Micro 2 Final Exam, Moed A (20 February 2023) · Dr. Boaz Zik · worked-solution

True-False Mix, Production Economy & Subsidy vs Price Cap (2023 Moed A)

True-False Mix, Production Economy & Subsidy vs Price Cap (2023 Moed A)

Part of: Micro-Economics II Micro 2 Final Exam, Moed A (20/02/2023), Micro-Economics II, Dr. Boaz Zik. 3 hours, cheat sheet and calculator allowed. Part 1: ten true/false statements (5 points each, 50 total). Part 2: two open questions (Q2 = 25 points, Q3 = 25 points). Builds on: Lec_04-Theory of the Firm, Partial & General Equilibrium, Lec_05-Production Economy & Externalities Practice: PS_01-Profit Maximisation & Properties of the Profit Function, PS_02-Isoquants, RTS & Properties of the Profit Function, PS_03-Cost Minimisation & Properties of the Cost Function, PS_04-Partial Equilibrium, Welfare & Price Ceilings, PS_05-Pareto Efficiency, Dominance & Social Welfare, PS_06-Exchange Equilibrium, Walrasian Prices & the Social Planner, PS_08-Production Economy, Efficiency Locus & Input Allocation

Source

Answers follow the official solutions. The true/false key is a 2-page PDF that repeats the statements with the correct word underlined (no explanations), so the reasons below are written for revision, not copied. The open-question solution is 6 pages. Where the official text has a typo or a gap, the answer is kept and a > [!warning] flags it.

What this paper tests

Intuition: the one move behind each part
  • True/false: for each statement ask "what is held fixed, and what does the statement quietly assume?". Most false statements fail on a hidden "for any" or "at every level".
  • Q2: equalise the two firms' RTS to get the efficiency locus, plug into the outputs to get the PPF, then set RPT equal to the consumer's MRS.
  • Q3: price equals marginal cost on the supply side, price equals marginal utility on the demand side. A subsidy shifts a curve; a price cap makes the short side of the market decide the quantity.

Part 1 — True / False (50 points)

Verdicts in the table are the official underlined answers. The explanations follow.

# Topic Official verdict
1 Endowment of YY rises by one unit: old contract curve True
2 Wage multiplied by tt: profit multiplied by tt False
3 Quasilinear production: ll never rises with output False
4 p>C′(0)p > C'(0) and C′′>0C''>0 give C′(x∗)=pC'(x^*)=p True
5 Prices multiplied by 3: same input bundle for x0x_0 True
6 Diminishing RTS: (8,16)(8,16) strictly above the isoquant of (10,20)(10,20) and (5,10)(5,10) False (the key marks both words, see the warning)
7 The Edgeworth box allocation drawn in the exam False
8 aa Pareto dominates bb: every social planner prefers aa True
9 On the PPF, RTS1=RTS2RTS_1=RTS_2 for any bundles on the two isoquants False
10 Lump-sum tax on firms fixes an externality False

(1) Contract curve after the endowment of YY rises — True

Answer

True. Almost every old contract-curve allocation stops being Pareto efficient.

Take α=a1−a\alpha = \frac{a}{1-a} and β=b1−b\beta = \frac{b}{1-b}. On the contract curve MRS1=MRS2MRS_1 = MRS_2:

α y1x1=β Y−y1X−x1⟹y1=β Y x1αX+(β−α) x1\alpha\,\frac{y_1}{x_1} = \beta\,\frac{Y - y_1}{X - x_1} \quad\Longrightarrow\quad y_1 = \frac{\beta\, Y\, x_1}{\alpha X + (\beta - \alpha)\,x_1}

Replacing YY by Y+1Y+1 raises y1y_1 for every x1>0x_1 > 0, so the whole curve shifts up. An old allocation therefore fails in one of two ways:

  • If the extra unit of YY is left unallocated, giving it to either agent is a Pareto Dominance improvement (more of a good the agent likes, nobody loses).
  • If the extra unit goes to one agent, the two MRS values no longer match, so the agents can still trade to a Pareto improvement.
Edge case

The statement says "any allocation". The endpoints of the curve (one agent holds nothing, the other everything, including the new unit) remain efficient. The exam intends the interior points, which is why the official answer is True.

(2) Wage multiplied by tt — False

Answer

False. Homogeneity of Degree 1 needs both prices scaled: Π∗(tp,tw)=t Π∗(p,w)\Pi^*(tp, tw) = t\,\Pi^*(p,w). Scaling only the wage does not scale profit by tt.

Counterexample

Take f(l)=lf(l)=\sqrt{l}. Then l∗=(p2w)2l^* = \left(\frac{p}{2w}\right)^2 and

Π∗(p,w)=p24w.\Pi^*(p,w) = \frac{p^2}{4w}.

With wage twtw the profit is p24tw=1tΠ∗(p,w)\frac{p^2}{4tw} = \frac{1}{t}\Pi^*(p,w), which is divided by tt, not multiplied.

Intuition

The statement mixes up homogeneity (everything scales) with a one-price change. A higher wage never raises profit (the profit function is decreasing in ww by the Envelope Theorem: ∂Π∗/∂w=−l∗\partial \Pi^*/\partial w = -l^*), so a multiple of t>1t>1 cannot be right.

(3) Quasilinear production f(k,l)=g(l)+kf(k,l) = g(l) + k — False

Answer

False. Not "at any level of production": at low output, more output does need more labour.

For cost minimisation the RTS is MPl/MPk=g′(l)/1=g′(l)MP_l / MP_k = g'(l)/1 = g'(l) and the tangency condition sets it equal to plpk\frac{p_l}{p_k}. So there is one cost-efficient labour level l∗l^* with g′(l∗)=pl/pkg'(l^*) = p_l/p_k, independent of output.

  • High output (above g(l∗)g(l^*)): the firm holds l=l∗l=l^* and gets all extra output from capital kk. Labour does not increase.
  • Low output (below g(l∗)g(l^*), assuming g(0)=0g(0)=0): using l∗l^* would already overshoot the target, so the optimum is the corner k=0k=0 with g(l)=xg(l)=x and l=g−1(x)l=g^{-1}(x), which does rise with xx.
Intuition

The statement is true only above a threshold output. Because it says "at any level", one low-output counterexample makes it false.

(4) p>C′(0)p > C'(0) and C′′>0C''>0 — True

Answer

True. C′′>0C''>0 means marginal cost is strictly rising, and p>C′(0)p > C'(0) means the first unit is profitable. Profit Π(x)=px−C(x)\Pi(x) = px - C(x) is strictly concave with Π′(0)=p−C′(0)>0\Pi'(0) = p - C'(0) > 0, so the optimum is interior, where Π′(x∗)=0\Pi'(x^*) = 0, i.e. C′(x∗)=pC'(x^*) = p.

Hidden assumption

This needs marginal cost to climb above pp eventually (otherwise the firm never stops). With C′′>0C''>0 and a normal cost function that is automatic; the official key does not discuss it.

(5) Prices multiplied by 3, same output x0x_0 — True

Answer

True. For a fixed output x0x_0 the firm solves Cost Minimisation, and the solution depends only on the price ratio and the technology.

Multiplying both input prices by 3 leaves plpk\frac{p_l}{p_k} unchanged, so the tangency RTS=plpkRTS = \frac{p_l}{p_k} and the isoquant constraint pick the same bundle. Only the level of cost triples.

Wording

The exam says "prices of both outputs" although there is a single product XX; "both inputs" (capital and labour) is clearly meant. The verdict is True either way: with output held at x0x_0, the chosen input bundle does not depend on the output price at all.

(6) Diminishing RTS and the bundle (8,16)(8,16) — False (with a caveat)

Answer

Official key: the line shows both True and False underlined, with False in bold. The consistent answer is False.

Note (8,16)=0.6⋅(10,20)+0.4⋅(5,10)(8,16) = 0.6\cdot(10,20) + 0.4\cdot(5,10), a convex combination of the two bundles.

  • If isoquants are only weakly convex (Convex Isoquant), the midpoint lies on the same or a higher isoquant: "weakly higher", not "strictly".
  • But (10,20)(10,20) and (5,10)(5,10) are on the same ray, and a straight line through them lying in one isoquant means the RTS is constant on that stretch, so the isoquant is linear there and (8,16)(8,16) is on the same isoquant, not a strictly higher one.
Ambiguous key

The official key underlines both words for (6). Taking "diminishing RTS" as strictly diminishing would make the premise impossible (two collinear points cannot both lie on a strictly convex isoquant), so the statement could not be shown true. False is the defensible answer; flagging that the key is ambiguous.

Intuition

The "strictly higher" claim is the Cobb-Douglas case from PS_02-Isoquants, RTS & Properties of the Profit Function only when the two bundles are different directions. Here the examiner picked proportional bundles on purpose: it is the perfect-substitutes (flat isoquant) knife edge.

(7) The Edgeworth box allocation drawn in the exam — False

Answer

False. The allocation is not Pareto efficient.

In the picture, agent 1's bundle (x1a,y1a)(x_1^a, y_1^a) is measured from the bottom-left corner and agent 2's bundle (x2a,y2a)(x_2^a, y_2^a) is measured from the top-right corner, and the two dots are different points. Agent 2's dot is above and to the right of agent 1's dot, so

x1a+x2a<X,y1a+y2a<Y.x_1^a + x_2^a < X, \qquad y_1^a + y_2^a < Y .

Some of both goods is left over. Giving the leftovers to either agent makes that agent strictly better off and the other no worse off (with monotone preferences, see Local Nonsatiation), so the allocation is Pareto dominated by a feasible one.

Intuition

A point in an Edgeworth box has to be a single point, because an allocation must hand out all the goods. Two dots mean wasted resources, and wasteful allocations are never Pareto efficient.

(8) Pareto dominance and social planners — True

Answer

True. If aa Pareto dominates bb, nobody is worse off and somebody is strictly better off. A social planner with welfare function W(U1,U2)W(U_1, U_2) increasing in each agent's utility therefore ranks aa above bb.

Intuition

"Any social planner" means any planner who cares about every agent's utility in the standard way (strictly increasing WW, e.g. the Cobb-Douglas planner of the course). A planner such as min⁡(U1,U2)\min(U_1,U_2) is only weakly increasing and could be indifferent, but that is not what the exam has in mind. See PS_05-Pareto Efficiency, Dominance & Social Welfare.

(9) RTS equal on the PPF — False

Answer

False. On the PPF the efficient input split has RTS1=RTS2RTS_1 = RTS_2 (that is the Efficiency Locus). The statement says this holds for any bundles on the isoquants f1=x0f_1 = x_0 and f2=y0f_2 = y_0.

Those isoquants contain many bundles, most of which do not use the economy's resources efficiently (for example, bundles that do not add up to the total labour and capital, or that sit away from the tangency). Only the particular pair of bundles that uses all resources and is tangent satisfies RTS1=RTS2RTS_1 = RTS_2.

Intuition

"Any" is the word to circle. The condition describes the best input split for producing the point on the PPF, not every input pair that gives the same outputs.

(10) Lump-sum tax and an externality — False

Answer

False. A lump-sum tax does not depend on what the firm produces, so it does not change the firm's marginal cost or its choice. The inefficiency comes from a wedge at the margin between private and social cost.

The fix is a per-unit Pigouvian Tax equal to the marginal external damage, which moves the firm's marginal cost up to the social marginal cost.

Intuition

Lump-sum taxes are the good tool for redistribution because they leave margins alone (Second Welfare Theorem). For the same reason they are useless for correcting a distorted margin.


Q2 — Production economy (25 points)

Firms: f1(l1,k1)=l1ak11−af_1(l_1,k_1) = l_1^{a} k_1^{1-a} and f2(l2,k2)=b⋅(l2ak21−a)f_2(l_2,k_2) = b\cdot\big(l_2^{a} k_2^{1-a}\big). Representative consumer: U(x,y)=xcy1−cU(x,y) = x^{c} y^{1-c}. Resources: k1+k2=l1+l2=Dk_1 + k_2 = l_1 + l_2 = D. Parameters: a=12a=\tfrac12, b=1b=1, c=12c=\tfrac12, D=50D=50.

2a — Efficiency locus (6 points)

Answer
k1=l1k_1 = l_1

(equivalently k2=l2k_2 = l_2): both firms use the same capital-labour ratio.

Working. Marginal products of firm 1:

MPl1=a l1a−1k11−a,MPk1=(1−a) l1ak1−aMP_{l_1} = a\,l_1^{a-1}k_1^{1-a}, \qquad MP_{k_1} = (1-a)\,l_1^{a}k_1^{-a}

so

RTS1=MPl1MPk1=a k1(1−a) l1RTS_1 = \frac{MP_{l_1}}{MP_{k_1}} = \frac{a\,k_1}{(1-a)\,l_1}

Firm 2 has the same shape with an extra factor bb in both marginal products, which cancels:

RTS2=b a l2a−1k21−ab (1−a) l2ak2−a=a k2(1−a) l2RTS_2 = \frac{b\,a\,l_2^{a-1}k_2^{1-a}}{b\,(1-a)\,l_2^{a}k_2^{-a}} = \frac{a\,k_2}{(1-a)\,l_2}

Efficiency needs RTS1=RTS2RTS_1 = RTS_2:

k1l1=k2l2\frac{k_1}{l_1} = \frac{k_2}{l_2}

Substitute k2=D−k1k_2 = D - k_1 and l2=D−l1l_2 = D - l_1, cross-multiply, and the k1l1k_1 l_1 terms cancel:

k1(D−l1)=l1(D−k1)  ⟹  k1=l1k_1(D - l_1) = l_1(D - k_1) \;\Longrightarrow\; \boxed{k_1 = l_1}
Typos in the official text

The key writes f1f_1 for firm 2's function and, in the RTS2RTS_2 line, uses bb as an exponent (l2b−1k21−bl_2^{b-1}k_2^{1-b}) where the exponent should be aa. The final result is unaffected.

Intuition

Both firms use the same Cobb-Douglas shape, so their RTS depends only on their own input ratio. The factor bb is just a productivity multiplier and does not change the shape of the isoquants, so it drops out of the locus. In the Edgeworth box for inputs the locus is the 45 degree diagonal.

2b — Production possibility frontier (6 points)

Answer
y=50−xy = 50 - x

Working. On the locus k1=l1k_1 = l_1 the firm-1 output is x=l1al11−a=l1x = l_1^{a}l_1^{1-a} = l_1. Firm 2 gets l2=k2=50−l1l_2 = k_2 = 50 - l_1, and with b=1b=1:

y=(50−l1)a(50−l1)1−a=50−l1=50−xy = (50-l_1)^{a}(50-l_1)^{1-a} = 50 - l_1 = 50 - x

So the PPF is a straight line, and RPT=1RPT = 1 everywhere (constant returns and identical technology, since b=1b = 1).

2c — Optimal allocation (5 points)

Answer

x∗=y∗=25x^* = y^* = 25 and l1∗=k1∗=l2∗=k2∗=25l_1^* = k_1^* = l_2^* = k_2^* = 25.

Working. The optimum has RPT=MRSRPT = MRS. The consumer's marginal rate of substitution is

MRS=c y(1−c) x=yxwhen c=12MRS = \frac{c\,y}{(1-c)\,x} = \frac{y}{x} \quad\text{when } c = \tfrac12
RPT=1=yx=MRS  ⟹  y=xRPT = 1 = \frac{y}{x} = MRS \;\Longrightarrow\; y = x

and on the PPF x=50−xx = 50 - x, so x=y=25x = y = 25. The input split follows from x=l1=k1=25x = l_1 = k_1 = 25 and the remaining 2525 of each input goes to firm 2.

PP01_q2_efficiency_locus_ppf_optimum

Left: the efficiency locus is the diagonal of the input box, and the optimum sits in the middle at (25,25)(25,25). Right: the straight-line PPF is tangent to the indifference curve xy=625xy = 625 (utility 2525) at (25,25)(25,25), where RPT=MRS=1RPT = MRS = 1.

2d — A higher aa (2 points)

Answer

Nothing changes: the quantities and the ratio y/xy/x stay the same.

The efficiency locus k1/l1=k2/l2k_1/l_1 = k_2/l_2 does not contain aa at all, and with b=1b=1 the outputs on the locus are still x=l1x = l_1 and y=50−l1y = 50 - l_1. So the PPF, the optimum, and y/x=1y/x = 1 are unchanged.

Intuition

aa changes how much firms rely on labour versus capital, but both firms share it, and the total supply of the two inputs is the same (DD each). On the diagonal the two inputs are used in equal amounts, so the labour share parameter never matters.

2e — A higher bb (2 points)

Answer

The ratio y/xy/x rises.

Now y=b(50−x)y = b(50 - x): the PPF has a steeper slope (RPT=bRPT = b) and a higher yy-intercept (50b50b). The optimum needs MRS=y/x=bMRS = y/x = b, so with c=12c=\tfrac12 the ratio y/xy/x equals bb and is higher.

Intuition

Checking the numbers myself: y=bxy = bx and y=b(50−x)y = b(50-x) give x=25x = 25 and y=25by = 25b. So with c=12c=\tfrac12, the quantity of XX stays at 25 and the quantity of YY grows with bb, which is why the ratio grows. This is my own check, the official solution states only the direction of the ratio.

2f — A higher cc (2 points)

Answer

The ratio y/xy/x falls: the consumer spends a larger share on XX and a smaller share on YY.

Intuition

The condition RPT=MRSRPT = MRS with RPT=1RPT = 1 gives c y(1−c) x=1\frac{c\,y}{(1-c)\,x} = 1, so yx=1−cc\frac{y}{x} = \frac{1-c}{c}, which falls in cc. On the PPF y=50−xy = 50-x this gives x=50cx = 50c and y=50(1−c)y = 50(1-c): more XX, less YY. Again my own check of the official direction.

2g — A higher DD (2 points)

Answer

The ratio is unchanged; quantities of both goods rise.

The efficiency locus and the RPT do not depend on DD, so the ratio is the same. The PPF shifts outwards (income rises) and both quantities rise in the same proportion.

Intuition

With constant returns, doubling the resources doubles output of everything along the ray with the same ratio. Using the same working, x=y=D/2x = y = D/2.


Q3 — Partial equilibrium, subsidy and price cap (25 points)

Quasilinear consumer U=100x−x2+yU = 100x - x^2 + y (yy is money, no corner solutions). Firm f(k,l)=l1/4k1/4f(k,l) = l^{1/4}k^{1/4} with pk=pl=12p_k = p_l = \tfrac12.

Numbering

The exam labels the last part (e), there is no (d); the official solution calls it (d). They are the same question (subsidy versus price cap).

3a — Cost function (8 points)

Answer
C(x)=x2C(x) = x^2

Working. Cost Minimisation: set the RTS equal to the input price ratio:

RTS=kl=plpk=1  ⟹  k=lRTS = \frac{k}{l} = \frac{p_l}{p_k} = 1 \;\Longrightarrow\; k = l

Plug into the isoquant for output xx:

x=l1/4k1/4=k1/4k1/4=k1/2  ⟹  k(x)=l(x)=x2x = l^{1/4}k^{1/4} = k^{1/4}k^{1/4} = k^{1/2} \;\Longrightarrow\; k(x) = l(x) = x^2
C(x)=pl l(x)+pk k(x)=12x2+12x2=x2C(x) = p_l\, l(x) + p_k\, k(x) = \tfrac12 x^2 + \tfrac12 x^2 = x^2

Supply is marginal cost: PS(x)=C′(x)=2xP^S(x) = C'(x) = 2x.

3b — Equilibrium price and quantity (7 points)

Answer

xe=25x^e = 25 and Pe=50P^e = 50.

For a Quasilinear Utility consumer, demand is marginal utility of xx (the marginal utility of money is 1):

PD(x)=100−2xP^D(x) = 100 - 2x
100−2x=2x  ⟹  xe=25,Pe=50100 - 2x = 2x \;\Longrightarrow\; x^e = 25,\quad P^e = 50

3c — Consumer subsidy of 20 (5 points)

Answer
  • ΔCS=275\Delta CS = 275 and Δ\Delta firm surplus =275= 275.
  • Government pays 20×30=60020 \times 30 = 600.
  • Added welfare is 550<600550 < 600, so the gain does not cover the cost (net loss of 5050).

Before the subsidy (Consumer Surplus and Firm Surplus):

CS=∫025(100−2x−50) dx=[50x−x2]025=1250−625=625CS = \int_0^{25} (100 - 2x - 50)\,dx = \Big[50x - x^2\Big]_0^{25} = 1250 - 625 = 625
FS=∫025(50−2x) dx=[50x−x2]025=625FS = \int_0^{25} (50 - 2x)\,dx = \Big[50x - x^2\Big]_0^{25} = 625

After the subsidy. The official solution adds the subsidy to the demand curve, PD(x)=120−2xP^D(x) = 120 - 2x (the price the producer effectively receives is the consumer's willingness to pay plus 20):

120−2x=2x  ⟹  xe=30,Pe=60120 - 2x = 2x \;\Longrightarrow\; x^e = 30,\quad P^e = 60
CS=∫030(120−2x−60) dx=[60x−x2]030=1800−900=900CS = \int_0^{30} (120 - 2x - 60)\,dx = \Big[60x - x^2\Big]_0^{30} = 1800 - 900 = 900
FS=∫030(60−2x) dx=900FS = \int_0^{30} (60 - 2x)\,dx = 900
ΔCS=900−625=275,ΔFS=900−625=275\Delta CS = 900 - 625 = 275,\qquad \Delta FS = 900 - 625 = 275
Subsidy cost=20×30=600,added welfare=275+275=550<600\text{Subsidy cost} = 20 \times 30 = 600,\qquad \text{added welfare} = 275 + 275 = 550 < 600
Intuition

Here "60" is the price the firm receives; the consumer pays 60−20=4060 - 20 = 40 (the point on the original demand curve at x=30x=30). The consumer surplus calculated with 120−2x120 - 2x against the price 60 equals the surplus on the original demand against the price 40, so the same 900. The net loss of 5050 is the standard deadweight triangle: 12×20×(30−25)=50\tfrac12 \times 20 \times (30-25) = 50, because the subsidy pushes xx past the efficient quantity where value (marginal utility 4040 at x=30x=30) falls below marginal cost (6060).

PP01_q3_subsidy_surplus

Left: surpluses at the competitive equilibrium. Right: with the subsidy, producers receive 60 and consumers pay 40 at x=30x = 30; each surplus rises from 625 to 900.

3d — Which policy gives higher welfare? (5 points)

Answer

The subsidy gives higher welfare: 1,200>4501{,}200 > 450 (price cap at 10).

Price cap at 10. At price 1010 supply is 2x=102x = 10, so only x=5x = 5 is produced, and demand exceeds supply (excess demand). Quantity is the short side, supply. Welfare is total value minus cost on the units traded:

∫05(100−2x−2x) dx=[100x−2x2]05=500−50=450\int_0^{5} \big(100 - 2x - 2x\big)\,dx = \Big[100x - 2x^2\Big]_0^{5} = 500 - 50 = 450

Subsidy. Welfare is the surplus gain less the cost of the subsidy: 1800−600=12001800 - 600 = 1200.

450<1,200450 < 1{,}200

PP01_q3_price_cap

Under the cap at 10 only 5 units are traded; the shaded green area is the 450450 of welfare kept, and the large red triangle is the surplus lost compared with the efficient quantity.

Intuition

The price cap is far below the equilibrium price of 50, so it chokes supply down to 5 units (against an efficient 25). The subsidy overshoots to 30, but the overshoot costs only a small triangle of 50 compared with the efficient welfare of 1,250.

Assumption in the official welfare of 450

The figure 450=∫(v′−C′)450 = \int (v' - C') treats the 5 units as going to the consumers who value them most (efficient rationing). With random queuing the welfare would be even lower, so the ranking does not change.


Recap table

Part Tool One-line answer
1 Contract curve shifts when YY changes True
2 Homogeneity needs both prices scaled False, profit is divided by tt in the example
3 Quasilinear: g′(l∗)=pl/pkg'(l^*) = p_l/p_k False, at low output labour rises
4 Strictly convex cost, p>C′(0)p > C'(0) True, interior optimum C′(x∗)=pC'(x^*) = p
5 Cost minimisation uses only price ratio True
6 Collinear bundles on one isoquant False (key marks both words)
7 Two different dots, resources wasted False
8 Planner with increasing WW True
9 RTS equal only at the efficient input split False
10 Lump-sum tax does not change margins False
2a RTS1=RTS2RTS_1 = RTS_2 k1=l1k_1 = l_1
2b Outputs on the locus y=50−xy = 50 - x
2c RPT=MRSRPT = MRS x∗=y∗=25x^* = y^* = 25
2d aa not in the locus No change
2e RPT =b= b y/xy/x rises
2f y/x=(1−c)/cy/x = (1-c)/c y/xy/x falls
2g Scale economy Ratio same, quantities up
3a RTS = price ratio C=x2C = x^2
3b PD=PSP^D = P^S x=25x = 25, P=50P = 50
3c Shift demand up by 20 x=30x = 30, ΔCS=ΔFS=275\Delta CS = \Delta FS = 275, cost 600, net −50-50
3d Welfare under a cap 450<1,200450 < 1{,}200, subsidy better