When Spending More Does Not Improve the Choice
Consumer choice is often introduced with a budget line and an optimum on that line. That result depends on an important feature of preferences: more of at least one good must improve the consumer’s well-being whenever affordable income remains. A consumer who has already reached a satiation point is different. If the preferred bundle is affordable, spending additional income on goods can leave utility unchanged or even reduce it.
Let \(x=(x_1,\ldots,x_n)\) be a bundle of \(n\) goods, let \(p_j>0\) be the price of good \(j\), and let \(m\) be income. The consumer chooses a nonnegative bundle subject to the budget constraint \(\sum_{j=1}^n p_jx_j\leq m\). The inequality allows income to remain unspent.
If a satiation bundle is affordable, meaning \(\sum_{j=1}^n p_jx_j^s\leq m\), it is a utility-maximizing choice. The consumer does not need to spend the remaining income to reach the highest attainable utility. If there are several bundles that all yield maximum utility, some may use more income than others; what matters is that spending more does not make the consumer strictly better off.
An Affordable Satiation Bundle
A common way to represent a single ideal bundle is a utility function that falls as consumption moves away from a target. The following example uses a strictly concave quadratic utility function, so its unique global maximum is easy to identify.
Worked Example: An Affordable Ideal Bundle
A consumer chooses quantities \(x\geq0\) and \(y\geq0\), with income \(m=\$100\). Prices are \(p_x=\$10\) and \(p_y=\$5\). Preferences are represented by \(U(x,y)=-(x-4)^2-(y-6)^2\). Find the optimum, the amount spent, and the income left unspent.
Method. First find the unconstrained utility maximum, then check its affordability. The negative squared terms are each at most zero, so \(U(x,y)\leq0\) for every bundle. Equality holds only when \(x=4\) and \(y=6\). Thus \((4,6)\) is the unique satiation bundle.
Affordability check. Its cost is \(10(4)+5(6)=40+30=\$70\). Because \(\$70\leq\$100\), the bundle is feasible and achieves the greatest possible utility.
Unspent income. The consumer spends \(\$70\), so the amount left unspent is \(\$100-\$70=\$30\). At the chosen bundle, utility is \(-(4-4)^2-(6-6)^2=0\), the highest attainable value.
Check. Any feasible alternative has utility at most zero, while the chosen bundle attains zero. Therefore it is optimal, and its budget constraint is nonbinding. The method avoids assuming that an optimum must lie on the budget line.
This example also shows why the budget line alone cannot identify the optimum. A point on that line requires all \(\$100\) to be spent, but the consumer’s utility is already maximized at a bundle costing \(\$70\). Spending the remaining \(\$30\) would move the consumer away from the ideal bundle.
Why Strictly Increasing Preferences Use the Full Budget
The familiar full-budget conclusion follows from a simple feasibility argument. It does not require a particular utility function or the use of a tangency condition.
Justification. Suppose instead that a candidate bundle leaves \(s=m-\sum_{j=1}^n p_jx_j>0\) dollars unspent. Choose any good \(j\), and increase its quantity by \(\Delta=s/(2p_j)>0\). This extra amount costs \(p_j\Delta=s/2\), so the new bundle is still affordable. Strictly increasing preferences make the consumer strictly better off after the increase. The original bundle therefore cannot have been optimal. This contradicts the assumption that it was an optimum, so an optimum under these conditions must use the full budget.
The assumptions are essential. If preferences are satiated, an increase in a good may not improve utility; beyond a satiation point it may lower utility. If some good has zero price, spending the budget is not needed to add that good. The full-budget result therefore applies to the stated model with positive prices and strictly increasing preferences, not automatically to every consumer-choice problem.
Worked Example: Comparing Strictly Increasing and Satiated Preferences
A consumer has income \(m=\$30\), with \(p_x=\$4\) and \(p_y=\$2\). Consider first the bundle \((x,y)=(3,5)\) under utility \(U(x,y)=x+y\). Is it optimal? Then compare with the same prices and income under \(U(x,y)=-(x-3)^2-(y-2)^2\).
Strictly increasing case. The bundle costs \(4(3)+2(5)=12+10=\$22\), leaving \(\$8\). Since \(U=x+y\) strictly increases in either good, the bundle cannot be optimal. For example, spending the remaining \(\$8\) on good \(x\) increases \(x\) by \(\$8/\$4=2\) units. The new bundle is \((5,5)\), costs \(4(5)+2(5)=\$30\), and has utility \(5+5=10\), compared with \(3+5=8\). The original bundle is strictly worse than an affordable alternative.
Satiated case. The quadratic utility is maximized at \((3,2)\), where both squared terms equal zero. This bundle costs \(4(3)+2(2)=12+4=\$16\), so it is affordable and leaves \(\$30-\$16=\$14\) unspent. Its utility is zero, and no other bundle can exceed zero because each squared term is nonnegative.
Check. Under \(U=x+y\), the proposed bundle is not optimal because the affordable bundle \((5,5)\) gives greater utility. Under the quadratic utility, \((3,2)\) is the unique optimum and leaves \(\$14\) unspent. The preferences, not the income and prices alone, determine whether the budget must bind.
When the Ideal Bundle Is Not Affordable
If the satiation bundle costs more than income, the consumer cannot attain it. The optimum must instead be the feasible bundle that gives the greatest utility. In some problems the budget then binds, but this is a consequence of the problem’s preferences and constraints, not a universal rule for satiated consumers.
For a quadratic utility function of the form \(U(x,y)=-(x-a)^2-(y-b)^2\), maximizing utility is equivalent to minimizing the squared distance \((x-a)^2+(y-b)^2\) from the ideal point \((a,b)\). If that ideal point is unaffordable, the constrained optimum is the feasible bundle closest to it. The next example computes this choice and verifies the result.
Worked Example: An Unaffordable Satiation Bundle
A consumer has \(m=\$10\), prices \(p_x=\$3\) and \(p_y=\$2\), and utility \(U(x,y)=-(x-4)^2-(y-2)^2\). Find the utility-maximizing bundle for \(x,y\geq0\), and determine whether the budget binds.
Method and ideal-bundle check. The utility maximum without the budget restriction is \((4,2)\). Its cost is \(3(4)+2(2)=\$16\), which exceeds the available \(\$10\). The ideal bundle is infeasible. Since the utility function is the negative of squared distance from \((4,2)\), we minimize \((x-4)^2+(y-2)^2\) over the affordable set.
The ideal point lies above the budget line \(3x+2y=10\). The closest feasible point must be on that line: a point strictly inside the budget set could be moved a short distance toward \((4,2)\), reducing the squared distance while remaining feasible. To find the closest point on the line, minimize \((x-4)^2+(y-2)^2\) subject to \(3x+2y=10\). The first-order conditions for this strictly convex minimization are \(x-4+3\lambda=0\) and \(y-2+2\lambda=0\), where \(\lambda\) is the multiplier. Thus \(x=4-3\lambda\) and \(y=2-2\lambda\).
Substitute these expressions into the budget equation: \(3(4-3\lambda)+2(2-2\lambda)=10\). Expanding gives \(12-9\lambda+4-4\lambda=10\), so \(16-13\lambda=10\), and therefore \(\lambda=6/13\). The resulting quantities are \(x=4-18/13=34/13\) and \(y=2-12/13=14/13\).
Verification. Both quantities are positive, and expenditure is \(3(34/13)+2(14/13)=102/13+28/13=130/13=\$10\). Utility is \(-((34/13-4)^2+(14/13-2)^2)=-((-18/13)^2+(-12/13)^2)=-(324+144)/169=-468/169\). The squared-distance objective is strictly convex, so the computed point is its unique constrained minimum and hence the unique utility maximum. The budget binds in this case.
Common mistake avoided. The unconstrained ideal bundle costs \(\$16\), so reporting \((4,2)\) as the choice ignores the budget. The correct choice is the feasible point closest to that ideal bundle, \((34/13,14/13)\), which uses all \(\$10\).
Satiation Can Produce More Than One Optimum
Not every satiated preference has one unique ideal bundle. Utility can reach a maximum once quantities pass particular thresholds and then remain constant. In that case the set of utility-maximizing bundles may contain many points. If the least-cost bundle that reaches all thresholds is affordable, it is one optimum with slack; other affordable bundles may also be optimal.
Worked Example: A Plateau of Maximum Utility
A consumer has \(m=\$23\), prices \(p_x=\$2\) and \(p_y=\$3\), and utility \(U(x,y)=\min\{x,4\}+\min\{y,3\}\). Find the maximum utility, a least-cost maximizing bundle, and another maximizing bundle that also leaves income unspent.
Maximum utility. The first term cannot exceed \(4\), and the second cannot exceed \(3\). Thus \(U(x,y)\leq4+3=7\). Any bundle with \(x\geq4\) and \(y\geq3\) attains utility \(7\).
Least-cost maximizing bundle. To reach both thresholds at the lowest cost, choose \(x=4\) and \(y=3\). Its cost is \(2(4)+3(3)=8+9=\$17\). Since \(\$17\leq\$23\), it is feasible and leaves \(\$23-\$17=\$6\) unspent. It attains utility \(\min\{4,4\}+\min\{3,3\}=4+3=7\), so it is optimal.
Another optimum. The bundle \((5,3)\) costs \(2(5)+3(3)=\$19\), leaving \(\$4\). Its utility is still \(\min\{5,4\}+\min\{3,3\}=4+3=7\). Both bundles maximize utility even though they cost different amounts. Spending additional income on \(x\) after reaching \(x=4\) does not improve utility.
Check. No bundle can exceed utility \(7\), and both displayed bundles attain \(7\). The choice is not unique, and the budget is nonbinding at both examples of an optimum.
Common Mistakes and Practical Checks
- Assuming every optimum spends all income. That conclusion needs strictly increasing preferences and positive prices. Check the preference assumptions before imposing budget equality.
- Treating an unaffordable satiation bundle as feasible. Calculate its cost first. If its cost exceeds income, solve the constrained problem rather than reporting the ideal point.
- Assuming extra spending must raise utility. With satiation, quantities beyond a threshold may add no utility; with preferences that decline away from an ideal point, extra consumption can reduce utility.
- Confusing a nonbinding constraint with an incorrect solution. A bundle with slack can be optimal if it already attains the highest available utility. Verify optimality by comparing utility, not by checking whether all income was spent.
- Assuming the optimum is always unique. A utility plateau can make many bundles equally good. State whether the problem identifies a unique best bundle or a set of best bundles.
Check Your Understanding
For each problem, identify whether the budget binds and use the stated preferences and prices.
- A consumer has \(m=\$30\), \(p_x=\$3\), and \(p_y=\$2\), with \(U(x,y)=-(x-2)^2-(y-5)^2\). Find the optimum and unspent income. (Answer: \((2,5)\); expenditure \(\$16\); \(\$14\) unspent.)
- A consumer has \(m=\$12\), \(p_x=\$2\), \(p_y=\$4\), and \(U(x,y)=-(x-3)^2-(y-4)^2\). Find the constrained optimum. (Answer: \((2,2)\); the budget binds at \(\$12\).)
- Let \(U(x,y)=\min\{x,5\}+2\min\{y,2\}\), with \(m=\$25\), \(p_x=\$2\), and \(p_y=\$3\). Give the least-cost bundle attaining maximum utility and the unspent income. (Answer: \((5,2)\); maximum utility \(9\); \(\$9\) unspent.)
- At \((x,y)=(3,5)\), a consumer has \(m=\$30\), \(p_x=\$4\), \(p_y=\$2\), and \(U(x,y)=x+y\). Is this an optimum? Give an affordable improvement. (Answer: No; \((5,5)\) costs \(\$30\) and raises utility from \(8\) to \(10\).)
- A three-good consumer has ideal bundle \((1,2,3)\), utility \(U(x_1,x_2,x_3)=-\sum_{j=1}^3(x_j-x_j^s)^2\), prices \((\$5,\$2,\$1)\), and income \(\$20\), where \(x^s=(1,2,3)\). Find the optimum and unspent income. (Answer: \((1,2,3)\); expenditure \(\$12\); \(\$8\) unspent.)