Skip to content

Instantly share code, notes, and snippets.

@MaxGhenis

MaxGhenis/org_validation_2024.ipynb

Last active March 31, 2026 12:01
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save MaxGhenis/634977ffccb65a288f54d602a2244ed9 to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save MaxGhenis/634977ffccb65a288f54d602a2244ed9 to your computer and use it in GitHub Desktop.
Download ZIP
ORG validation for policyengine-us-data PR
Display the source blob
Display the rendered blob
Raw
org_validation_2024.ipynb
{
"cells": [
{
"cell_type": "markdown",
"id": "d422ab93",
"metadata": {},
"source": [
"# ORG Validation: 2024 Annual Donor Build\n",
"\n",
"This notebook summarizes the 2024 CPS ORG donor build used for `hourly_wage`, `is_paid_hourly`, and the BLS-based `is_union_member_or_covered` assignment.\n",
"\n",
"It is intended as PR support material rather than repo source."
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "7e359ce0",
"metadata": {
"execution": {
"iopub.execute_input": "2026-03-31T11:50:17.241889Z",
"iopub.status.busy": "2026-03-31T11:50:17.241799Z",
"iopub.status.idle": "2026-03-31T11:50:20.605167Z",
"shell.execute_reply": "2026-03-31T11:50:20.604764Z"
}
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/Users/maxghenis/.codex-worktrees/us-data-org-impute/.venv/lib/python3.13/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
" from .autonotebook import tqdm as notebook_tqdm\n"
]
},
{
"data": {
"text/plain": [
"PosixPath('/Users/maxghenis/.codex-worktrees/us-data-org-impute/policyengine_us_data/storage/census_cps_org_2024_wages.csv.gz')"
]
},
"execution_count": 1,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import sys\n",
"from pathlib import Path\n",
"\n",
"import numpy as np\n",
"import pandas as pd\n",
"from IPython.display import display\n",
"\n",
"US_INPUTS_WORKTREE = Path(\"/Users/maxghenis/.codex-worktrees/us-org-inputs\")\n",
"if str(US_INPUTS_WORKTREE) not in sys.path:\n",
" sys.path.insert(0, str(US_INPUTS_WORKTREE))\n",
"\n",
"from policyengine_us_data.datasets.org.org import (\n",
" ORG_FILENAME,\n",
" ORG_MONTHS,\n",
" ORG_PREDICTORS,\n",
" CPS_BASIC_MONTHLY_ORG_COLUMNS,\n",
" _cps_basic_org_month_url,\n",
" _predict_union_coverage_from_bls_tables,\n",
" load_org_training_data,\n",
")\n",
"from policyengine_us_data.storage import STORAGE_FOLDER\n",
"\n",
"pd.set_option(\"display.max_columns\", 20)\n",
"pd.set_option(\"display.float_format\", lambda x: f\"{x:,.4f}\")\n",
"\n",
"cache_path = STORAGE_FOLDER / ORG_FILENAME\n",
"cache_path"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "d2d63c8a",
"metadata": {
"execution": {
"iopub.execute_input": "2026-03-31T11:50:20.606378Z",
"iopub.status.busy": "2026-03-31T11:50:20.606291Z",
"iopub.status.idle": "2026-03-31T11:50:20.710134Z",
"shell.execute_reply": "2026-03-31T11:50:20.709665Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>metric</th>\n",
" <th>value</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>donor rows</td>\n",
" <td>119,237.0000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>weighted weight sum (millions)</td>\n",
" <td>16,523,600.2140</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>weighted mean hourly wage</td>\n",
" <td>35.1700</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>weighted p10 hourly wage</td>\n",
" <td>14.4000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>weighted p50 hourly wage</td>\n",
" <td>25.0000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>weighted p90 hourly wage</td>\n",
" <td>62.5000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>weighted p99 hourly wage</td>\n",
" <td>222.3700</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>weighted paid-hourly share</td>\n",
" <td>0.5532</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>weighted union share</td>\n",
" <td>0.1101</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>weighted share wage &gt; $100</td>\n",
" <td>0.0276</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <td>weighted share wage &gt; $200</td>\n",
" <td>0.0155</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11</th>\n",
" <td>weighted share wage &lt; $7.25</td>\n",
" <td>0.0107</td>\n",
" </tr>\n",
" <tr>\n",
" <th>12</th>\n",
" <td>weighted share wage &lt; $10</td>\n",
" <td>0.0216</td>\n",
" </tr>\n",
" <tr>\n",
" <th>13</th>\n",
" <td>weighted p10 weekly hours</td>\n",
" <td>25.0000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>14</th>\n",
" <td>weighted p50 weekly hours</td>\n",
" <td>40.0000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>15</th>\n",
" <td>weighted p90 weekly hours</td>\n",
" <td>50.0000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>16</th>\n",
" <td>max hourly wage</td>\n",
" <td>2,890.6700</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" metric value\n",
"0 donor rows 119,237.0000\n",
"1 weighted weight sum (millions) 16,523,600.2140\n",
"2 weighted mean hourly wage 35.1700\n",
"3 weighted p10 hourly wage 14.4000\n",
"4 weighted p50 hourly wage 25.0000\n",
"5 weighted p90 hourly wage 62.5000\n",
"6 weighted p99 hourly wage 222.3700\n",
"7 weighted paid-hourly share 0.5532\n",
"8 weighted union share 0.1101\n",
"9 weighted share wage > $100 0.0276\n",
"10 weighted share wage > $200 0.0155\n",
"11 weighted share wage < $7.25 0.0107\n",
"12 weighted share wage < $10 0.0216\n",
"13 weighted p10 weekly hours 25.0000\n",
"14 weighted p50 weekly hours 40.0000\n",
"15 weighted p90 weekly hours 50.0000\n",
"16 max hourly wage 2,890.6700"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"org = load_org_training_data().copy()\n",
"org[\"is_union_member_or_covered\"] = _predict_union_coverage_from_bls_tables(\n",
" org[ORG_PREDICTORS]\n",
").astype(np.float32)\n",
"\n",
"w = org[\"sample_weight\"].astype(float).to_numpy()\n",
"wage = org[\"hourly_wage\"].astype(float).to_numpy()\n",
"hours = org[\"weekly_hours_worked\"].astype(float).to_numpy()\n",
"paid = org[\"is_paid_hourly\"].astype(float).to_numpy()\n",
"union = org[\"is_union_member_or_covered\"].astype(float).to_numpy()\n",
"\n",
"def wmean(x):\n",
" return float(np.average(x, weights=w))\n",
"\n",
"def wquantile(x, quantiles):\n",
" order = np.argsort(x)\n",
" xs = x[order]\n",
" ws = w[order]\n",
" cdf = np.cumsum(ws) / ws.sum()\n",
" return [float(xs[np.searchsorted(cdf, q, side=\"left\")]) for q in quantiles]\n",
"\n",
"summary = pd.DataFrame(\n",
" {\n",
" \"metric\": [\n",
" \"donor rows\",\n",
" \"weighted weight sum (millions)\",\n",
" \"weighted mean hourly wage\",\n",
" \"weighted p10 hourly wage\",\n",
" \"weighted p50 hourly wage\",\n",
" \"weighted p90 hourly wage\",\n",
" \"weighted p99 hourly wage\",\n",
" \"weighted paid-hourly share\",\n",
" \"weighted union share\",\n",
" \"weighted share wage > $100\",\n",
" \"weighted share wage > $200\",\n",
" \"weighted share wage < $7.25\",\n",
" \"weighted share wage < $10\",\n",
" \"weighted p10 weekly hours\",\n",
" \"weighted p50 weekly hours\",\n",
" \"weighted p90 weekly hours\",\n",
" \"max hourly wage\",\n",
" ],\n",
" \"value\": [\n",
" len(org),\n",
" round(w.sum() / 1e6, 3),\n",
" round(wmean(wage), 2),\n",
" round(wquantile(wage, [0.1])[0], 2),\n",
" round(wquantile(wage, [0.5])[0], 2),\n",
" round(wquantile(wage, [0.9])[0], 2),\n",
" round(wquantile(wage, [0.99])[0], 2),\n",
" round(wmean(paid), 4),\n",
" round(wmean(union), 4),\n",
" round(wmean((wage > 100).astype(float)), 4),\n",
" round(wmean((wage > 200).astype(float)), 4),\n",
" round(wmean((wage < 7.25).astype(float)), 4),\n",
" round(wmean((wage < 10).astype(float)), 4),\n",
" round(wquantile(hours, [0.1])[0], 2),\n",
" round(wquantile(hours, [0.5])[0], 2),\n",
" round(wquantile(hours, [0.9])[0], 2),\n",
" round(float(wage.max()), 2),\n",
" ],\n",
" }\n",
")\n",
"summary"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "90f78e2a",
"metadata": {
"execution": {
"iopub.execute_input": "2026-03-31T11:50:20.711194Z",
"iopub.status.busy": "2026-03-31T11:50:20.711117Z",
"iopub.status.idle": "2026-03-31T11:50:20.726474Z",
"shell.execute_reply": "2026-03-31T11:50:20.726003Z"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Lowest weighted union states (rows >= 500)\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"/var/folders/9l/_wztzgbx7mgc7l1r0416cy7m0000gn/T/ipykernel_2281/292089446.py:3: FutureWarning: DataFrameGroupBy.apply operated on the grouping columns. This behavior is deprecated, and in a future version of pandas the grouping columns will be excluded from the operation. Either pass `include_groups=False` to exclude the groupings or explicitly select the grouping columns after groupby to silence this warning.\n",
" .apply(\n"
]
},
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>state_fips</th>\n",
" <th>rows</th>\n",
" <th>union_share_weighted</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>33</th>\n",
" <td>37.0000</td>\n",
" <td>2,664.0000</td>\n",
" <td>0.0297</td>\n",
" </tr>\n",
" <tr>\n",
" <th>41</th>\n",
" <td>46.0000</td>\n",
" <td>1,490.0000</td>\n",
" <td>0.0385</td>\n",
" </tr>\n",
" <tr>\n",
" <th>40</th>\n",
" <td>45.0000</td>\n",
" <td>1,913.0000</td>\n",
" <td>0.0406</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <td>13.0000</td>\n",
" <td>2,502.0000</td>\n",
" <td>0.0432</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>5.0000</td>\n",
" <td>2,115.0000</td>\n",
" <td>0.0440</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" state_fips rows union_share_weighted\n",
"33 37.0000 2,664.0000 0.0297\n",
"41 46.0000 1,490.0000 0.0385\n",
"40 45.0000 1,913.0000 0.0406\n",
"10 13.0000 2,502.0000 0.0432\n",
"3 5.0000 2,115.0000 0.0440"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Highest weighted union states (rows >= 500)\n"
]
},
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>state_fips</th>\n",
" <th>rows</th>\n",
" <th>union_share_weighted</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>11</th>\n",
" <td>15.0000</td>\n",
" <td>1,623.0000</td>\n",
" <td>0.2688</td>\n",
" </tr>\n",
" <tr>\n",
" <th>32</th>\n",
" <td>36.0000</td>\n",
" <td>4,475.0000</td>\n",
" <td>0.2163</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>2.0000</td>\n",
" <td>1,287.0000</td>\n",
" <td>0.1945</td>\n",
" </tr>\n",
" <tr>\n",
" <th>30</th>\n",
" <td>34.0000</td>\n",
" <td>2,422.0000</td>\n",
" <td>0.1765</td>\n",
" </tr>\n",
" <tr>\n",
" <th>47</th>\n",
" <td>53.0000</td>\n",
" <td>2,340.0000</td>\n",
" <td>0.1760</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" state_fips rows union_share_weighted\n",
"11 15.0000 1,623.0000 0.2688\n",
"32 36.0000 4,475.0000 0.2163\n",
"1 2.0000 1,287.0000 0.1945\n",
"30 34.0000 2,422.0000 0.1765\n",
"47 53.0000 2,340.0000 0.1760"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"state_summary = (\n",
" org.groupby(\"state_fips\", group_keys=False)\n",
" .apply(\n",
" lambda g: pd.Series(\n",
" {\n",
" \"rows\": int(len(g)),\n",
" \"union_share_weighted\": float(\n",
" np.average(\n",
" g[\"is_union_member_or_covered\"],\n",
" weights=g[\"sample_weight\"],\n",
" )\n",
" ),\n",
" }\n",
" )\n",
" )\n",
" .reset_index()\n",
")\n",
"state_summary = state_summary[state_summary[\"rows\"] >= 500]\n",
"\n",
"print(\"Lowest weighted union states (rows >= 500)\")\n",
"display(state_summary.nsmallest(5, \"union_share_weighted\"))\n",
"\n",
"print(\"Highest weighted union states (rows >= 500)\")\n",
"display(state_summary.nlargest(5, \"union_share_weighted\"))"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "87b339b3",
"metadata": {
"execution": {
"iopub.execute_input": "2026-03-31T11:50:20.727534Z",
"iopub.status.busy": "2026-03-31T11:50:20.727453Z",
"iopub.status.idle": "2026-03-31T11:50:59.653532Z",
"shell.execute_reply": "2026-03-31T11:50:59.653057Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>cutoff</th>\n",
" <th>rows</th>\n",
" <th>weighted_share</th>\n",
" <th>direct_share_within_tail_weighted</th>\n",
" <th>derived_share_within_tail_weighted</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>100</td>\n",
" <td>3208</td>\n",
" <td>0.0276</td>\n",
" <td>0.0000</td>\n",
" <td>1.0000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>200</td>\n",
" <td>1774</td>\n",
" <td>0.0155</td>\n",
" <td>0.0000</td>\n",
" <td>1.0000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>500</td>\n",
" <td>26</td>\n",
" <td>0.0002</td>\n",
" <td>0.0000</td>\n",
" <td>1.0000</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" cutoff rows weighted_share direct_share_within_tail_weighted \\\n",
"0 100 3208 0.0276 0.0000 \n",
"1 200 1774 0.0155 0.0000 \n",
"2 500 26 0.0002 0.0000 \n",
"\n",
" derived_share_within_tail_weighted \n",
"0 1.0000 \n",
"1 1.0000 \n",
"2 1.0000 "
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tail_frames = []\n",
"for month in ORG_MONTHS:\n",
" raw = pd.read_csv(\n",
" _cps_basic_org_month_url(2024, month),\n",
" usecols=CPS_BASIC_MONTHLY_ORG_COLUMNS,\n",
" low_memory=False,\n",
" )\n",
" month_org = raw.loc[\n",
" raw[\"HRMIS\"].isin([4, 8])\n",
" & (raw[\"pworwgt\"] > 0)\n",
" & (raw[\"prerelg\"] == 1)\n",
" & raw[\"pemlr\"].isin([1, 2])\n",
" & raw[\"peio1cow\"].isin([1, 2, 3, 4, 5])\n",
" & raw[\"peernhry\"].isin([1, 2])\n",
" & (raw[\"gestfips\"] > 0)\n",
" & (raw[\"prtage\"] >= 16)\n",
" & (raw[\"pehruslt\"] > 0)\n",
" & (raw[\"pternwa\"] > 0)\n",
" ].copy()\n",
" weekly_earnings = month_org[\"pternwa\"].astype(float) / 100\n",
" direct = month_org[\"pternhly\"].astype(float) / 100\n",
" hourly = np.where(\n",
" (month_org[\"peernhry\"] == 1) & (direct > 0),\n",
" direct,\n",
" weekly_earnings / month_org[\"pehruslt\"].astype(float),\n",
" )\n",
" src = np.where(\n",
" (month_org[\"peernhry\"] == 1) & (direct > 0),\n",
" \"direct\",\n",
" \"derived\",\n",
" )\n",
" tail_frames.append(\n",
" pd.DataFrame(\n",
" {\n",
" \"hourly\": hourly,\n",
" \"source\": src,\n",
" \"sample_weight\": month_org[\"pworwgt\"].astype(float),\n",
" }\n",
" )\n",
" )\n",
"\n",
"tail = pd.concat(tail_frames, ignore_index=True)\n",
"tail_summary = []\n",
"for cutoff in [100, 200, 500]:\n",
" mask = tail[\"hourly\"] > cutoff\n",
" subset = tail.loc[mask]\n",
" tail_summary.append(\n",
" {\n",
" \"cutoff\": cutoff,\n",
" \"rows\": int(mask.sum()),\n",
" \"weighted_share\": round(\n",
" float(np.average(mask.astype(float), weights=tail[\"sample_weight\"])),\n",
" 4,\n",
" ),\n",
" \"direct_share_within_tail_weighted\": round(\n",
" float(\n",
" np.average(\n",
" (subset[\"source\"] == \"direct\").astype(float),\n",
" weights=subset[\"sample_weight\"],\n",
" )\n",
" ),\n",
" 4,\n",
" ) if len(subset) else None,\n",
" \"derived_share_within_tail_weighted\": round(\n",
" float(\n",
" np.average(\n",
" (subset[\"source\"] == \"derived\").astype(float),\n",
" weights=subset[\"sample_weight\"],\n",
" )\n",
" ),\n",
" 4,\n",
" ) if len(subset) else None,\n",
" }\n",
" )\n",
"\n",
"pd.DataFrame(tail_summary)"
]
},
{
"cell_type": "markdown",
"id": "1f1918e0",
"metadata": {},
"source": [
"## Reading Guide\n",
"\n",
"- The annual wage distribution is the primary donor distribution used for `hourly_wage` and `is_paid_hourly`.\n",
"- The state union table reflects the BLS-based assignment used instead of the sparse public-use CPS union microdata fields.\n",
"- The tail audit shows whether very high hourly wages are coming from direct hourly reports or from the weekly-earnings divided by hours fallback."
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python (us-data-venv)",
"language": "python",
"name": "us-data-venv"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.13.12"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
Raw
org_validation_2024_summary.md

ORG Validation: 2024 Annual Donor Build

Validation artifact for the ORG donor build used in policyengine-us-data.

Files:

  • org_validation_2024.ipynb: executed notebook

Key annual results:

  • donor rows: 119,237
  • weighted mean hourly wage: $35.17
  • weighted hourly wage quantiles: p10 $14.40, p50 $25.00, p90 $62.50, p99 $222.37
  • weighted is_paid_hourly: 55.32%
  • weighted is_union_member_or_covered: 11.01%

State union-rate pattern also looks directionally right:

  • low states: NC 2.96%, SD 3.88%, SC 4.11%, GA 4.31%, AR 4.40%
  • high states: HI 26.71%, NY 21.67%, AK 19.27%, WA 17.64%, NJ 17.59%

Main caveat:

  • the high hourly-wage tail is still worth watching; values above $100/hr are largely driven by reconstructing hourly wage as weekly earnings divided by hours when no direct hourly rate is reported.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment