EC 270 - 2026-09-29 - Lecture 06

EC 270 - Lecture 06 - 2026-09-29

Intermediate Microeconomics I · Logan McLeod · LH 1011

Week 4 · 11:30-12:38 · 68 min · Full transcript · Flashcards

Overview

This lecture carried on with Chapter 5: Choice. McLeod first solved the Cobb-Douglas consumer's problem with the Lagrange method. It gives the same demands as direct substitution from last Thursday, and the demands come out as formulas rather than numbers. The class then looked at interior solutions where the indifference curves have kinks (complementary preferences, which only direct substitution can solve). It finished with corner solutions, where the consumer buys none of one good: perfect substitutes and the more general substitutable preferences are solved by comparing the slope of the indifference curve with the slope of the budget line, case by case.

The Lagrange method

Last Thursday the Cobb-Douglas demands were found by direct substitution. Today the same problem, maxx1,x2x1αx2β such that p1x1+p2x2=m, was solved the second way.

Setting up the Lagrangian

The Lagrangian (called L in class, because "Lagrangian's too long of a word to keep saying") is a new function: the objective function, minus a new variable λ times the constraint. λ is called the Lagrange multiplier. For the consumer, the objective is the utility function and the constraint is the budget constraint.

Why: Adding λ gives a third thing to differentiate with respect to. That third first-order condition hands back the budget constraint, so one set of conditions covers both "most preferred" and "affordable".

L=u(x1,x2)−λ(p1x1+p2x2−m)

Here u is the utility function, λ is the Lagrange multiplier, p1 and p2 are the prices, and m is income. The m is moved to the left so the bracket equals zero. "I'm not changing the constraint at all", but it has to be written this way to go into the Lagrangian.

The three first-order conditions

The optimal bundle sets the partial derivative of L with respect to each of x1, x2 and λ equal to zero. These are the first-order conditions (FOCs).

Why: L is the objective minus the constraint, and the two parts are separate, so you can differentiate each part on its own. The third FOC just returns the budget constraint. Utility has no λ in it, so its derivative is 0, and the derivative of −λ(…) with respect to λ is −1 times the bracket.

∂L∂x1=∂u∂x1−λp1=0∂L∂x2=∂u∂x2−λp2=0∂L∂λ=−(p1x1+p2x2−m)=0

Here ∂u/∂xi is the marginal utility of good i. In the budget part, the terms without the variable you are differentiating with respect to drop to 0.

Cobb-Douglas demands by Lagrange

Solving the Cobb-Douglas problem with the Lagrange method gives the ordinary demands below.

Why: Each demand spends a fixed share of income on each good: p1x1∗=αα+βm goes on good 1 and p2x2∗=βα+βm goes on good 2.

x1∗=αα+βmp1,x2∗=βα+βmp2

Here α and β are the exponents in the utility function, m is income, and p1 and p2 are the prices.

Example: Cobb-Douglas demands by the Lagrange method (done on the board)

Maximise u=x1αx2β subject to p1x1+p2x2=m.

  1. L=x1αx2β−λ(p1x1+p2x2−m) -- write the Lagrangian
  2. ∂L∂x1=αx1α−1x2β−λp1=0 -- first FOC: the power rule on x1, and the budget part gives λp1
  3. ∂L∂x2=βx1αx2β−1−λp2=0 -- second FOC, the same idea for x2
  4. ∂L∂λ=−(p1x1+p2x2−m)=0⇒p1x1+p2x2=m -- third FOC, which is just the budget constraint
  5. λ=αx1α−1x2βp1 and λ=βx1αx2β−1p2 -- solve each of the first two FOCs for λ (board equations 1 and 2)
  6. αx1α−1x2βp1=βx1αx2β−1p2 -- set them equal to get rid of λ
  7. αβ⋅x1α−1x2βx1αx2β−1=p1p2 -- move the prices to the right and everything else to the left
  8. αx2βx1=p1p2 -- the powers cancel down to x1−1 and x21 (board equation 4)
  9. x2∗=βαp1p2x1∗ -- solve for x2 in terms of x1 (board equation 5)
  10. p1x1∗+p2βαp1p2x1∗=m -- substitute equation 5 into the budget constraint
  11. p1x1∗(1+βα)=m -- cancel the p2's and factor out p1x1∗
  12. p1x1∗α+βα=m -- write 1 as αα to get a common denominator
  13. x1∗=αα+βmp1 -- divide by p1 and multiply by αα+β (board equation 6)
  14. x2∗=βαp1p2⋅αα+βmp1 -- substitute equation 6 into equation 5
  15. x2∗=βα+βmp2 -- the p1's and the α's cancel

Answer: x1∗=αα+βmp1 and x2∗=βα+βmp2 -- exactly the demands found last Thursday by direct substitution.

The step labels on slide 23 point at the wrong equation numbers. For example, "Set (7) and (8) equal" means the two FOCs for x1 and x2.

Made-up example: Cobb-Douglas demands with numbers

Take α=2, β=3, m=100, p1=4 and p2=6.

  1. x1∗=22+3⋅1004 -- put the numbers into the demand for good 1
  2. x1∗=0.4⋅25=10 -- simplify
  3. x2∗=32+3⋅1006 -- put the numbers into the demand for good 2
  4. x2∗=0.6⋅1006=606=10 -- simplify
  5. 4(10)+6(10)=40+60=100 -- check: the budget is exhausted
  6. αx2βx1=2(10)3(10)=23 and p1p2=46=23 -- check: the tangency holds

Answer: (x1∗,x2∗)=(10,10) -- the consumer spends 40% of income on good 1 and 60% on good 2, matching the shares 25 and 35.

Why both methods agree

The Lagrange method and direct substitution give the same demands. That is because both solve the same two conditions: the bundle is on the budget line, and it sits at a tangency.

Why: The third FOC is the budget constraint. Getting rid of λ gives board equation 4, αx2βx1=p1p2, and the left side is the size of the Cobb-Douglas MRS. So equation 4 is the tangency condition. Setting the two λ's equal is also the "bang for the buck" rule from last lecture, MU1p1=MU2p2, with λ as the common value.

MU1MU2=αx2βx1=p1p2

Here MUi is the marginal utility of good i, and p1p2 is the size of the budget line's slope.

Exam

Repeated several times: "It doesn't matter which method you use, you'll end up with the same answer." The reason given: "they're coming from exactly the same place... we're identifying the tangency, and it's on the budget constraint." McLeod walked through this link "so that you can see that they're not entirely different approaches."

Why the demands are formulas

The demands come out as expressions, not a number like x1∗=2, because no values were given for income, prices or the preference parameters. Plug those in and you get a specific bundle.

Why: Each input moves a different part of the graph, so the formula tracks the optimum for any values:

Interior solutions

Cobb-Douglas gave an interior solution at a tangency. The same idea of the "best affordable bundle" also works when there is no smooth tangency.

When a solution is interior

An interior solution buys a positive amount of both goods. The ordinary demands then come from two conditions: the bundle exhausts the budget, and the MRS equals the budget line's slope. The third condition, which slide 25 lists and McLeod called the "subtext" behind the tangency, is that the indifference curves have no kinks.

Why: A tangency needs a smooth curve that you can "roll around" until it just touches the budget line. Cobb-Douglas curves get closer and closer to the axes but never touch them, so they never give a zero. L-shaped curves don't touch the axes either, so they also give interior solutions, just not at a tangency.

x1∗>0,x2∗>0,p1x1∗+p2x2∗=m

Here x1∗ and x2∗ are the optimal amounts, and the last equation says the budget is used up.

Complementary preferences

Complementary preferences mean the consumer always wants the two goods in a fixed proportion, which need not be one-to-one. Perfect complements are the special case a=1, a one-to-one ratio. The indifference curves are still L-shaped, with their kinks (vertices) on the ray x2=ax1 from the origin.

Why: Past the kink, extra of only one good adds nothing, because the consumer lacks the partner good to go with it. That is why the curve goes flat or vertical. It also leaves only three possible slopes, which "makes it difficult to find a tangency."

u(x1,x2)=min{ax1,x2}MRS={−∞if x2>ax10if x2<ax1undefinedif x2=ax1

Here a is the number of units of x2 wanted with each unit of x1. The MRS is −∞ on the vertical part of the L, 0 on the horizontal part, and undefined at the kink.

Exam

"Be a little bit careful here with the word perfect." In class, "perfect complements" means a one-to-one ratio. Dropping "perfect" (complementary preferences) means still a fixed proportion, "it's just not a one-to-one proportion." The slides still title the min{ax1,x2} pages "Perfect Complements".

Solving complementary preferences

The best affordable bundle is at the kink of an L-shaped curve, where that kink sits on the budget line. So solve the budget constraint together with x2∗=ax1∗, by direct substitution.

Why: In class, three bundles on the budget line all cost the same. Avery picked the middle one, at the kink, because "it's on the highest indifference curve." If they all cost the same, pick the one that makes you best off. The price intuition: one "pack" of 1 unit of x1 and a units of x2 costs p1+ap2, so the consumer can afford mp1+ap2 packs.

x1∗=mp1+ap2,x2∗=amp1+ap2

Here m is income, p1 and p2 are the prices, and a is the fixed ratio of x2 to x1.

Example: Complementary demands by direct substitution (done on the board)

Maximise min{ax1,x2} subject to p1x1+p2x2=m.

  1. p1x1∗+p2x2∗=m -- the optimum is on the budget line (equation 1)
  2. x2∗=ax1∗ -- the optimum is at the kink, on the ray (equation 2)
  3. p1x1∗+p2(ax1∗)=m -- substitute equation 2 into equation 1
  4. x1∗(p1+ap2)=m -- factor out x1∗
  5. x1∗=mp1+ap2 -- divide both sides (equation 3)
  6. x2∗=ax1∗=amp1+ap2 -- substitute equation 3 into equation 2 (equation 4)

Answer: (x1∗,x2∗)=(mp1+ap2,amp1+ap2) -- this is "a lot shorter" than the Cobb-Douglas case. With a=1 (perfect complements) it is mp1+p2 of each good.

Made-up example: Complementary demands with numbers

Take a=2, p1=1, p2=3 and m=70.

  1. p1+ap2=1+2(3)=7 -- the cost of one pack (1 unit of x1 and 2 units of x2)
  2. x1∗=707=10 -- the number of packs, which is also the amount of good 1
  3. x2∗=ax1∗=2(10)=20 -- two units of good 2 per pack
  4. 1(10)+3(20)=10+60=70 -- check: the budget is exhausted

Answer: (x1∗,x2∗)=(10,20) -- ten packs, bought in the exact 1-to-2 ratio the consumer wants.

Exam

Stated twice: "With complementary preferences, there's only one method to use. There's only the method of direct substitution. You can't use the Lagrange method... because there's no tangency to depend on."

Corner solutions

Cobb-Douglas and L-shaped curves never touch the axes. Corner solutions come from indifference curves that do touch the axes.

What a corner solution is

A corner solution buys none of one good: x1∗=0 or x2∗=0. The budget is still exhausted, but at the optimum the MRS need not equal −p1p2.

Why: When an indifference curve reaches an axis, the highest curve you can afford may touch the budget line right at the axis, where the two slopes differ.

p1x1∗+p2x2∗=m,MRS≠−p1p2 is allowed

Here the first equation says the budget is used up even at the corner.

Why slopes must be compared

With straight-line indifference curves you can't make a tangency happen, so you compare the two slopes instead.

Why: The budget line's slope, −p1p2, comes from prices, and the consumer takes prices as given ("the market generates the prices"). The indifference curve's slope comes from preferences, and we take preferences as given too. Neither slope can change, so moving x1 and x2 around never produces a tangency. The comparison is between willingness to trade (the MRS) and ability to trade (the price ratio), the same idea as last lecture.

Perfect substitutes

Perfect substitutes mean the consumer sees the two goods as equivalent. Only the total matters, as with Diet Coke versus Diet Pepsi. The indifference curves are parallel straight lines with slope −1, so the consumer is always willing to make a one-for-one trade.

Why: If you like both the same, you buy whichever is cheaper, and you spend your whole income on it. If the prices are equal, every mix is equally good.

u(x1,x2)=x1+x2,(x1∗,x2∗)={(mp1,0)if p1<p2(0,mp2)if p1>p2any bundle on the budget lineif p1=p2

Here mpi is how much of good i you get by spending all of income m on it.

Example: Willing to trade versus able to trade (slides)

Perfect substitutes, so MRS=−1. Compare a budget line with slope −2 to one with slope −0.5.

  1. |−1|<|−2| -- the consumer is willing to give up 1 unit of x2 for an extra x1, but is only able to get it by giving up 2
  2. Cut x1 and raise x2 until x1=0 -- each trade in that direction makes the consumer better off, ending at (0,mp2)
  3. |−1|>|−0.5| -- the consumer is willing to give up 1 unit of x2 for an extra x1, and the market gives 2 units of x1 for it
  4. Raise x1 and cut x2 until x2=0 -- ending at (mp1,0)

Answer: a steeper budget line gives the corner with only good 2, and a flatter one gives the corner with only good 1.

Example: Diet Coke or Diet Pepsi (class)

You like Diet Coke (good 1, horizontal axis) and Diet Pepsi (good 2, vertical axis) the same. You shop on three different days with income m.

  1. p1>p2 -- Coke costs more: "You like them the same. Diet Pepsi costs less. What are you buying?" You buy only Pepsi, so (0,mp2)
  2. p1<p2 -- Coke is now cheaper, so you buy only Coke: (mp1,0)
  3. p1=p2 -- the budget line has slope −1 and lies right on top of an indifference curve

Answer: buy only the cheaper cola. When the prices tie, every bundle on the budget line is optimal. McLeod said "they're all the most preferred bundle", and slide 43 says they are "equally the most preferred affordable".

left=0; right=16; bottom=0; top=16
---
y=8-x|x>0
y=12-x|x>0
y=15-x|x>0

Here x is good 1 and y is good 2. The lines are the slides' indifference curves for u=8, 12 and 15, all with slope −1.

Exam

"With perfect substitutes, we need to be careful on... thinking about the specific cases of the relative prices." Is good 1 cheaper, is good 2 cheaper, or are they priced the same? Don't forget the equal-price case, which the slides pose as a question: "But, what about the case when p1=p2?"

Substitutable preferences

Substitutable preferences drop the word "perfect". The consumer is always willing to make the same trade, giving up δ units of x2 for one more unit of x1, but the trade need not be one-for-one. The indifference curves are still straight lines, with slope −δ, and δ=1 is perfect substitutes.

Why: It works the same way as perfect substitutes, except that the consumer's own trade rate δ replaces 1. Which corner you land on depends only on δ, p1 and p2. The exact amounts also depend on m.

u(x1,x2)=δx1+x2,MRS=−δ(x1∗,x2∗)={(mp1,0)if δ>p1p2(0,mp2)if δ<p1p2any bundle on the budget lineif δ=p1p2

Here δ is the size of the indifference curve's slope and p1p2 is the size of the budget line's slope. Case 1 (the indifference curve is steeper) gives only good 1. Case 2 (the budget line is steeper) gives only good 2. In case 3 the two lines lie on top of each other.

Example: The MRS of substitutable preferences (class)

Utility is u=δx1+x2.

  1. MU1=∂u∂x1=δ -- the derivative of δx1 is the constant times 1, and x2 drops out
  2. MU2=∂u∂x2=1 -- δx1 drops out, and x2 with respect to itself is 1
  3. MRS=−MU1MU2=−δ1=−δ -- the MRS formula from Chapter 4

Answer: MRS=−δ -- the indifference curves are straight lines with slope −δ.

Made-up example: Which corner?

Take δ=3, p1=4, p2=2 and m=20.

  1. |MRS|=3 and p1p2=42=2 -- the willingness to trade versus the market rate
  2. 3>2 -- this is case 1: an extra unit of good 1 is worth 3 units of good 2 to the consumer but costs only 2
  3. x1∗=mp1=204=5 and x2∗=0 -- spend all income on good 1
  4. u(5,0)=3(5)+0=15 -- the utility at this corner
  5. u(0,202)=3(0)+10=10 -- the other corner gives less

Answer: (x1∗,x2∗)=(5,0) -- the consumer buys only good 1.

Exam

Slopes are compared in absolute value, which caused confusion in class. McLeod explained: "I'm using absolute values because it's so much easier to say the slope of the MRS is steeper than the slope of the budget constraint in absolute values, cuz the number is bigger." Without the bars, −2 looks "lower" than −1 even though it is steeper. McLeod added: "sometimes I play a little fast and loose with the math... if that ever happens, please stop me and ask questions."

Exam checklist

Admin & deadlines


Transcribed automatically from the lecture recording. Slides used: EC 270 - Ch. 05 - Choice (slides).pdf. Course info used: EC 270 - Math Appendix (slides).pdf, EC 270 - Rules of Derivatives.pdf, EC 270 - Syllabus - Fall 2026.pdf. Audio archived at /mnt/porsche/configs/lectures/archive/2026-09/EC 270 - 2026-09-29 - Lecture 06.opus.