Transcript March_2009_NABE_SilVal

The Subprime Financial Crisis
Roots of the Crisis
Susan Woodward, Sand Hill Econometrics
1
© 2004 Sand Hill Econometrics. All rights reserved.
Prelude
1980 – change in Usury laws Depository Institutions Deregulatory and Monetary Control Act
1990 to 1996 – introduction of credit scores
1998 – clearing-house for derivatives killed by bi-partisan leaders
(clearing-house would know gross and net exposure of buyers & sellers)
2004 – SEC eliminated capital rules for investment banks
Average ibank ratios of capital to assets:
before 2004:
1 to 12
after 2004:
1 to 33
2
© 2004 Sand Hill Econometrics. All rights reserved.
Real Residential Construction, 1968-2008
Residential Construction, 1968-2008
2,500
2,000
1,500
1,000
500
20
06
20
04
20
02
20
00
19
98
19
96
19
94
19
92
19
90
19
88
19
86
19
84
19
82
19
80
19
78
19
76
19
74
19
72
19
70
19
68
0
Year
3
© 2004 Sand Hill Econometrics. All rights reserved.
Residential Construction
Has always been volatile.
Is very sensitive to the level of nominal interest rates – when rates rise,
construction contracts
Construction is such a large sector that when it contracts, it can create
a recession all by itself
Has had a major role in most recessions (obvious in the chart), but not
in 2001
As of 2008q3, more than all of the shortfall in GDP was accounted for
by the decline in residential construction (roughly $350bn)
4
© 2004 Sand Hill Econometrics. All rights reserved.
Mortgage Interest Rates, 1987-2009
by 2003, the big refi boom was over
5
© 2004 Sand Hill Econometrics. All rights reserved.
Housing market experience, 1990-2003
1990-2004: high LTV loans to borrowers with good credit had low default rates
50 years (maybe more) of rising dollar house prices
20 years of low, 2-3%, and mean-reverting inflation,
20 years of (mostly) declining mortgage interest rates
A recession in 2001 with virtually no housing component
Lower volatility everywhere: stock market, the bond market, business income,
residential construction, real activity, personal income, and more…
By 2003, when 30-year fixed mortgage rates reached 5.25%, the mortgage refi boom had
to end, and new lending either had to contract a lot, or lenders had to think of
something new to do.
6
© 2004 Sand Hill Econometrics. All rights reserved.
Mortgage Originations, 2001-2007, $ trillions
4,000
subprime
3,500
prime
3,000
2,500
2,000
1,500
1,000
500
0
2001
2002
2003
2004
2005
2006
2007
7
© 2004 Sand Hill Econometrics. All rights reserved.
Note that 2003
Was the last year of the refi boom. The decline in
originations would have been even bigger in 2004 if
lenders had not begun chasing subprime borrowers
aggressively. In 2006, subprime loans were nearly 40%
of all originations.
8
© 2004 Sand Hill Econometrics. All rights reserved.
Little signs that all was not well…
9
© 2004 Sand Hill Econometrics. All rights reserved.
Vacancies
The vacancy data are very telling. Note that we have quite
a long history for this series– back to 1956. In the fourth
quarter of 2005, the vacancy rate for owner-occupied
homes reached 2% for the first time since we began
keeping track of vacancies.
These vacancy rates are for owner-occupied dwellings.
Vacancy rates for rentals are much higher (5-8%) and
also more volatile.
10
© 2004 Sand Hill Econometrics. All rights reserved.
Who had a problem?
Many investment banks were big holders bad pieces of subprime
loans and became insolvent. All lenders stopped trusting them.
Some large commercial banks were big holders of subprime loans,
and close to insolvent. Other banks stopped trusting them.
Sellers of insurance (AIG) (especially insurance not regulated as
insurance, such as credit default swaps) to subprime lenders and
investors were under-capitalized and close to insolvent, lenders
stopped trusting them, too.
Some households defaulted on their loans.
Some households felt poorer because the value of their houses and
portfolios were down, so they spent less.
11
© 2004 Sand Hill Econometrics. All rights reserved.
Crisis was mainly a Large Bank crisis
12
© 2004 Sand Hill Econometrics. All rights reserved.
Was the financial crisis the precipitating
event of the recession?
No – it was the crash in house prices. Residential construction – just
under 5% of GDP. If it falls 40%, that’s a recession.
As of 2008q3, more than all of the shortfall in GDP was accounted for
by residential construction. Q4 shortfall in GDP is broader.
Credit spreads always widen in recessions, bank crisis or not.
But the bank crisis cannot have made things better. It could have made
things much worse if the Fed had not moved to shore up bank
capital.
13
© 2004 Sand Hill Econometrics. All rights reserved.
Stuff done right
Inject capital into banks
increased inter-bank trust
interbank lending resumed
14
© 2004 Sand Hill Econometrics. All rights reserved.
Stuff not done right
Initial Focus on auctioning (pricing) assets
Allowing markets to suspect Fannie and
Freddie were without federal backing
F&F borrowing rates were 140 basis points> treasuries
No fiscal stimulus put in place in 2008
15
© 2004 Sand Hill Econometrics. All rights reserved.
What can we do now?
Short-run: Fiscal Stimulus -expected GDP shortfall $900 bn in 2009
tax cuts or rebates
investment tax credit
public works
subsidy (or tax relief) to employers
negative (or zero, with Federal govt paying states) sales tax
inflate
16
© 2004 Sand Hill Econometrics. All rights reserved.
Tax cuts or rebates
Tried in 2008
- no detectable impact
- consistent with theory (people smooth consumption
over time, if they can)
- consistent with other studies of windfalls
17
© 2004 Sand Hill Econometrics. All rights reserved.
Investment Tax Credit
PRO
Experience suggests it does move investment forward in
time
CON
Money goes mainly to manufacturers of capital goods
and to skilled labor
18
© 2004 Sand Hill Econometrics. All rights reserved.
Public works
PRO
Large body of research suggests multiplier lies between 1 and 1.5
(for each $1 spent, GDP rises by $1 to $1.50)
CON
Impact cannot be as quick as sales tax cuts or employment
subsidies
Benefits go mainly to contractors and skilled labor
Government is generally not too good at choosing construction
projects. Let the States do it.
19
© 2004 Sand Hill Econometrics. All rights reserved.
Subsidies to Employers
PRO
impact is directly on employment
can be done quickly
numbers are big: Businesses pay $500 bn/yr in payroll (social security)
taxes
CON
Money goes to businesses directly, not households
Impact on GDP is much disputed, multiplier could be well under 1.0,
suggesting disproportionate benefits to business owners and a net cost to
taxpayers
20
© 2004 Sand Hill Econometrics. All rights reserved.
Negative -- or zero -- Sales Tax
PRO
Evidence says it will raise GDP
Can be done quickly (all but 5 states have a sales tax)
Numbers are big – about $400 bn per year
CON
Political bonus points are smaller because beneficiaries are so
diffuse
Works best if it is a surprise
21
© 2004 Sand Hill Econometrics. All rights reserved.
Inflate
PRO
Higher inflation does stimulate real activity
CON
It takes about 1.5 years for the impact to be complete
Inflation above 4% would ruin our reputation for a
national commitment to low and stable inflation, a
reputation we have spent 20 years building.
22
© 2004 Sand Hill Econometrics. All rights reserved.
The short evaluation …
Timid
– GDP shortfall is forecast to be more than $900 bn in 2009.
– only $200 bn of the stimulus is to be spent in 2009
Inappropriate
– most of the money is spent in 2010 rather than 2009
– money is spent on public works, much of which is pork, instead of
by people
23
© 2004 Sand Hill Econometrics. All rights reserved.
The Details
Increased outlays by federal agencies
Grants to individuals and state-local governments
Tax cuts
Total
2009
2010
2011
34.8
110.7
76.3
85.3
108.6
49.9
64.8
180.1
8.2
184.9
399.4
134.4
Stimulus, billions of dollars, by fiscal year (October to September),
figures from CBO
24
© 2004 Sand Hill Econometrics. All rights reserved.
After another several months of bad employment figures,
they will be back for more…
25
© 2004 Sand Hill Econometrics. All rights reserved.
How bad do things look now?
26
© 2004 Sand Hill Econometrics. All rights reserved.
What would be a more direct stimulus?
1. Federal govt pays state sales taxes, 100% for the first
quarter, 2/3 next quarter, 1/3 the next, then done. ($400 bn
per year, only 6 states without a sales tax)
2. Increase level and duration of unemployment benefits
3. Investment tax credit to businesses
4. Pay employees’ part of payroll (social security) tax
($450 bn), perhaps also employers’ part
27
© 2004 Sand Hill Econometrics. All rights reserved.
What about banks?
Good bank/Bad bank proposals
most proposals suggesting separating floundering institutions into
two entirely separate entities. Put assets of questionable value and
all long-term debt into the “bad” bank.
Two problems:
1. The bad bank will almost surely be insolvent
2. The debt-holders would unquestionably be worse off, because
they would see a smaller pool of assets from which to recover.
28
© 2004 Sand Hill Econometrics. All rights reserved.
Another idea…
Make the good bank an asset of the bad bank.
Debtholders still can recover from the total pool of assets
29
© 2004 Sand Hill Econometrics. All rights reserved.
Example: Citi
Citi now
Good bank
Bad bank
1,325
610
Assets
Loans
Equity in other bank
Total assets
1,935
-
-
427
1,935
1,325
1,037
780
780
Bonds
1,144
118
1,026
Equity
11
427
11
Liabilities
Deposits
Capital ratio
0.5%
32.2%
-
1.0%
30
© 2004 Sand Hill Econometrics. All rights reserved.
What’s the point?
Change incentives – low risk v. high risk lending
Make the next steps for reorganization clear
Reassure depositors (including foreign deposits) that the good bank is
solvent so that they don’t run
PROBLEMS
Still a blow to the long-term debt-holders (are they pension plans and
insurance companies?) though not as large as other good/bad bank plans
Ambiguity of standing of long-term debt-holders and foreign deposits
Provides little guidance regarding entities like AIG
31
© 2004 Sand Hill Econometrics. All rights reserved.
What next?
Short-term
another stimulus bill, more targeted at consumption and employment
Longer-term
less levered financial institution, especially depositories, likely
there are any
ibanks also, if
some arrangement for keeping track of CDS
re-thinking of retirement schemes (British usage)
limitations on activities of insured depositories, both active (subprime lending) and
passive (purchase of CDS on assets held)
Regulate everything that is insurance (like CDS) as insurance, with capital
requirements
32
© 2004 Sand Hill Econometrics. All rights reserved.