Arnold Kling  

Why Repo Exists

Brink Lindsey, in case you mis... The President's Roanoke Speech...

Nick Rowe asks the question. My answer is based on Marcia Stigum's Money Market, a book that more economists need to read.

Securities dealers are dealers, sort of like auto dealers. If you go to an auto dealer, you expect to find cars on the lot to buy. The dealer does not have the capital to have $1 million tied up in inventory. So the dealer takes out a loan from the bank, using the inventory of cars as collateral.

Now think of a securities dealer. It has an inventory of Treasury securities. It finances them with a loan, using the securities as collateral. If firm A lends to dealer B on a one-day repo, firm A actually has ownership of the security for the day. Dealer B buys the security back from dealer A one day later, at a higher price. The price difference represents the repo loan rate.

Now, at some point, if lots of dealers are financing portfolios of 10-year bonds with one-week repo loans, a fall in the value of 10-year bonds is going to cause some serious problems. Until something like that happens, though, you don't know whether there is too much repo out there. I wish Gary Gorton, who is in love with repo, would at least admit the possibility that there is such a thing as too much of it.

Comments and Sharing

COMMENTS (4 to date)
MG writes:

Let's not make the perfect the enemy of the good. Compared with most other forms of financing, the OTC repo market is technically equipped to handle a lot of stress. There is the equivalent of an initial downpayment (margin) which is based on the expectation that collateral proces can (actually) fall; if market volatility persists, variation margin calls require that additional collateral be posted; having a lot of dealers involved in the market helps because netting of credit exposure is the practice, thus mimimizaing the overall credit risk of any single transaction; and securities lawyers have actually done a decent job of designing a system that minimizes value destruction even in bankrupcy. This market allows for the financing of an incredibly hetergogenous amount of collateral. But, as the OTC repo market moves to tighten its pricing to reflect these outlier risks Arnold must be alluding to, the bulk of plain vanilla directional speculation/hedging would gravitate to exchanges, which may not be a bad thing anyway. Frankly, with a bit more transparency in dealer positions, this market could be as credit tight as one would need.

JeffM writes:

I agree with MG, but let's follow the analogy with floor plan loans a bit further.

If the demand for cars tanks, a lot of car dealers will not be able to pay off their loans, and banks will incur losses. If banks incur losses that are big enough, they will fail. So, if repo is a bad thing, so must floor plans be a bad thing. Similarly, mortgage loans must be a bad thing because, as we have just seen, many mortgages will default if demand for houses craters. If loans that may default are bad, then all credit to private parties is bad.

The fact is that all lending involves risk. I'd argue from experience that the risk of a lender taking a loss on a loan secured by actively traded securities, granted at a discount to market, and supported by the right to demand extra collateral if market values fall is very low. Moreover, if the loan defaults and a loss must be taken, the amount of loss is low compared to the aggregate loss on a repossessed car or a foreclosed house or an inventory of cheese. (In the case of the cheese, which had spoiled when the electricity was cut off by the power company, the loss was 100 cents on the dollar.) Finally, if the market for the collateral freezes and the bank must take permanent or prolonged possession of the collateral, a bank gets an income flow from a bond whereas it seldom gets any income from a repossessed car or foreclosed house. The bank may not have wanted a 10-year bond in its portfolio, but the marginal expense of keeping such an asset on its books is virtually zero, which is certainly not true of a house, which requires outlays for taxes, insurance, and maintenance.

In 32 years of banking, I do not remember taking a loss on a loan over-collaterlized with actively traded securities and supported by a right to call or to demand extra collateral. On the other hand, I certainly do remember losses on unsecured loans or loans secured by houses, cars, equipment, buildings, physical inventory, or accounts receivable.

Mike Rulle writes:

Repo's are the Dealers' equivalent of the Futures Exchanges. As long as cash calls are made daily, it is very difficult for there to be systemic problems. Problems arise when exceptions are made.

Another analogy is interest rate swaps. As long as collateral rules are in place it is almost impossible for a dealer to get over leveraged.

I use this analogy often. Amaranth lost $9 billion in a few days in the nat gas futures market. The reason it was not more is they were limited in size by the amount of cash they had to post. When they ran out of cash their positions were unwound whole to their counter parties. Not a ripple occurred in the market.

The reason AIG caused mark to market losses of double digit billions to a variety of Firms is the latter did not require cash collateral---and AIG was able to leverage themselves to the moon. The dealers should have been forced to band together and bail themselves out.

Norman Pfyster writes:

Repos are just a form of short-term secured lending using industry-standard terms for efficiency purposes. Is someone really asking why short-term secured lending exists?

Gorton certainly does not believe that repo markets can't cause problems: his whole argument is that the current financial crisis was caused by "runs on the bank" in the repo, securities lending and derivatives market (i.e., the collateralized financing market), and thus that it needs regulation.

Comments for this entry have been closed
Return to top