Some margin notes for the 2004 JHS text

by Dr. Griffin

(last revised December 2015)
This is a great text in many ways, but there are places where more accuracy or illumination is helpful.

Ch pp Line NOTE
1 5 1 Some elements of this paragraph should be quoted and attributed to Graaff.
1 8 1 Here is perhaps the first time the Kaldor and Hicks criteria are lumped in this textbook, so as to make them appear to be a single criterion. Readers should keep in mind that these are separate criteria (regardless of what they are named), and they are two members of a family of compensation tests. I find it very helpful to distinguish Kaldor & Hicks criteria as 2 separate things (as Graaff does; book cited by JH&S). In the interest of not merely historical accuracy but in understanding the underpinnings of these concepts, it appears that Hicks fathered the term "compensating variation" (1939, Value and Capital) but did not think about its use in a hypothetical compensation test as did Kaldor (1939, Economic J.). Hicks was initially focused on the correctness of Marshall's version of consumer surplus for the individual. Baumol later called this first compensation test "Mr. Kaldor's principle" and "Mr. Kaldor's criterion" (1946/7, Rev.Econ.Studies). For the individual, Henderson recognized that there were more compensated measures than just CV, considering price and quantity changes, but did not discuss a welfare test (1941, Rev.Econ.Studies). Scitovsky saw that CV for a price change was different than -CV for the opposite change, and he was interested in the welfare significance of this result via the compensation test (1941, Rev.Econ.Studies). With some stated debt to Henderson, Hicks named EV in 1942 (Rev.Econ.Studies), and his later paper (1943, "The Four Consumer's Surpluses" Rev.Econ.Studies) clearly identifies CV, EV, CS, ES, & their relation to each other and the "Marshall measure". At that time Hicks named the 4 as follows: price-compensating variation, price-equivalent variation, quantity-compensating variation, and quantity-equivalent variation. Hick's 1956 book, A Revision of Demand Theory credits Henderson with the "discovery" of the quantity change welfare measure. One of the bibles on valuation, The Measurement of Environmental and Resource Values (Freeman 1993; now Freeman, Herriges, & Kling 2014 where still more care is exercised) is careful to develop CV & CS as Kaldor measures and EV & ES as Hicks measures, but the text occasionally lumps them as "Hicks-Kaldor" when undertaking general discussion. The graduate text by Chris Jones (Applied Welfare Economics, 2005, Oxford) distinguishes the Kaldor test from the Hicks test (p. 11) as do I, but he says some people call the Kaldor test a Kaldor-Hicks test (fn. 12). It would be somewhat comforting if this was true, but it is difficult to be confident about what people mean when they say K-H; most just don't understand the details. In my opinion, when professionals say or write K-H in their work, they risk having their work questioned. There's not just one compensation test, and there are normative questions that should not be swept under a rug by inferring a single test.
2 23 13 and appropriate price vector Once the correct initial endowments (initial income) is selected, market activity will resolve the appropriate prices.
2 23 last A CE isn't merely the prices. It's the consumption bundles and production vectors for all agents too.
2 24 The assumptions listed in footnotes 2 and 3 do not appear to be exhaustive.
2 28 6-9 It is worrisome and unnecessary to define "first best" as both/either competitive equilibrium or Pareto optimum. Pareto optimality is the efficiency standard. "Market failures" are not departures from competitive equilibria; they are departures from Pareto optimality. Market failures commonly result in competitive equilibria; they just aren't equilibria that are Pareto optimal due to the unmet assumptions of the First Theorem.
2 28 14 and a price vector. One doesn't prescribe prices for markets, and the markets will produce the correct prices if the appropriate distribution of factors has been selected.
2 29 Fig 2.7 Because C and C' are Scitovsky indifference curves, tangencies at Ob and Ob' are not assured.
2 30 1-2 This sentence is true for a Pareto Improvement criterion, but it is untrue for a Pareto (optimality) criterion. If we're sitting at a PO state (presumably because the CE they mention is achieved in a scenario where the First Thm holds true), then no other state is a Pareto Improvement so no other state can be considered. But there are infinitely many PO which are just as good as the beginning PO in the eyes of the Pareto criterion, so other "distributional considerations" can indeed be entertained.
2 30 Last-3 The last full sentence on this page again confuses PO and PI. There is no relevant "status quo" for the Pareto criterion. There is for each Pareto Improvement criteria.
3 33 Figure 3.1 has some oddly shaped (backward-bending) indifference curves in it.
3 34 mid "Figure 3.2 where the indifference curve C" Be careful. Scitovsky indifference curves aren't the same things as welfare indifference curves.
3 35 1-7 This is false, as shown by Griffin (1995). Portions of this paragraph might be correct if JHS are only talking about the Kaldor test.
3 36 11-13 Note that the first sentence of this paragragh refers to the Kaldor test, and second sentence is suggesting a Hicks test.
3 38 4-5 "In other words, ..." This sentence is false.
3 39 last Note that they are not calling W0 a Scitovsky indifference curve.
4 59 25 Here's where producer surplus gets relabelled as consumer surplus because of a visual similarity. This is a debatable use of the term consumer surplus.
4 60 One of Figure 4.6's axes is mislabelled.
6 132 30-33 A problem with the term "Hicksian consumer surplus" is that it seems to imply that there's one such measure for a given price change, like there is for the Marshallian measure. But there are many Hicksian tests, indeed infinitely many unless we decide to confine our attention to the big two (CV, EV) for a price change.
6 170-1 § 6.C Incorrect: "Compensating variation supports the Kaldor-Hicks compensation test and equivalent variation supports the Scitovsky reversal test." [As noted in class and other readings, Kaldor: CV and Hicks: EV.] Incorrect if we reserve the term "variation" for price changes: "... compensating variation for the income change ..."
7 203 last ¶ It's unclear why a price change needs to be discussed here when the topic pertains to quantity changes.
7 204 12 CV is not defined for quantity changes. Compensating surplus is the relevant concept.
7 206 17 Hicks introduced these concepts in 1943 but not these terms.
7 206 19 CS and ES are not defined for price changes in my opinion, just as CV & EV are not defined for quantity changes.
8 255 5th from bottom I'd use a word other than "efficiency" in a text such as this, given that efficiency has been previously assigned a Pareto-based meaning.
8 257 last sent. Value judgments by policy makers are still value judgments.
8 258-9 The purpose of Figure 8.5 and accompanying discussion appears to be to justify a "Sum of WTP" approach to policy analysis. (1) More formality for this important topic is warranted. (2) It would be better to develop this matter fully in Chapter 3 instead of here. (3) Why not "Sum of WTA" or "Sum of ΔS & ΔR"?
8 263 7th from bottom Shouldn't the statement of the prior section "assuming efficiency in rationing" be repeated here, although here it would be pertaining to rationing among producers?
8 269 4 This statement can be made stronger. "... a potentially Pareto-superior fashion ..." is correct. Since all of the prior deviations from competitive market conditions cause deadweight losses (i.e. summed welfare measures are negative), they all fail to pass the Marshallian compensation test. They also are not Pareto improvements because some group is a loser, but that's not as strong as saying summed welfare changes are negative.
8 270 9-10 Scitovsky indifference curves aren't just iso-SW curves, they're iso-individual-utility for more than one individual. So, if the graph is right, "a" -> "c" accomplishes a Pareto improvement; no compensation is needed.
8 272 center para. This seems to imply that compensation criteria (not criterion) are not the same thing as potential Pareto criteria.
8 286 I'm wondering why the ΔR notation and lingo of prior chapters has been abandoned in this section in favor of "producer surplus" change.
13 530 1 "Internalization" is misdefined here, since the word has evolved some. Internalization means to enact a corrective policy for the externality in question. Whereas the externality existed because causal agents didn't care about their influence on other agents (unpriced good), a corrective policy restrains the behavior of causal agents or provides them with a reason to care, thereby internalizing the interrelationship into the decisionmaking of the causal agents. All kinds of policies are potential internalizations.
13 535 7th from bottom The Coase contribution is not well characterized in the initial sentence here. The Pareto achievement of markets was already known. The more central contribution of Coase had to do with the question of who should get the initially assigned property rights if we are to achieve efficiency.
13 537 6 Lbar and L are reversed.
13 546 34-35 Odd statement here beginning with "Second, ...". Transactions costs do not prohibit a competitive equilibrium from being obtained. I would add that transaction costs do not even prohibit competitive equilibria from being Pareto optimal. Transaction costs are, however, a challenge in policy design, as different policy approaches may invoke differing amounts and distributions of these costs among agents.
13 550 15 The inclusion of "clean air" in this list of allegedly pure public goods is inaccurate. Many others make this mistake too. Many uses of clean air are rival. [To what uses do you put "clean air"? Which ones are nonrival?]
13 554 7-9 "Suppose the amount ... visitors." This is an unacceptable supposition. It is obvious that any water taken by anyone is subsequently unavailable for others. The authors are confusing nonrivalness with abundance. If there is so much water in the spring that rival usage does not call for any kind of rationing, then optimal price is zero. There are no opportunity costs for use because the good is not scarce. The rival/nonrival dichotomy does not apply here.
14 574 6 The words "unambiguously Pareto superior" constitute a gross misrepresentation of the hypothetical finding that B>C in every time period or even that NPV>0. Due to the aggregation going on in all of these measures (B,C,NPV), the word "potential" should precede Pareto in this passage.
14 578 9 "current" should be "future"