Case 089Credit derivatives and counterparty riskHard
A bank's five-year swap with a mid-sized client has expected positive exposure of Rs 8, 12, 14, 11 and 5 crore by year. Default probability is 2% a year, loss given default 60%. Compute the CVA and say how collateral would change it.
1The situation
Harsil Bank has a five-year interest rate swap with a mid-sized manufacturing client. There is no collateral agreement. The bank's exposure model gives the expected positive exposure, the average amount the client would owe the bank if it defaulted, for each year: Rs 8, 12, 14, 11, 5 crore.
The credit team puts the client's default probability at 2% a year and the loss given default at 60%. Discount factors for years one to five are 0.94, 0.88, 0.83, 0.78, 0.73. The trading desk wants the credit valuation adjustment to charge into the swap price, and the client has asked whether signing a collateral agreement with a Rs 3 crore threshold would improve its price.
2Your task
What is the CVA, which year contributes most, and how much would the collateral agreement save?
Quick check
Which year contributes the most to the CVA?
Worked solution
Try it on paper, then open one step at a time.
30-second answerThe answer to give first
The CVA is about Rs 0.49 crore, with year 3 the largest contributor, and a Rs 3 crore threshold collateral agreement would cut it to roughly Rs 0.14 crore. Each year's contribution is loss given default times expected exposure times the chance of default in that year times the discount factor. The default chance in year t is 2% times the 98% survival of each earlier year. Collateral leaves the default odds alone and shrinks the exposure above the threshold.
Step 1What is the CVA actually measuring?
If you lend a friend money that he repays only when you ask, your risk is how much he owes on the day he disappears, times the chance he disappears, times how much you would not get back. The credit valuation adjustmentThe price of the chance that a derivatives counterparty defaults while owing you money. It is deducted from the trade value and charged to the client. is that same product, summed year by year and discounted to today. The swap's value to the bank moves with rates, so the amount owed is an average over scenarios where the client owes the bank: the expected positive exposure.
| LGD | loss given default, 60% |
| EPE_t | expected positive exposure in year t, Rs crore |
| q_t | the chance of defaulting in year t, having survived to it |
| D_t | discount factor for year t |
Step 2How does the sum come out year by year?
Work the marginal default chance first: 2.00%, 1.96%, 1.92%, 1.88%, 1.84%. The chance falls slightly each year because the client must survive the earlier years to default in a later one. Then multiply across. Year 1 gives 0.6 times 8 times 2.00% times 0.94, Rs 0.090 crore. Year 3 gives 0.6 times 14 times 1.92% times 0.83, Rs 0.134 crore, the largest piece, because the exposure peaks there. The total is Rs 0.4856 crore, about Rs 0.49 crore. Using a flat 2% every year, ignoring survival, gives Rs 0.503 crore, a small overstatement worth naming.
| Year | EPE, Rs crore | Default chance that year | Discount factor | CVA, Rs crore |
|---|---|---|---|---|
| 1 | 8 | 2.00% | 0.94 | 0.0902 |
| 2 | 12 | 1.96% | 0.88 | 0.1242 |
| 3 | 14 | 1.92% | 0.83 | 0.1339 |
| 4 | 11 | 1.88% | 0.78 | 0.0969 |
| 5 | 5 | 1.84% | 0.73 | 0.0404 |
| Total | 0.4856 |
Step 3Why does the exposure hump in the middle?
Two forces pull against each other. Early on, rates have had little time to move, so the swap's value cannot have drifted far from zero. Late in the life, few payments are left, so even a large rate move changes the value by little. The peak sits where rates have had time to move and enough payments remain for the move to matter, typically a third to a half of the way through. A cross-currency swap with a final exchange of principal would not hump; its exposure would keep climbing to maturity.
Step 4How much does collateral save, and what is left?
Under a collateral agreement the client posts collateral for whatever it owes above Rs 3 crore. Collateral does not change the client's default chance; it shrinks the exposure bars, here to at most Rs 3 crore a year, which cuts the CVA from Rs 0.49 crore to roughly Rs 0.14 crore, about 70% less. The figure is approximate: capping the average at the threshold slightly overstates the remaining exposure, and it leaves out the gap risk, the days between the client's last margin payment and the bank closing out after a default, during which rates can move again. A desk would model that margin period of risk explicitly.
Close with what the client hears. The swap price carries a credit charge of about Rs 0.49 crore today, spread into the fixed rate. Signing the collateral agreement would cut that to about Rs 0.14 crore, in exchange for the client finding liquid collateral when rates move against it. For a mid-sized company that liquidity cost can be larger than the saving, which is why many such clients stay uncollateralised and pay the charge.
Where candidates lose it
The common loss is multiplying total exposure by the five-year default probability in one step, which ignores both the shape of the exposure and when default is likely. The sum must be built year by year.
The second is saying collateral reduces the probability of default. It does not; it reduces how much is owed when default happens, and it leaves a residual exposure for the days between a missed margin call and the close-out.
What the interviewer asks next
- How would the CVA change if the client's credit spread doubled?
- What is wrong-way risk, and would it apply to this swap if the client borrows at a floating rate?
- How would the bank's own default risk enter, and why is that controversial?
Company names and figures are illustrative.
