The Euler equation is the heart of the Ramsey–Cass–Koopmans model and of modern macroeconomics. This page derives it twice: first in two periods with a Lagrangian, which shows the economics, then in continuous time with a Hamiltonian, which is what the growth model uses. Every symbol is defined before it is used.
A household has lifetime wealth W (in today’s units) to spend on consumption c₁ today and c₂ tomorrow. Any saving earns a net return r, so a unit not consumed today buys 1 + r units tomorrow. The budget constraint, in present value, is
The household maximises with β = 1/(1 + ρ). Form the Lagrangian with multiplier μ on the budget:
Take (log utility when θ = 1). Its marginal utility is u′(c) = c−θ, so the Euler equation reads c₁−θ = β(1 + r)c₂−θ, or
Take logs and use ln(1 + x) ≈ x for small x: the growth rate of consumption is approximately (r − ρ)/θ. The exponent 1/θ is the elasticity of intertemporal substitution: how much consumption growth responds to the gap between the return and impatience. For r = 7% and ρ = 3%, log utility (θ = 1) gives growth of about 4%, and θ = 2 gives about 2%.
In the Ramsey model time is continuous and the household’s problem is to choose a path of consumption per head c(t) to maximise
The factor e−(ρ−n)t combines impatience ρ with the growth of the household’s own size n. The tool is the current-value Hamiltonian, with the costate μ(t), the value of an extra unit of capital in utility terms:
The equation compares two rates: the return that saving earns and the household’s impatience. If the return is higher, a household willing to wait is rewarded by consumption that rises over time; if lower, consumption declines. The equation says nothing about the level of consumption. That is set by the budget constraint, and the two together, in the growth model, pin down the saddle path of Lesson 9.
Almost every modern macro model has a household whose saving is described by an Euler equation. In the New Keynesian model it becomes the dynamic IS curve, linking output to the real interest rate. The same logic appears in consumption-smoothing and permanent-income theory.
The consumption Euler equation links consumption in adjacent periods along an optimal plan. In two periods it is u′(c₁) = β(1 + r)u′(c₂): the utility lost by saving one unit today equals the discounted utility of the 1 + r units it yields tomorrow. In continuous time with CRRA utility it becomes ċ/c = (r − ρ)/θ.
Write the household's problem as maximising lifetime utility subject to a lifetime budget constraint. In two periods, form the Lagrangian u(c₁) + βu(c₂) + μ(W − c₁ − c₂/(1 + r)) and take first-order conditions. In continuous time, form a current-value Hamiltonian with a costate for capital, take the first-order conditions for consumption and capital, and eliminate the costate.
θ is the coefficient of relative risk aversion, and 1/θ is the elasticity of intertemporal substitution. A larger θ means the household dislikes consumption that changes over time, so a given gap between the return and impatience tilts consumption growth less.
It is the end-of-time condition that rules out over-saving: the present value of the capital stock must go to zero, lim e^(−(ρ−n)t) μ(t) k(t) = 0. A plan that keeps capital accumulating at a rate that outpaces discounting leaves resources unconsumed and cannot be optimal.