Gorton & Pennacchi (1990)

Citation: Gorton, G., & Pennacchi, G. (1990). “Banks and Loan Sales: Marketing Non-Marketable Assets.” NBER Working Paper No. 3551. (Published in Journal of Monetary Economics, 1995)

Gorton and Pennacchi address a puzzle at the heart of financial intermediation theory: if banks exist because they produce information and monitor borrowers—activities that require holding loans to maturity for incentive compatibility—then why did a massive market for selling commercial loans emerge in the 1980s? Loan sales grew from $26.7 billion in 1983 to $290.9 billion by 1989.

The theoretical problem:

Standard intermediation theory (Diamond 1984, Boyd & Prescott 1986) implies that bank loans should be non-marketable. If banks could sell loans, they would lose incentive to evaluate and monitor borrowers. Loan buyers would recognize this moral hazard and discount accordingly—creating a lemons market. Yet loan sales exist and are growing. What changed?

flowchart TD
  subgraph Traditional Theory
    B[Bank monitors borrower]
    B --> H[Must hold loan to have incentive]
    H --> NM[Loans non-marketable]
  end
  
  subgraph Observed Reality 1980s
    LS[Loan sales market emerges]
    LS --> Q[Why is this possible?]
  end

Possible mechanisms for incentive compatibility:

The authors test two implicit contract features that could make loan sales work despite the moral hazard problem:

Implicit guarantees. Banks might implicitly promise to repurchase loans that deteriorate, providing de facto credit enhancement even though explicit guarantees are prohibited by regulation.
Retained interest. Banks might retain a fraction of each loan sold, maintaining “skin in the game” that preserves monitoring incentives.

Empirical findings:

Using 872 individual loan sales from a major money center bank (1987-1988), they find:

Little evidence of implicit guarantees. The data do not support the hypothesis that banks are providing implicit credit enhancement to loan buyers.
Some evidence of retained interest. The fraction of loans sold is negatively related to loan risk—banks retain larger shares of riskier loans. This is consistent with maintaining monitoring incentives.
Technological change is key. The most compelling explanation is that information technology has reduced the cost for loan buyers to verify bank monitoring. If buyers can observe whether the selling bank is doing its job, the moral hazard problem diminishes.

flowchart LR
  subgraph Why Loan Sales Work
    TC[Tech progress] --> LC[Lower info costs]
    LC --> V[Buyers can verify monitoring]
    V --> IC[Incentive compatible]
  end
  
  subgraph Evidence
    E1[Most buyers are other banks]
    E2[Banks retain riskier loans]
    E3[No evidence of implicit guarantees]
  end

The technological change hypothesis:

The paper argues that advances in computing and telecommunications have fundamentally altered the information economics of banking:

Institutional investors (banks, insurance companies, pension funds) invested heavily in information technology
This reduced their cost of gathering, analyzing, and transmitting credit information
Most loan buyers are other banks—institutions with comparative advantage in verifying bank monitoring activities
The same forces explain the growth of junk bond markets (held 75% by institutions) and commercial paper markets

Implications for bank regulation:

The existence of liquid markets for bank loans challenges the fundamental rationale for deposit insurance and bank regulation. Traditional justification assumes bank assets are illiquid and depositors cannot value them. If loans can be sold in liquid markets, this information asymmetry is resolved, and the case for government intervention weakens.

Key contributions:

Documents the empirical puzzle of loan sales and tests alternative explanations
Develops a formal model of optimal loan sale contracts with moral hazard
Provides early evidence that technology was transforming the marketability of bank assets
Raises fundamental questions about the continued need for bank regulation

The paper is foundational for understanding securitization. It shows that the boundary between “relationship lending” (illiquid, information-intensive) and “transaction lending” (liquid, standardized) is not fixed but depends on information costs. When technology reduces these costs, previously non-marketable assets become marketable—a dynamic that would accelerate dramatically in subsequent decades.

The emergence of loan sales was enabled by technological progress that reduced information asymmetries between sellers and buyers. Banks retain fractions of riskier loans to maintain incentives, but the primary driver was buyers’ growing ability to verify seller behavior—a shift that transformed the economics of financial intermediation.