[ge_arrow] Follow-up review items for the rational expectations section (PR #1084) #1085

PR #1084 adds a rational expectations reading of Arrow's sequential equilibrium to ge_arrow, with two new exercises (Arrow's two trading arrangements with many goods, and Hart's examples). Tom would like it merged as is, so this issue collects the review findings for a follow-up PR. All eight new code cells run cleanly and the numerical claims check out; the items below are corrections and polish.

Fixes

1. Exercise 4, part 6 overstates what happens under a misforecast. The portfolios helper keeps security prices at the equilibrium q* while changing the forecast, so the securities market does not clear (aggregate demand for the two securities is about 12.35 and 14.88 against a supply of about 13.80). The solution text then says "The plans are feasible and the securities market clears", which is not true as coded. Either solve for the q that clears the securities market given the wrong forecast (roughly 0.401 and 0.599; the realized spot prices still contradict the forecast, so the conclusion survives), or drop that sentence and say only that the realized spot prices differ from the forecast.

2. Notation clash on n. The lecture uses K consumers and n Markov states, and the new section says sequential trading needs "only $n$ one-period markets at each date". Exercise 4 then introduces "$n$ consumers". Rename the exercise's consumer count to K.

3. Inconsistent dating of Muth. The overview says Muth named the hypothesis "almost a decade later" than Arrow's 1952 talk, while the concluding remarks say "eight years before" Muth 1961. 1952 to 1961 is nine years, or eight counting from the 1953 French publication. Pick one and use it in both places.

4. The kihlstrom2019 reference is unverifiable. It is entered as unpublished Wharton class notes with no URL and I could not locate it online. The "we learned this reading from" attribution and the Bellman anecdote both rest on it. Please check with Tom whether there is a link or a stable citation; if not, consider wording such as "unpublished notes shared with the authors".

5. Check the Hart note against the paper. The note states that an equilibrium is "Pareto optimal relative to the set of such equilibria when there is one good and the market structure is complete up to the next-to-last date". Hart's constrained-optimality notion is usually stated relative to allocations attainable through the existing markets rather than relative to other equilibria, so the wording should be checked against Hart (1975).

Optional polish

• Exercise 4, part 2 says the economy is "specified in the solution below". Moving the parameter values into the exercise statement would let readers attempt it without opening the solution.
• The contingent-claims solver works on a system that is homogeneous of degree zero, so the Jacobian is singular. It converges cleanly today, but fixing the numeraire inside the solver would be more robust.

Context

• Related PR: [ge_arrow] Add the rational expectations reading of Arrow's sequential equilibrium #1084 (branch tom_arrow2)
• New bib entries hart1975, kihlstrom2019, diamond1967 introduce no duplicate keys.