Monday, 21 July 2008

US banks - how they got into such a mess

Some loans always go bad. It is an unavoidable fact that all bankers must confront. Sensible bankers take precautions; they build up loan provisions, so when borrowers default, the loss is covered.

Loan provisioning directly affects bank profitability. If a bank stashes some cash away to cover a loss, it can not also distribute that cash as profits. Although loan provisioning creates some complex tax issues, in general banks don't like to defer large amounts of cash to cover potential losses.

How much cash has the troubled US banking system saved up to cover potential losses? Back in the early 1990s, US banks put away the equivalent of 2.6 percent of total loans as loan loss provisions. Since then, the ratio has fallen steadily. It reached a low in 2007, when the ratio fell to just a fraction over 1 percent.

However, the subprime crash has caused banks to suddenly reverse this policy. In the last couple of months, banks have been rapidly increasing their provisioning. Presumably, the Fed has advised US banks to prepare for an uptick in losses.

However, where the banks right to economise on their loan loss provisions? Did loan losses fall during the early 2000s, thus justifying the reduction in loan provisioning?

Not really, after the S&L crisis subsided in the early 1990s, net loan losses fell. However, by the mid-1990s, net loan losses as a proportion of total loans fluctuated somewhere between 0.5 and 1 percent of total loans. For most of this decade, net loan losses were either higher or at comparable levels to the 1990s. Actually, net loan losses peaked at over 1 percent around 2002. In short, net loan loss rates gave the banks little justification for economising on loan provisioning.

Subtracting the two numbers gives some idea about how capacity of US banks to absorb losses. As the chart below illustrates, the margin between loss provisioning and net loan losses has narrowed considerably over the last ten years. Take a look at the last data points; it shows that the provisioning buffer has narrowed quitely sharply in the last couple of months. This indicates that despite the increase in provisioning, US banks are finding it difficult to cover the upswing in net loan losses.

Presumably, behind these numbers is the well-known story of mortgage securitization, increased subprime lending, and self-certification loans. If banks could off-load risk to others, then lending volumes could increase.

The numbers also show that US banks are struggling to absorb losses. The really frightening thing is that the existing level of net loan losses were accumulated when US economic growth was high, while unemployment was low. When the US economy begins to seriously slow, then net loan losses will skyrocket. US banks are simply not ready for this.

No wonder the Fed is so keen to keep interest rates low.


Josh said...

Do you have similar charts for UK banks?


Good story.

Edward Harrison said...

Great story, Alice. It does show how bank profitability is goosed two fold on the way up and chopped two-fold on the way down.

Not only do profits rise temporarily because of risky loans, but banks under-provision for these loans making profits seem even larger.

Very dodgy accounting, if you ask me.

dearieme said...

Seems to me to be a natural reflection of the incentives for the executives - in the bubble you get rich and in the bust you get a very handsome pay off.

Anonymous said...

Brilliant as usual, Alice, and prescient in view of Wachopvia's nearly nine billion loss announced today. B. in C.

Anonymous said...

Brilliant as usual, Alice, and prescient in view of Wachopvia's nearly nine billion loss announced today. B. in C.

MAB said...


Good blog. This is not a new story, although the scale is huge this time around. Bad loans are made during good times as they used to say.

Low interest rates will be of little help. Low rates help borrowers, but discourage actual lending. See the paradox? I'm a saver, but I will not lend at low rates foolishly again. Unfortunately, the U.S. government will lend on my behalf.

I see more & more government bailout schemes coming which are inherently inflationary.

Inflation in needs, deflation in wants and assets like stocks & bonds (especially leveraged assets). Bad investments don't become good investments just because Uncle Sam takes an interest.

Just my 2 cents. Keep up the good work. Cheers.

Anonymous said...

Alice, this may explain why USB "only" reinsured 3% (40M) of its sub-prime debt(1.3B). If it only "expected" to lose 1% of its loans, then 3% was a massive provision. Happy concidence - the reinsurance also kept the sub-prime loans off the balance sheet as doubtful items.

Your post of a week or so ago refers


Alice Cook said...


An excellent point.

I think the answer to your point is a complex one. You are also pushing me to the limits of my knowledge of banking regulations. But here goes.

A credit default swap allows a bank to zero rate any loan that is insured. In effect, it takes the loan off the bank's balance sheet. However, not all loans disappear like this. Some loans remain on the balance sheet and must be provisioned. This is what the charts are picking up.


Cahya said...

An interesting article.
See you and have a nice blogging.