Envelope theorem

Envelope theorem. In mathematical economics and the calculus of variations, the envelope theorem is a major result about the differentiation of value functions in optimization problems. It provides a method for computing how the optimal value of an objective function changes when an exogenous parameter changes, holding the choice variables at their optimal values. The theorem simplifies analysis by showing that only the direct effect of the parameter on the objective function needs to be considered, as the indirect effect through adjustments of the choice variables is zero for small changes. Its development is closely associated with the work of Paul Samuelson in his Foundations of Economic Analysis and has become a fundamental tool in microeconomic theory, comparative statics, and dynamic programming.

Statement of the theorem

Consider a standard optimization problem where an agent chooses a vector \( x \) to maximize a function \( f(x, \alpha) \), where \( \alpha \) is a parameter. Let \( x^*(\alpha) \) be the optimal choice and \( V(\alpha) = f(x^*(\alpha), \alpha) \) be the value function. Under suitable regularity conditions, including differentiability and an interior optimum, the envelope theorem states that the derivative of the value function with respect to the parameter is given by the partial derivative of the objective function, evaluated at the optimum: \( \frac{dV}{d\alpha} = \frac{\partial f}{\partial \alpha} \big|_{x = x^*(\alpha)} \). This result holds because the first-order conditions for optimization imply that the derivative of \( f \) with respect to \( x \) is zero at \( x^* \), so any indirect effect via changes in \( x^* \) vanishes. The theorem can be extended to constrained problems using Lagrange multipliers, where the derivative of the value function with respect to a parameter equals the partial derivative of the Lagrangian.

Intuitive interpretation

The core intuition is that at an optimum, the decision maker is already choosing variables to maximize the objective, so a marginal change in a parameter does not induce a first-order change in the objective through re-optimization of those variables. This is analogous to the fact that at the peak of a hill, a small shift in the landscape's foundation does not require you to move to remain at the peak—you are already at a stationary point. In economic terms, if a firm is maximizing profit given a price parameter, a tiny change in that price does not alter the profit contribution from the firm's input choices, as those choices were already optimal. This principle is why in consumer theory, the derivative of the indirect utility function with respect to price yields the Hicksian demand via Roy's identity, and in producer theory, Hotelling's lemma relates the profit function to the supply function.

Applications in economics

The theorem is ubiquitous in economic modeling for conducting comparative statics without fully resolving the optimization problem anew. In consumer choice theory, it underpins the derivation of Shephard's lemma from the expenditure function and the relationship between the cost function and conditional factor demands. Within the theory of the firm, it is used to derive Hotelling's lemma from the profit function and to analyze the effects of parameter shifts in production function models. The theorem is also critical in contract theory and mechanism design for assessing how changes in informational parameters affect the optimal contract's value. Furthermore, it is employed in public economics to evaluate the welfare impact of tax changes and in international trade models to understand the effects of tariff adjustments on national welfare, as seen in analyses building on the Heckscher–Ohlin model.

Generalizations

Several important generalizations extend the basic envelope theorem to broader contexts. The Benveniste–Scheinkman theorem provides conditions for the differentiability of value functions in dynamic programming and optimal control theory, which is foundational for the Hamilton–Jacobi–Bellman equation. In problems with inequality constraints, the Karush–Kuhn–Tucker conditions and the concept of the value function lead to envelope results that involve the multipliers on binding constraints. The theorem has also been extended to problems with multiple parameters, leading to the envelope formula for the Hessian matrix of the value function, which is used in sensitivity analysis. For optimization in Banach spaces and variational analysis, more abstract versions exist, connecting to the theory of subdifferentials and the work of R. Tyrrell Rockafellar in convex analysis.

Derivation and proof

A formal proof for the unconstrained case starts by differentiating the value function \( V(\alpha) = f(x^*(\alpha), \alpha) \) with respect to \( \alpha \), applying the chain rule: \( \frac{dV}{d\alpha} = \sum_i \frac{\partial f}{\partial x_i} \frac{dx_i^*}{d\alpha} + \frac{\partial f}{\partial \alpha} \). At the optimum \( x^* \), the first-order condition requires \( \frac{\partial f}{\partial x_i} = 0 \) for all \( i \), provided the optimum is interior. Thus, the summation term vanishes, leaving \( \frac{dV}{d\alpha} = \frac{\partial f}{\partial \alpha} \big|_{x = x^*(\alpha)} \). For constrained problems, consider maximizing \( f(x, \alpha) \) subject to \( g_j(x, \alpha) = 0 \). The Lagrangian is \( \mathcal{L}(x, \lambda, \alpha) = f(x, \alpha) + \sum_j \lambda_j g_j(x, \alpha) \). The envelope theorem then states \( \frac{dV}{d\alpha} = \frac{\partial \mathcal{L}}{\partial \alpha} \big|_{x = x^*(\alpha), \lambda = \lambda^*(\alpha)} \), as the terms involving derivatives of \( x \) and \( \lambda \) vanish due to the first-order conditions. This result is pivotal in proofs of duality theorems in economics, such as those connecting the Slutsky equation to expenditure minimization. Category:Mathematical theorems Category:Mathematical economics Category:Microeconomic theory