cass_fiscal#

  • Series: lecture-python.myst

  • File: lectures/cass_fiscal.md

  • Audit date: 2026-08-26

  • Corpus snapshot: e25fdf2345

  • Categories audited: writing, math, code, figures, references, links, admonitions (JAX out of scope)

  • Overall score: 7.6 / 10

  • Priority: HIGH

Score breakdown#

Category

Score

One-line note

Writing

3/10

qe-writing-006 ×12; qe-writing-001 ×3; qe-writing-005 ×3, +4 more.

Math

9/10

qe-math-001 ×1.

Code

7.5/10

qe-code-001 ×4; qe-code-003 ×1.

JAX

out of scope

JAX rules target lecture-jax.

Figures

4/10

qe-fig-003 ×10; qe-fig-005 ×6; qe-fig-008 ×11, +1 more.

References

10/10

no mechanical violations detected.

Links

10/10

no mechanical violations detected.

Admonitions

10/10

no mechanical violations detected.

Issues#

Critical#

None found.

High severity#

  • [qe-fig-001] — Do not set figure size unless necessary. Count: 9. Lines: 828, 875, 910, 978, 1057, 1386, 1653, 1708, 1734. Example: figsize=.

  • [qe-fig-003] — No matplotlib embedded titles. Count: 10. Lines: 774, 779, 786, 792, 798, 1022, 1028, 1034, 1043, 1048. Example: .set_title.

  • [qe-fig-005] — Descriptive figure names for cross-referencing. Count: 6. Lines: 809, 900, 1054, 1383, 1631, 1686. Example: code-cell figure without mystnb figure metadata.

  • [qe-fig-008] — Use lw=2 for line charts. Count: 11. Lines: 772, 777, 784, 790, 796, 1020, 1026, 1027, 1032, 1039, …. Example: plot() without lw=.

  • [qe-writing-006] — Capitalize lecture titles properly. Count: 12. Lines: 39, 112, 169, 232, 275, 283, 331, 598, 1208, 1525, …. Example: H2 Title Case: ‘The Economy’ (Economy).

  • [qe-writing-008] — Remove excessive whitespace between words. Count: 56. Lines: 18, 20, 22, 24, 32, 34, 75, 87, 102, 125, …. Example: 3 spaces.

Medium severity#

  • [qe-code-001] (reviewer) — Follow PEP8 unless closer to mathematical notation. Count: 4. Lines: 313, 420, 969, 1395. Example: 74 code lines carry trailing whitespace (313, 420, 473, 531-533, 578, 620, 679, 716, 819-820, 831, 860, 905-906, 914-915, 922-924, 927-929, 933, 954, 968, 987-989, 995, 1015, 1023, 1029, 1039, 1065-1067, 1114-1116, 1150, 1153-1155, 1190-1191, 1225, 1228, 1232, 1239, 1373-1374, 1379, 1394-1395, 1409-1410, 1420, 1423-1425, 1441-1442, 1474-1475, 1641, 1656-1657, 1696, 1711, 1729, 1733, 1738); ten lines exceed 79 characters, the worst being 122 at line 1395 (fig.legend(...) with six keyword arguments on one line), 94 at 969 and 970, and 93 at 345; and the namedtuple continuation at 314 is indented to column 13 under a paren opened at column 22.

  • [qe-code-003] — Package installation at lecture top. Count: 1. Lines: 1. Example: non-Anaconda import with no install cell: [‘mpmath’].

  • [qe-math-001] — Prefer UTF-8 unicode for simple parameter mentions, be consistent. Count: 1. Lines: 1771. Example: LaTeX \mu outside math delimiters.

  • [qe-writing-001] — Use one sentence per paragraph. Count: 3. Lines: 1173, 1585, 1614. Example: 2 sentences in one paragraph.

  • [qe-writing-002] (reviewer) — Keep writing clear, concise, and valuable. Count: 3. Lines: 24, 65, 236. Example: line 24 is a 44-word sentence that also contains “with an sequences of several distorting flat-rate taxes”; lines 64-65 end one line with “to represent it as” and begin the next with “as”, so the word is duplicated across the break; line 236 reads “Let \(U_1 = \frac{\partial U}{\partial c}, U_2 = \ldots = -\frac{\partial U}{\partial n}.\), we can derive FOC from the Lagrangian” - a full stop inside the math followed by a comma, and a “Let …, we can derive” construction with no main clause.

  • [qe-writing-003] (reviewer) — Maintain logical flow. Count: 3. Lines: 302, 1101, 1609. Example: the note at 302-306 asks the reader to accept growth machinery 1200 lines before it is explained (“we include routines to handle the growth component, which will be discussed further in the section {ref}`growth_model`… to avoid code duplication”), so μ_t=1, μ_ss=None and A_path=None appear in every signature from 342 onward with no meaning attached; then line 1609 tells the reader the growth section’s routines are “in the section {ref}`cass_fiscal_shooting`”, which is not where a growth routine is - it is where the ordinary routines were pre-loaded with the growth arguments. Lines 1101-1103 are unreadable as written: “a foreseen in-” ends one line and “crease in \(\tau_{ct}\)” begins the next, and the parenthetical “(i.e., a decrease in \((1+\tau_{ct})\) \((1+\tau_{ct+1})\))” has lost the operator between the two factors.

  • [qe-writing-005] (reviewer) — Use bold for definitions, italic for emphasis. Count: 3. Lines: 164, 405, 1094. Example: the definition at 164-166 bolds four separate phrases in one sentence, articles included - budget-feasible government policy, a feasible allocation, a price system on top of the term being defined - so more of the sentence is bold than is not. The eight quantity labels at 405-501 (Price:, Capital rental rate, Labor rental rate:, Gross one-period return on capital:, One-period discount factor:, Net one-period rate of interest:) use italic as pseudo-headings. And the six experiment headers at 1094, 1141, 1177, 1364, 1400 and 1414 are bold pseudo-headings, while the growth section’s equivalents at 1625 and 1682 are real H4s - the same structural element formatted two different ways in one file.

Low severity#

  • [qe-writing-007] (reviewer) — Use visual elements to enhance understanding. Count: 1. Lines: 600. Example: the shooting algorithm is stated as six steps (600-612) whose whole content is a search over \(c_0\) - “Compute the difference \(\hat k_S - \bar k\)… adjust \(c_0\) and repeat”, “Adjust \(c_0\) iteratively using the bisection method” - and the family of candidate capital paths that bracket \(\bar k\), some diverging up and some crashing to zero, is never plotted. That figure is the one that makes shooting comprehensible, and it is also what would show the reader why mpmath is needed at all (line 300: “in cases where the solution diverges due to numerical instability”). All six figures in the lecture are post-solution transition panels.

Strengths#

  • Two independent solution methods are implemented for the same model - the shooting algorithm (616-700) and residual minimisation (1225-1341) - and the exercises ask the reader to replicate all four experiments with the second method, so the agreement between them is demonstrated rather than claimed.

  • Every equilibrium object gets its own labelled equation and a matching one-line implementation, in order: \(q_t\) at {eq}`eq:equil_q` then compute_q_path, \(\eta_t\) then compute_η_path, \(w_t\) then compute_w_path, \(\bar R\) at {eq}`eq:gross_rate` then compute_R_bar, \(r_{t,t+s}\) then compute_rts_path (405-537).

  • The no-arbitrage restriction is derived rather than imposed: 174-201 rearranges the household budget constraint to isolate the coefficient on \(k_t\), then argues both directions (buy unlimited capital if positive, short-sell synthetic capital if negative) before concluding the coefficient must vanish.

  • Each experiment’s economics is narrated as a dated sequence with the mechanism at each date - the \(\tau_c\) experiment at 1123-1137 walks through \(t=0\) (anticipatory jump and consumption binge), \(t \in (0, T)\) (falling capital raising \(\bar R\)), \(t = T\) (the jump depressing \(\bar R\) below 1) and \(t > T\) (austerity and gradual recovery).

  • The growth extension reuses the whole apparatus by rescaling to effective labour units (1525-1607), and each modified equation is presented explicitly as “the counterpart to” its constant-technology original ({eq}`eq:feasi_capital` → {eq}`eq:feasi_mod`, {eq}`eq:diff_second` → {eq}`eq:diff_mod`, {eq}`eq:consume_R` → its modified form).