Would better data analysis reduce the financial bail-out?

By Andy Oram
September 26, 2008 | Comments: 12

I don't pretend to understand the finance system any better than the average Wall Street trader, but it seems to me the White House and Congress are stuck in 1930s-era thinking and are bulking up their bail-out unnecessarily. Couldn't we get the data to apply our precious public funds more effectively?

As I understand it, the proposed bail-out is shooting seven hundred billion dollar-sized bullets into the dark. We'll hit our targets, but most bullets will be wasted.

We're being asked to prop up the most egregious companies with the worst history of decision-making. But it's not because we care about those companies--it's because their demise would drag down other companies that invested in them. Or so we are told.

The danger is real. This crisis draws the government together because it's about protecting the whole financial system, not just a few cronies. Washington Mutual collapsed just yesterday because it had invested inordinately in subprime mortgage bonds.

But that underscores my point: the danger is not spread around evenly. The White House and Congress think they have to inject one huge investment--huge beyond the scope of human comprehension--to protect the whole system. Why can't we intervene more selectively and over a longer period of time?

book coverbook cover

Banks know where their money is. I believe a lot of that information is already public--and if the financial institutions want bail-outs, they can make more of it public.

We could figure out exactly how risky each asset is, exactly how much exposure each institution has to bad loans or collapsing stocks and bonds, and what the overall health of each institution is. Even Freddie Mac and Fannie Mae are important are constituted of hundreds of thousands of individual, small loans.

We don't necessarily have to shore up a large institution that has acted recklessly (not to mention unethically). We could trace the other companies that depend on that institution and apply the same analysis to them. If someone acted in good faith and has demonstrable signs of distress, we could provide relatively small loans to keep them from failing. This applies to everyone from a condo owner with a skyrocketing mortgage rate to a bank or corporation. We could spider the net of financial investments the way search engines spider the Web, and triage our efforts.

I don't know what criteria the Treasurer wants to use to invest the nation's money, but I bet the criteria could be a whole lot smarter. If we're creating a small board with the power to hand out billions of dollars, transparency is a top concern--and rigorous data analysis will support that.

Naturally, financial institutions want to hide as much of their operations and holdings as they can. But they can't afford to do that any longer. Our nation certainly can't.

You might also be interested in:


Their plan address the risk.. you are right it is spread around unevenly, but the Dutch auction let's the risk find it's own level.

We need this now not a great debate by unqualified congressmen. They should have to take a test to become qualified to even run for congress. If we wait.. the cost goes up.

I don't think you really understand the problem at all, let alone its magnitude. You're trying to approach this like an engineer, but the variable are neither known, nor fixed, which is why we have markets to try to solve them. Smart people with a lot of skin in the game couldn't resolve the price/risk issue, especially when small changes in assumptions will have a make or break effect on institutions.

That doesn't mean the bailout plan is right or that it will even work, but playing cute with some "analysis" isn't going to help matters.

Maybe I need to be more specific. Alex, did you think I was asking someone to predict how risky a stock, bond, or loan is? There's no attempt at prediction in my proposal.

A bank knows what percentage of its money is in mortgage loans. One of the problems with the system that led to this crash is that the loans (and especially subprime and fluctuating rate loans) were bundled with other assets and resold. So many institutions do not know what percentage of their money is in loans, but if we trace it back to the institutions they bought from, we should be able to find out.

That kind of reasoning applies to other risky investments too, such as hedge funds. I don't pretend to be able to predict what will fail. But institutions do know how the money is divided up. If you have a money market, just look at the prospectus.

The fix is already in. It's clear enough by now that the US government will bail out the financial institutions by buying up their "toxic waste." So it's way too late to target the individual problems and avoid the giant expense.

Nonetheless, those nasty assets are about to become the property of the American people. Analysis has to be applied to maximize return down the line, probably beginning with moves that can sustain any real equity in the underlying mortgaged property. Andy's certainly right about the importance of tracing down the origin of the assets. It should have been done earlier, but it still needs to be done. The more granular the analysis, the better prepared the new asset holder (the American people) will be to: 1) model various stratifications of the portfolio, and; 2) drill down to the elements which need to be dealt with case-by-case.

As the information is consolidated, I'd recommend paying close attention to any local tax liens that might be outstanding on the underlying property. There's an intensely active market in such lien sales because the lien purchaser's rights have priority over any others', including the mortgage holder's. You'll need to know whether satisfying those liens (and the associated penalties) is a good bet, or is simply an exercise in throwing good money after bad.

Andy, a lot of the problem can be found in http://blogs.wsj.com/deals/2008/09/22/why-has-the-credit-crunch-been-so-bad-look-to-washington/?mod=msn_money_ticker

(Sorry. I'm at Panera; they block tinyurl.com ).

The problem is that the current accounting rules require that each asset be revalued quarterly. If some other broker has a fire sale on something and you have a "comparable" asset in your portfolio, you must mark down your asset immediately.

I believe this is all SOX legislation, which was congress's reaction to Enron's accounting.

The funny thing about that is now the presidential candidates are saying we need more regulation! Ugh.

When I was a kid my Mom took me to the doctor and got me inoculated for a bunch of diseases that were rare or for most purposes, did not exist (smallpox was about extinct at the time). She was a good Mother, but let's be extreme for a second and extend your logic. Shouldn't we just stop, do some analysis and then make decisions? Perhaps. If my Mom had did the same thing and tried to determine with absolute certainty the risks of inoculating me - as an individual - irrespective of the risks, there would be a possibility that I exposed to something like polio. A devastating disease.

Some kids who are inoculated get sick or even die - but intelligent and caring parents take that small risk and do the right think. "...but I bet the criteria could be a whole lot smarter..." given time to do said analysis. Of course you are correct, but if every parent lapsed into analysis paralysis with every inoculation decision, then diseases that we had hoped to wipe out would menace us once again.

Now, having said all that: Of course there will be analysis done on these instruments, but none of them are the same. Each instrument is like doing analysis on a separate company - no easy task. So, in the near-term decisions will need to be made quickly and with less refined intelligence - else there is the possibility that the arterial contagion that has frozen the credit markets leaps from Wall Street to Main Street.

Someone should invent headphones that stick fingers in your ears and go 'lalalalalalala'.
Then when you ask your broker, fund manager, pension etc
'Isnt there something wrong about this financial bubble?' as many had been doing for years, you could see the headphones and know there is quite a high risk involved in this game.

Only people with no maths at all were surprised by this - unfortunately they seem to be running the financial system so a risk calculator would be of no use to them

Andy Oram: "did you think I was asking someone to predict how risky a stock, bond, or loan is? There's no attempt at prediction in my proposal."

and, "So many institutions do not know what percentage of their money is in loans, but if we trace it back to the institutions they bought from, we should be able to find out."

Let me clarify. What I think you are saying is that if we could trace the bad loans to the institutions and target those loans directly, we could reduce the size of the bailout. That is what I am arguing as a misunderstanding of the problem.

The banks already know what their loans compositions are, but that is only a fraction of the problem. The total size of sub-prime loans is in the $T's of dollars, most of which might be performing. However, the performance of those loans is dependent on the economy and the trajectory of house prices. No-one knows how big the write offs need to be, and there is no way to accurately know that. However, we do know that the bailout funds would not cover the total risk of bad loans. has been suggested as a more realistic number. The next problem, and arguably a worse one is the derivative products, such as CDS's. These were used for hedging and speculation. The problem here is that the CDS market is ~ $60T. This is so big that Bear Stearns was forcibly bought to prevent chaos in trying to unwind positions. AIGs insurance in this area created liabilities which could have wiped out its equity.

So what we have is a situation where highly levered banks and institutions could suffer a catastrophic loss of equity if they are on the wrong side of a trade. This loss is only partially dependent on risky loans. It is no use trading vanilla commercial loans with a counter party bank if you don't know what their positions are with unregulated derivatives or even how far they have have correctly marked to market their loan portfolio.

The $700bn, an educated guess at the size of the immediate liquidity problem due to the worst of the loans may or may not solve this problem.

Then we have to really worry about 2nd and 3rd order effects. If the bailout inadequately restores bank capital ratios, or even if it does, there is an expectation that the banks may have to delever, This will make credit much tighter, contracting the economy which could bring on new waves of loan write downs in un-related areas. The nightmare scenario is another depression.

The Fed and treasury policy is to try to restore bank's confidence in each other by removing the perceived counter party risk.(This is indicated by the TED spread). Whether the $700bn plan is more than sufficient or less, we just don't know. But we do know that the problem is not just about the sub-prime loans and their structured nature, which is just part of the financial instability we are experiencing.

If I have mis-characterized what you were saying with a straw-man argument, I apologize, however, that is my understanding of what you were saying - that fundamentally if we could identify just the "bad" loans, the problem would be smaller than $700bn. O'Reilly has access to Bill Janeway, perhaps he can be of mor ehelp in explaining this issue.

Thanks for the fuller explanation. Yes, I think you caught my thinking, and your message seems to be that simple classifications such as "fixed-rate versus "adjustable-rate" or "hedge fund" doesn't offer enough information to assess how risky a portfolio is. That would explain why the market is so instable (because nobody knows whom to entrust with their money) and would also invalidate my suggestion.

I'm into the site this morning to download some cookbook code, and happened to see the title of this discussion. I was thinking about this subject riding into work this morning.

What bothers me most, from my perspective as a developer/engineer, is how can so many intelligent people at these institutions state that they don't know the valuations of the financial instruments and investments that they hold at any particular moment?

My point is that these institutions have data analysis capabilities comparable to those of the NSA. They have the data, but the financial management lemmings interpreting it were thinking with their wallets, not their brains.


Given your logic that a contagion might spread so we must act quickly for better or worse, why not just shoot and bury anyone who is suspected of having the disease?

That's what Congress is about to do.

But sometimes it's difficult to determine who is actually sick and who is not, especially when the illness affects the brain.

The gun is pointed not at the fat cats, but at us - the workers. The workers who will be laid off and will not have health insurance. The workers who will not have a pension, a balanced 401k portfolio, or a social security check that will buy a month's worth of dog food.

Or a roof over our heads.


@Eric -

You said:

"What bothers me most, from my perspective as a developer/engineer, is how can so many intelligent people at these institutions state that they don't know the valuations of the financial instruments and investments that they hold at any particular moment?"

Here is the deal.

What is your house worth right now?

Okay, you have to sell it in 5 minutes. What is it worth?

What if you had to sell your house in the next 5 minutes and there are no buyers? What is your house worth now?

This is what the intelligent people are facing, which makes it impossible to determine the value of these instruments.

News Topics

Recommended for You

Got a Question?