{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8d3c15f9",
   "metadata": {},
   "source": [
    "# Dynamic Panel Data Models in Python: From Nickell Bias to System GMM\n",
    "\n",
    "This notebook is a runnable companion to the blog post [*Dynamic Panel Data Models in Python: From Nickell Bias to System GMM*](https://carlos-mendez.org/post/python_dynamic_panel/).\n",
    "\n",
    "When this year's outcome depends on last year's outcome — employment, debt, capital, habits — ordinary panel methods break in ways that are invisible on the printed output. Using the classic Arellano and Bond (1991) panel of 140 UK manufacturing firms observed 1976–1984, we estimate the persistence $\\rho$ of firm employment with the full estimator ladder: pooled OLS gives 0.962 (biased up), fixed effects gives 0.626 (Nickell bias, down), Anderson-Hsiao IV is consistent but useless (1.233, SE 0.478), and difference GMM hugs the biased bound at 0.679. Blundell-Bond system GMM with 32 collapsed instruments delivers the defensible headline — $\\hat{\\rho} = 0.927$, clean AR(2) and Hansen diagnostics — and the toolchain replicates the published `pydynpd` benchmark digit for digit."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a42a63d5",
   "metadata": {},
   "source": [
    "**Learning objectives:**\n",
    "\n",
    "- Understand why a lagged dependent variable plus a fixed effect breaks both pooled OLS (bias up) and the within estimator (Nickell bias, down), and how the two wrong answers bracket the truth.\n",
    "- Implement the full estimator ladder in Python — OLS and fixed effects with `pyfixest`, Anderson-Hsiao IV, and difference and system GMM with `pydynpd`.\n",
    "- Estimate employment persistence on the classic Arellano-Bond (1991) UK panel and diagnose weak instruments using Bond's (2002) bracket check.\n",
    "- Assess GMM credibility with the AR(1)/AR(2) serial-correlation tests and the Hansen overidentification test, including the counterintuitive \"p close to 1 is a red flag\" reading.\n",
    "- Compare instrument-proliferation choices (lag windows, collapsing) and verify the toolchain against the package's published replication benchmark."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c84a357",
   "metadata": {},
   "source": [
    "## Setup and imports\n",
    "\n",
    "We need two workhorse libraries: [`pyfixest`](https://py-econometrics.github.io/pyfixest/pyfixest.html) for the OLS, fixed-effects, and IV benchmarks, and [`pydynpd`](https://github.com/dazhwu/pydynpd) (Wu, Hua and Xu 2023) for difference and system GMM, with a command syntax deliberately close to Stata's `xtabond2`. The version pin matters: the compatibility fix in the next cell targets exactly the 0.2.2 release."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "52cbd9ae",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:43.480308Z",
     "iopub.status.busy": "2026-06-11T09:01:43.480016Z",
     "iopub.status.idle": "2026-06-11T09:01:44.073845Z",
     "shell.execute_reply": "2026-06-11T09:01:44.073203Z"
    }
   },
   "outputs": [],
   "source": [
    "!pip install pydynpd==0.2.2 pyfixest --quiet"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8c7fbbe",
   "metadata": {},
   "source": [
    "One practical wrinkle deserves a friendly explanation rather than a silent workaround. `pydynpd` 0.2.2 was written before NumPy 2.0, which removed the alias `np.in1d` and stopped allowing `float()` and `math.sqrt()` to be called directly on 1x1 matrices. Rather than downgrading NumPy or forking the package, we apply a six-line *compatibility shim*: restore the `np.in1d` alias, and inject tiny wrapper functions into the one `pydynpd` module that does the offending scalar conversions (`specification_tests`). Because Python module globals shadow builtins, the injected `float` and `math.sqrt` wrappers are picked up only inside that module — the rest of the session is untouched. The digit-for-digit replication check near the end of this notebook confirms the shim does not perturb any estimate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "f1c6ce7c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:44.076070Z",
     "iopub.status.busy": "2026-06-11T09:01:44.075935Z",
     "iopub.status.idle": "2026-06-11T09:01:45.476199Z",
     "shell.execute_reply": "2026-06-11T09:01:45.475641Z"
    }
   },
   "outputs": [],
   "source": [
    "import contextlib\n",
    "import io\n",
    "import math\n",
    "import types\n",
    "import warnings\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# -- pydynpd 0.2.2 / NumPy 2.x compatibility shim ---------------------\n",
    "# pydynpd 0.2.2 predates NumPy 2.0, which removed np.in1d and forbids\n",
    "# float()/math.sqrt() on 1x1 matrices. Module globals shadow builtins,\n",
    "# so injecting wrappers into pydynpd.specification_tests restores both\n",
    "# behaviors without forking the package.\n",
    "if not hasattr(np, \"in1d\"):\n",
    "    np.in1d = np.isin\n",
    "import pydynpd\n",
    "\n",
    "if getattr(pydynpd, \"__version__\", \"0.2.2\") not in (\"0.2.2\",):\n",
    "    warnings.warn(\"Compat shim was written for pydynpd 0.2.2 - \"\n",
    "                  \"re-test before trusting results on a newer version\")\n",
    "from pydynpd import specification_tests as _st\n",
    "\n",
    "\n",
    "def _scalar(v):\n",
    "    return np.asarray(v).item() if np.ndim(v) else v\n",
    "\n",
    "\n",
    "_st.float = lambda v: float(_scalar(v))\n",
    "_st.math = types.SimpleNamespace(sqrt=lambda v: math.sqrt(_scalar(v)))\n",
    "\n",
    "from pydynpd import regression  # import after shim\n",
    "import pyfixest as pf"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7d680a53",
   "metadata": {},
   "source": [
    "Next, the configuration block: the random seed (used only to pick which firms appear in Figure 1 — every estimator below is closed-form and deterministic), the site color palette, the dark-theme matplotlib settings, and two strings that define our *running specification*. `SPEC_MAIN` is the pydynpd formula for the AR(1) labor-demand model — `n` on its first lag plus current and lagged `w` and `k` — and `GMM_FULL` declares that all lags from $t-2$ back to the start of the sample (`2:99`) of `n`, `w`, and `k` are available as GMM-style instruments."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bed4b181",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:45.477690Z",
     "iopub.status.busy": "2026-06-11T09:01:45.477556Z",
     "iopub.status.idle": "2026-06-11T09:01:45.508720Z",
     "shell.execute_reply": "2026-06-11T09:01:45.508199Z"
    }
   },
   "outputs": [],
   "source": [
    "RANDOM_SEED = 42\n",
    "np.random.seed(RANDOM_SEED)\n",
    "\n",
    "# Site color palette\n",
    "STEEL_BLUE = \"#6a9bcc\"\n",
    "WARM_ORANGE = \"#d97757\"\n",
    "NEAR_BLACK = \"#141413\"\n",
    "TEAL = \"#00d4c8\"\n",
    "GRAY = \"#999999\"\n",
    "\n",
    "# Dark theme palette\n",
    "DARK_NAVY = \"#0f1729\"\n",
    "GRID_LINE = \"#1f2b5e\"\n",
    "LIGHT_TEXT = \"#c8d0e0\"\n",
    "WHITE_TEXT = \"#e8ecf2\"\n",
    "\n",
    "plt.rcParams.update({\n",
    "    \"figure.facecolor\": DARK_NAVY,\n",
    "    \"axes.facecolor\": DARK_NAVY,\n",
    "    \"axes.edgecolor\": DARK_NAVY,\n",
    "    \"axes.linewidth\": 0,\n",
    "    \"axes.labelcolor\": LIGHT_TEXT,\n",
    "    \"axes.titlecolor\": WHITE_TEXT,\n",
    "    \"axes.spines.top\": False,\n",
    "    \"axes.spines.right\": False,\n",
    "    \"axes.spines.left\": False,\n",
    "    \"axes.spines.bottom\": False,\n",
    "    \"axes.grid\": True,\n",
    "    \"grid.color\": GRID_LINE,\n",
    "    \"grid.linewidth\": 0.6,\n",
    "    \"grid.alpha\": 0.8,\n",
    "    \"xtick.color\": LIGHT_TEXT,\n",
    "    \"ytick.color\": LIGHT_TEXT,\n",
    "    \"xtick.major.size\": 0,\n",
    "    \"ytick.major.size\": 0,\n",
    "    \"text.color\": WHITE_TEXT,\n",
    "    \"font.size\": 12,\n",
    "    \"legend.frameon\": False,\n",
    "    \"legend.fontsize\": 11,\n",
    "    \"legend.labelcolor\": LIGHT_TEXT,\n",
    "    \"figure.edgecolor\": DARK_NAVY,\n",
    "    \"savefig.facecolor\": DARK_NAVY,\n",
    "    \"savefig.edgecolor\": DARK_NAVY,\n",
    "})\n",
    "\n",
    "VAR_LABELS = {\n",
    "    \"n\": \"Log employment\",\n",
    "    \"w\": \"Log real wage\",\n",
    "    \"k\": \"Log capital stock\",\n",
    "    \"ys\": \"Log industry output\",\n",
    "}\n",
    "\n",
    "SPEC_MAIN = \"n L(1:1).n L(0:1).w L(0:1).k\"\n",
    "GMM_FULL = \"gmm(n, 2:99) gmm(w, 2:99) gmm(k, 2:99)\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ffe97e95",
   "metadata": {},
   "source": [
    "Finally, two small helpers we will reuse constantly. `run_abond` wraps `pydynpd.regression.abond` — the package's single entry point, which takes a Stata-style command string, the dataframe, and the panel identifiers `[\"id\", \"year\"]` — and optionally swallows the table it prints (useful in the proliferation grid, where we run six models and want a readable log). `gmm_summary` pulls the headline numbers out of a fitted model: $\\hat{\\rho}$ and its standard error, a 95 percent confidence interval, the Hansen and AR-test p-values, and the instrument count."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0b7f6438",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:45.510054Z",
     "iopub.status.busy": "2026-06-11T09:01:45.509968Z",
     "iopub.status.idle": "2026-06-11T09:01:45.513246Z",
     "shell.execute_reply": "2026-06-11T09:01:45.512821Z"
    }
   },
   "outputs": [],
   "source": [
    "def run_abond(command_str, df, quiet=False):\n",
    "    \"\"\"Run pydynpd and return the first fitted model.\n",
    "\n",
    "    pydynpd prints its regression table to stdout as a side effect; quiet=True\n",
    "    suppresses that (used in the proliferation grid to keep the log readable).\n",
    "    \"\"\"\n",
    "    if quiet:\n",
    "        with contextlib.redirect_stdout(io.StringIO()):\n",
    "            return regression.abond(command_str, df, [\"id\", \"year\"]).models[0]\n",
    "    return regression.abond(command_str, df, [\"id\", \"year\"]).models[0]\n",
    "\n",
    "\n",
    "def gmm_summary(model, label):\n",
    "    \"\"\"Extract headline numbers from a fitted pydynpd model.\"\"\"\n",
    "    rt = model.regression_table\n",
    "    rho = rt.loc[rt.variable == \"L1.n\", \"coefficient\"].iloc[0]\n",
    "    se = rt.loc[rt.variable == \"L1.n\", \"std_err\"].iloc[0]\n",
    "    return {\n",
    "        \"estimator\": label,\n",
    "        \"rho1\": rho,\n",
    "        \"se\": se,\n",
    "        \"ci_lo\": rho - 1.96 * se,\n",
    "        \"ci_hi\": rho + 1.96 * se,\n",
    "        \"hansen_p\": model.hansen.p_value,\n",
    "        \"ar1_p\": model.AR_list[0].P_value,\n",
    "        \"ar2_p\": model.AR_list[1].P_value,\n",
    "        \"n_instruments\": model.z_information.num_instr,\n",
    "    }"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2ad5a61f",
   "metadata": {},
   "source": [
    "## Data loading and panel structure\n",
    "\n",
    "The dataset is the original Arellano and Bond (1991) panel: an unbalanced sample of UK manufacturing firms observed annually from 1976 to 1984 — the canonical teaching dataset of this literature, so every number we produce can be checked against forty years of published output. The columns we use are already in logs: `n` (employment), `w` (real wage), `k` (gross capital), and `ys` (industry output). We load it from the blog's GitHub repository, falling back to a local copy if the notebook runs next to the post bundle.\n",
    "\n",
    "Before estimating anything, we want three facts: how big the panel is, how unbalanced it is, and — most importantly — *where the variation lives* (between firms or within firms over time)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "0c878e75",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:45.514319Z",
     "iopub.status.busy": "2026-06-11T09:01:45.514245Z",
     "iopub.status.idle": "2026-06-11T09:01:45.915379Z",
     "shell.execute_reply": "2026-06-11T09:01:45.914743Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset shape: (1031, 10)\n",
      "Firms: 140, years: 1976-1984\n",
      "\n",
      "Observations per firm (unbalanced panel):\n",
      "                n_firms\n",
      "years_observed         \n",
      "7                   103\n",
      "8                    23\n",
      "9                    14\n",
      "\n",
      "Summary statistics (log variables used in estimation):\n",
      "              n         w         k        ys\n",
      "count  1031.000  1031.000  1031.000  1031.000\n",
      "mean      1.056     3.143    -0.442     4.638\n",
      "std       1.342     0.263     1.514     0.094\n",
      "min      -2.263     2.082    -4.431     4.465\n",
      "25%       0.166     3.027    -1.510     4.576\n",
      "50%       0.827     3.178    -0.658     4.611\n",
      "75%       1.949     3.314     0.406     4.706\n",
      "max       4.687     3.812     3.852     4.855\n",
      "\n",
      "Between-firm SD of log employment: 1.339\n",
      "Within-firm SD of log employment:  0.195\n"
     ]
    }
   ],
   "source": [
    "URL = \"https://raw.githubusercontent.com/cmg777/starter-academic-v501/master/content/post/python_dynamic_panel/abdata.csv\"\n",
    "try:\n",
    "    df = pd.read_csv(URL)\n",
    "except Exception:\n",
    "    df = pd.read_csv(\"abdata.csv\")  # local fallback (post bundle)\n",
    "\n",
    "print(f\"Dataset shape: {df.shape}\")\n",
    "print(f\"Firms: {df['id'].nunique()}, years: {df['year'].min()}-{df['year'].max()}\")\n",
    "\n",
    "obs_per_firm = df.groupby(\"id\").size()\n",
    "print(\"\\nObservations per firm (unbalanced panel):\")\n",
    "print(obs_per_firm.value_counts().sort_index().rename_axis(\"years_observed\")\n",
    "      .to_frame(\"n_firms\").to_string())\n",
    "\n",
    "print(\"\\nSummary statistics (log variables used in estimation):\")\n",
    "print(df[[\"n\", \"w\", \"k\", \"ys\"]].describe().round(3).to_string())\n",
    "\n",
    "firm_mean_n = df.groupby(\"id\")[\"n\"].transform(\"mean\")\n",
    "print(f\"\\nBetween-firm SD of log employment: {df.groupby('id')['n'].mean().std():.3f}\")\n",
    "print(f\"Within-firm SD of log employment:  {(df['n'] - firm_mean_n).std():.3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e78754a7",
   "metadata": {},
   "source": [
    "**Interpretation.** The panel holds 1,031 firm-year observations on 140 firms — *short and wide*: many firms (large N) observed for only 7 to 9 years each (small T). That geometry is exactly the setting dynamic-panel GMM was designed for, and exactly where Nickell bias — which shrinks at rate $1/T$ — bites hardest. The decisive number is the variance decomposition: the between-firm SD of log employment (1.339) is nearly **seven times** the within-firm SD (0.195). Employment differences live almost entirely *across* firms, so the unobserved firm level $\\alpha_i$ is the dominant feature of these data — and any estimator that mishandles it will be wrong by a lot, not a little.\n",
    "\n",
    "A picture makes the same point more vividly: 40 randomly chosen firms (the only place the seed is used), the median across all 140 firms, and one example firm."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "43eaf7e2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:45.917136Z",
     "iopub.status.busy": "2026-06-11T09:01:45.917004Z",
     "iopub.status.idle": "2026-06-11T09:01:46.013141Z",
     "shell.execute_reply": "2026-06-11T09:01:46.012644Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x550 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rng = np.random.default_rng(RANDOM_SEED)\n",
    "sample_ids = rng.choice(df[\"id\"].unique(), size=40, replace=False)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(9, 5.5))\n",
    "fig.patch.set_linewidth(0)\n",
    "for fid in sample_ids:\n",
    "    firm = df[df[\"id\"] == fid].sort_values(\"year\")\n",
    "    ax.plot(firm[\"year\"], firm[\"n\"], color=STEEL_BLUE, alpha=0.35, lw=1.2)\n",
    "median_path = df.groupby(\"year\")[\"n\"].median()\n",
    "ax.plot(median_path.index, median_path.values, color=WARM_ORANGE, lw=3,\n",
    "        label=\"Median firm (all 140)\")\n",
    "big = df[df[\"id\"] == obs_per_firm.idxmax()].sort_values(\"year\")\n",
    "ax.plot(big[\"year\"], big[\"n\"], color=TEAL, lw=2.2, label=\"One example firm\")\n",
    "ax.set_xlabel(\"Year\")\n",
    "ax.set_ylabel(VAR_LABELS[\"n\"])\n",
    "ax.set_title(\"Firm employment paths are persistent and parallel-ish:\\n\"\n",
    "             \"each firm orbits its own level - a firm fixed effect\",\n",
    "             fontsize=13)\n",
    "ax.legend(loc=\"upper right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4c7ecb7d",
   "metadata": {},
   "source": [
    "**Interpretation.** Each blue line is one firm, and the picture shows the two ingredients that make $\\rho$ hard to estimate, simultaneously. First, the lines are roughly *parallel* and rarely cross: each firm orbits its own level — the visual signature of a large firm fixed effect $\\alpha_i$. Second, each line is *smooth*: employment this year looks a lot like last year — the visual signature of high persistence. The orange median drifts gently downward after 1980 (the early-1980s UK manufacturing recession, a common shock the year dummies will absorb). An estimator that ignores $\\alpha_i$ will mistake orbit differences for persistence; one that removes $\\alpha_i$ clumsily will damage the persistence signal it is trying to measure."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "10c7517c",
   "metadata": {},
   "source": [
    "## Data preparation: lags and first differences\n",
    "\n",
    "Every estimator below runs on transformed variables: one-period lags of `n`, `w`, and `k` for the level equations, and first differences for the Anderson-Hsiao regression. Two details matter. The transformations must respect firm boundaries — which is why everything goes through `groupby(\"id\")` — and each lag is *expensive*: it costs every firm its first observed year."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "d38593d4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:46.014448Z",
     "iopub.status.busy": "2026-06-11T09:01:46.014357Z",
     "iopub.status.idle": "2026-06-11T09:01:46.022175Z",
     "shell.execute_reply": "2026-06-11T09:01:46.021563Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Full panel rows: 1031\n",
      "Estimation sample after requiring one lag: 891 rows (140 firms)\n"
     ]
    }
   ],
   "source": [
    "d = df.sort_values([\"id\", \"year\"]).copy()\n",
    "g = d.groupby(\"id\")\n",
    "for v in [\"n\", \"w\", \"k\"]:\n",
    "    d[f\"{v}_lag1\"] = g[v].shift(1)\n",
    "d[\"n_lag2\"] = g[\"n\"].shift(2)\n",
    "for v in [\"n\", \"w\", \"k\"]:\n",
    "    d[f\"d_{v}\"] = g[v].diff()\n",
    "d[\"d_n_lag1\"] = g[\"d_n\"].shift(1)\n",
    "d[\"d_w_lag1\"] = g[\"d_w\"].shift(1)\n",
    "d[\"d_k_lag1\"] = g[\"d_k\"].shift(1)\n",
    "\n",
    "est_sample = d.dropna(subset=[\"n_lag1\", \"w_lag1\", \"k_lag1\"]).copy()\n",
    "print(f\"Full panel rows: {len(d)}\")\n",
    "print(f\"Estimation sample after requiring one lag: {len(est_sample)} rows \"\n",
    "      f\"({est_sample['id'].nunique()} firms)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dac58014",
   "metadata": {},
   "source": [
    "**Interpretation.** Requiring a single lag shrinks the sample from 1,031 to 891 rows — a 13.6 percent cut — while keeping all 140 firms. This is the first lesson in dynamic-panel frugality: with T as small as 7, every additional lag or difference burns a meaningful share of the data. Watch the observation counts fall as the methods get hungrier — the GMM estimators below run on 751 observations, and the two-lag replication specification at the end on just 611."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b76a66dd",
   "metadata": {},
   "source": [
    "## The bias bracket: pooled OLS vs fixed effects\n",
    "\n",
    "We now run the *same regression* twice — log employment on its lag, wages, capital, and year dummies — changing only how it treats the firm effect $\\alpha_i$. Both treatments fail, but in *opposite, theoretically known directions*, and that is what makes the failure useful.\n",
    "\n",
    "**Why pooled OLS is biased upward.** Pooled OLS ignores $\\alpha_i$, leaving it in the error term. But last year's employment $n_{i,t-1}$ depends on $\\alpha_i$ — high-level firms had high employment last year too — so the regressor is positively correlated with the error. Persistently large firms look like firms with enormous persistence, and $\\hat{\\rho}$ gets pushed *up* toward 1.\n",
    "\n",
    "**Why fixed effects is biased downward.** The within estimator subtracts each firm's sample average, which removes $\\alpha_i$ exactly — but that average contains the firm's *future* shocks, creating a mechanical *negative* correlation between the demeaned lag and the demeaned error: the Nickell (1981) bias, of order $1/T$. With T of 7–9 it is large, and adding more firms does not help.\n",
    "\n",
    "Two failures with known signs are a measurement instrument in their own right: Bond (2002) turned them into a diagnostic *bracket* that any consistent estimator must land inside. In the `pf.feols` calls, `| year` absorbs year fixed effects (`| id + year` absorbs both), and `vcov={\"CRV1\": \"id\"}` requests cluster-robust standard errors by firm."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "7a925f4c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:46.023434Z",
     "iopub.status.busy": "2026-06-11T09:01:46.023334Z",
     "iopub.status.idle": "2026-06-11T09:01:55.632406Z",
     "shell.execute_reply": "2026-06-11T09:01:55.631954Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pooled OLS (year dummies, SEs clustered by firm):\n",
      "             Estimate  Std. Error   t value  Pr(>|t|)    2.5%   97.5%\n",
      "Coefficient                                                          \n",
      "n_lag1         0.9617      0.0084  115.0717    0.0000  0.9452  0.9782\n",
      "w             -0.4147      0.1600   -2.5915    0.0106 -0.7311 -0.0983\n",
      "w_lag1         0.3556      0.1559    2.2803    0.0241  0.0473  0.6639\n",
      "k              0.3997      0.0565    7.0710    0.0000  0.2879  0.5114\n",
      "k_lag1        -0.3675      0.0565   -6.4990    0.0000 -0.4793 -0.2557\n",
      "\n",
      "Fixed effects / within (firm + year dummies, clustered SEs):\n",
      "             Estimate  Std. Error  t value  Pr(>|t|)    2.5%   97.5%\n",
      "Coefficient                                                         \n",
      "n_lag1         0.6262      0.0515  12.1510    0.0000  0.5243  0.7281\n",
      "w             -0.5035      0.1450  -3.4729    0.0007 -0.7902 -0.2169\n",
      "w_lag1         0.2308      0.1077   2.1420    0.0339  0.0178  0.4438\n",
      "k              0.4078      0.0566   7.2024    0.0000  0.2959  0.5198\n",
      "k_lag1        -0.1648      0.0547  -3.0100    0.0031 -0.2730 -0.0565\n",
      "\n",
      "  rho_OLS = 0.9617 (se 0.0084)   <- upper bound (biased up)\n",
      "  rho_FE  = 0.6262 (se 0.0515)   <- lower bound (biased down)\n"
     ]
    }
   ],
   "source": [
    "FORMULA_RHS = \"n_lag1 + w + w_lag1 + k + k_lag1\"\n",
    "ols = pf.feols(f\"n ~ {FORMULA_RHS} | year\", data=est_sample,\n",
    "               vcov={\"CRV1\": \"id\"})\n",
    "fe = pf.feols(f\"n ~ {FORMULA_RHS} | id + year\", data=est_sample,\n",
    "              vcov={\"CRV1\": \"id\"})\n",
    "\n",
    "print(\"Pooled OLS (year dummies, SEs clustered by firm):\")\n",
    "print(ols.tidy().round(4).to_string())\n",
    "print(\"\\nFixed effects / within (firm + year dummies, clustered SEs):\")\n",
    "print(fe.tidy().round(4).to_string())\n",
    "\n",
    "rho_ols, se_ols = ols.coef()[\"n_lag1\"], ols.se()[\"n_lag1\"]\n",
    "rho_fe, se_fe = fe.coef()[\"n_lag1\"], fe.se()[\"n_lag1\"]\n",
    "print(f\"\\n  rho_OLS = {rho_ols:.4f} (se {se_ols:.4f})   <- upper bound (biased up)\")\n",
    "print(f\"  rho_FE  = {rho_fe:.4f} (se {se_fe:.4f})   <- lower bound (biased down)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "beb85f15",
   "metadata": {},
   "source": [
    "**Interpretation.** The two \"wrong\" estimators disagree by 0.336 — an enormous gap in economic terms. Pooled OLS says $\\hat{\\rho} = 0.9617$, so close to a unit root that an employment shock would have a half-life of about 18 years; fixed effects says $\\hat{\\rho} = 0.6262$, a half-life of about 1.5 years. Same data, same regression, opposite stories — and *neither is right*, but both errors have known sign. The bracket [0.626, 0.962] is sharply identified: the two cluster-robust confidence intervals do not even overlap. The bracket deserves its own picture, because it is the yardstick every later estimate will be measured against."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "1a2de52a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:55.633964Z",
     "iopub.status.busy": "2026-06-11T09:01:55.633857Z",
     "iopub.status.idle": "2026-06-11T09:01:55.721489Z",
     "shell.execute_reply": "2026-06-11T09:01:55.721137Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(9, 4.8))\n",
    "fig.patch.set_linewidth(0)\n",
    "ax.axvspan(rho_fe, rho_ols, color=GRID_LINE, alpha=0.55, zorder=0)\n",
    "ax.axvline(1.0, color=GRAY, lw=1.2, ls=\"--\", alpha=0.8)\n",
    "ax.text(1.0, 1.62, \"unit root\", color=GRAY, fontsize=10, ha=\"center\")\n",
    "for y, (lab, rho, se, col, note) in enumerate([\n",
    "    (\"Pooled OLS\", rho_ols, se_ols, STEEL_BLUE,\n",
    "     \"biased UP: L1.n correlated with firm effect\"),\n",
    "    (\"Fixed effects\", rho_fe, se_fe, WARM_ORANGE,\n",
    "     \"biased DOWN: Nickell bias (T is small)\"),\n",
    "]):\n",
    "    ax.errorbar(rho, y, xerr=1.96 * se, fmt=\"o\", color=col, ms=10,\n",
    "                capsize=5, lw=2.5, capthick=2.5)\n",
    "    ax.text(rho, y - 0.28, note, color=LIGHT_TEXT, fontsize=10, ha=\"center\")\n",
    "ax.text((rho_fe + rho_ols) / 2, 1.18,\n",
    "        \"consistent estimates\\nshould land in here\", color=WHITE_TEXT,\n",
    "        fontsize=11, ha=\"center\", style=\"italic\")\n",
    "ax.set_yticks([0, 1])\n",
    "ax.set_yticklabels([\"Pooled OLS\", \"Fixed effects\"])\n",
    "ax.set_ylim(-0.6, 1.8)\n",
    "ax.set_xlabel(r\"Estimate of employment persistence  $\\hat{\\rho}$  (L1.n)\")\n",
    "ax.set_title(\"Two wrong answers that bracket the truth\", fontsize=13)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f7bd02ea",
   "metadata": {},
   "source": [
    "**Interpretation.** The shaded band between 0.626 and 0.962 is the playing field for the rest of the notebook. Bond's (2002) rule: a candidate estimator landing *below* the band is suffering something Nickell-like; one landing *above* it is suffering something OLS-like or worse; and one landing just barely inside the band, hugging an edge, is probably being dragged toward that edge by a fixable weakness. With the yardstick built, we can try the first estimator that is actually *consistent* for $\\rho$."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c02311ff",
   "metadata": {},
   "source": [
    "## Anderson-Hsiao IV: consistent but imprecise\n",
    "\n",
    "Anderson and Hsiao (1981) proposed a two-step escape. **Step one: difference away the firm effect.** First-differencing eliminates $\\alpha_i$ exactly — but it makes the differenced lag $\\Delta n_{i,t-1}$ endogenous, because it contains $\\varepsilon_{i,t-1}$, which also lives inside $\\Delta\\varepsilon_{it}$. **Step two: instrument the differenced lag** with the *level* $n_{i,t-2}$: it helps predict the change $\\Delta n_{i,t-1}$, and — provided $\\varepsilon_{it}$ is not serially correlated — it predates and is therefore independent of both shocks inside $\\Delta\\varepsilon_{it}$. We estimate by 2SLS, which `pyfixest` expresses with a third formula part: `d_n_lag1 ~ n_lag2`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "6122a48b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:55.722810Z",
     "iopub.status.busy": "2026-06-11T09:01:55.722720Z",
     "iopub.status.idle": "2026-06-11T09:01:56.708873Z",
     "shell.execute_reply": "2026-06-11T09:01:56.708397Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Anderson-Hsiao 2SLS (differences, year dummies, clustered SEs):\n",
      "             Estimate  Std. Error  t value  Pr(>|t|)    2.5%   97.5%\n",
      "Coefficient                                                         \n",
      "d_w           -0.5243      0.2135  -2.4556    0.0153 -0.9465 -0.1021\n",
      "d_w_lag1       0.5808      0.3128   1.8566    0.0655 -0.0377  1.1992\n",
      "d_k            0.2463      0.0777   3.1693    0.0019  0.0927  0.4000\n",
      "d_k_lag1      -0.2925      0.1964  -1.4890    0.1388 -0.6808  0.0959\n",
      "d_n_lag1       1.2327      0.4782   2.5781    0.0110  0.2873  2.1781\n",
      "\n",
      "  rho_AH = 1.2327 (se 0.4782)\n",
      "  95% CI: [0.296, 2.170]\n"
     ]
    }
   ],
   "source": [
    "ah_sample = d.dropna(subset=[\"d_n\", \"d_n_lag1\", \"d_w\", \"d_w_lag1\",\n",
    "                             \"d_k\", \"d_k_lag1\", \"n_lag2\"]).copy()\n",
    "ah = pf.feols(\"d_n ~ d_w + d_w_lag1 + d_k + d_k_lag1 | year | d_n_lag1 ~ n_lag2\",\n",
    "              data=ah_sample, vcov={\"CRV1\": \"id\"})\n",
    "print(\"Anderson-Hsiao 2SLS (differences, year dummies, clustered SEs):\")\n",
    "print(ah.tidy().round(4).to_string())\n",
    "\n",
    "rho_ah, se_ah = ah.coef()[\"d_n_lag1\"], ah.se()[\"d_n_lag1\"]\n",
    "print(f\"\\n  rho_AH = {rho_ah:.4f} (se {se_ah:.4f})\")\n",
    "print(f\"  95% CI: [{rho_ah - 1.96 * se_ah:.3f}, {rho_ah + 1.96 * se_ah:.3f}]\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e3f34e95",
   "metadata": {},
   "source": [
    "**Interpretation.** The point estimate is $\\hat{\\rho} = 1.2327$ — *above* the unit root, outside the bracket — with a standard error of 0.4782, nearly 60 times the OLS standard error. The 95 percent confidence interval [0.296, 2.170] contains the entire OLS-FE bracket, the unit root, and explosive dynamics, all at once. Taken correctly, it says one instrument extracts far too little information from a highly persistent series to be useful. The fix is not a *better* witness but *more* of them: if $n_{i,t-2}$ is a valid instrument, then by the same logic so are $n_{i,t-3}$, $n_{i,t-4}$, and every deeper lag — the doorway to GMM."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "60ee5b2d",
   "metadata": {},
   "source": [
    "## Difference GMM (Arellano-Bond 1991)\n",
    "\n",
    "Arellano and Bond turn Anderson-Hsiao's single moment condition into a whole family. Under sequential exogeneity and no serial correlation in $\\varepsilon_{it}$, *every* level dated $t-2$ or earlier is uncorrelated with the differenced error:\n",
    "\n",
    "$$E[n_{i,t-s} \\Delta\\varepsilon_{it}] = 0 \\quad \\text{for all } s \\ge 2$$\n",
    "\n",
    "The same holds for lagged `w` and `k`, and GMM combines the dozens of resulting conditions optimally. The `pydynpd` command string encodes exactly this: `SPEC_MAIN` plus `gmm(n, 2:99) gmm(w, 2:99) gmm(k, 2:99)` (use every lag from depth 2 onward), `timedumm` (year dummies), and `nolevel` (differenced equation only — that is what makes it *difference* GMM). GMM comes in *one-step* and *two-step* flavors; two-step is asymptotically more efficient, and `pydynpd` reports the Windmeijer (2005) finite-sample correction for its standard errors, so the two-step column is the one to quote."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "21a11a5f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:56.710073Z",
     "iopub.status.busy": "2026-06-11T09:01:56.709998Z",
     "iopub.status.idle": "2026-06-11T09:01:56.800391Z",
     "shell.execute_reply": "2026-06-11T09:01:56.799831Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "One-step difference GMM:\n",
      " Dynamic panel-data estimation, one-step difference GMM\n",
      " Group variable: id                               Number of obs = 751     \n",
      " Time variable: year                              Min obs per group: 5    \n",
      " Number of instruments = 91                       Max obs per group: 7    \n",
      " Number of groups = 140                           Avg obs per group: 5.36 \n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|     n     |   coef.    | Corrected Std. Err. |     z      |   P>|z|   |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|    L1.n   | 0.7074701  |      0.0841788      | 8.4043723  | 0.0000000 | *** |\n",
      "|     w     | -0.7087965 |      0.1171020      | -6.0528147 | 0.0000000 | *** |\n",
      "|    L1.w   | 0.5000149  |      0.1113282      | 4.4913603  | 0.0000071 | *** |\n",
      "|     k     | 0.4659776  |      0.1010440      | 4.6116293  | 0.0000040 | *** |\n",
      "|    L1.k   | -0.2151309 |      0.0858525      | -2.5058195 | 0.0122168 |  *  |\n",
      "| year_1978 | 0.0057636  |      0.0166077      | 0.3470434  | 0.7285587 |     |\n",
      "| year_1979 | 0.0136366  |      0.0193748      | 0.7038306  | 0.4815383 |     |\n",
      "| year_1980 | -0.0071557 |      0.0213479      | -0.3351935 | 0.7374791 |     |\n",
      "| year_1981 | -0.0340692 |      0.0264327      | -1.2889049 | 0.1974311 |     |\n",
      "| year_1982 | -0.0059175 |      0.0272325      | -0.2172936 | 0.8279795 |     |\n",
      "| year_1983 | 0.0187213  |      0.0288529      | 0.6488533  | 0.5164332 |     |\n",
      "| year_1984 | 0.0352279  |      0.0331578      | 1.0624312  | 0.2880400 |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "Hansen test of overid. restrictions: chi(79) = 88.797 Prob > Chi2 = 0.211\n",
      "Arellano-Bond test for AR(1) in first differences: z = -5.50 Pr > z =0.000\n",
      "Arellano-Bond test for AR(2) in first differences: z = -0.14 Pr > z =0.891\n",
      "\n",
      "\n",
      "Two-step difference GMM (Windmeijer-corrected SEs):\n",
      " Dynamic panel-data estimation, two-step difference GMM\n",
      " Group variable: id                               Number of obs = 751     \n",
      " Time variable: year                              Min obs per group: 5    \n",
      " Number of instruments = 91                       Max obs per group: 7    \n",
      " Number of groups = 140                           Avg obs per group: 5.36 \n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|     n     |   coef.    | Corrected Std. Err. |     z      |   P>|z|   |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|    L1.n   | 0.6787867  |      0.0890781      | 7.6201324  | 0.0000000 | *** |\n",
      "|     w     | -0.7198296 |      0.1221408      | -5.8934431 | 0.0000000 | *** |\n",
      "|    L1.w   | 0.4626914  |      0.1134755      | 4.0774568  | 0.0000455 | *** |\n",
      "|     k     | 0.4539046  |      0.1275537      | 3.5585358  | 0.0003729 | *** |\n",
      "|    L1.k   | -0.1914923 |      0.1044671      | -1.8330393 | 0.0667967 |     |\n",
      "| year_1978 | 0.0052583  |      0.0156783      | 0.3353880  | 0.7373324 |     |\n",
      "| year_1979 | 0.0081292  |      0.0188528      | 0.4311953  | 0.6663264 |     |\n",
      "| year_1980 | -0.0125121 |      0.0213151      | -0.5870084 | 0.5571981 |     |\n",
      "| year_1981 | -0.0389696 |      0.0259642      | -1.5008968 | 0.1333823 |     |\n",
      "| year_1982 | -0.0110198 |      0.0280678      | -0.3926153 | 0.6946036 |     |\n",
      "| year_1983 | 0.0188432  |      0.0307305      | 0.6131750  | 0.5397606 |     |\n",
      "| year_1984 | 0.0346064  |      0.0336964      | 1.0270068  | 0.3044173 |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "Hansen test of overid. restrictions: chi(79) = 88.797 Prob > Chi2 = 0.211\n",
      "Arellano-Bond test for AR(1) in first differences: z = -4.46 Pr > z =0.000\n",
      "Arellano-Bond test for AR(2) in first differences: z = -0.17 Pr > z =0.866\n",
      "\n",
      "\n",
      "  one-step: rho = 0.7075 (se 0.0842)\n",
      "  two-step: rho = 0.6788 (se 0.0891), 91 instruments\n"
     ]
    }
   ],
   "source": [
    "print(\"One-step difference GMM:\")\n",
    "diff_one = run_abond(f\"{SPEC_MAIN} | {GMM_FULL} | timedumm nolevel onestep\", d)\n",
    "print(\"\\nTwo-step difference GMM (Windmeijer-corrected SEs):\")\n",
    "diff_two = run_abond(f\"{SPEC_MAIN} | {GMM_FULL} | timedumm nolevel\", d)\n",
    "\n",
    "s1 = gmm_summary(diff_one, \"Diff GMM (one-step)\")\n",
    "s2 = gmm_summary(diff_two, \"Diff GMM (two-step)\")\n",
    "print(f\"\\n  one-step: rho = {s1['rho1']:.4f} (se {s1['se']:.4f})\")\n",
    "print(f\"  two-step: rho = {s2['rho1']:.4f} (se {s2['se']:.4f}), \"\n",
    "      f\"{s2['n_instruments']} instruments\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "82a01b59",
   "metadata": {},
   "source": [
    "**Interpretation.** The machinery works exactly as advertised — 91 instruments across 751 usable observations, a two-step estimate of $\\hat{\\rho} = 0.6788$ (SE 0.0891), and every printed diagnostic passes: AR(1) rejects as it mechanically must, AR(2) is nowhere near rejecting (p = 0.866), and Hansen accepts (p = 0.211). And yet this estimate should *not* be trusted, for a reason no printed test reveals: it sits only 0.053 above the FE lower bound of 0.626 — within one standard error of it, in the bottom sixth of the bracket. This is Bond's (2002) informal diagnostic failing loudly. When the true series is highly persistent, the level two years ago barely predicts this year's *change* — so all 91 instruments are individually weak, and weak-instrument bias in this design points toward the within estimator. Blundell and Bond (1998) demonstrated precisely this failure *on precisely this dataset* — and their fix is next."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9406acb8",
   "metadata": {},
   "source": [
    "## System GMM (Blundell-Bond 1998): the headline model\n",
    "\n",
    "If lagged *levels* are weak instruments for *differences*, what about the reverse? Lagged *differences* turn out to be strong instruments for *levels*. System GMM estimates a stacked *system* of both equations at once: the differenced equation keeps its Arellano-Bond instruments, and the levels equation gets instrumented by lagged differences via the new moment conditions\n",
    "\n",
    "$$E[\\Delta n_{i,t-1}(\\alpha_i + \\varepsilon_{it})] = 0$$\n",
    "\n",
    "That is a genuinely new assumption — *mean stationarity*: firms may sit at wildly different steady-state levels, but their initial deviations from those steady states must be unrelated to the levels themselves. Two practical choices complete the headline specification: *collapsed* instruments (`collapse` combines each lag depth into a single instrument column, holding the count to 32, safely below Roodman's rule that instruments should not outnumber the 140 firms), and two-step Windmeijer-corrected results. Dropping `nolevel` from the command string is what switches `pydynpd` from difference to system GMM."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "63fdb8c5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:56.801708Z",
     "iopub.status.busy": "2026-06-11T09:01:56.801619Z",
     "iopub.status.idle": "2026-06-11T09:01:56.865419Z",
     "shell.execute_reply": "2026-06-11T09:01:56.865076Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Two-step system GMM, collapsed instruments:\n",
      " Dynamic panel-data estimation, two-step system GMM\n",
      " Group variable: id                               Number of obs = 751     \n",
      " Time variable: year                              Min obs per group: 5    \n",
      " Number of instruments = 32                       Max obs per group: 7    \n",
      " Number of groups = 140                           Avg obs per group: 5.36 \n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|     n     |   coef.    | Corrected Std. Err. |     z      |   P>|z|   |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|    L1.n   | 0.9269913  |      0.0785085      | 11.8075341 | 0.0000000 | *** |\n",
      "|     w     | -0.8155041 |      0.2763832      | -2.9506278 | 0.0031713 |  ** |\n",
      "|    L1.w   | 0.6331152  |      0.3327639      | 1.9025958  | 0.0570933 |     |\n",
      "|     k     | 0.5894690  |      0.1715356      | 3.4364236  | 0.0005894 | *** |\n",
      "|    L1.k   | -0.4888581 |      0.1969821      | -2.4817381 | 0.0130743 |  *  |\n",
      "| year_1978 | 0.0215577  |      0.0280611      | 0.7682429  | 0.4423429 |     |\n",
      "| year_1979 | 0.0437393  |      0.0373670      | 1.1705321  | 0.2417869 |     |\n",
      "| year_1980 | 0.0338457  |      0.0400282      | 0.8455475  | 0.3978052 |     |\n",
      "| year_1981 | 0.0147678  |      0.0544092      | 0.2714210  | 0.7860672 |     |\n",
      "| year_1982 | 0.0577658  |      0.0555424      | 1.0400310  | 0.2983255 |     |\n",
      "| year_1983 | 0.0848864  |      0.0471996      | 1.7984554  | 0.0721049 |     |\n",
      "| year_1984 | 0.0799244  |      0.0411493      | 1.9423035  | 0.0521004 |     |\n",
      "|    _con   | 0.6404202  |      0.4628017      | 1.3837897  | 0.1664229 |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "Hansen test of overid. restrictions: chi(19) = 18.918 Prob > Chi2 = 0.462\n",
      "Arellano-Bond test for AR(1) in first differences: z = -4.49 Pr > z =0.000\n",
      "Arellano-Bond test for AR(2) in first differences: z = -0.01 Pr > z =0.994\n",
      "\n",
      "\n",
      "One-step system GMM, collapsed instruments:\n",
      " Dynamic panel-data estimation, one-step system GMM\n",
      " Group variable: id                               Number of obs = 751     \n",
      " Time variable: year                              Min obs per group: 5    \n",
      " Number of instruments = 32                       Max obs per group: 7    \n",
      " Number of groups = 140                           Avg obs per group: 5.36 \n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|     n     |   coef.    | Corrected Std. Err. |     z      |   P>|z|   |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|    L1.n   | 0.9024603  |      0.0634304      | 14.2275598 | 0.0000000 | *** |\n",
      "|     w     | -0.7727454 |      0.2227516      | -3.4690908 | 0.0005222 | *** |\n",
      "|    L1.w   | 0.6431681  |      0.2380502      | 2.7018173  | 0.0068962 |  ** |\n",
      "|     k     | 0.6235219  |      0.1489092      | 4.1872635  | 0.0000282 | *** |\n",
      "|    L1.k   | -0.5111490 |      0.1762255      | -2.9005388 | 0.0037252 |  ** |\n",
      "| year_1978 | 0.0123814  |      0.0230174      | 0.5379150  | 0.5906357 |     |\n",
      "| year_1979 | 0.0334238  |      0.0284310      | 1.1756114  | 0.2397502 |     |\n",
      "| year_1980 | 0.0252671  |      0.0328913      | 0.7682020  | 0.4423672 |     |\n",
      "| year_1981 | 0.0130379  |      0.0426232      | 0.3058873  | 0.7596905 |     |\n",
      "| year_1982 | 0.0522858  |      0.0424192      | 1.2325966  | 0.2177263 |     |\n",
      "| year_1983 | 0.0663111  |      0.0381680      | 1.7373465  | 0.0823260 |     |\n",
      "| year_1984 | 0.0649112  |      0.0356630      | 1.8201276  | 0.0687396 |     |\n",
      "|    _con   | 0.5111072  |      0.4130100      | 1.2375178  | 0.2158949 |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "Hansen test of overid. restrictions: chi(19) = 18.918 Prob > Chi2 = 0.462\n",
      "Arellano-Bond test for AR(1) in first differences: z = -4.83 Pr > z =0.000\n",
      "Arellano-Bond test for AR(2) in first differences: z = -0.06 Pr > z =0.949\n",
      "\n",
      "\n",
      "  two-step: rho = 0.9270 (se 0.0785), 32 instruments\n",
      "  one-step: rho = 0.9025 (se 0.0634)\n"
     ]
    }
   ],
   "source": [
    "print(\"Two-step system GMM, collapsed instruments:\")\n",
    "sys_two = run_abond(f\"{SPEC_MAIN} | {GMM_FULL} | timedumm collapse\", d)\n",
    "print(\"\\nOne-step system GMM, collapsed instruments:\")\n",
    "sys_one = run_abond(f\"{SPEC_MAIN} | {GMM_FULL} | timedumm collapse onestep\", d)\n",
    "\n",
    "s3 = gmm_summary(sys_two, \"Sys GMM (two-step, collapsed)\")\n",
    "s4 = gmm_summary(sys_one, \"Sys GMM (one-step, collapsed)\")\n",
    "print(f\"\\n  two-step: rho = {s3['rho1']:.4f} (se {s3['se']:.4f}), \"\n",
    "      f\"{s3['n_instruments']} instruments\")\n",
    "print(f\"  one-step: rho = {s4['rho1']:.4f} (se {s4['se']:.4f})\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a7cf9662",
   "metadata": {},
   "source": [
    "**Interpretation.** This is the headline: $\\hat{\\rho} = 0.9270$ (SE 0.0785) — inside the bracket, in its upper half, 0.25 above the weak-instrument difference-GMM estimate, with 32 collapsed instruments and textbook diagnostics. Substantively, about 93 percent of an employment shock survives into the next year — a shock half-life of roughly nine years — versus the 1.5 years fixed effects would have claimed. One honest caveat: the 95 percent CI [0.773, 1.081] includes 1.0, so a unit root cannot be rejected; the defensible claim is the point estimate and its lower bound.\n",
    "\n",
    "**Reading the diagnostics** (each has a non-obvious reading):\n",
    "\n",
    "| Test | Correct reading | Our headline value |\n",
    "|---|---|---|\n",
    "| AR(1) in differences | **Must reject.** Differencing makes adjacent errors share $\\varepsilon_{i,t-1}$, so rejection is mechanical and *good news* | z = -4.49, p = 0.000 — rejects, as required |\n",
    "| AR(2) in differences | **Must not reject.** This is the test that validates the $t-2$ instruments: serial correlation in $\\varepsilon_{it}$ would contaminate them | z = -0.01, p = 0.994 — clean |\n",
    "| Hansen J | **Two-tailed in spirit.** p < 0.05 means instruments look invalid; p near 1 means the test has been overwhelmed by too many instruments | p = 0.462 — comfortable on both sides |\n",
    "\n",
    "That last claim — that the Hansen p-value responds to the instrument *count*, not just instrument *validity* — is testable on our own data. Let us run the experiment."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "69d7bac8",
   "metadata": {},
   "source": [
    "## Instrument proliferation: lag windows vs collapse\n",
    "\n",
    "We re-estimate the *identical* system-GMM model six times, varying only two plumbing choices: the lag window (`2:3` = use lags 2 and 3 only; `2:5`; `2:99` = use everything) and whether the instrument matrix is collapsed. If the Hansen test were a pure validity meter, its p-value would be roughly stable across these cells. It is not."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "e3d5d6f3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:56.867118Z",
     "iopub.status.busy": "2026-06-11T09:01:56.867039Z",
     "iopub.status.idle": "2026-06-11T09:01:57.105221Z",
     "shell.execute_reply": "2026-06-11T09:01:57.104745Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  lags 2:3   collapse=False ->  68 instruments, rho = 0.956, Hansen p = 0.035\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  lags 2:3   collapse=True  ->  17 instruments, rho = 0.921, Hansen p = 0.096\n",
      "  lags 2:5   collapse=False ->  95 instruments, rho = 0.935, Hansen p = 0.186\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  lags 2:5   collapse=True  ->  23 instruments, rho = 0.937, Hansen p = 0.255\n",
      "  lags 2:99  collapse=False -> 113 instruments, rho = 0.930, Hansen p = 0.235\n",
      "  lags 2:99  collapse=True  ->  32 instruments, rho = 0.927, Hansen p = 0.462\n"
     ]
    }
   ],
   "source": [
    "grid_specs = [\n",
    "    (\"2:3\", False), (\"2:3\", True),\n",
    "    (\"2:5\", False), (\"2:5\", True),\n",
    "    (\"2:99\", False), (\"2:99\", True),\n",
    "]\n",
    "grid_rows = []\n",
    "for window, collapsed in grid_specs:\n",
    "    gmm_part = f\"gmm(n, {window}) gmm(w, {window}) gmm(k, {window})\"\n",
    "    opts = \"timedumm collapse\" if collapsed else \"timedumm\"\n",
    "    model = run_abond(f\"{SPEC_MAIN} | {gmm_part} | {opts}\", d, quiet=True)\n",
    "    row = gmm_summary(model, f\"sys GMM lags {window}\"\n",
    "                      + (\", collapsed\" if collapsed else \"\"))\n",
    "    row[\"lag_window\"] = window\n",
    "    row[\"collapsed\"] = collapsed\n",
    "    grid_rows.append(row)\n",
    "    print(f\"  lags {window:5s} collapse={str(collapsed):5s} -> \"\n",
    "          f\"{row['n_instruments']:3d} instruments, rho = {row['rho1']:.3f}, \"\n",
    "          f\"Hansen p = {row['hansen_p']:.3f}\")\n",
    "\n",
    "grid = pd.DataFrame(grid_rows)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "643460d5",
   "metadata": {},
   "source": [
    "**Interpretation.** Two findings, both reassuring and unsettling at once. The reassuring one: $\\hat{\\rho}$ barely moves — all six cells land in the narrow range [0.921, 0.956], so the *point estimate* is robust to the plumbing. The unsettling one: the *test we would use to defend it* is not. Reading the uncollapsed rows top to bottom, the instrument count climbs 68, 95, 113 (approaching the 140-firm ceiling) and the Hansen p-value drifts 0.035, 0.186, 0.235 — even though the model never changes. That upward drift is the overfitting trajectory whose endpoint is the notorious \"p = 1.000\" red flag. The grid also catches proliferation distorting the test in the *other* tail: the uncollapsed 2:3 specification is outright *rejected* by Hansen (p = 0.035) while its collapsed twin passes (p = 0.096) — same lag window, same data, opposite verdicts, driven purely by instrument count."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "716a49c2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:57.106515Z",
     "iopub.status.busy": "2026-06-11T09:01:57.106443Z",
     "iopub.status.idle": "2026-06-11T09:01:57.165855Z",
     "shell.execute_reply": "2026-06-11T09:01:57.165375Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x550 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(9, 5.5))\n",
    "fig.patch.set_linewidth(0)\n",
    "for collapsed, col, marker, lab in [(False, STEEL_BLUE, \"o\", \"Full instrument matrix\"),\n",
    "                                    (True, TEAL, \"D\", \"Collapsed instruments\")]:\n",
    "    sub = grid[grid[\"collapsed\"] == collapsed]\n",
    "    ax.scatter(sub[\"n_instruments\"], sub[\"hansen_p\"], s=140, color=col,\n",
    "               marker=marker, edgecolors=DARK_NAVY, lw=1.5, zorder=3, label=lab)\n",
    "    for _, r in sub.iterrows():\n",
    "        ax.annotate(f\"lags {r['lag_window']}\",\n",
    "                    (r[\"n_instruments\"], r[\"hansen_p\"]),\n",
    "                    textcoords=\"offset points\", xytext=(0, 12),\n",
    "                    color=LIGHT_TEXT, fontsize=9.5, ha=\"center\")\n",
    "ax.axhline(0.05, color=WARM_ORANGE, lw=1.5, ls=\"--\")\n",
    "ax.text(137, 0.022, \"p = 0.05: instruments rejected below this line\",\n",
    "        color=WARM_ORANGE, fontsize=9.5, ha=\"right\", va=\"top\")\n",
    "ax.axvline(140, color=GRAY, lw=1.5, ls=\":\")\n",
    "ax.text(138, 0.78, \"N = 140 firms\\n(Roodman's ceiling)\", color=GRAY,\n",
    "        fontsize=9.5, ha=\"right\")\n",
    "ax.set_xlabel(\"Number of instruments\")\n",
    "ax.set_ylabel(\"Hansen test p-value\")\n",
    "ax.set_ylim(-0.06, 1.0)\n",
    "ax.set_title(\"Instrument proliferation: more is not better\", fontsize=13)\n",
    "ax.legend(loc=\"upper left\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ac793521",
   "metadata": {},
   "source": [
    "**Interpretation.** The teal diamonds (collapsed) buy nearly identical point estimates with a quarter of the instruments, at the honest price of larger standard errors — 0.0785 collapsed versus 0.0274 uncollapsed at the 2:99 window. That uncollapsed SE looks like a precision triumph, but it is precisely the too-good-to-be-true precision Roodman warns about: 113 instruments fitted to 140 firms are partly fitting noise, and the same overfitting that flatters the SE is what disarms the Hansen test. Our headline deliberately takes the larger, more honest standard error."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9975f403",
   "metadata": {},
   "source": [
    "## Replication check: the pydynpd documentation example\n",
    "\n",
    "Good practice with any estimation package — especially one we patched with a compatibility shim — is to replicate its published benchmark before trusting novel output. The `pydynpd` README estimates the *original* Arellano-Bond (1991) two-lag specification on this same dataset: two lags of `n`, a restricted instrument window `gmm(n, 2:4)`, wages treated as predetermined with `gmm(w, 1:3)`, capital as a standard exogenous instrument `iv(k)`, and difference GMM (`nolevel`). The package authors validated that output against Stata's `xtabond2`. Note this is a *different model* from our running specification — the point here is toolchain verification, not a second opinion on $\\rho$. The hard assertion below aborts the run on any discrepancy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "c259bcc5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:57.167088Z",
     "iopub.status.busy": "2026-06-11T09:01:57.167004Z",
     "iopub.status.idle": "2026-06-11T09:01:57.192764Z",
     "shell.execute_reply": "2026-06-11T09:01:57.192394Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Dynamic panel-data estimation, two-step difference GMM\n",
      " Group variable: id                               Number of obs = 611     \n",
      " Time variable: year                              Min obs per group: 4    \n",
      " Number of instruments = 42                       Max obs per group: 6    \n",
      " Number of groups = 140                           Avg obs per group: 4.36 \n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|     n     |   coef.    | Corrected Std. Err. |     z      |   P>|z|   |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "|    L1.n   | 0.2710675  |      0.1382542      | 1.9606462  | 0.0499203 |  *  |\n",
      "|    L2.n   | -0.0233928 |      0.0419665      | -0.5574151 | 0.5772439 |     |\n",
      "|     w     | -0.5668527 |      0.2092231      | -2.7093219 | 0.0067421 |  ** |\n",
      "|     k     | 0.3613939  |      0.0662624      | 5.4539824  | 0.0000000 | *** |\n",
      "| year_1979 | 0.0011898  |      0.0092322      | 0.1288765  | 0.8974554 |     |\n",
      "| year_1980 | -0.0316432 |      0.0116155      | -2.7242254 | 0.0064453 |  ** |\n",
      "| year_1981 | -0.0900163 |      0.0206593      | -4.3571693 | 0.0000132 | *** |\n",
      "| year_1982 | -0.0996210 |      0.0296036      | -3.3651654 | 0.0007650 | *** |\n",
      "| year_1983 | -0.0693308 |      0.0404276      | -1.7149347 | 0.0863572 |     |\n",
      "| year_1984 | -0.0614505 |      0.0475525      | -1.2922666 | 0.1962648 |     |\n",
      "+-----------+------------+---------------------+------------+-----------+-----+\n",
      "Hansen test of overid. restrictions: chi(32) = 32.666 Prob > Chi2 = 0.434\n",
      "Arellano-Bond test for AR(1) in first differences: z = -1.29 Pr > z =0.198\n",
      "Arellano-Bond test for AR(2) in first differences: z = -0.31 Pr > z =0.760\n",
      "\n",
      "\n",
      "  Published vignette values: L1.n = 0.2710675, Hansen chi2 = 32.666,\n",
      "  42 instruments.\n",
      "  Our run:                   L1.n = 0.2710675, Hansen chi2 = 32.666, 42 instruments.\n",
      "  Exact match: True\n"
     ]
    }
   ],
   "source": [
    "ab_repl = run_abond(\n",
    "    \"n L(1:2).n w k | gmm(n, 2:4) gmm(w, 1:3) iv(k) | timedumm nolevel\", d)\n",
    "rt = ab_repl.regression_table\n",
    "rho_repl = rt.loc[rt.variable == \"L1.n\", \"coefficient\"].iloc[0]\n",
    "print(f\"\\n  Published vignette values: L1.n = 0.2710675, Hansen chi2 = 32.666,\")\n",
    "print(\"  42 instruments.\")\n",
    "print(f\"  Our run:                   L1.n = {rho_repl:.7f}, Hansen chi2 = \"\n",
    "      f\"{ab_repl.hansen.test_value:.3f}, {ab_repl.z_information.num_instr} instruments.\")\n",
    "match = (abs(rho_repl - 0.2710675) < 1e-6\n",
    "         and abs(ab_repl.hansen.test_value - 32.666) < 1e-3\n",
    "         and ab_repl.z_information.num_instr == 42)\n",
    "print(f\"  Exact match: {match}\")\n",
    "if not match:\n",
    "    raise AssertionError(\"Replication check failed - investigate before publishing\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bc288881",
   "metadata": {},
   "source": [
    "**Interpretation.** The replication is exact to all printed digits: `L1.n` = 0.2710675, Hansen $\\chi^2(32)$ = 32.666, 42 instruments — `Exact match: True` under the hard assertion. This verifies the entire toolchain, NumPy-2 shim included, against the package's own benchmark (itself validated against `xtabond2`). Two teaching points hide in the output. First, the much lower persistence coefficient here (0.271) is *not* a contradiction of our headline 0.927: this is difference GMM — subject to the same weak-instrument drag diagnosed above — on a two-lag specification with a restricted instrument window; \"the\" persistence estimate is always joint with the specification and estimator that produced it. Second, AR(1) here does *not* reject (p = 0.198): with two lags of `n` soaking up the dynamics, even the mechanical differencing correlation is muted — diagnostics must be read against the model, not against a universal rulebook."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14cfdb16",
   "metadata": {},
   "source": [
    "## Synthesis: seven estimators, one parameter\n",
    "\n",
    "Each section produced a number; the story only snaps into focus when they share an axis. We assemble the summary table and draw the forest plot the whole notebook has been building toward."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "ee0e26b5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:57.193928Z",
     "iopub.status.busy": "2026-06-11T09:01:57.193848Z",
     "iopub.status.idle": "2026-06-11T09:01:57.197759Z",
     "shell.execute_reply": "2026-06-11T09:01:57.197382Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                    estimator   rho1     se  ci_lo  ci_hi  hansen_p  ar1_p  ar2_p  n_instruments\n",
      "                   Pooled OLS 0.9617 0.0084 0.9453 0.9781       NaN    NaN    NaN            NaN\n",
      "                Fixed effects 0.6262 0.0515 0.5252 0.7272       NaN    NaN    NaN            NaN\n",
      "            Anderson-Hsiao IV 1.2327 0.4782 0.2955 2.1699       NaN    NaN    NaN            1.0\n",
      "          Diff GMM (one-step) 0.7075 0.0842 0.5425 0.8725    0.2113    0.0 0.8913           91.0\n",
      "          Diff GMM (two-step) 0.6788 0.0891 0.5042 0.8534    0.2113    0.0 0.8660           91.0\n",
      "Sys GMM (one-step, collapsed) 0.9025 0.0634 0.7781 1.0268    0.4621    0.0 0.9492           32.0\n",
      "Sys GMM (two-step, collapsed) 0.9270 0.0785 0.7731 1.0809    0.4621    0.0 0.9944           32.0\n"
     ]
    }
   ],
   "source": [
    "summary_rows = [\n",
    "    {\"estimator\": \"Pooled OLS\", \"rho1\": rho_ols, \"se\": se_ols,\n",
    "     \"ci_lo\": rho_ols - 1.96 * se_ols, \"ci_hi\": rho_ols + 1.96 * se_ols,\n",
    "     \"hansen_p\": np.nan, \"ar1_p\": np.nan, \"ar2_p\": np.nan,\n",
    "     \"n_instruments\": np.nan},\n",
    "    {\"estimator\": \"Fixed effects\", \"rho1\": rho_fe, \"se\": se_fe,\n",
    "     \"ci_lo\": rho_fe - 1.96 * se_fe, \"ci_hi\": rho_fe + 1.96 * se_fe,\n",
    "     \"hansen_p\": np.nan, \"ar1_p\": np.nan, \"ar2_p\": np.nan,\n",
    "     \"n_instruments\": np.nan},\n",
    "    {\"estimator\": \"Anderson-Hsiao IV\", \"rho1\": rho_ah, \"se\": se_ah,\n",
    "     \"ci_lo\": rho_ah - 1.96 * se_ah, \"ci_hi\": rho_ah + 1.96 * se_ah,\n",
    "     \"hansen_p\": np.nan, \"ar1_p\": np.nan, \"ar2_p\": np.nan,\n",
    "     \"n_instruments\": 1},\n",
    "    s1, s2, s4, s3,\n",
    "]\n",
    "summary = pd.DataFrame(summary_rows)\n",
    "print(summary.round(4).to_string(index=False))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "38ace59a",
   "metadata": {},
   "source": [
    "**Interpretation.** Read as a single table, the ladder is unambiguous. The two naive estimators (0.9617 and 0.6262) define the bracket. Anderson-Hsiao (1.2327) is the only estimator *outside* it, with a confidence interval wider than everyone else's put together. The two difference-GMM rows (0.7075 and 0.6788) sit in the bracket's bottom sixth with 91 instruments each — formally admissible, substantively suspect. The two system-GMM rows (0.9025 and 0.9270) sit in the upper half with 32 collapsed instruments, the cleanest AR(2) values in the table, and Hansen p-values in the comfortable middle."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "ebae2a8d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-11T09:01:57.198946Z",
     "iopub.status.busy": "2026-06-11T09:01:57.198877Z",
     "iopub.status.idle": "2026-06-11T09:01:57.261071Z",
     "shell.execute_reply": "2026-06-11T09:01:57.260714Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 950x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "order = [\"Pooled OLS\", \"Fixed effects\", \"Anderson-Hsiao IV\",\n",
    "         \"Diff GMM (one-step)\", \"Diff GMM (two-step)\",\n",
    "         \"Sys GMM (one-step, collapsed)\", \"Sys GMM (two-step, collapsed)\"]\n",
    "colors = {\"Pooled OLS\": GRAY, \"Fixed effects\": GRAY,\n",
    "          \"Anderson-Hsiao IV\": STEEL_BLUE,\n",
    "          \"Diff GMM (one-step)\": WARM_ORANGE,\n",
    "          \"Diff GMM (two-step)\": WARM_ORANGE,\n",
    "          \"Sys GMM (one-step, collapsed)\": TEAL,\n",
    "          \"Sys GMM (two-step, collapsed)\": TEAL}\n",
    "plot_df = summary.set_index(\"estimator\").loc[order].reset_index()\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(9.5, 6))\n",
    "fig.patch.set_linewidth(0)\n",
    "ax.axvspan(rho_fe, rho_ols, color=GRID_LINE, alpha=0.55, zorder=0)\n",
    "ax.text((rho_fe + rho_ols) / 2, -0.75, \"OLS-FE credible bracket\",\n",
    "        color=LIGHT_TEXT, fontsize=10, ha=\"center\", style=\"italic\")\n",
    "ax.axvline(1.0, color=GRAY, lw=1.2, ls=\"--\", alpha=0.8)\n",
    "ypos = np.arange(len(plot_df))[::-1]\n",
    "for y, (_, r) in zip(ypos, plot_df.iterrows()):\n",
    "    ax.errorbar(r[\"rho1\"], y, xerr=1.96 * r[\"se\"], fmt=\"o\",\n",
    "                color=colors[r[\"estimator\"]], ms=9, capsize=4, lw=2.2,\n",
    "                capthick=2.2, zorder=3)\n",
    "    ax.text(r[\"rho1\"], y + 0.28, f\"{r['rho1']:.3f}\", color=WHITE_TEXT,\n",
    "            fontsize=10, ha=\"center\")\n",
    "ax.set_yticks(ypos)\n",
    "ax.set_yticklabels(plot_df[\"estimator\"])\n",
    "ax.set_xlabel(r\"Employment persistence  $\\hat{\\rho}$  (L1.n) with 95% CI\")\n",
    "ax.set_ylim(-1.1, len(plot_df) - 0.4)\n",
    "ax.set_title(\"Seven estimators, one parameter:\\n\"\n",
    "             \"system GMM lands inside the bracket with tight precision\",\n",
    "             fontsize=13)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "638da30c",
   "metadata": {},
   "source": [
    "**Interpretation.** The forest plot is the tutorial in one image. The grey markers (OLS at 0.962, FE at 0.626) define the shaded credible band. Anderson-Hsiao's enormous blue whisker straddles everything, including the dashed unit-root line. The orange difference-GMM pair sits at the bottom of the band, hugging the FE bound exactly as Bond's diagnostic predicts under weak instruments. And the teal system-GMM pair (0.902 and 0.927) lands in the upper half of the band with usable precision. Notice what separates the winner from the losers: *nothing on any single printed line of output*. Only the bracket logic, the weak-instrument reasoning, the proliferation experiment, and the replication check — the workflow, not a p-value — identify 0.927 as the defensible answer."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "560faec3",
   "metadata": {},
   "source": [
    "## Summary and next steps\n",
    "\n",
    "**Key takeaways:**\n",
    "\n",
    "1. **The bracket.** OLS (0.962) and FE (0.626) bracket the truth from above and below; their disagreement of 0.336 is not noise but a diagnostic measuring stick.\n",
    "2. **Consistency is not enough.** Anderson-Hsiao is consistent but imprecise (SE 0.48), and difference GMM (0.679, 91 instruments) passed every printed test while hugging the biased FE bound — the textbook weak-instrument symptom when the series is persistent.\n",
    "3. **System GMM delivers the defensible answer.** $\\hat{\\rho} = 0.927$ (SE 0.079), inside the bracket, with AR(2) p = 0.994, Hansen p = 0.462, and 32 collapsed instruments against 140 firms.\n",
    "4. **Read the diagnostics correctly.** AR(1) must reject (mechanical); AR(2) must not reject (it protects the instruments); Hansen is two-tailed in spirit — p drifting toward 1 as instruments pile up is overfitting, not validity.\n",
    "5. **Verify the toolchain.** The replication of the `pydynpd` published benchmark was exact to all printed digits (L1.n = 0.2710675, 42 instruments), NumPy-2 shim included.\n",
    "\n",
    "Substantively: about 93 percent of an employment shock survives into the next year — a shock half-life of roughly nine years — though the headline CI [0.773, 1.081] includes the unit root, so the defensible claim is the point estimate and its lower bound.\n",
    "\n",
    "For the full discussion (practitioner checklist, limitations, exercises, and references), see the blog post: [Dynamic Panel Data Models in Python: From Nickell Bias to System GMM](https://carlos-mendez.org/post/python_dynamic_panel/).\n",
    "\n",
    "**References (selection):** Arellano & Bond (1991, *REStud*); Anderson & Hsiao (1981, *JASA*); Nickell (1981, *Econometrica*); Blundell & Bond (1998, *J. Econometrics*); Bond (2002, *Portuguese Economic Journal*); Windmeijer (2005, *J. Econometrics*); Roodman (2009, *Stata Journal*); Wu, Hua & Xu (2023, *JOSS* — `pydynpd`)."
   ]
  }
 ],
 "metadata": {
  "colab": {
   "provenance": [],
   "toc_visible": true
  },
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3"
  },
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}