Theorems on Implicit Functions and Inverse Mappings

Problem Statement

Most important mathematical relationships are defined implicitly: by equations, systems of equations, equilibrium conditions. The equation of state of a gas links pressure, volume, and temperature, but we wish to know how pressure changes as temperature changes at a fixed volume. The equation of a curve $F(x,y) = 0$ defines a manifold, but to compute tangents we need to understand at which points it resolves “smoothly” as the graph of a function.

Theorems on implicit and inverse mappings answer a fundamental question: when does a nonlinear system “locally” behave like a linear one? The answer is: when the corresponding linear approximation (the Jacobian matrix) is invertible. This is essentially the principle of linearization, which permeates all of mathematical analysis.

The Implicit Function Theorem: One-Dimensional Case

Theorem (Cauchy, 1831): Let $F: U \subset \mathbb{R}^2 \rightarrow \mathbb{R}$ be continuously differentiable, $F(x_0, y_0) = 0$ and $\frac{\partial F}{\partial y}(x_0, y_0) \neq 0$. Then there exist neighborhoods $V \ni x_0$ and $W \ni y_0$ such that for each $x \in V$, the equation $F(x, y) = 0$ has a unique solution $y = \varphi(x) \in W$. The function $\varphi$ is continuously differentiable, and $ \varphi'(x) = -\frac{\partial F}{\partial x}(x, \varphi(x)) \big/ \frac{\partial F}{\partial y}(x, \varphi(x)). $

The condition $\frac{\partial F}{\partial y} \neq 0$ is key. Geometrically, this means: the level surface $F = 0$ “intersects” the vertical direction nontransversally. Analogy: the linear equation $ax + by = 0$ is solvable for $y$ if and only if $b \neq 0$.

Example 1: Circle. $F(x, y) = x^2 + y^2 - 1$. $F_x = 2x$, $F_y = 2y$.

At the point $(\sqrt{3}/2, 1/2)$: $F_y = 1 \neq 0$ → the theorem applies, $y = \varphi(x) = \sqrt{1-x^2}$ near this point. $\varphi'(x) = -x / \sqrt{1-x^2}$. At $x = \sqrt{3}/2$: $\varphi' = -\sqrt{3}$. This is the slope of the tangent to the circle at the point $(\sqrt{3}/2, 1/2)$.

At the point $(1, 0)$: $F_y = 0$ → the theorem does not apply. Indeed, at the point $(1,0)$ the circle has a vertical tangent and is not the graph of a function $y(x)$.

Example 2: The Kepler Equation. $M = E - e \sin E$ implicitly relates mean anomaly $M$ to eccentric anomaly $E$ ($e$ is the orbital eccentricity). Take $F(M, E) = E - e \sin E - M$. $\frac{\partial F}{\partial E} = 1 - e \cos E \neq 0$ for $e < 1$ (elliptical orbits). Therefore, $E = E(M)$ is a smooth function, which ensures correctness of numerical solution.

The Derivative of Implicit Functions via the Total Differential

An alternative, quicker way: from $F(x, \varphi(x)) = 0$ take the total differential:

$ dF = F_x dx + F_y d\varphi = 0 \rightarrow d\varphi = -\frac{F_x}{F_y} dx. $

For functions of many variables $F(x_1, \ldots, x_n, y) = 0$ with $\frac{\partial F}{\partial y} \neq 0$:

$ \frac{\partial y}{\partial x_k} = -\frac{\partial F}{\partial x_k} \big/ \frac{\partial F}{\partial y}. $

Example 3: Thermodynamics. The van der Waals equation: $(P + a/V^2)(V - b) = RT$, where $P$ is pressure, $V$ is molar volume, $T$ is temperature, $a, b, R$ are constants. This is an implicit equation $F(P, V, T) = 0$.

$\frac{\partial P}{\partial T}|_V = \frac{RT}{V-b} \cdot \frac{1}{V-b} \big/ -\frac{\partial F}{\partial P}$. Calculation: $\frac{\partial F}{\partial T} = R(V-b)$, $\frac{\partial F}{\partial P} = (V-b)^2$, hence $\frac{\partial P}{\partial T}|_V = \frac{R}{V-b}$. This is the coefficient of pressure change during isochoric heating.

Multidimensional Implicit Function Theorem

Let $F: \mathbb{R}^{n+m} \rightarrow \mathbb{R}^m$ be continuously differentiable, $F(x_0, y_0) = 0$, and the matrix $\frac{\partial F}{\partial y}$ of size $m \times m$ is nondegenerate at the point $(x_0, y_0)$. Then in the neighborhood of $x_0$ the system $F(x, y) = 0$ has a unique solution $y = \varphi(x)$ with $D\varphi = - \left(\frac{\partial F}{\partial y}\right)^{-1} \cdot \frac{\partial F}{\partial x}$.

Application — Comparative Statics in Economics. The system of equilibrium equations $G(q, p, \alpha) = 0$, where $q$ are quantities, $p$ are prices, $\alpha$ are policy parameters. The matrix $\frac{\partial G}{\partial (q, p)}$ is the Jacobian with respect to endogenous variables. If it is nondegenerate, the theorem guarantees existence of an equilibrium function $(q, p) = \varphi(\alpha)$, and $D\varphi = - \left(\frac{\partial G}{\partial (q, p)}\right)^{-1} \cdot \frac{\partial G}{\partial \alpha}$ — multipliers: how equilibrium changes when parameters change.

The Inverse Mapping Theorem

Theorem: Let $F: U \subset \mathbb{R}^n \rightarrow \mathbb{R}^n$ be continuously differentiable and $\det(DF(a)) \neq 0$. Then $F$ is bijective in some neighborhood of the point $a$, the inverse function $G = F^{-1}$ is continuously differentiable, and $DG(F(a)) = (DF(a))^{-1}$.

The condition of nonzero Jacobian is a multidimensional analogue of the monotonicity condition $f'(x) \neq 0$ for one-dimensional inversion. The Jacobian is the “algebraic volume” under the mapping. If it is not zero, the mapping locally does not “collapse” space.

Geometric meaning: if the linear approximation $DF(a)$ is invertible, then $F$ itself is invertible in a small neighborhood. This is the principle: “what is true for the best approximation is (locally) true for the function itself.”

Example: Polar $\rightarrow$ Cartesian. $F(r, \theta) = (r \cos \theta, r \sin \theta)$. Jacobian: $\det DF = r$. Invertibility is guaranteed for $r \neq 0$, that is, away from the origin. At $r = 0$ the Jacobian becomes zero—and indeed, at the origin polar coordinates degenerate.

Connection with the Lagrange Method and Applications to Optimization

The implicit function theorem underlies the first-order optimality conditions under constraints. The Lagrange multiplier $\lambda$ has a direct meaning: $\lambda = dV^/d\beta$, where $V^$ is the optimal value of the objective function, $\beta$ is the right-hand side of the constraint $g(x) = \beta$. This follows from applying the theorem to the system $\nabla f = \lambda \nabla g,\ g(x) = \beta$.

Envelope theorem: If $V^(\beta) = \max f(x)$ subject to $g(x) = \beta$, then $dV^/d\beta = \lambda^*(\beta)$ — the Lagrange multiplier at the optimum point. This is a powerful tool of sensitivity analysis: the value of a resource equals its shadow price.

Question for Reflection: Consider a market with demand function $D(p, I)$ ($p$ is price, $I$ is income) and supply $S(p, t)$ ($t$ is tax). Equilibrium: $D(p^, I) = S(p^, t)$. Using the implicit function theorem, find $\frac{\partial p^*}{\partial t}$ — how the equilibrium price changes as the tax increases. What does the sign of this derivative mean?