Docstoc

Introduction to Bond Markets_ Analysis_ and Strategies

Document Sample
Introduction to Bond Markets_ Analysis_ and Strategies Powered By Docstoc
					Mortgage Pass-Through
Securities


    Chapter 11
Pass-Through Securities

   created when one or more mortgage holders form a
    collection (pool) of mortgages and sell shares
    (participation certificates) in the pool
       pass-throughs are then basis for other derivatives (CMOs
        and stripped mortgage-backed securities)
   CFs come from mortgage pmts from pool of
    mortgages
       timing differs from mortgage pmts
       amt differs from mortgage pmts – difference due to
        servicing fees and other fees for guaranteeing issue
       CFs not known with certainty because of prepayments
WAC and WAM

   WAC – weighted-average coupon
       weight mortgage rate of each mortgage in pool by
        amt of mortgage outstanding
   WAM – weighted-average maturity
       weighting remaining # of months to maturity for
        each loan in pool by amt of mortgage outstanding
Agency Pass-Throughs

   Ginnie Mae
   Freddie Mac
   Fannie Mae
   types of guarantees
       fully-modified PTs – timely pmt of principal and
        interest
       modified PTs – guarantees interest and principal
        but only timely payment of interest
Agency Pass-Throughs

   Ginnie Mae PTs
       guaranteed by full faith and credit of US
        government
       essentially default risk free
       Ginnie Mae guarantees security referred to as
        mortgage-backed security (MBS)
           fully modified pass-throughs
           only include mortgages guaranteed by Rural Housing
            Service, Veteran’s Association, or Farmers Home
            Assoc.
Agency Pass-Throughs

   Freddie Mac PTs
       their pass-through is called a participation certificate
       not guaranteed by US government but …
       modified pass-throughs
       Gold PC – introduced in 1990
           fully modified PT
           eventually only PC that Freddie will issue
   Fannie Mae PTs
       MBS
       not obligation of government because government-
        sponsored agency not government agency
       fully modified pass-throughs
Non-Agency Pass-Throughs

   issued by commercial banks, thrifts, and
    private conduits
       purchase nonconforming mortgages, pool, and
        sell pass-throughs which have underlying pool as
        collateral
       same thing as Agency except not guarantee of
        government
   registered with SEC
   rated by same firms that rate debt
Credit Enhancements

   rating companies consider
       type of property
       type of loan
       term of loan
       geographical dispersion of loan
       loan size
       purpose of loan
   rating given but can be changed by credit
    enhancement (this has been key to growth of this
    type of security in market)
Credit Enhancements
   external
       3rd party guarantees that provide first-loss protection
        against losses up to a point (say 10%)
       bond insurance – same as muni bond insurance
       pool insurance – covers losses from defaults and
        foreclosures
           usually for $ amt for life of pool
           some written so $ amt declines as pool seasons as long as
               credit performance is better than expected
               ratings agencies approve
           need additional insurance to cover losses from bankruptcy or
            fraud
       rating of 3rd party must be at least as high as rating sought*
Credit Enhancements

   internal – may change CFs even with no default
       reserve funds
           cash – deposits of cash generated from issuance proceeds
           excess spread accounts – allocation of excess spread or cash
            into separate reserve account after paying net coupon
       overcollateralization
           principal amt of issue < principal amt of pool of loans
       senior/subordinate structure – most widely used
           subordinate class absorbs all losses up to amt in class
           subordinate class has higher yield
           shifting interest structure – redirects prepayments from
            subordinate class to senior class according to given schedule
            (want to maintain insurance)
Prepayment

   value of any security is what?
       issue for PTs why?
   prepayment speed
       CF yield – yield calculated based on projected CF
   prepayment conventions
       FHA experience – no longer used since
        prepayment rates are related to interest rate cycle
       conditional prepayment rate
       PSA prepayment benchmark
Conditional Prepayment Rate

   single-monthly mortality
    rate
       CPR is annual so convert
        to monthly rate to find
        amt of monthly
        prepayments
   SMM rate and
                                   SMM  1  (1  CPR)   1/ 1 2



    prepayment
       assume that
        approximately x% of
        remaining balance
        prepays at beginning of
        month
PSA Prepayment Benchmark

   Public Securities Association (PSA) benchmark
       expressed as monthly series of annual prepmt rates
       assumes prepayment rate increases as loans become
        more seasoned
       assumes following CPRs for 30-year mortgages
           CPR of 0.2% for 1st month and increased by 0.2% per year
            each month for next 30 months
           6% CPR per year for remaining years
       benchmark referred to as 100 PSA
           if t<= 30, CPR = 6%(t/30)
           if t>30, CPR = 6%
PSA Benchmark

   50 PSA – assuming prepayment rate of half
    the CPR of the benchmark
   150 PSA – rate 1.5 times CPR of PSA
    benchmark
   SMM for month 5 assuming 100 PSA
       CPR = 6%(5/30) = 1%

           SMM  1  (1  0.01)   1/1 2




        = 0.000837
Monthly CF Construction

   assume underlying mortgages are fixed-rate
    level payment with WAC = 8.125%
       pass-through rate is 7.5% and WAM of 357
        months
       assume 100 PSA
       in second example, assume 165 PSA
Prepayment Models

   models statistical relationships among factors
    expected to affect prepayments
   models used to view borrowers as generic
       more data available now so models are more
        complex
       models differ for agency and nonagency MBS
   book presents models developed by Bear
    Stearns
Agency Prepayment Models

   not as much data available on individual loans so
    models done at “pool” level
   components in Bear Stearns model
       housing turnover – existing home sales
           family relocation due to changes in employment and family
            status (change in family size, divorce)
           trade-up and trade-down activity due to changes in interest
            rates, income, and home prices
           insensitive to level of mortgage rates
       cash-out refinancing
       rate/term refinancing
Agency Prepayment Models

   cash-out refinancing
       refinancing by borrower in order to monetize the price
        appreciation of the property
       depends on increase in housing prices in region where
        property is located
       favorable tax law regarding capital gains adds incentives to
        monetize price appreciation (exempts gains up to
        $500,000)
       may be economical even if mortgage rates are rising and
        with transaction costs
       cash-out refinancing is tied to housing prices and
        insensitive to mortgage rates
Agency Prepayment Models

   rate/term refinancing
       means borrower has gotten new mortgage on same
        property to save either on interest cost or shortening life of
        mortgage with no increase in the monthly payment
       decision whether or not to refinance is due to PV of $
        interest savings from lower rate after subtracting estimated
        costs to refinance
       proxy for rate/term refinancing for model:
           difference between prevailing rate and note rate – not good
            proxy
           better on is refinancing ratio – note rate to current rate
           WAC is numerator
           ratio < 1 – note rate less than current so no incentive to
            refinance
           ratio > 1 – some incentive possibly to refinance
Housing Turnover in Agency Prepayment
Models
   factors used by Bear Stearns
       seasoning effect – (see graph on next slide)
           idea is that you must recognize the homeowner’s tenure in the
            house – may not be same as age of loan because of possible
            refinancing
       housing price appreciation effect
           over time LTV of home changes due to either amortization of
            loan or change in value of home – incentive to refinance if
            value of home goes up
           need to estimate prepayments due to housing appreciation
           Bear Stearns used Home Appreciation Index (HPI) – (see
            slide)
       seasonality effect
           home buying increases in spring and peaks in late summer –
            buying declines in fall and winter – prepayments follow similar
            pattern but may lag a bit with peak closer to early fall
Bear Stearns Baseline HTO Prepayments
for Agency MBS
Effect of Housing Price Appreciation of
Agency Prepayments
Cash-Out Refinancing

   driven by price appreciation since loan origination –
    need proxy for this
       Bear Stearns uses HPI
       see slide to show cash-out refinancing incentives for 4
        assumed rates of appreciation
       according to Bear Stearns model, projected prepayments
        due to cash-out refinancing
           exist for all ratios greater than 0.6
           prepayments increase as the ratios increase
           the greater the price appreciation for a given ratio, the greater
            the projected prepayments
Effect of Housing Price Appreciation on
Cash-Out Refinancing on Agency
Prepayments
Rate/Term Refinancing Component

   decision to refinance not based totally on note rate
    relative to current rate
       S-curve for prepayments
           if totally based on ratio, why does curve flatten out (or
            prepayment rate flatten out)
               because borrowers left in pool can not get refinancing or some
                have other reasons why refinancing does not make sense
       S-curve not sufficient for modeling rate/term refinancing –
        ignores 2 things that affect decision:
           burnout effect – Bear Stearns use some pool variables as
            proxies:
               original term, loan purpose, WAC rate, weighted average loan
                age, loan size, rate premium over benchmark, yield curve slope
                (see slide)
           threshold media effect
Baseline Refinancing Function for Bear Stearns’ Agency
Prepayment Model for an “Ordinary” Pool of Agency Borrowers




  original loan of $125,000, age of 12 months, no rate premium at origination, no prior option to
  refinance, 3.5% annual home price appreciation
Baseline S-Curve for Agency Borrowers
Based on Loan Amount




  shows model’s S-curves for $25,000 loan size increments – relative to loans with balances less than
  $100,000, loan balances that exceed $150,000 are about 1.5 to 2.5 times more sensitive to refinancing
Nonagency Prepayment Models

   same components as in Agency models
       but more info. on these loans so prepayment model
        estimated for each type of loan (rather than for pool of
        loans)
       Bear Stearns gives projected prepayment rates based on
           size of loan
           rate premium
           documentation
           occupancy status
           current LTV
Baseline Projected Prepayment Rate Across a Range of
Refinancing Incentives for 3 Loan Types
Cash Flow for Nonagency PTs

   CFs not affected by default and delinquency
    for agency PTs
   PSA has issued a standardized benchmark
    for default rates
   SDA benchmark gives annual default rate for
    a mortgage pool as a function of the
    seasoning of the mortgages
       can use multiples of default rate similar to
        prepayment benchmark – so can have 200 SDA
Cash Flow Yield
   rate that makes the PV of expected CF equal to the price
     bond-equivalent yield

           market convention for annualizing yield on fixed-income security that
            pays interest more than once a year
           found by doubling a semi-annual yield
           semi-annual yield for PT is
            semiannual cash flow yield = (1 + y M)6 – 1
            where yM is monthly interest rate that equates PV of projected monthly CF to
               price of PT
             bond-equivalent yield = 2[(1 + yM)6 – 1]

       must specifiy underlying prepayment assumption
       to realize this yield, investor must reinvest CFs at yield, investor
        must hold PT until all mortgages paid off, and assumed
        prepayment rate must actually occur
           so be careful in placing too much confidence in yields!
Average Life

   average life of MBS
       average time to receipt of principal payments
        (scheduled and prepayments) weighted by
        amount of principal expected



       average life depends on prepayment assumption
Prepayment Risk

   assume you buy 10% coupon GNMA with market
    rates of 10%
       assume rates in market fall to 6%
       consequences?? - contraction risk
           price increase but not as large as increase for option-free
            bond
               negative convexity
           CF reinvested at lower rate
       assume rates rise to 15% - extension risk
           price of PT falls but will fall more because rate increase slows
            down rate of prepayments
           problem for investor is this is exact time that they want rate of
            prepayments to increase so they have money to reinvest at
            higher market rate of 15%
Asset / Liability Management

   PTs unattractive to some institutions
       depository institutions exposed to extension risk when they
        invest – they want to lock in spread over cost of funds – PT
        is longer term than their liabilities
       life insurance companies are exposed to extension risk
        when using PTs – they might issue 4 year GIC – problem is
        uncertainty about CF from PTs that they will receive to
        have to pay off GIC
       pension fund is exposed to contraction risk using PTs –
        they have long-term liabilities and want to lock in current
        rates – exposed to risk that prepayments will speed up and
        maturity of investments will shorten (happens when rates
        fall) and they will have to reinvest prepayments at lower
        rate
Secondary Market Trading

   quoted same manner as Treasuries
       94-05 means 94 and 5/32nds of par or
        94.15625% of par
   PTs identified by pool prefix and pool number
       prefix tells type of PT – 20 for Freddie Mac PC
        means underlying pool of conventional mortgages
        with original maturity of 15 years
       prefix of AR for GNMAs means ARMs
       pool number tells specific mortgages underlying
        PT and issuer of PT

				
wanghonghx wanghonghx http://
About