Optimization in Economics

Where the Derivative Says Stop

Economics runs on optimization: firms maximize profit, consumers maximize utility, planners minimize cost. Calculus turns 'find the best' into a two-step ritual.

First-order condition (FOC) — at an interior optimum, the derivative vanishes:

$$f'(x^*) = 0$$

At the top of a hill, the ground is momentarily flat. If $f'(x) > 0$ you gain by moving right; if $f'(x) < 0$, by moving left. Only where the derivative is zero is there nothing left to grab.

Second-order condition (SOC) — flat ground could be a peak, a valley, or a saddle-flat shelf:

$$f''(x^*) < 0: \text{maximum} \qquad f''(x^*) > 0: \text{minimum}$$

The economic translation is marginal analysis. Profit $\pi(q) = R(q) - C(q)$ peaks where

$$MR = MC$$

marginal revenue equals marginal cost. While one more unit brings in more than it costs, produce it; the optimum is exactly where the last unit breaks even at the margin. The FOC isn't a math trick — it is the formal version of 'expand until the next step stops paying.'

Discipline the ritual enforces:

• Check the SOC: $MR = MC$ also holds at profit minima.
• Check boundaries: if the feasible range is $[0, capacity]$, the best point may be a corner where no derivative vanishes.
• Compare candidate points by their actual values — the derivative finds candidates, the objective crowns the winner.

Marginal reasoning is the single most transferable idea in this course: hire until the marginal worker pays for herself, advertise until the marginal euro returns a euro, and stop everywhere the margin hits zero.

The optimization ritual

Common pitfall: Stopping at the FOC. $MR = MC$ also holds at profit minima — without the second-order check and a boundary scan, the "optimum" you report may be the worst point on the curve.

The Margin Is a Crossing, Not a Feeling

Here is the picture professional economists carry: two curves — marginal revenue sloping down, marginal cost sloping up — and the optimum living at their crossing.

$$q^*: \quad MR(q^*) = MC(q^*)$$

Why the crossing, seen three ways:

• Left of $q^*$: $MR > MC$ — the next unit adds more revenue than cost. The gap between the curves is pure profit being left on the table by stopping here.
• Right of $q^*$: $MR < MC$ — each extra unit destroys value. The gap is now a loss per unit of overreach.
• At $q^*$: the gaps vanish. Total profit equals the accumulated area between the curves up to the crossing — and that area is as large as it will ever be.

$$\pi(q^*) = \int_0^{q^*}\big(MR - MC\big)\,dq$$

The classic error this picture kills: 'we're profitable, so produce more.' Total profit being positive says nothing about the margin. Past $q^*$ you remain profitable for a while — while shrinking. Firms don't feel the peak; totals still look good on the way down. Only the margin knows.

Reading real decisions with it:

• A consultant works overtime until the marginal hour's fee no longer beats its (rising) personal cost.
• A retailer adds opening hours until the marginal hour's contribution hits its staffing cost.
• An ad budget grows until the marginal euro of spend returns exactly one euro of margin.

Every one of these is the same crossing with different axis labels. When you can see any business question as two curves hunting for their intersection, you are doing economics rather than arithmetic — and the flat-derivative ritual of e01 becomes something you could re-derive on a napkin.

Tip: Read the two curves as a conversation: left of the crossing, every unit is money left on the table; right of it, every unit is self-inflicted loss. Total profit is the accumulated area between the curves — largest exactly at the crossing.

Common pitfall: Maximizing revenue instead of profit. Revenue peaks where $MR = 0$ — well past the profit optimum at $MR = MC$. The units between those two quantities all sell, and all lose money.