My Photo
Note: Jeff does not accept guest blog posts on A Dash of Insight.

For inquiries regarding advertising and republication, contact

Follow Jeff on Twitter!

Enter your email address:

Delivered by FeedBurner


  • Seeking Alpha
    Seeking Alpha Certified
  • AllTopSites
    Alltop, all the top stories
  • iStockAnalyst
Talk Markets
Forexpros Contributor
Copyright 2005-2014
All Rights Reserved

« Sentiment is Slow to Change: a Basketball Lesson, Part One | Main | Sector Update and A Personal Note »

April 09, 2008


Turley Muller

I get bothered when people freak out over level 3 assets. It's a not a new concept, it's a part of the accounting discipline and I don't see any better way. It may not be perfect, but what are really the alternatives?? I want someone to answer that question for me.

I think the models work well for the most-part, It's the fact that input assumptions can be manipulated thus valuations can be fudged in the desired of desired results. Market transactions are objective- every can agree that A bought asset(y) from B paying x amount.

Models are up for interpretation, however, banks are strongly interested in them being as accurate and robust as possible. Investors, largely spurred by the media - "Mark to Myth" "mark to make-believe" etc to distrust these models, and take a position that if there are no observable prices, then the assets are worthless and any model that says otherwise is lying.

Many models employ current market pricing to back into variables that can be applied to assets that are not trading. Volatility and term structure being the most important, - OAS can be implied from securities that do trade then used in valuation with the addition of pre-payment and default models.

Certainly default probability assumptions were too low before the meltdown, but that's why we wouldn't expect those assets to trade at par today. Maybe default PDEs were way off, but not to the degree where current value is zero. Maybe 5-10% is now 25-30%, or whatever, values should be 50 or more, not 10 or less.

Banks didn't want to sell these assets, because no one wanted to buy due to own plate being full. Only at a significant discount would an institution want to buy, and that would cause huge writedowns since there is market pricing.

Here is my logic:

If models suggest values are not as scant as fear has lead investors to believe, then institutions wouldn't want to sell and get a worse price. If they know it to be worth more then they will hold unless forced to sell. If values were really that bad, I think we would have seen lots of unloading, why not? If they are already deemed to be near worthless and that's about all the market will pay, then no reason to keep those assets, - lose both ways.

lack of volume and liquidity causes the value that could be achieved by selling lower than the more realistic, intrinsic value that could be achieved in a liquid market. So, if a bank wants to sell an asset in an illiquid market, then wouldn't other buyers think there must be something drastically wrong for one to sell into an illiquid market knowing that significant value will be left on the table?

I get a fell from the media and reader comments that lean towards the theory that banks won't sell assets because that will show they are worthless, so all players are colluding not to trade these questionable securities, so they can use models and make up their own values and fool everyone. It also seems some aren't happy until the models reflect near zero values, I think that's where a lot write downs stemmed from- pressure to adjust modeled values lower. "They're too high" "It's Make-believe"

Certainly I think at the onset modeled values may hay been to high, just because the unexpected, however, the pendulum looks to have swung the other way- way too far

Pressure from skeptics (no expertise, very limited info) are exerting too much influence on the experts, having all available information.

I expect some write-ups. In this ere, how could a CFO not be as conservative as humanly possible? Compared to what happened to Ebbers& Fastow & Koz. I think if one of these CFOs continued to mislead, public would be crying for the death penalty.

Bill aka NO DooDahs!

Oh, we're all familiar (or should be) with accruals, abnormal investment, and earnings quality having an impact on future stock price returns. No quibble there, read the research, you'll see links to it on my blog, blah blah blah.

David's assertion appears to be that the various "anomalae" act differently in a particular industry, i.e., financials. Not all financials are "bags of accruals" when accruals are defined per the research regarding the anomalous effects ... many non-financials have higher levels of accruals than many financials ... etc.

THAT'S what I want to see. Industry-specific data on the accrual issue. Until then, I'll have to take that assertion with a big spoonful of salt.


If an investor is underweight financials, then by definition the "rest of the market" is actually more bullish on financials than the investor is.

Also, maybe David's referencing...

The Real Effects of Investor Sentiment.

Polk (LSE) Sapieza (NU)

"...positive relation between abnormal investment and discretionary accruals... that firms with high abnormal investment subsequently have low stock returns..."

Bill aka NO DooDahs!

David, do you have a link to a backtest, or to an academic study, that shows the accrual "anomaly" impact on relative stock price performance is industry-specific?

If not, then then what is your basis for stating it can't be applied to financials?

David Merkel

Jeff, you're probably aware of the accrual anomaly, where companies with high normalized accruals tend to underperform.

With financials, that model can't be applied, because financials are a bag of accruals. Yet, we know that some financials have higher quality accruals than others. Higher quality accruals are shorter, more transparent, and less credit-sensitive, or, less volatile.

I worked in insurance financial reporting for years. Every life reserve is a level 3 asset, though they don't call it that. We do advanced DCF models to estimate results. The process has been refined over decades, and is pretty smooth now.

With the investment banks, transparency is
low, and we don't know how well the level 3 models work yet. I've worked on this issue myself for clients. It's tough, and there is often no dominant model to set the prices, and sometimes, a previously reputable firm issuing Pollyanna-ish opinions that an anxious management grabs onto.

Level 3 has growing pains that it has to work out. We're not even to the level of insurance reserving yet, and look at the discounts the insurers have to suffer in their P/Es. Perhaps the investment banks deserve similar multiples (adjusted higher for their higher ROEs versus insurers).

The comments to this entry are closed.