I'm trying to backtest some strategies which include 3-5 years single name EM corporate bonds. Some of them don't even have a meaningful historical data. Some others are pretty illiquid. Other than these, my portfolio contains ETFs, equities, FX, and Commodities - pretty much multi asset class portfolio.

For risk management purposes I separated these bonds and looking at them from a different angle. But leaving them out in portfolio optimization will be a bad idea as these bonds account for 20-30% of the overall portfolio.

I guess my question is is it possible to somehow model EM single name corporate bonds? Has anybody come across a similar problem/project? Any resources that you can recommend?