A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
The Bellman equation decomposes a dynamic optimization problem into current period and future components: V(x) = max[u(c,x) + βV(x')]. This recursive formulation enables solving infinite-horizon problems and characterizing optimal consumption, investment, and labor supply decisions over time.
From your work with constrained optimization and Lagrangian methods, you know how to find the best outcome when an agent faces a single decision with constraints. Dynamic optimization extends this to sequential decisions over time — but writing out an infinite sequence of first-order conditions and solving them simultaneously is impractical. The Bellman equation cuts through this complexity with a single, elegant insight: if you know the value of being in any future state, then today's optimal decision reduces to a one-period problem. This is the principle of optimality — an optimal plan has the property that, whatever your current state and decision, the remaining decisions must also be optimal given the state resulting from your current choice.
Concretely, define V(x) as the value function — the maximum total discounted payoff an agent can achieve starting from state x and behaving optimally forever after. The Bellman equation states that V(x) = max_c [u(c, x) + βV(x')], where c is the current control (like consumption), u(c, x) is the current-period payoff, β is the discount factor (between 0 and 1, reflecting time preference), and x' is the next-period state determined by a transition equation x' = g(x, c). The equation says: the value of being in state x equals the best you can do by choosing c today — enjoying immediate payoff u(c, x) — plus the discounted value of wherever that choice puts you tomorrow. The recursive structure means the same equation holds at every point in time, collapsing an infinite-horizon problem into a single functional equation.
Solving the Bellman equation means finding the value function V and the associated policy function c*(x) that tells the agent what to do in every state. From your study of fixed-point iteration, you can appreciate the main computational approach: value function iteration starts with an initial guess V₀(x), plugs it into the right-hand side of the Bellman equation to compute V₁(x), and repeats. Under standard conditions (bounded payoffs, β < 1), the contraction mapping theorem guarantees this process converges to the unique fixed point V*. Each iteration improves the approximation, and the policy function converges alongside it. Alternatively, you can take the first-order condition of the maximization on the right-hand side to derive the Euler equation, which characterizes the optimal tradeoff between consuming today and saving for tomorrow — a condition you will encounter repeatedly in macroeconomic models.
The power of the Bellman equation in economics is its universality. A household choosing how much to consume and save each period, a firm deciding when to invest in new capital, a job seeker deciding whether to accept a wage offer or keep searching — all of these problems share the same recursive structure. The state variable changes (wealth, capital stock, current wage offer), the payoff function changes, and the transition equation changes, but the logic is identical: decompose the infinite future into "today" and "the value of tomorrow," optimize today's choice, and let the recursion handle the rest. This framework is the backbone of modern macroeconomic theory, from growth models to asset pricing to labor economics.