CCP, CLH and PNC all provide an overview of collections based on...

CCP, CLH and PNC all provide an overview of collections based on age-range, but without detailed amortisation calculations or breakdown of the PDL book value, there's no way to calculate the changing ROIC. So in summary, it's not possible to do the calculations you're after.

A debt purchasing company is like a retailer that buys long-term stock upfront and tries to get a decent return over 6 years. In their defence, they're no different to other "upfront capex" (I'm using this term very loosely here) companies - the same ROIC questions arise:
* Electronic retailers (eg. JBH) - How do we know if the current stock purchased are getting acceptable ROIC rates, that the company is amortising old stock properly, and the warehouse is not full of unsellable junk?
* Mining service companies - Whether the equipment purchased in the current year will assist with sustaining profitability or turn into rust buckets when contracts don't eventuate?
* Software companies (eg. XRO) - What's the quality of the R&D spend - is it boosting or detrimental to ROE? What if they're building the wrong thing?

So one needs to look at other measures to make sense of the picture. Below is my mental model of how dependable the figures for debt purchasing companies are - from most to least.

1st level numbers (concrete): Collections, operating expenses (excluding PDL purchases), original price of PDL bought.
2nd level numbers (significant management discretion): Amortisation charges on PDLs.
3rd level numbers (impacted by 2nd level): NPAT, EBIT, etc. (profit = collections - amortisation - opex)
4th level numbers (impacted by an accumulation of years of 2nd level numbers): PDL book values, equity, return on equity, debt/equity ratio.

Management's discretion over amortisation (2nd level) turns the numbers impacted by it (3rd - 4th level) into mush for investor's with a cynical view.

One way to get a picture of the underlying performance is to form ratios between concrete numbers (1st level) the more flexible numbers (2nd - 4th level) and compare them across all the players. You'll get a better appreciation for the different qualities and sustainability of NPAT, book values and ROEs. Of course you'll also need to adjust for each company's different circumstances. CCP AGM 2015 Presentation page 6 is worth looking at as it's essentially a page of peer comparison doing just this.

Note: I'm not saying any of the companies are doing anything wrong - this is just an consequence of the industry they're in that makes some of the conventional numbers quite fluffy. Detailed value investors will need find other ways of making sense of things. Same thing applies with having to dive into ARPU, CAC, CLV, churn rate, for fast growing SaaS companies.

Hope that helps!