Any good proxy for embedded value???

[question]

For those companies do not calculate EV (traditional EV, EEV or MCEV), is there any good proxy for that? If so, please advise. Thanks so much...

[Freebird]

How about the average of the NY 7 scenarios for Cash Flow Testing? You would have to be using a discount rate of the after-tax net investment yield for each scenario to calculate the present value of ending market surplus (PVEMS) for CFT.

Here's my thinking on this:

MCEV is basically the market value of the company, calculated using risk neutral scenarios, or in other words present value of distributable earnings (PVDE). It can be shown (I have seen a demonstration in TAS) that PVEMS using the after-tax investment yield as the discount rate is equivalent to PVDE.

Consider that the NY 7 scenarios are somewhat symmetrical in that they have both up and down movements amongst the 7. So, if you average them, you might get something akin to the mean of a risk neutral set of scenarios, which is what is used for MCEV. Also, because EV is calculated by projecting and discounting at the same rate, the effect of the timing of reserves and capital is eliminated. This is also true for PVEMS in CFT, if you discount at the after-tax net investment yield; i.e., effects of timing are eliminated. Now it's true that the NY 7 are deterministic real-world scenarios, so they ostensibly have no comparison to risk-neutral scenarios. However, if you consider that the mean of the risk-neutral scenarios is essentially the swap curve, that would be comparable to the Treasury used as the basis for the NY 7. One problem with the analogy is that rates are so low right now, the "pop-down" scenario does not pop-down very much in the NY7, since it is floored at 0. Risk-neutral scenarios can be negative and thus more symmetrical than the NY7 would be right now.

However, I believe that PVEMS, or PVDE, discounted at the after-tax investment yield would be a decent proxy for the current market value of the company, given that you really "believe" what your models are producing.

Anyway, if any of this makes sense to anyone, I'd appreciate your comments.

[Freebird]

Hmmm... 265 views and no responses. I am thinking that my explanation may be a little too obtuse. It could also be that the idea of a deterministic process (NY 7) being a proxy for a risk-neutral process (MCEV) might be a little "outlandish". My background is in cash flow testing, with only a small amount of recent exposure to MCEV. This was just a thought I had, and I was hoping to get some knowledgeable folks to poke holes in this idea.

Anybody?

[glassjaws]

Take a balance sheet and income statement and calculate distributable earnings from them. PV of distributable earnings is a basic definition of embedded value. Or so I've been lead to believe.

[Freebird]

Originally Posted by glassjaws

Take a balance sheet and income statement and calculate distributable earnings from them. PV of distributable earnings is a basic definition of embedded value. Or so I've been lead to believe.

Yes, you are correct that MCEV amounts to PVDE discounted at the risk-free rate, or the average of the results from a risk-neutral set of economic scenarios. What I am suggesting is that you essentially have PVDE at a "risk-free" rate in your Cash Flow Testing results by taking PVEMS at the after-tax earned yield.

Your surplus in the CFT projection is earning the after-tax yield, so for all intents and purposes you are projecting and discounting at the same rate, like you do for MCEV. By doing this (projecting and discounting at the same rate), you take the effects of the timing of reserves and capital out of your PVDE result. I may be missing something important, but this idea seems logical to me.

[Olrich]

My knowledge on this is thin at best, but wouldn't the results of CFT and MCEV be dramatically different (i.e. different signs) for any kind of spread based business?

In your CFT models, the assets will be earning a "real world" yield (Treasuries + Credit Spreads - Expected Defaults) so you can actually earn the targeted spread, while in the MCEV model you might be crediting guaranteed rates and still have significant spread compression. Thoughts?

[Freebird]

Originally Posted by Olrich

In your CFT models, the assets will be earning a "real world" yield (Treasuries + Credit Spreads - Expected Defaults) so you can actually earn the targeted spread, while in the MCEV model you might be crediting guaranteed rates and still have significant spread compression. Thoughts?

I am not involved with the EV models that much, just the results. However, I think the modeling software (ALFA in this case) takes care of the spread issue by bumping up the expected defaults so that the modeled market value of assets at the start of the run matches the actual market value of assets at the risk-free rate.

Any comments from ALFA or TAS modelers?