5
$\begingroup$

I just joined a support team for an equity exotic trading desk in a bank, I am looking for a high level overview of how exotic trading works in a bank.

For my questions let's take a common product: an Autocall on a stock with barriers: KO 100% and KI 70% with valuation every 6 months. Here are a few questions I currently have.

I never studied Finance an have some trouble to understand a few basic topics, any answer even partial is welcome.


  1. Pricing

What does it mean to "price" an Autocall. When the client buy an Autocall does he pay a premium like for an option ? or is it the KI barrier paying for the structure ? what is the price then ?

  1. Structuration

What does it mean to structure an Autocall ? I understood an autocall can be composed of a put down and in / barrier options, bonds for the coupon, etc.

Does it mean if the client buy an Autocall the bank will buy a put down and in and a bond to hedge, how does it work and what is structuration ?

  1. Risk management

On a high level how do traders hedge their risk, I understand they compute the greeks like for options.

What is the impact of the current positions and risks on the price of a product, can some products / autocall you sell to client help you hedge your risk, is it common ?

Do they usually lose money on the hedge (e.g. they sell the product and then need to hedge during the whole lifetime) ? I've heard of gamma trading, can they make money with the hedge, etc.

  1. PnL: Main Source

How is the trading desk in a bank making money ? with a premium, by managing risks, do they take directional position and don't hedge themselves completely ? what usually drives the PnL of an exotic desk in a bank ?

Is this correct:

  • Hedge: costly ? I've heard of gamma trading, can they make money with the hedge, etc.

  • Directional positions: possible but very limited ?

  • Margin: main revenue ?

How does it work for daily PnL, is the margin making most of the profits.


I just would like to have a high level overview of how it works. If you have any material to get started don't hesitate to send a link.

Thanks a lot for your help !

$\endgroup$
4
$\begingroup$
  1. "pricing" am autocallables is simply working out what it's worth. This is done by having some model (Google local vol / stochastic local vol) which is calibrated to the market (ie listed vanilla options, broker quotes for less liquid tenors, and some light exotics (ie American barriers, cliquetes, etc.)), and then simulating the underlying many times, calculating tee underlying value on each path, and then averaging them (monte carlo). Typically, you won't calculate the price of a particular autocallables, rather you'll find the value of the coupon/barrier such that the structure is worth 100%. A client will normally ask their advisor/wealth manager for a level on a structure, and they will ask multiple issuers, each will give a barrier/coupon for the same structure, and the client will pick the one that is most attractive.

  2. Structuring can mean a few different things. Be it adding features to a structure in order to tailor it to a clients wishes, or maybe better achieve the exposure they're after, or it may simply be choosing the strike level. It may go farther and involve creating a whole new style of structured product (ie. Kick in goals, mountain range structures, cliquetes, accumulators, range accruals, prdcs, tarns, lizards - they're all different structured products created for various reasons).

  3. Typically the exact trade is not hedged, as its extremely expensive. Instead, the Greeks up so some desired order msy be hedged. In reality, the position is added to the portfolio of hundreds of other structers, and the risk from all amalgamated. Some risks offset, this is good - your new structures (at least at the moment hedge your old ones). Others increase your risk. As the market changes, your risk will change. Managing the risk simply means to keep your risk inside sensible limits, and minimise it where its financially viable (it is possible to get it to zero, but very expensive. Risk management also must take into account the costs of the hedges. Leaving some risk on the table is acceptable, if you're getting paid for it).

  4. A structured desk makes money by selling things for more than they are worth. It is that simple - in point one, I mention that you solve for the couoin such that the price is 100%. This is not quite true, in reality, you solve for the couoin such that the value is (say ) 99%, amd sell it for 100%. If you then hedge the trade for zero cost (unrealistic) through the life of the trade, you make 1%. What actually happens is that your hedges incur trading costs. Provided those costs are lower than the value of the trade, and the total of the position and the hedges' pnls are greater than the remaining amount, you make money. You could take more pnl on day 1, to give you more leeway in the hedging costs, but then your price will be less competitive and you reduce your odds of winning the trade.

Hope that helps.

$\endgroup$
  • $\begingroup$ Thanks a lot for your answer Will it is much more clear. How does it works for the coupon payment, how do they hedge against it ? Do you have any useful website, book or coursera you would recommend to get started ? $\endgroup$ – XTrading Jul 21 at 22:42
  • $\begingroup$ Hi Will / XTrading ... for the Autocallable product, is it always true that - (i) KO is always good for the bank, since we already made our Day1 P&L and KO means no more hedging needed (so, sooner KO happens, lower the hedging costs)?, (ii) What are the signs of the usual greeks on Day 1 (in a typical autocall trade) from the banks perspective - (A) Is it true that Delta is negative (since client is delta positive - he bets a yield enhancement based on upside in the equity), albeit as the KO barrier is approached delta for the bank becomes positive, (B) Vega is also negative for the bank ?? $\endgroup$ – bhutes Jul 22 at 2:26
  • 1
    $\begingroup$ @bhutes i. KO is not always good no, it depends on what the structure has done throughout the trade and how it has been hedged. You are correct though that an early ko results in less dynamic hedging. ii. The structures are usually short delta and short vol (though "bearish" autocallables exist, where the autocall occurs if the underlying price is lower than the barrier, these are short delta and still short vol). How delta changes as you approach the barrier depends on the vol and shape of the fwd curve, it can be either negative or positive. $\endgroup$ – will Jul 22 at 6:09
  • $\begingroup$ Hi will, one last question. What will the traders do with the notional they receive when they sell an Autocall ? How do they invest it, do they use a term repo to invest the cash until the next valuation, what do they do ? $\endgroup$ – XTrading Jul 25 at 6:58
  • $\begingroup$ @XTrading When trades are priced, the discounting that's used is not the risk free rate, libor, or some externally observable curve. Instead, it is discounted using an internal rate curve specific to the seller (i.e. the rate the issuer expects to be able to achieve on cash) - you can expect this rate to be close to a risky discount curve using the issuer's credit rating... $\endgroup$ – will Jul 25 at 12:27

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.