7
$\begingroup$

Since no one was able to answer my previous question, I am going to try to backtest a modern convertible bond arbitrage strategy myself (perhaps I'll report the results back here, if you're lucky :)). What suite of R packages should I use? Where can I get the data (free or cheap, preferably)?

Updated to add:

The main question here is what pre-packaged routines or packages are available to test convertible bond arbitrage models (see linked question for description of models), not simply quantitative models in general. I presume they are most likely to be available in R, but I'd take Matlab, too. The cbprice function in the Fixed-Income Toolbox leaves much to be desired, here.

Improve this question
$\endgroup$
5
  • $\begingroup$ Does it matter what security generates the price? I have only used the backtest in R, but any of these should let you generate portfolios along with a periodicity and condition for portfolio entry/exit. $\endgroup$
    – Richard Herron
    Commented Aug 4, 2011 at 11:08
  • $\begingroup$ @richardh The tough part of these strategies is figuring out the Greeks (like options pricing). $\endgroup$
    – Tal Fishman
    Commented Aug 4, 2011 at 11:58
  • $\begingroup$ OK. I guess I consider that more a Greeks question than a backtesting question. Have you considered RQuantLib? $\endgroup$
    – Richard Herron
    Commented Aug 4, 2011 at 16:09
  • $\begingroup$ @richardh, right, hence my update above. I have tried looking through RQuantLib, but as I commented to Brian Peterson's post below, the documentation is not exactly user friendly, hence my question. $\endgroup$
    – Tal Fishman
    Commented Aug 4, 2011 at 16:18
  • $\begingroup$ @TalFishman My personal point of view is that the availability of the data is the first and foremost problem. If you look at the stylized equation cb = bond + eq option (in some cases, the optionality is very complex and you have to do it bond by bond), you already need issuer credit curve, equity price, volatility surface for the option - ideally at the same time and preferrably both sides of the market. If I were to try something like this, I would start with one bond, keeping the general setting in mind and see where you go from there. I dont know of a packaged solution to do things. $\endgroup$
    – vanguard2k
    Commented Nov 8, 2017 at 10:52

1 Answer 1

Reset to default
4
$\begingroup$

I can't answer your question about the data, since my sources for data like that are Reuters and Bloomberg, neither of which are cheap.

For testing trading strategies, I'll separate them into three camps.

For 'simple' optimization schemes, RMetrics fPortfolio can likely do the trick. This would be for weight-based asset allocation. RMetrics also has some pricing functionality for bonds.

For 'complex' optimization schemes, still weight-based, PortfolioAnalytics can construct complex portfolio optimization problems using any arbitrary R function as part of your objective.

If you want to use indicators to define signals and construct trading rules, the quantstrat package will likely fit.

If you want additional pricing functionality for bonds, look at RQuantLib

Improve this answer
$\endgroup$
1
  • $\begingroup$ Thanks for the info, but I was looking for specific info about convertible bonds (see update). I have tried searching the various R tools, but they are not exactly user friendly, hence my question. $\endgroup$
    – Tal Fishman
    Commented Aug 2, 2011 at 14:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.