The Mortgage Professor

Spreading Risk Around Isn't the Answer

By Jack Guttentag
Saturday, March 22, 2008

You have heard the complaint: In our home loan system, nobody worries about the risk because they pass it on to the next player in the chain. If everyone in the chain had skin in the game -- something to lose if the loan went bad -- we wouldn't be in the mess we are in now.

A long chain of risk transfers is certainly a feature of our housing finance system. In a typical scenario whereby loans end up as collateral for a mortgage security, a mortgage broker shifts the risk to a wholesale lender, who transfers it to an investment banker, who transfers it to multiple investors. The broker at the head of the chain usually knows the most about the risk, while the investor who ultimately bears the risk often knows the least.

Some knowledgeable observers have suggested that the government should enact the following rule: Every player in the chain must have skin in the game. I think this is a bad idea.

Economists call a situation in which one party makes a decision for another, without having the same stake in the outcome, the "principal-agent problem." It is pervasive in our society. Principals protect themselves by establishing rules that their agents must follow and by control mechanisms designed to ensure that the rules are being observed. Sometimes these mechanisms work, and sometimes they don't.

At the top of the mortgage chain, the possibility of loss, or skin in the game, is one of the control mechanisms used by principals. Investors purchasing a mortgage security may require that the investment bank issuing the security retain a piece of it. The investment bank, in turn, will require that the lenders from whom it purchases loans agree to repurchase those loans that don't meet the investment bank's standards.

The rules and their enforcement became lax during the boom years, but now they are extremely stringent -- wherever the market continues to function, that is. Even if government intrusion were warranted here, which I don't believe is, the timing is terrible.

At the bottom of the chain, lenders set underwriting rules -- conditions that loans must meet to be approved -- that agents (brokers and loan officers) must follow. A lender employee called an underwriter, or an automated underwriting system, must sign off on a loan before it is approved. The control mechanisms for enforcing compliance, however, are weak.

Wholesale lenders can't require that mortgage brokers have skin in the game. The typical broker does not have the money to buy back loans that don't meet the lenders' standards. About the only thing a lender can do is take a miscreant broker off its approved list, but that won't prevent the broker from doing business elsewhere.

The inability of lenders to control brokers is one reason why some observers would not be unhappy to see brokers disappear, and some of their proposals seem to have that as an unstated objective. They appear to believe that eliminating brokers would eliminate the principal-agent problem at the point of sale.

But that is not the case. The alternative to mortgage brokers is loan officers, with whom lenders have a principal-agent relationship that is basically the same as the one that exists in their relationship with mortgage brokers. Loan officers are lender employees who are compensated on a commission basis, whose income depends entirely on the number and amount of loans they produce, and who have no stake in whether a loan turns out to be good or bad.

I remember vividly that soon after I joined the board of a large savings and loan, the chief executive gave the board a list of the most highly compensated employees of the company. To my surprise, the chief executive's name was only third on the list, and I did not recognize the two names at the top. They turned out to be loan officers.

I found that there was no way to cap the compensation of loan officers, or to force them to accept greater risk, even though they legally are employees. Lenders who have tried to base commissions on loan performance have found that their best producers quit to join other lenders.

The only way to get skin in the game at the bottom of the financing chain, where loans are originated, is to revert to a system similar to those found in many less-developed countries. In such a system, lenders are either individuals or very small firms in which the owner is the chief executive and makes all loan decisions. Every player in the chain would have risk, and the potential for boom and bust would be negligible. The trouble is, systems of this type typically offer just one type of loan, with high interest rates, short terms and large down payments.

Jack Guttentag is professor of finance emeritus at the Wharton School of the University of Pennsylvania. He can be contacted through his Web site,

Copyright 2008 Jack Guttentag

Distributed by Inman News Features

© 2008 The Washington Post Company