Credit derivatives and securitisation

Credit derivatives and securitisation transformed credit markets in the 1990s and 2000s. They allow risk to be priced, traded, and transferred independent of the underlying loan. This is good — risk goes to whoever can bear it cheapest. It's also dangerous when the chain becomes so long that no one can see the underlying credit clearly. The 2008 crisis was largely a securitisation crisis.

Credit Default Swaps (CDS)

A CDS is insurance against a credit event (default, restructuring, bankruptcy) on a specific reference entity. The buyer of protection pays a periodic premium (the CDS spread, expressed in basis points per year of notional); the seller of protection agrees to make a payout if the reference entity experiences a credit event during the contract's life. CDS spreads trade actively for major sovereigns and large corporates and are quoted on Bloomberg and Tradeweb at multiple tenors (1Y, 3Y, 5Y, 10Y), with 5Y being the most liquid.

The CDS premium leg and contingent leg

Annual premium leg (buyer pays seller):

    Premium $  =  s  ×  N  ×  ( days in period / 360 )

Contingent leg (seller pays buyer on a credit event):

    Payout $   =  ( 1  −  R )  ×  N

where:
  s     =  the CDS spread, expressed as a decimal (500 bps = 0.0500)
  N     =  the notional amount of protection in currency units
            — the face amount the contract references; the protection
            buyer can hold less, more, or none of the actual bonds.
  R     =  the recovery rate, expressed as a decimal
            — the post-default value of $1 of the reference's bonds
            — historically ≈ 40% for senior unsecured corporate,
              25-40% for sovereigns, lower for subordinated paper
  1 − R =  the loss given default — what the seller actually pays out
            in 'cash settlement' terms once a credit event triggers
  360   =  the money-market day-count convention used in standard
            ISDA CDS documentation

The buyer's economic position is short the credit; the seller is long. The CDS spread s is the market's price for that credit risk, paid annually until either a credit event or maturity. If no event occurs over the contract's life, the seller has accumulated all the premium payments and paid nothing. If a credit event occurs, the seller delivers (1 − R) × N in cash to the buyer (or, less commonly, takes physical delivery of defaulted bonds and pays N).

A worked CDS trade

Trade: Buy $10m notional of 5-year Kenya sovereign CDS
Market: 5Y Kenya CDS quoted at 500 bp
ISDA-standard recovery assumption: R = 25% (sovereign)

Annual premium leg:
  Premium = 0.0500 × $10,000,000 × (360/360) = $500,000 per year
  Paid quarterly: $125,000 per quarter

Scenario 1: Kenya does NOT default within 5 years
  Buyer has paid 5 × $500,000 = $2.5m in premiums
  Seller keeps it all, pays nothing.
  Buyer's loss = $2.5m (an explicit cost for protection that was unused).

Scenario 2: Kenya defaults in year 3
  Premiums paid before default: 3 × $500,000 = $1.5m
  Payout received: (1 − 0.25) × $10,000,000 = $7.5m
  Buyer's net P&L: $7.5m − $1.5m = +$6.0m
  (Plus the partial-period accrued premium pro-rated to default date.)

The CDS spread (500 bp here) is the market's annual price for Kenya credit
risk. It is typically a few points wider than Kenya's bond spread over
Treasuries (the 'basis'), reflecting CDS-market-specific liquidity and
structural factors plus the convention difference on R.

What CDS gives you

• Synthetic exposure to credit risk without owning the bond. Useful when the bond is hard to source (off-the-run, thinly-traded) or when funding the bond purchase is expensive.
• Hedging — a bank with KES 10bn of Kenyan corporate loans can buy Kenya sovereign CDS to partly hedge the country-risk component. The hedge is imperfect (corporate names move differently from sovereign in stress) but better than nothing.
• Speculation — express a view on credit without committing capital to buy the underlying bond. Useful when leverage and capital efficiency matter.
• Price discovery — CDS markets often move before bond markets, particularly for sovereigns, because the CDS market draws different participants and trades faster.

Securitisation — pooling and tranching

Securitisation takes a pool of loans (mortgages, auto loans, credit-card receivables, corporate loans) and issues new securities backed by the cash flows from that pool. The new securities are tranched by seniority: senior tranches get paid first and absorb losses last (low yield, high rating); equity and mezzanine tranches absorb losses first and earn higher yields. The transformation is mechanical: a single pool of mostly-investment-grade-like credit produces a small AAA slice and a small first-loss slice, with the relative sizes determined by the historical loss distribution and the desired ratings on each tranche.

The tranching math

Loss allocation rules in a typical securitisation:

  Pool experiences cumulative losses of L (in % of pool size)
  Tranche thicknesses: equity = e%, mezzanine = m%, senior = s%
  By construction: e + m + s = 100%

Loss absorbed by each tranche, computed bottom-up:
  Equity tranche loss   =  min( L, e )                  / e         × 100%
  Mezzanine tranche loss = min( max(0, L − e), m )       / m         × 100%
  Senior tranche loss   =  max( 0, L − e − m )           / s         × 100%

where:
  L     =  cumulative pool losses, expressed as a percent of pool notional
  e     =  equity tranche thickness, expressed as a percent of pool
            — typically 2-8% in modern CLOs / MBS deals
  m     =  mezzanine tranche thickness, expressed as a percent of pool
            — typically 8-15% across one or more mezz tranches
  s     =  senior (AAA) tranche thickness, expressed as a percent of pool
            — typically 75-90% in modern deals
  min, max  =  the floor and ceiling functions ensuring losses don't
               'wrap around' a tranche once it is fully wiped out

Example: $500m pool with 8% equity ($40m), 12% mezz ($60m), 80% senior ($400m).
If cumulative losses reach $50m (10% of pool):
  Equity absorbs:    min(50, 40) = 40   →  100% loss on equity tranche
  Mezz absorbs:      min(max(0, 50−40), 60) = 10  →  10/60 = 16.7% loss
  Senior absorbs:    max(0, 50 − 40 − 60) = 0     →  no loss on senior

The senior tranche is only at risk if cumulative losses exceed the subordinate cushion (e + m). With e + m around 20% in a typical pre-2007 MBS deal, the senior tranche received AAA ratings under historical loss distributions that bottomed out around 5-10%. The 2008 crisis broke that assumption — nationwide housing-price declines correlated defaults across regions, pool losses exceeded the historical worst case, and AAA tranches absorbed real losses.

African securitisation

Africa has limited securitisation markets — partly because mortgage and credit-card lending is shallow, partly because the legal infrastructure (trust law, special-purpose vehicles, secured-creditor enforcement) is uneven. South Africa has the most developed market. Kenya has done occasional securitisations (utility receivables, vehicle finance) but they're rare. The Capital Markets Authority Act now permits Asset-Backed Securities formally; activity should grow.