rate, holding FICO
constant.\8\
---------------------------------------------------------------------------
\7\ For Fannie/Freddie loans risk is even lower for high LTV loans
because most of the losses have been covered by private mortgage
insurance. http://fcic-static.law.stanford.edu/cdn_media/fcic-docs/
2009-06-04%20Freddie%20Mac-%20Cost%20of%20Affordable%20Housing
%20Mission.pdf.
\8\ This is probably because when dealing with prices falling by 40
percent in some regions, even downpayments of 20 percent provide much
less protection than might be thought at loan origination.
High FICO scores did matter, as did risk layering. This
was true for both channels; moving northeast from southwest in
---------------------------------------------------------------------------
the pictures lowered the lift from 2003 to 2006.
Loans with LTV from 75 percent to 85 percent had the
biggest increase for every level of FICO. This might be because
loans involving moral hazard were more likely to have LTVs
right at 80 percent and these loans were more sensitive to
declines in property values. This ``hump'' in the risk profile
is entirely from the 2006 vintage (see Table 2); there was no
such hump in the profile of loans originated in 2003.
The channel mattered; PLS loans had much bigger increases
across loan characteristics, by roughly twice.
The last two points are suggestive of moral hazard being associated
with 80 percent LTV loans after 2003, being in PLS pools, and being
sensitive to property value changes.
Low income and targeted lending
The above does not separate out low income and other types of
``affordable'' lending products that have been blamed for defaults.
Again, the question is not whether they had higher default rates (they
did), but whether they were riskier. Here I look at some data and
analysis provided to the Financial Crisis Inquiry Commission,\9\ which
compares actual performance with expected for various loan types, using
Freddie Mac data.
---------------------------------------------------------------------------
\9\ See ``Cost of Freddie Mac's Affordable Mission,'' presented to
Freddie Mac Board, June 4, 2009. See http://fcic-
static.law.stanford.edu/cdn_media/fcic-docs/2009-06-04%20Freddie%20
Mac-%20Cost%20of%20Affordable%20Housing%20Mission.pdf.
---------------------------------------------------------------------------
The picture looks at three types of loans: those that did not
qualify for housing goals (blue line), those that did qualify but were
not done via special programs (yellow line), and those done via
programs designed to attract goals-rich loans (green line). The
horizontal axis has default rates estimates before the fact (from
Freddie Mac models) and the vertical axis is corresponding actual
default rates.
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
The lines show that all three types did considerably worse than
predicted. However, the blue and yellow lines (regular business and
special affordable programs) are very close, indicating that the
reaction to the Great Recession shock was the same for regular as it
was for affordable loans. This bit of evidence suggests that the
housing goals added little to the risk of the GSEs.
In summary, the news from the two sets of pictures is that two
types of loans that might be thought to have been risky, low
downpayment and ``affordable,'' have not been especially risky. Risk,
in the Great Recession stress test was largely due to economic
conditions, the channel through which the loans were made and layered
risk loans. http://business.gwu.edu/creua/research-papers/files/
FHA2011Q3.pdf.