Fixed-rate bond pricing can yield different results compared to Excel's built-in functions

Posted by Gavin08 on Oct 27, 2011; 6:07pm
URL: http://quantlib.414.s1.nabble.com/Fixed-rate-bond-pricing-can-yield-different-results-compared-to-Excel-s-built-in-functions-tp6340.html

Hi,
when I price a fixed-rate coupon bond with Quantlibxl, sometimes I get slight differences in the clean price compared to what Excel 2010's PRICE formula gives. For example, Quantlib returns 99.995 and PRICE returns 100.005. However, on other occasions, results look identical up to the the 6 or 7 digits I check.

Of course I use the same day count convention (30/360 US-style, for US bonds) with both functions.
I use the NullCalendar as calendar in Quantlib, but have also tried other calendar settings. As for the other bond settings, I keep them all at their Quantlib default values (although tweaking around with them didn't improve results robustly).

Any ideas as to where these calculation differences come from? Which one to trust more - quantlibxl or Excel that's been around for decades now? ;)

Best,
Gavin