wilmott.com

Hi everybody!I currently work at ... (since 2 weeks), a company developing portfolio simulation tools. I combined from many websites Inflation-Indexed bond pricing formulas (OATi ,..). Basically I calculate either the IPC ratio or the RPI ratio (retail prices index). Then, a simple multiplication with the bond's price is needed. My question is, would you know other ways to incorporate inflation in an investment strategy through these indexed bonds(day by day view, for example, assuming a CPI and adjusting the nominal) ? Thanks in advance,Michaël

i always thought that it was the real rate of return that mattered as the Principal (and coupon payments) grow by the CPI and you are discounting those nominal flows by the nominal curve leaving you with the real rate.

Hi dave and thanks again. I thought the ratio was something like this: REFCPIissuedate = CPIi + ( (t-1)/D * (CPIi+1 - CPIi) ) with i=issuemonth-4 D=number of day of issuemonth t=issueday then RATIO=REFCPIsettlementdate/REFCPIissuedate and finally PRICE = PRICEvanilla * RATIO To your mind the CPI discount directly the real rate of return in the Inflation-Indexed bond pricing formla? Please lead me through the way of light..

Most inflation-linked bonds follow the notional adjustment rule that you specified.Here are a few exceptions:- Brazilian NTN-B and NTN-C coupons depend on both historical CPIs (until t - 1 month) and forecast CPIs (for current month)Check out http://www.andima.com.br/merc_sec/arqs/ ... guide.pdf- Australian Treasury Indexed Bonds use a "Capitalized Index" which depends on all previous CPIs fixed every quarterSee http://www.aofm.gov.au/content/borrowin ... p?NavID=14 Also, you may fund structured notes that depend on Libors, CMSs, equity indices, inflation.....

QuoteOriginally posted by: MichaosssHi dave and thanks again. I thought the ratio was something like this: REFCPIissuedate = CPIi + ( (t-1)/D * (CPIi+1 - CPIi) ) with i=issuemonth-4 D=number of day of issuemonth t=issueday then RATIO=REFCPIsettlementdate/REFCPIissuedate and finally PRICE = PRICEvanilla * RATIO To your mind the CPI discount directly the real rate of return in the Inflation-Indexed bond pricing formla? Please lead me through the way of light..All I am saying is that there are two ways to skin this catyou can forecast your nominal cash flows and discount them using the nominal curve or you just work in real return space. if you chose the latter then you will use your inflation linkers to figure out the term structure of real yields.

Sorry.. The pricing formula described above is for Sout Africa IIB. You're right Dave..(& thanks Quentin)..We must do at the end coupon = couponvanilla * RATIO and the same thing for facial value adjustment. Then inflation indexed Bond pricing formula is the same as the vanilla bond (sum of PV of flows)..Do everybody agree with me ?

Ok for the non stochastic formula. I found pricing methods for european indexed bonds (with and without accrued interest). Concerning stochastic modeling, the inflation index process follows the stoch. B&S formula and the solution after Ito adjustment is something like this.. Ct = C0 exp ( a*t - t*b²/2 + b*Wt ).. So now, I'm looking for best calibration (with a and b non stochastic).. I will take empirical averages for a and b, and generate W (wiener process) with Mersenne Twister + BoxMuller. Has anyone any improvments?