I am trying to price of a Down-and-Out Barrier call option with leverage. When the price of the underlying asset hits a certain barrier (B), the option becomes worthless. The issuer of these options indicates that their price is calculated as follows: $$P = \frac{S - F}{ratio} $$

(Note: the ratio is used in case the price of the underlying asset is high, like in the case of Google stock which is around $1,500. The ratio is often 10 or 100.)

The issuer of the Turbo charges an interest of around 2% on the financing level $F$, which is paid daily by increasing the level of $F$ and consequently $B$ everyday. So that:

$$ F_t = F_0 (1+r)^t$$

The diagram below explains the structure.

enter image description here Closed form formula Assume I want to estimate the value of this option after 1 year. As far as I understand, the Black & Scholes Model cannot be used. This is because this Barrier resembles an American option, as the Barrier option can be exercised at any time (i.e. when the price of the stock hits the Barrier). As the Black & Scholes Model applies only to European options which can be exercised only at expiration, a closed form model seems difficult to me.

Price approximation My question is if the price of this Barrier option could be approximated as follows.

  1. Simulate the price of the underlying stock using Brownian motion with say n simulations.
  2. Take all price paths that never touch the Barrier and list their end prices.
  3. Estimate the value of the option by calculating the option value for each of these end prices $S_i$ using $(S_i - F_t)$, sum them and discount to the present day, and then divide by the total number of price paths $n$:

$$ P = \sum_i\frac{(S_i - F_t)e^{-rt}}{n} $$

Does this make sense? Any suggestions on improving this, to make the approximation more realistic?

  • $\begingroup$ Black Scholes should not be used in practise because it's extremely oversimplified. There exist closed-form formulae for barrier option in the Black Scholes world (see Hull's book). This has nothing to do with American-style options though. Your Monte Carlo approach sounds fine. You may want to get a good book on Monte Carlo simulations and then ask a specific question on the topic (variance reduction, computing Greeks etc)? $\endgroup$
    – Alex
    Commented Jul 30, 2020 at 21:49
  • $\begingroup$ Thanks a lot. Could you give the title of Hull's book? And can you recommend a good book on Monte Carlo simulations? $\endgroup$
    – twhale
    Commented Jul 30, 2020 at 21:55
  • 2
    $\begingroup$ Hull: Option, Futures and other Derivatives and Glasserman: Monte Carlo Methods in Financial Engineering $\endgroup$
    – Alex
    Commented Jul 30, 2020 at 22:07
  • $\begingroup$ A word of caution: there are continuously monitored barriers and discretely monitored barriers, be clear which kind you are evaluating. $\endgroup$
    – nbbo2
    Commented Jul 31, 2020 at 2:12
  • $\begingroup$ @noob2: I think they are continuously monitored. How does this influence my steps outlined above? The options are traded between 8.00 and 22:00 on weekdays. However, they can only be knocked out when the underlying asset is being traded, which is between 15:30 and 22:00h. During that time it can be knocked out at any time. $\endgroup$
    – twhale
    Commented Jul 31, 2020 at 5:48

1 Answer 1


Following the notation in Hull, let $H$ be the barrier level. I list the prices of European-style down-and-out barrier options with continuously observed barrier.

  • If $H\leq K$, then $$c_{di}=S_0e^{-qT}(H/S_0)^{2\lambda}N(y)-Ke^{-rT}(H/S_0)^{2\lambda-2}N(y-\sigma\sqrt{T})$$ and $$c_{do}=c-c_{di}.$$
  • If $H>K$, then $$c_{do}=S_0N(x_1)e^{-qT}-Ke^{-rT}N(x_1-\sigma\sqrt{T})-S_0e^{-qT}(H/S_0)^{2\lambda}N(y_1)+Ke^{-rT}(H/S_0)^{2\lambda-2}N(y_1-\sigma\sqrt{T})$$ and $$c_{di}=c-c_{do}.$$

Here, \begin{align} d_{1,2} &= \frac{\ln(S_0/K)+(r-q\pm0.5\sigma^2)T}{\sigma\sqrt{T}}, \\ \lambda &= \frac{r-q+0.5\sigma^2}{\sigma^2}, \\ y&=\frac{\ln(H^2/(S_0K))}{\sigma\sqrt{T}}+\lambda\sigma\sqrt{T},\\ x_1 &= \frac{\ln(S_0/H)}{\sigma\sqrt{T}}+\lambda\sigma\sqrt{T}, \\ y_1 &= -\frac{\ln(S_0/H)}{\sigma\sqrt{T}}+\lambda\sigma\sqrt{T}. \end{align}


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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