Goulet-Coulombe (2021) — bundled paper baseline

Status: bundled paper-baseline replication. The recipe runs end-to-end on a stock install with bit-exact replicate. Real FRED-MD figure replication requires swapping the bundled 36-month sample panel for the official 1960-onwards FRED-MD vintage (one-line edit in the recipe; see “Real FRED-MD” section).

This page documents what the bundled examples/recipes/goulet_coulombe_2021_replication.yaml recipe does today. It is not a full paper replication on its own — the bundled panel is a 36-month sample (1960-1962) for smoke testing — but the recipe shape, layer choices, and replicate contract are paper-faithful.

Paper

Title: “The Macroeconomy as a Random Forest”
Author: Goulet-Coulombe, P. (2021)
Source: Working paper / JAE 2024
Recipe ports the ridge baseline of the paper’s FRED-MD INDPRO forecasting horse race. The full MRF / GTVP route lives at examples/recipes/l4_mrf_placeholder.yaml.

Recipe shape

Layer	Choice
L0	`failure_policy=fail_fast`, `reproducibility_mode=seeded_reproducible`, `random_seed=42`
L1	`target=INDPRO`, `target_horizons=[1, 3, 6, 12]`, `information_set_type=final_revised_data`, custom_panel_inline (smoke; see below)
L2	McCracken-Ng official tcodes (`apply_official_tcode`)
L3	Lag features + `target_construction(direct, h)`
L4	Ridge (`alpha=1.0`) + AR(BIC) benchmark, expanding-window walk-forward, `refit_policy=every_origin`, `forecast_strategy=direct`
L5	MSE primary + relative metrics (vs AR benchmark)
L6	DM (HLN-corrected) horse-race test, `dependence_correction=newey_west`
L8	Manifest + per-cell artifacts

Today’s reproduce + replicate

macroforecast run examples/recipes/goulet_coulombe_2021_replication.yaml -o out/gc2021
macroforecast replicate out/gc2021/manifest.json

Verified on v0.8.6 (server1, 2026-05-06):

run:        1.40s, cells=1, all_ok=True
replicate:  recipe_match=True, sink_hashes_match=True

cell sink hashes (deterministic across runs):
  l1_data_definition_v1:   d164d15be5b3135c
  l1_regime_metadata_v1:   41cb7c29ec77d7c1
  l2_clean_panel_v1:       effec065444a0ac4
  l3_features_v1:          139853d1515cab8a
  l3_metadata_v1:          7c2368d756c4a66a
  l4_forecasts_v1:         395df3d0f5979dac
  l4_model_artifacts_v1:   0a2b7f63e1df2a31
  l4_training_metadata_v1: 62abd040e502ebdd
  l5_evaluation_v1:        f5ac78c21ae68f7e
  l6_tests_v1:             bc383c20f93143dd

Smoke regression: covered by tests/test_examples_smoke.py::test_recipe_runs_end_to_end[goulet_coulombe_2021_replication.yaml].

Real FRED-MD (full paper replication)

To replicate the actual paper figures (1960-onwards walk-forward over the full FRED-MD vintage), swap two YAML keys:

 1_data:
   fixed_axes:
-    custom_source_policy: custom_panel_only
+    custom_source_policy: official_only
+    dataset: fred_md
     frequency: monthly
     horizon_set: custom_list
     ...
   leaf_config:
     target: INDPRO
     target_horizons: [1, 3, 6, 12]
-    custom_panel_inline:
-      ...  # 36-month smoke sample
+    sample_start_rule: max_balanced
+    sample_end_rule: latest_available

After the swap, the L1 loader pulls real FRED-MD via the vintage manager (see docs/architecture/layer1/index.md for vintage configuration) and the recipe runs the actual paper horse race.

To extend to MRF (the paper’s headline result), add another L4 fit node with family: macroeconomic_random_forest alongside the ridge (is_benchmark: true). See encyclopedia L4 for the full family list.

What this page proves vs. what is still TBD

✅ Recipe shape is paper-faithful at the schema level (L0/L1/L2/L3/L4/L5/L6 axes match the paper’s experimental design).
✅ End-to-end runtime is wired: every layer materialises, every sink hashes deterministically, replicate bit-exact succeeds.
✅ Stock-install runnable: no extras (pip install macroforecast is enough).
⏳ Full paper figures from real FRED-MD: requires the one-line swap above plus a real vintage configured. Not gated by this page; the user runs it.
⏳ Numerical match against the paper’s reported MSEs: requires the real-FRED run plus access to the paper’s appendix tables for cross-check.