question about mortgage related securities

[myqee]

I am no expert in MBSs and CDOs but I had a novice question. I know the argument that gets repeated time and time again in the press is that if a security is illiquid, then it's value cannot be defined. I'm just curious though, why isn't there another straightforward way to estimate the value of a security based on how much cash-flow it is generating and the loan default rate. From a basic business standpoint, if I have an object (security, business, asset) that is giving me cash every month, and I know the probability / rate at which this cash flow may decline in the future, then I should be able to estimate the value of my asset now. One could always take the WORST case scenario for the default rate (say 80% of all subprime mortgages will default) to estimate the value of these assets. Am I missing something completely? Or is that what people are doing to estimate their asset valuations now?

 

[charlesdwright]

Or is that what people are doing to estimate their asset valuations now?

This is the rub. MTM accounting requires a "market value." What happens when the market is wide? Assets are marked down.

If you're a regulated entity with statutory capital ratios to maintain, then a marking of assets to "worst case" means you're going to have a "worst case" balance sheet with regard to those assets; it means you're going to have to raise "worst case" capital; it means, in other words, that a "worst case" scenario -- a crisis -- has been created simply because of an arbitrary decision to value your assets on a "worst case" basis.

It's a textbook case of unintended consequences.

 

[myqee]

I think the worst case scenario defined in terms of projected default rates is useful because it connects the value of an asset with an interpretable figure -- i.e., is it ridiculous to assume that 80% of subprimes will default, or 50%, or 25%? That way at least we can make a gut-feeling evaluation, rather than throwing up our hands and saying we have no idea. The latter leads to panic and complete lock up of the markets (as we are seeing) -- in which case, the asset is pretty much marked to zero.

 

[charlesdwright]

I think the worst case scenario defined in terms of projected default rates is useful because it connects the value of an asset with an interpretable figure -- i.e., is it ridiculous to assume that 80% of subprimes will default, or 50%, or 25%? That way at least we can make a gut-feeling evaluation, rather than throwing up our hands and saying we have no idea. The latter leads to panic and complete lock up of the markets (as we are seeing) -- in which case, the asset is pretty much marked to zero.

It's my fault -- I didn't anwer your question; I assumed my response indicated a "yes" answer.

The only variable is the arbitrary markdown. That I have no way of knowing.

 

[doug reich]

As charles implied, you are talking about "mark to model" accounting. Yes, we can come up with a good approximate value of an asset if we held it to maturity.

But these are not investment securities; they are trading securities, and as such they are marked to market. That is the rule. If the market values the security at 10, you have to value it at 10, even though the security is actually worth 70 on an investing basis. That's a big problem, because if you hold it in a margin account or you otherwise borrowed money with this as collateral, you have to raise more cash buy selling this or other securities irrespective of their value.

If someone (the government?) bought up these securities at current, depressed market prices and held them until the market came back and valued them as a function of intrinsic value, then they could do quite well.

EDIT: Look at that! I hadn't even read this article, but they say exactly what I did: http://www.quantnet.com/forum/showthread.php?t=3655

 

[myqee]

Thanks for the comment. My point though is not simply to promote mark to model, but to promote a more straightforward and interpretable model. For example, I think one reason the press and politicians can't even start to imagine talking about models is because the ones being used by professionals are too mathematically sophisticated, e.g., Gaussian cupola. I think there is a lot more straightforward ways to value an asset that anybody with a business background can grasp. I think that kind of modeling should be quoted more often in the press and in congress.

 

[MNB]

But these are not investment securities; they are trading securities, and as such they are marked to market. That is the rule. If the market values the security at 10, you have to value it at 10, even though the security is actually worth 70 on an investing basis.

Something to think about:

The people who make these decisions are running a business. They have to take into account a multitude of factors, and usually their end goal is a stable and increasing share price.

So, do they really HAVE to mark to "market" or "model"? Suppose they consider the consequences of not abiding by the rules (such as being sued by regulators, reputational damage) as not being as serious as revealing the true extent of their losses.

They can suggest that they are marking to "model", and then put this "worst case" default scenario at, say, 30% defaults instead of 80%. Since there is no market, they may just get away with it...

 

[newtoquant]

Market doesn't like uncertainty

Let me reiterate the question posted by myqee

" I know the argument that gets repeated time and time again in the press is that if a security is illiquid, then it's value cannot be defined."

As doug reich pointed out that these are lesser than the intrinsic value as these are distressed assetes. So the press intention of writing this is not meaning that you can't define a value of security with a simple model or any complex model, their intention is no bank or Institution would like to sell the distressed assets at lesser than than the intrinsic value as they don't know how much discounted from intrinsic value (Here Default rate is uncertainty).


market doesn't like uncertainty

-naga

 

[doug reich]

So, do they really HAVE to mark to "market" or "model"? Suppose they consider the consequences of not abiding by the rules (such as being sued by regulators, reputational damage) as not being as serious as revealing the true extent of their losses.

They'll be shut down by the regulators or IRS if they don't mark to market. They'll be both lampooned for being incompetent as investors as well as businessmen. They may go to prison for fraud.

That's not to say they don't try to game the system, but once investors are getting harmed by the managers breaking the law, the government cracks down hard.

 

[MNB]

I shall wait and see with interest.....

Can you really shut down a large bank? It doesn't look like they are that keen to do that - nor would it be advisable at all! Still, would it be any worse than following in Lehman's footsteps?

 

[myqee]

So the press intention of writing this is not meaning that you can't define a value of security with a simple model or any complex model, their intention is no bank or Institution would like to sell the distressed assets at lesser than than the intrinsic value as they don't know how much discounted from intrinsic value (Here Default rate is uncertainty).

I think if we look at the time course of default rates for subprime mortgages, one can at least get a ball park projection -- for example, 2008 2nd quarter delinquency rates for all single-family residential mortgages was around 4.3% and this figure was around 2% in 2006 so by simple linear extrapolation one could easily come up with projections (I know linear is an oversimplification but exponential is overkill and usually cannot be sustained). These numbers are for ALL mortgages so one would have to do the analysis on subprimes or stratified by FICO score. But in any case I don't see why everybody has to throw up their hands and say it is unknowable. I think it would be good if Paulson does just that and assigns the best guess projection and be done with it.

I guess another thing I would be curious to know is what is the bid-ask spread for these assets. We could translate those bids and asks into the projected default rates that the prices correspond to and see who is being more realistic.

 

[Andy Nguyen]

You simply can't choose how you mark your book. If compliance people said that you have to MTM everyday, that's what you do.
If what you are holding is illiquid and you don't see a mark on the market, then prepare to explain why you use that mark. They in return will use some different mark to mark your book. You will mark to show that you have positive paper profit and they may mark in a way that you are a few millions in the red. The only way to know is to liquidate your portfolio.
Prepare to argue for hours with risk people for hours, everyday, for simpliest, stupiest thing.
That's why traders and risk/compliance people don't see each others eye to eye. But together they have to co-exist like white on rice.
Yeah, people have tried to work around this (Jerome Kerviel at SocGen is the latest example) but saying there aren't certain rules to do things on Wall Street is simply naive and false.

 

[MNB]

Andy, I can tell you that you are totally right, but potentially in the wrong direction...

Probably not a good idea to talk anymore about this!