Quantitative Methods and Economic Optimization

Introduction

Managerial economics converts economic reasoning into decisions about price, output, capacity, advertising, inventory, labour, investment and resource allocation. In practice, managers rarely choose between purely qualitative alternatives. They compare measurable consequences under constraints: revenue versus cost, growth versus risk, service quality versus capacity, and short-term returns versus long-term capability.

Quantitative methods provide the language for these comparisons. Functions describe relationships among variables; graphs reveal direction and curvature; elasticity measures responsiveness; optimization identifies the best feasible choice; linear programming allocates scarce resources; sensitivity analysis tests whether a recommendation remains valid when assumptions change; and spreadsheet models make the analysis operational.

A model is not reality. It is a deliberately simplified representation of reality. Its usefulness depends on logical consistency, appropriate data, transparent assumptions and careful interpretation. The central managerial discipline is therefore not merely finding a numerical optimum, but determining whether the optimum is economically meaningful and robust.

Learning Objectives

By the end of this chapter, you should be able to:

1. Express managerial relationships using functions, tables and graphs.
2. Interpret slopes, intercepts, curvature and marginal relationships.
3. Calculate and apply price, income and cross-price elasticity.
4. Distinguish unconstrained from constrained optimization.
5. Use first- and second-order conditions to evaluate optima.
6. Formulate and interpret linear programming models and shadow prices.
7. Conduct sensitivity analysis and identify binding constraints.
8. Build a transparent spreadsheet-based decision model.
9. Evaluate assumptions, uncertainty, ethics and limitations in quantitative recommendations.

1. Functions and Graphical Economic Reasoning

A function maps an input to an output. In managerial economics, the output may be demand, cost, profit or productivity. For example, a linear demand function can be written as:

Here, Q is quantity demanded, P is price, a is the quantity intercept and b is the absolute change in quantity associated with a one-unit change in price. The usual assumption is that other determinants of demand—income, preferences, prices of related goods and expectations—are held constant. This is the ceteris paribus condition.

A cost function may be represented by:

where C(Q) is total cost, F is fixed cost, v is constant variable cost per unit and Q is output. This is a short-run linear approximation. It does not imply that costs remain linear at all production levels.

Revenue and profit are defined as:

where R is total revenue, P(Q) is price as a function of output, C(Q) is total cost and π is profit. Units must be consistent: if price is measured in dollars per unit and output in units, revenue is measured in dollars.

Slopes and marginal concepts

For a differentiable function y = f(x), its instantaneous slope is:

The derivative measures the change in y caused by a very small change in x. In economics, marginal cost is MC = dC/dQ, marginal revenue is MR = dR/dQ, and marginal product of labour is MP_L = dQ/dL. Marginal quantities are rates, so their units must be stated. For example, marginal cost is currency per additional unit of output.

The second derivative measures curvature:

A positive second derivative indicates increasing slope or convexity; a negative second derivative indicates decreasing slope or concavity. Curvature is important because optimization requires distinguishing maxima from minima.

Economic interpretation of graphs

A graph should be treated as an argument, not decoration. The horizontal axis normally displays the choice variable, while the vertical axis displays the outcome or objective. A downward-sloping demand curve indicates that, under stated conditions, higher price is associated with lower quantity demanded. It does not prove that every consumer behaves identically or that the relationship is causal without further evidence.

The intercepts show values implied by the model when the other variable equals zero. They may be economically irrelevant outside the observed range. Managers should therefore avoid extrapolating a fitted function far beyond the data used to estimate it.

2. Elasticity and Responsiveness

A slope is measured in physical units, whereas elasticity is unit-free. Elasticity measures the percentage response of one variable to a percentage change in another.

Point elasticity

The point elasticity of demand with respect to price is:

where dQ/dP is the marginal response of quantity to price. For a standard demand curve, ε_P is negative. Managers often discuss its absolute value, |ε_P|. Demand is elastic when |ε_P| > 1, unit elastic when |ε_P| = 1, and inelastic when |ε_P| < 1.

For the demand function Q = 100 − 2P, at P = 20, quantity is Q = 60 and dQ/dP = −2. Therefore:

Demand is inelastic at this point. A small percentage increase in price is associated with a smaller percentage decrease in quantity, within the model.

Arc elasticity

When changes are relatively large, arc elasticity avoids dependence on the initial observation:

This measure is useful for comparing two observed points. It should not be confused with a causal estimate: observed changes may also reflect advertising, competitor actions or seasonality.

Revenue and elasticity

Total revenue is TR = P × Q. Along a downward-sloping demand curve, an increase in price raises revenue when demand is inelastic, lowers revenue when demand is elastic, and leaves revenue locally unchanged at unit elasticity. This rule assumes other determinants remain constant and ignores costs, competitive response and customer lifetime value.

Other elasticities

Income elasticity is:

where Y is income. A positive value generally indicates a normal good; a negative value indicates an inferior good, subject to the context and data.

Cross-price elasticity is:

A positive value commonly indicates substitutes, while a negative value commonly indicates complements. These interpretations require stable definitions and careful empirical analysis.

3. Optimization Without and With Constraints

Optimization asks which feasible choice maximizes or minimizes an objective. The objective may be profit, contribution, output, customer coverage or cost. The decision variable may be price, production quantity, labour, advertising or a product mix.

Unconstrained optimization

Suppose a manager chooses x to maximize f(x). An interior candidate optimum satisfies the first-order condition:

The second-order condition for a strict local maximum is:

For a strict local minimum, f″(x*) > 0. These conditions are sufficient for local classification under differentiability, but they do not automatically establish a global optimum. A global conclusion requires examining the feasible domain, boundary points and, where relevant, concavity or convexity.

Worked example: profit maximization

Let inverse demand be P = 100 − 2Q and total cost be C(Q) = 20Q + 100. Then:

The first derivative is:

Setting it equal to zero gives Q* = 20. The second derivative is d²π/dQ² = −4, so the profit function is strictly concave and this is a unique global maximum over the nonnegative feasible range. Price is P* = 100 − 2(20) = 60. Profit is:

The marginal condition provides the same result: MR = 100 − 4Q and MC = 20; setting MR = MC gives Q = 20. The condition is valid here because the firm faces the specified inverse demand and differentiable cost function, and the resulting quantity is feasible.

Boundary solutions

A first-order condition may produce an infeasible value. If a production decision must satisfy Q ≥ 0 and the objective is decreasing at zero, the optimum may be Q = 0. Similarly, capacity, legal limits or minimum service requirements may force a boundary solution. Always compare interior candidates with feasible boundaries.

Constrained optimization and the Lagrangian

Consider maximizing f(x,y) subject to g(x,y) = b. The Lagrangian is:

The first-order conditions are:

The multiplier λ is the shadow value of relaxing the constraint, under regularity conditions. It measures the approximate change in the optimized objective from a one-unit increase in b, while other parameters remain fixed. Its units are objective units per unit of the constraint.

For an interior consumer optimum, utility maximization subject to a budget constraint produces the tangency condition:

where MU_X and MU_Y are marginal utilities and P_X, P_Y are prices. The condition assumes an interior solution, differentiability and locally nonsatiated preferences; corner solutions require separate analysis.

4. Linear Programming for Resource Allocation

Linear programming (LP) is appropriate when the objective and constraints can be represented as linear relationships. A general maximization model is:

subject to:

Here, x_j is the level of activity or product j, c_j is its contribution per unit, a_ij is the amount of resource i required per unit of activity j, b_i is available resource i, and Z is total contribution. The coefficients must use compatible units.

Worked LP example

A factory produces products A and B. Contribution is $40 per unit of A and $30 per unit of B. Each A requires 2 machine hours and 1 labour hour; each B requires 1 machine hour and 1.5 labour hours. Available capacity is 100 machine hours and 90 labour hours. Let x and y denote units of A and B.

subject to:

The machine constraint intersects the axes at (50, 0) and (0, 100); the labour constraint intersects at (90, 0) and (0, 60). Their intersection solves:

From the first equation, y = 100 − 2x. Substitution gives x + 1.5(100 − 2x) = 90, hence x = 30 and y = 40. Contribution is Z = 40(30) + 30(40) = 2,400. The axis corner (50, 0) yields $2,000, and (0, 60) yields $1,800. Thus (30, 40) is optimal among the feasible corner points. Both resources bind at the solution.

The corner-point result follows from the fundamental structure of an LP: when an optimum exists, at least one optimal solution occurs at an extreme point of the feasible region. In larger models, the simplex method or spreadsheet Solver identifies the relevant corner computationally.

Shadow prices and sensitivity

A shadow price estimates the improvement in the objective from one additional unit of a binding resource, within the allowable range over which the current optimal basis remains valid. A nonbinding resource generally has zero marginal value in the current model because extra availability does not improve the objective immediately.

Shadow prices are not market prices. They are conditional values generated by a model. They can change when coefficients, constraints or the active set changes. Managers should not purchase additional capacity at a price above the relevant shadow value without considering nonlinear effects, fixed costs and uncertainty.

5. Sensitivity Analysis and Robust Decisions

Sensitivity analysis examines how results change when assumptions change. It should be conducted for:

• demand and price estimates;
• variable and fixed costs;
• capacity and labour availability;
• contribution margins;
• discount rates and exchange rates;
• forecast errors and implementation delays.

One-at-a-time analysis changes one parameter while holding others constant. It is easy to communicate but may understate risk when variables move together. Scenario analysis changes several coherent assumptions, such as a base, adverse and favourable market environment. Break-even analysis identifies the threshold at which a decision changes.

For a single product with selling price P, variable cost per unit V and fixed cost F, the accounting break-even quantity is:

The denominator is unit contribution. This formula assumes constant price and variable cost, one relevant product, and no capacity or demand limitation. It is not a complete investment-risk measure.

A useful managerial practice is to report a recommendation together with a range, the key drivers and the conditions under which the recommendation reverses. Robust optimization asks not only “What is best under the forecast?” but also “What performs acceptably across plausible forecasts?”

6. Spreadsheet-Based Economic Models

A spreadsheet is an implementation environment, not a substitute for economic reasoning. A well-designed model separates:

1. Inputs: prices, costs, capacities, probabilities and assumptions.
2. Calculations: revenue, cost, constraints, elasticities and objective values.
3. Outputs: recommended decisions, sensitivity tables, charts and warnings.

Best practices include using named input cells, consistent units, visible formulas, data validation, protected calculation cells, scenario controls and an assumptions sheet. Avoid hard-coding numbers inside formulas, circular references, hidden rows and unexplained overrides. Include checks such as “resource used ≤ resource available” and “quantity ≥ 0.”

Spreadsheet Solver can optimize a target cell by changing decision cells subject to constraints. The model must identify whether the objective is to maximize or minimize, whether variables are continuous or integer, and whether constraints are linear. An advertising or staffing decision may require integer restrictions; a linear Solver model may otherwise recommend fractional employees.

Spreadsheet outputs should be independently tested. Recalculate a small instance by hand, inspect boundary solutions, change inputs dramatically, and compare Solver’s result with economic logic. Version control, access permissions and audit trails are important where models influence pricing, credit, employment or public-resource decisions.

7. Illustrative Case: FreshBite Meal Kits

Background

FreshBite, an illustrative meal-kit company, must allocate weekly kitchen capacity between vegetarian kits (V) and family kits (F). The contribution margins are $12 and $18 respectively. Each vegetarian kit uses 1 packing hour and 2 ingredient units; each family kit uses 2 packing hours and 3 ingredient units. Weekly capacity is 1,000 packing hours and 1,500 ingredient units. Demand is limited to 600 vegetarian and 500 family kits.

Problem formulation

subject to:

The first two constraints represent resources; the next two represent market demand. The formulation assumes constant contribution margins, divisible output for planning purposes, known capacities and no interaction between products beyond the listed resources.

Analysis and approach

The ingredient constraint is comparatively restrictive. If both resource constraints bind, solving V + 2F = 1000 and 2V + 3F = 1500 gives F = 500 and V = 0. This point satisfies demand limits and generates contribution of $9,000. However, the point (600, 200) uses 1,000 packing hours and 1,800 ingredient units, so it is infeasible. The demand corner (0, 500) is feasible and also produces $9,000, while the ingredient-only combination (600, 100) produces $8,? Wait: 12(600)+18(100) = $9,000. Thus multiple optimal solutions may exist along the relevant boundary, depending on the contribution-resource trade-off.

Indeed, contribution per ingredient unit is $6 for V and $6 for F. Ingredient capacity therefore generates the same contribution for any mix using 1,500 ingredient units, provided packing capacity and demand limits are respected. Feasible optimal combinations include (600, 100) and (0, 500), and mixtures between them that use the ingredient limit while satisfying packing capacity. This is a valuable managerial insight: the model does not identify one uniquely superior product mix because the contribution rates per scarce ingredient unit are equal.

Outcome and managerial interpretation

FreshBite should not treat Solver’s single displayed solution as the only answer. Management can choose among alternative optimal mixes using qualitative criteria such as customer retention, operational complexity, nutritional commitments or strategic growth. If family-kit demand rises, or if ingredient requirements change, the optimal set may change. A sensitivity analysis should test ingredient availability, contribution margins and the reliability of demand caps.

Learning points

The case illustrates four principles. First, formulation quality precedes computation. Second, binding constraints determine economic value. Third, multiple optima require managerial judgment beyond the objective function. Fourth, a mathematically optimal plan can be operationally poor if it ignores uncertainty, perishability, labour skills or customer relationships.

8. Assumptions, Limitations and Governance

Quantitative models commonly assume stable relationships, accurate measurements, rational responses, known constraints and independence among variables. These assumptions may fail. Demand may be endogenous to price; competitors may react; costs may be nonlinear; capacity may involve setup times; and data may reflect historical discrimination.

Correlation should not automatically be interpreted as causation. An elasticity estimated from historical price variation may be biased if firms raise prices when demand is expected to strengthen. Managers should distinguish accounting contribution from economic profit, and forecast precision from decision usefulness.

Ethical governance is essential when optimization affects people. Pricing models should be tested for discriminatory outcomes; staffing models should not use productivity proxies that unfairly penalize protected groups; credit models should document variables and provide review mechanisms; and public-resource allocation should make trade-offs transparent. Privacy, data minimization, explainability and human oversight are especially important when machine-learning systems estimate demand or recommend decisions.

Contemporary decision systems increasingly combine spreadsheets, cloud analytics, optimization software and AI forecasting. AI can improve prediction, but forecast accuracy does not guarantee optimization quality. A biased forecast, an omitted constraint or an objective that rewards the wrong behaviour can produce a precise but harmful recommendation. Human review, stress testing and post-decision monitoring remain necessary.

Chapter Summary

Functions represent relationships among managerial variables and provide the basis for revenue, cost and profit analysis. Derivatives measure marginal change, while second derivatives describe curvature and assist in classifying optima. Elasticity expresses percentage responsiveness and supports pricing, revenue and market analysis.

Unconstrained optimization uses first- and second-order conditions, while constrained optimization incorporates feasibility and shadow values. Linear programming allocates scarce resources when objectives and constraints are linear; its corner solutions, binding constraints and shadow prices provide practical managerial insight. Sensitivity and scenario analysis reveal how fragile a recommendation is. Spreadsheet models make analysis accessible but require disciplined structure, validation and governance.

The strongest managerial recommendation combines mathematical optimality with economic interpretation, operational feasibility, uncertainty analysis and ethical judgment.

Key Terms

Function: A rule that maps input values to output values.

Ceteris paribus: The assumption that other relevant conditions remain constant.

Derivative: The instantaneous rate of change of one variable with respect to another.

Marginal cost: The change in total cost from an additional unit of output.

Marginal revenue: The change in total revenue from an additional unit sold.

Elasticity: A unit-free percentage measure of responsiveness.

Point elasticity: Elasticity evaluated at a specific point.

Arc elasticity: Elasticity calculated between two observations.

Profit function: Revenue minus total cost.

First-order condition: A zero-derivative condition identifying an interior candidate optimum.

Second-order condition: A curvature condition used to classify a candidate optimum.

Constraint: A restriction on feasible choices.

Lagrangian: An expression combining an objective and constraint with a multiplier.

Shadow price: The marginal value of relaxing a constraint.

Linear programming: Optimization with linear objective and constraint relationships.

Binding constraint: A constraint satisfied as an equality at the solution.

Sensitivity analysis: Examination of how outputs change when assumptions vary.

Scenario analysis: Evaluation under coherent alternative sets of assumptions.

Break-even quantity: Output at which total revenue equals total cost under specified assumptions.

Solver: Spreadsheet or software functionality that searches for an optimal feasible decision.

Review Questions

Short Answer Questions

1. What is the economic meaning of a derivative?
2. Distinguish between slope and elasticity.
3. State the first- and second-order conditions for an interior local maximum.
4. What does a shadow price measure in a constrained optimization model?
5. Why should spreadsheet models separate inputs, calculations and outputs?

Long Answer Questions

1. Explain how a manager can use price elasticity to evaluate a proposed price increase, including the assumptions and limitations of the analysis.
2. Derive the profit-maximizing output condition MR = MC and explain when it may not identify the global optimum.
3. Formulate a linear programming model for allocating two products across two scarce resources and interpret its binding constraints.
4. Compare one-at-a-time sensitivity analysis, scenario analysis and break-even analysis.
5. Discuss the ethical and governance risks of using optimization and AI-based forecasts in managerial decisions.

Analytical / Application Questions

1. A firm faces inverse demand P = 120 − 3Q and total cost C(Q) = 20Q + 60. Determine the profit-maximizing output, price and profit, and verify the second-order condition.
2. A company has 80 machine hours and 60 labour hours. Product X contributes $25 and uses 2 machine hours and 1 labour hour; product Y contributes $20 and uses 1 machine hour and 2 labour hours. Formulate and solve the linear programming problem, then identify which resources bind.
3. A spreadsheet recommends a price increase based on an elasticity estimated from historical data. Design a sensitivity and governance review that tests demand uncertainty, competitor response, data bias and the conditions under which the recommendation should be rejected.

Multiple Choice Questions with Answers

Question 1

For the demand function Q = 100 − 2P, what is quantity demanded when P = 20?

A. 40

B. 60

C. 80

D. 140

Correct Answer: B. 60

Question 2

If price elasticity of demand has absolute value less than 1, demand is described as:

A. Perfectly elastic

B. Unit elastic

C. Inelastic

D. Perfectly inelastic

Correct Answer: C. Inelastic

Question 3

For an interior local maximum of a differentiable function, which condition is required?

A. First derivative equals zero and second derivative is negative

B. First derivative is positive and second derivative is zero

C. First derivative is negative and second derivative is positive

D. First derivative equals one and second derivative is negative

Correct Answer: A. First derivative equals zero and second derivative is negative

Question 4

In the profit example with inverse demand P = 100 − 2Q and marginal cost $20, the profit-maximizing output is:

A. 10

B. 20

C. 40

D. 60

Correct Answer: B. 20

Question 5

In a linear programming model, a constraint is binding when it:

A. Has a negative coefficient

B. Is omitted from the objective

C. Is satisfied as an equality at the solution

D. Always has a zero shadow price

Correct Answer: C. Is satisfied as an equality at the solution

Question 6

For Q_BE = F/(P − V), the denominator represents:

A. Fixed cost

B. Total revenue

C. Unit contribution

D. Average fixed cost

Correct Answer: C. Unit contribution

Question 7

A nonbinding resource constraint generally has what immediate marginal value in a linear program?

A. A positive infinite value

B. Zero shadow value

C. The same value as total profit

D. A value equal to fixed cost

Correct Answer: B. Zero shadow value

Question 8

Which spreadsheet design practice most improves auditability?

A. Hard-code assumptions inside formulas

B. Hide all calculation cells

C. Separate inputs from calculations and outputs

D. Remove feasibility checks

Correct Answer: C. Separate inputs from calculations and outputs

Question 9

Cross-price elasticity is commonly positive when two goods are:

A. Complements

B. Substitutes

C. Inferior goods

D. Public goods

Correct Answer: B. Substitutes

Question 10

A shadow price should be interpreted as:

A. The observed market price of every resource

B. The approximate objective improvement from relaxing a constraint within a valid range

C. The fixed cost per unit of output

D. The elasticity of demand

Correct Answer: B. The approximate objective improvement from relaxing a constraint within a valid range