0
$\begingroup$

I followed all the procedure in Quantlib to process interest rate swap valuation through Python Quantlib. I valued more than a million records. All the valuation is almost the expected amount. But 'Interest rate Swap' of these currencies ( CNY, KRW, THB, TWD) are way off than the expected valuation. Almost 100% variation in amount has been observed from the reported value.

Any suggestion on above currencies why their valuation are so different ? I read all the theories on Non Deliverable Swap currencies but I want to implement something in Quantlib to get the correct valuation. Please suggest.

$\endgroup$
1
$\begingroup$

The problem is quantlib supports only IRS. But you're trying to find NDS valuation.

  • IRS valuation:

    PV (present value) of the interest payments on the fixed leg: \begin{equation} f(x) = \sum \limits_{i=1}^{n} \delta({T}_{j-1},{T}_{j}) \cdot K \cdot P(0, T_j) \end{equation} PV (present value) of the interest payments on the floating leg:

    \begin{equation} f(x) = \sum \limits_{i=1}^{n} \delta({T}_{j-1},{T}_{j}) \cdot F({T}_{j-1},{T}_{j}) \cdot P(0, T_j) \end{equation} where:

    $\delta({T}_{j-1},{T}_{j})$ - day-count fraction

    $K$ , $F({T}_{j-1},{T}_{j}$ - fixed and forward rates

    $P(0, T_j)$ - discount factor (price of zero coupon bonds)

  • For NDS (non-deliverable swap) formula:

    \begin{equation} F(0,T) - S_0 = S_0 \bigg(\frac {1 + r_d \cdot T}{1 + r_f \cdot T} - 1 \bigg) = \\ = \frac {S_0(r_d - r_f) \cdot T}{1 + r_f \cdot T} \approx S_0(r_d - r_f) \cdot T \end{equation}

    But for NDS(floating for floating, fixed for floating) are usually used in emerging markets where the currency is illiquid, subject to exchange restrictions, or even non-convertible the quanto correction must be apply to foreign currency forward rates. For instance, if you want to evaluate cross currency swap the effective forward rate used in pricing:

    \begin{equation} f = (0, t_j, t_{j+1}) \cdot (1 + \rho_j \cdot \sigma_{f_j} \cdot \sigma_{FX} \cdot t_j) \end{equation}

    where

    $\rho_j$ - FX/Forward rate correlation

    $\sigma_{FX}$ - forex rate volatility

    More information: Boenkost, Schmidt "Notes on convexity and quanto adjustments for interest rates and related options" (2003), SSRN 1375570

  • | improve this answer | |
    $\endgroup$

    Your Answer

    By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

    Not the answer you're looking for? Browse other questions tagged or ask your own question.