An embedded and charset-unspecified text was scrubbed... Name: not available URL: <https://stat.ethz.ch/pipermail/r-sig-finance/attachments/20110106/170308f9/attachment.pl>
A question on portfolio value calculation
4 messages · Megh Dal, Guy Green, ArdiaD +1 more
In any realistic portfolio you will have some starting equity, and you would also have the costs/proceeds of your long & short positions. The portfolio value would then be: Starting equity + $(m1-m2-m3) -cost(pos1) +proceeds(pos2) +proceeds(pos2) or to put it another way: Starting equity +/- unrealised gains/losses on your positions. Your position sizes will be linked to your starting equity, both in the sense that your starting equity is a real-world constraint on the sizes of the positions that your broker will allow you to enter into, and also in the more theoretical (but still real-world) sense that the relative sizes of your starting equity and your positions contribute to the likelihood of your strategy exhausting all your equity at some point in the future, even if it is a winning one over the long term. Guy
Megh Dal wrote:
Hi all, can somebody suggest me on what is the correct way to calculate value of a portfolio (i.e. mark-to-market value) with having both long and short position? For example, suppose I have 3 positions in my portfolio pos1, po2, and pos3 and type of transaction is long, short, short respectively. Say, m2m value of those 3 positions are m1, m2 and m3 in money term. Then should m2m value of this portfolio be $(m1-m2-m3)? If this is correct I feel there are some practical problem with this approach. Let say I calculated the volatility of this portfolio assuming some normal distribution of return, let say it is $X. Then if I want to answer, what is the volatility for per unit value of my entire portfolio the answer would be : $X/$(m1-m2-m3). However if it happenes that $(m1-m2-m3) = 0 then above calculation becomes undefined. This approach also may be problametic if I have all short, in this case unit SD for my portfolio becomes obviously negative. Or should I go with $(abs(m1)+abs(m2)+abs(m3)) to avoid above scenario? Any explanation would be highly appreciated. Thanks
View this message in context: http://r.789695.n4.nabble.com/A-question-on-portfolio-value-calculation-tp3177254p3177324.html Sent from the Rmetrics mailing list archive at Nabble.com.
Have a look at this note: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1675067 This may help Dave
On 01/06/2011 01:28 PM, Guy Green wrote:
In any realistic portfolio you will have some starting equity, and you would also have the costs/proceeds of your long & short positions. The portfolio value would then be: Starting equity + $(m1-m2-m3) -cost(pos1) +proceeds(pos2) +proceeds(pos2) or to put it another way: Starting equity +/- unrealised gains/losses on your positions. Your position sizes will be linked to your starting equity, both in the sense that your starting equity is a real-world constraint on the sizes of the positions that your broker will allow you to enter into, and also in the more theoretical (but still real-world) sense that the relative sizes of your starting equity and your positions contribute to the likelihood of your strategy exhausting all your equity at some point in the future, even if it is a winning one over the long term. Guy Megh Dal wrote:
Hi all, can somebody suggest me on what is the correct way to calculate value of a portfolio (i.e. mark-to-market value) with having both long and short position? For example, suppose I have 3 positions in my portfolio pos1, po2, and pos3 and type of transaction is long, short, short respectively. Say, m2m value of those 3 positions are m1, m2 and m3 in money term. Then should m2m value of this portfolio be $(m1-m2-m3)? If this is correct I feel there are some practical problem with this approach. Let say I calculated the volatility of this portfolio assuming some normal distribution of return, let say it is $X. Then if I want to answer, what is the volatility for per unit value of my entire portfolio the answer would be : $X/$(m1-m2-m3). However if it happenes that $(m1-m2-m3) = 0 then above calculation becomes undefined. This approach also may be problametic if I have all short, in this case unit SD for my portfolio becomes obviously negative. Or should I go with $(abs(m1)+abs(m2)+abs(m3)) to avoid above scenario? Any explanation would be highly appreciated. Thanks
On 01/06/2011 05:26 AM, Megh Dal wrote:
Hi all, can somebody suggest me on what is the correct way to calculate value of a portfolio (i.e. mark-to-market value) with having both long and short position? For example, suppose I have 3 positions in my portfolio pos1, po2, and pos3 and type of transaction is long, short, short respectively. Say, m2m value of those 3 positions are m1, m2 and m3 in money term. Then should m2m value of this portfolio be $(m1-m2-m3)? If this is correct I feel there are some practical problem with this approach. Let say I calculated the volatility of this portfolio assuming some normal distribution of return, let say it is $X. Then if I want to answer, what is the volatility for per unit value of my entire portfolio the answer would be : $X/$(m1-m2-m3). However if it happenes that $(m1-m2-m3) = 0 then above calculation becomes undefined. This approach also may be problametic if I have all short, in this case unit SD for my portfolio becomes obviously negative. Or should I go with $(abs(m1)+abs(m2)+abs(m3)) to avoid above scenario? Any explanation would be highly appreciated.
See the CFTC's guidelines and calculations for the '13 column report'. This is a widely used and copied format for portfolio reporting. Typically, both net and gross exposures are calculated. Regards, - Brian
Brian G. Peterson http://braverock.com/brian/ Ph: 773-459-4973 IM: bgpbraverock