Hidden Information Modelling: The Principal-Agent Problem with Adverse Selection

Imagine hiring a contractor to renovate your house. You know what work you need done and what you’re willing to pay, but there’s a catch: the contractor knows far more about their own efficiency and costs than you do. They might be a highly efficient professional who can complete the job quickly and cheaply, or they might be slower and less skilled, requiring more time and resources. This information gap creates a fascinating economic puzzle that has profound implications for contracts, pricing, and market efficiency.

This scenario illustrates what economists call the principal-agent problem with adverse selection, a cornerstone concept in modern contract theory that helps us understand how transactions work when one party has private information before a contract is signed.

The setup: when agents know more than principals 🔗

Hidden information modeling examines situations where an agent possesses private knowledge about their characteristics before entering into a contract with a principal. In economic terms, this private knowledge is called the agent’s “type.” The principal only knows the probability distribution of these different types but cannot observe which specific type they’re dealing with.

Consider a simple but powerful example. A company needs to hire a supplier to produce goods, but the supplier knows their own marginal cost while the company does not. There might be two types of suppliers in the market: an efficient type with low production costs and an inefficient type with higher costs. The company knows that both types exist in the market but cannot tell which type any particular supplier is just by looking at them.

This information asymmetry fundamentally changes how contracts must be designed. Unlike situations where both parties have complete information, the principal must now create contracts that account for what they don’t know.

Efficient versus inefficient agent types 🔗

In the standard model, we typically consider two agent types. The efficient agent has lower costs and can deliver the same output for less effort or expense. The inefficient agent faces higher costs for the same level of production. Mathematically, economists represent this as different marginal cost parameters.

Before any contract is signed, the agent knows exactly which type they are, but the principal only knows the proportion of each type in the population. This creates the fundamental information problem that drives the entire analysis.

The first-best benchmark: contracts under complete information 🔗

To understand the problem created by hidden information, we first need to establish what would happen if information were complete. This “first-best” scenario serves as our benchmark for efficiency.

Under complete information, the principal could observe each agent’s type directly and offer type-specific contracts that achieve efficient production levels for both types. The efficient agent would be assigned their optimal production quantity, and the inefficient agent would receive theirs. Neither agent would earn any surplus beyond their reservation utility-the minimum they need to participate in the contract.

This represents a Pareto-efficient outcome where no party can be made better off without making someone else worse off. The principal extracts all the surplus from the transaction, leaving agents with exactly their outside option value. Think of it as the principal having perfect negotiating power combined with perfect information.

However, this ideal scenario crumbles when we introduce information asymmetry. If the principal cannot observe agent types, offering two different contracts and hoping agents will truthfully reveal themselves doesn’t work. The inefficient agent would simply claim to be efficient to get the better contract terms.

Why the first-best fails: incentive compatibility and the single-crossing property 🔗

The first-best contract fails under asymmetric information because it lacks what economists call incentive compatibility. An incentive-compatible contract is one where each agent type voluntarily chooses the contract intended for them rather than trying to pretend to be a different type.

For a contract menu to work, it must satisfy the self-selection constraint. Each type must prefer their designated contract over the contract intended for the other type. This is where the elegant mathematical property known as single-crossing comes into play.

Understanding the single-crossing property 🔗

The single-crossing property, also called the Spence-Mirrlees condition, describes how different agent types’ preferences differ in a systematic way. Essentially, it means that an efficient agent’s willingness to trade output for payment differs from an inefficient agent’s willingness in a predictable manner.

Picture indifference curves on a graph where one axis represents output and the other represents payment. For any two types, their indifference curves cross at most once. More importantly, higher types have steeper indifference curves, meaning they value additional output differently than lower types.

This property is crucial because it allows the principal to design contracts that separate the types. The efficient agent finds it relatively less costly to produce additional output, so contracts can be structured to make high-output options attractive to efficient types but unappealing to inefficient ones.

Information rents: the cost of ignorance 🔗

Here’s where things get interesting from an economic perspective. To induce agents to truthfully reveal their types through their contract choices, the principal must give up something valuable: an information rent.

An information rent is the extra payoff an agent earns above their reservation utility because they possess private information. Think of it as a premium the agent extracts for having knowledge the principal lacks.

In the contractor example, if you want an efficient contractor to choose the contract designed for their type rather than pretending to be inefficient, you need to make the efficient contract sufficiently attractive. This means the efficient contractor ends up earning more than just their minimum acceptable profit-they earn a rent on their superior efficiency.

The principal’s dilemma: rent extraction versus efficiency 🔗

This creates a fundamental trade-off for the principal. On one hand, they want to achieve allocative efficiency by assigning the right production levels to each type. On the other hand, they want to minimize the information rent they must pay.

The principal’s optimization problem becomes a delicate balancing act. They must design contracts that are incentive-compatible while simultaneously trying to extract as much surplus as possible. This tension between rent extraction and efficiency leads to second-best outcomes that differ from the first-best ideal.

The information rent cannot be eliminated entirely without violating incentive compatibility. If the principal tries to extract all the rent from the efficient type, that type would simply choose the contract intended for the inefficient type instead. The principal must leave enough rent to keep the efficient agent honest about their type.

The optimal second-best contract: distortions and shutdown policies 🔗

Given these constraints, what does the optimal contract under asymmetric information look like? The solution reveals several fascinating features that have become fundamental principles in contract theory.

No distortion at the top 🔗

One striking result is that the efficient type receives their first-best production level in the optimal second-best contract. Economists call this the “no distortion at the top” principle. The most efficient agent produces exactly the socially optimal quantity, just as they would under complete information.

Why? Because distorting the efficient type’s output downward would reduce total surplus without providing any benefit in terms of reduced information rent. The efficient type is already receiving a rent, and reducing their output would only make things worse without solving the information problem.

Downward distortion for the inefficient type 🔗

The inefficient type’s contract, however, gets distorted. Their production level falls below the first-best level they would receive under complete information. This downward distortion serves a specific purpose: it makes the inefficient contract less attractive to the efficient type, thereby reducing the information rent the principal must pay.

Think of it this way. By making the contract intended for inefficient agents less generous in terms of production quantity, the principal widens the gap between what the efficient agent gets from their own contract versus what they would get from claiming to be inefficient. This makes it easier to keep types separated while paying less rent.

When shutdown becomes optimal 🔗

In some cases, the analysis reveals an even more dramatic solution: the principal might choose not to contract with the inefficient type at all. This “shutdown policy” occurs when the cost of providing information rent to the efficient type exceeds the benefit of transacting with the inefficient type.

Imagine a market where inefficient contractors are relatively common but contribute little value. The company might decide it’s better to hire only clearly efficient contractors and simply not serve the market segment that inefficient contractors would have served. While this reduces the total amount of trade, it can be profit-maximizing for the principal.

This result has important policy implications. It suggests that information asymmetries can lead to market exclusion, where certain types of transactions simply don’t occur even though they would be mutually beneficial under complete information. In markets like insurance or credit, this can manifest as certain groups being unable to access services.

Real-world applications and implications 🔗

Hidden information modeling isn’t just theoretical elegance-it provides powerful insights into real economic phenomena. Insurance markets offer a classic example. Health insurers cannot perfectly observe how healthy potential customers are, leading to the adverse selection problems where sicker individuals are more likely to purchase insurance. Insurers respond with deductibles and coverage limits that effectively create a menu of contracts designed to separate types.

Similarly, in procurement contracting, government agencies often don’t know the true costs firms face when bidding on projects. The optimal contract might involve distortions in project scope or payment structures to induce truthful revelation while minimizing information rents to efficient contractors.

Labor markets also exhibit these features. When employers cannot perfectly observe worker productivity before hiring, they may use signals like education credentials or create probationary periods that effectively function as screening contracts.

What do you think? Can you identify situations in your own business dealings where information asymmetry creates similar contract design challenges? How might understanding the trade-off between efficiency and rent extraction help you negotiate better agreements when you’re uncertain about the other party’s true characteristics or costs?

References

1. https://en.wikipedia.org/wiki/Principal%E2%80%93agent_problem
2. https://en.wikipedia.org/wiki/Contract_theory
3. https://faculty.haas.berkeley.edu/stadelis/Econ_206_notes_2006.pdf
4. https://en.wikipedia.org/wiki/Single-crossing_condition