Hello! I investigated the applications of MRJD models to pricing electricity derivatives in my Master thesis. Hope putting this on the web will generate exchanges of ideas. I worked on this topic with GogolaAnita (who was my supervisor for this project). We estimated the parameters of the MRJD model based on a recent hourly dataset and we used the method to price puts. Although I delivered the final version of the Master thesis to the university last February, this is work in progress as far as research ideas go and I would certainly appreciate comments. GogolaAnita and I are working on a revised version, expanding both the estimation method and extending the approach to pricing calls. Quick point on regime switching: one could follow the Hamilton approach (Time series analysis, ch 22) to do regime switching. In this form (ie fixed jumps) loglikelihood estimation is popular for the parameters. Question is if you want to expand this to a time varying parameter setting or towards non-fixed jumps.greetings allu

Dear Mr. Lui (=alex lui?),thanks for mentioning me, very honest, indeed. However, I did NOT speak of this thesis (it is NOT a thesis ), but I was speaking of that of Graydon Barz. In allu's thesis indeed some things were taken from Barz, with the exception, that he does not solve the PIDE, and he does not even motivate the form of the trial function he is searching the characteristic fuction in. What he does (and which is NOT in allu's thesis) is to value option prices on forward prices. In allu's thesis the option prices are valued with 0 convenience yield and on spot electricity prices. I think it can hardly be motivated to value option prices on a non-storable underlying [a wake argument i gave below]. There is a motivation : our boss, the chief risk manager had meager knowledge on stochastic calculus and financial maths (he even crossed out the Ito correction terms in my reprots and put a question mark on the margin ) and he insisted on valuing options on spot.However, the methodology can be used for mean reverting , storable assets (e.g. erdgas, currency, even orange juice ).BTW: if you read the thesis, you will see, that the FE-FD algorithm does not give the coolest results. I would appreciate , if you could help us to explain, why not.

Anita,Firstly, add my name to the list of those wanting a look at your thesis:gammashark2000@yahoo.comSecondly,I have had this discussion about storing input fuels with a number of traders. The reality of electricity is that it is always a forward, until produced. So you have to ask yourself what you mean by spot options? Do you mean day ahead? Do you mean an hour ahead? etc.Once you've got an answer to that, even if you have an asset, then operational constraints begin to bite. This is crucial, because you can't just flick the switch (with the exception of certain hydro and pumped storage units) to get your electricity. So, if you get inside of the operational capabilities of your plant, by which I mean you want to produce electricity in the next hour, but you need 4 hours for ramp up, synchronization, etc..storing the fuel isn't relevant. Thus the spot, as it were, optionality is different for each plant in your portfolio.Constraints also bite on the way out - you can't just shut it off, you have to ramp down. This might not be a problem if you're selling, say 12 or 16 hours peak...but you must account for the extra power produced/fuel taken on the way in and out, and the changing efficiencies at different levels of production (you'll see there is an efficiency curve that shows the plant is most efficient at full production and declines all the way to minimum stable generation - and, it isn't linear)Deng covers this in his thesis, to some extent. I also know that some players in the market have got their quants sequestered with the engineers to sort out the near real time optionality of plant. I'll give you a name or two if you contact me privately.Advice - talk to the operational engineers of the plant in the portfolio in question to get a better handle on what is and isn't possible.good luck,gammashark.

Hi Anita,would it be possible to receive a copy of your thesis...?here's my address andrea.odetti@bpv.itthanks a lot!ciaoandrea

Does any of you have experience using commercial pricing and valuation software for power option contracts. I´m especially referring to the software that is offered by FEA (financial engineering associates) or Lacima (Les Clelow, Chris Strickland).Would be happy to start a discussion about that.

Ive heard very very negative things about FEA models... and I don't hold lacima in that high regard.... pricing daily/swing optionality is very tricky... there are lots of very very big pitfalls to beware of... If you want some advice/anecdotes, send me a PM

HiI´m impressed by the quick response. I guess a Pm is a private message. Yes, sure I´m very curios to get some anecdotes about it. Do you have first hand info ? Do you also work in the power industry ? regardsKai

Hi Anita,I would like to have a copy of your thesis, could you send it to marcos.lyra@baydenet.com.br.By the way, do you know any site where I can download a spreadsheet or a program that can be used to simulate a Jump Diffusion with mean reversion model ? Or if you have one could you send me?

Dear AnitaIt would be nice if I could be included to the mail list of your thesispauldrei@yahoo.comregardspauldrei

Anita,If possible, I too would also like to be included in the distribution of your thesis.My email address is monkey@myrandomstuff.co.ukKindest regardsMonkey