{
  "generated_at": "2026-10-07T08:13:01.619164+00:00",
  "paper": {
    "label": "Professional commentary",
    "period": "1992–2024",
    "source": "WSJ, NYT and FT",
    "first_date": "1992-01-02",
    "last_date": "2024-11-29",
    "firms": 228,
    "summary": {
      "posts": 74525,
      "negative": 15587,
      "neutral": 40070,
      "positive": 18868,
      "missing_inference": 324
    },
    "annual": [
      {
        "year": 1992,
        "posts": 1593,
        "negative": 319,
        "neutral": 695,
        "positive": 579,
        "missing_inference": 0
      },
      {
        "year": 1993,
        "posts": 1814,
        "negative": 382,
        "neutral": 732,
        "positive": 700,
        "missing_inference": 0
      },
      {
        "year": 1994,
        "posts": 4174,
        "negative": 447,
        "neutral": 2698,
        "positive": 1029,
        "missing_inference": 0
      },
      {
        "year": 1995,
        "posts": 5248,
        "negative": 818,
        "neutral": 2866,
        "positive": 1564,
        "missing_inference": 0
      },
      {
        "year": 1996,
        "posts": 5731,
        "negative": 1072,
        "neutral": 2845,
        "positive": 1814,
        "missing_inference": 0
      },
      {
        "year": 1997,
        "posts": 6231,
        "negative": 1096,
        "neutral": 3405,
        "positive": 1730,
        "missing_inference": 0
      },
      {
        "year": 1998,
        "posts": 6104,
        "negative": 1219,
        "neutral": 3339,
        "positive": 1546,
        "missing_inference": 0
      },
      {
        "year": 1999,
        "posts": 6055,
        "negative": 1034,
        "neutral": 3435,
        "positive": 1586,
        "missing_inference": 0
      },
      {
        "year": 2000,
        "posts": 1018,
        "negative": 218,
        "neutral": 596,
        "positive": 204,
        "missing_inference": 0
      },
      {
        "year": 2001,
        "posts": 1713,
        "negative": 451,
        "neutral": 878,
        "positive": 384,
        "missing_inference": 0
      },
      {
        "year": 2002,
        "posts": 3011,
        "negative": 793,
        "neutral": 1487,
        "positive": 731,
        "missing_inference": 0
      },
      {
        "year": 2003,
        "posts": 1728,
        "negative": 427,
        "neutral": 877,
        "positive": 424,
        "missing_inference": 0
      },
      {
        "year": 2004,
        "posts": 1047,
        "negative": 239,
        "neutral": 572,
        "positive": 236,
        "missing_inference": 0
      },
      {
        "year": 2005,
        "posts": 1460,
        "negative": 292,
        "neutral": 807,
        "positive": 361,
        "missing_inference": 0
      },
      {
        "year": 2006,
        "posts": 1324,
        "negative": 270,
        "neutral": 729,
        "positive": 325,
        "missing_inference": 0
      },
      {
        "year": 2007,
        "posts": 1671,
        "negative": 372,
        "neutral": 912,
        "positive": 387,
        "missing_inference": 0
      },
      {
        "year": 2008,
        "posts": 2239,
        "negative": 693,
        "neutral": 1114,
        "positive": 432,
        "missing_inference": 0
      },
      {
        "year": 2009,
        "posts": 2668,
        "negative": 665,
        "neutral": 1320,
        "positive": 683,
        "missing_inference": 0
      },
      {
        "year": 2010,
        "posts": 2367,
        "negative": 505,
        "neutral": 1266,
        "positive": 596,
        "missing_inference": 0
      },
      {
        "year": 2011,
        "posts": 1935,
        "negative": 448,
        "neutral": 1051,
        "positive": 436,
        "missing_inference": 2
      },
      {
        "year": 2012,
        "posts": 1777,
        "negative": 426,
        "neutral": 977,
        "positive": 374,
        "missing_inference": 1
      },
      {
        "year": 2013,
        "posts": 1379,
        "negative": 320,
        "neutral": 781,
        "positive": 278,
        "missing_inference": 0
      },
      {
        "year": 2014,
        "posts": 1672,
        "negative": 356,
        "neutral": 1035,
        "positive": 281,
        "missing_inference": 2
      },
      {
        "year": 2015,
        "posts": 1430,
        "negative": 330,
        "neutral": 822,
        "positive": 278,
        "missing_inference": 6
      },
      {
        "year": 2016,
        "posts": 1252,
        "negative": 342,
        "neutral": 676,
        "positive": 234,
        "missing_inference": 34
      },
      {
        "year": 2017,
        "posts": 1478,
        "negative": 329,
        "neutral": 850,
        "positive": 299,
        "missing_inference": 47
      },
      {
        "year": 2018,
        "posts": 1418,
        "negative": 332,
        "neutral": 823,
        "positive": 263,
        "missing_inference": 78
      },
      {
        "year": 2019,
        "posts": 1095,
        "negative": 352,
        "neutral": 558,
        "positive": 185,
        "missing_inference": 34
      },
      {
        "year": 2020,
        "posts": 674,
        "negative": 169,
        "neutral": 327,
        "positive": 178,
        "missing_inference": 33
      },
      {
        "year": 2021,
        "posts": 760,
        "negative": 205,
        "neutral": 388,
        "positive": 167,
        "missing_inference": 20
      },
      {
        "year": 2022,
        "posts": 877,
        "negative": 206,
        "neutral": 488,
        "positive": 183,
        "missing_inference": 20
      },
      {
        "year": 2023,
        "posts": 804,
        "negative": 213,
        "neutral": 386,
        "positive": 205,
        "missing_inference": 27
      },
      {
        "year": 2024,
        "posts": 778,
        "negative": 247,
        "neutral": 335,
        "positive": 196,
        "missing_inference": 20
      }
    ],
    "sources": [
      {
        "source": "FT",
        "posts": 6469,
        "negative": 1491,
        "neutral": 3626,
        "positive": 1352,
        "missing_inference": 36
      },
      {
        "source": "NYT",
        "posts": 2515,
        "negative": 861,
        "neutral": 1297,
        "positive": 357,
        "missing_inference": 24
      },
      {
        "source": "WSJ",
        "posts": 65541,
        "negative": 13235,
        "neutral": 35147,
        "positive": 17159,
        "missing_inference": 264
      }
    ],
    "matrix": [
      {
        "tone": -1,
        "sentiment": -1,
        "count": 11090
      },
      {
        "tone": -1,
        "sentiment": 0,
        "count": 19175
      },
      {
        "tone": -1,
        "sentiment": 1,
        "count": 6018
      },
      {
        "tone": 0,
        "sentiment": -1,
        "count": 3827
      },
      {
        "tone": 0,
        "sentiment": 0,
        "count": 15501
      },
      {
        "tone": 0,
        "sentiment": 1,
        "count": 8739
      },
      {
        "tone": 1,
        "sentiment": -1,
        "count": 670
      },
      {
        "tone": 1,
        "sentiment": 0,
        "count": 5394
      },
      {
        "tone": 1,
        "sentiment": 1,
        "count": 4111
      }
    ],
    "tone_summary": {
      "-1": 36283,
      "0": 28067,
      "1": 10175
    },
    "agreement": 30702,
    "moments": {
      "mean": 0.04402549480040255,
      "std": 0.6785248789639028
    },
    "method": "Canonical 74,525-article sample. Raw article inference merged with the supplementary article inference, keeping the last label per article. The pipeline used for the working paper assigns neutral to 324 articles without a valid raw inference label. Dictionary tone uses (positive words − negative words)/(positive words + negative words), with threshold 0.25."
  },
  "forum": {
    "label": "Retail discussion",
    "period": "2010–2025",
    "source": "Eastmoney Guba stock boards",
    "last_date": "2025-08-07",
    "annual": [
      {
        "year": 2010,
        "posts": 204469,
        "negative": 21572,
        "neutral": 147714,
        "positive": 35183,
        "all_posts": 204572
      },
      {
        "year": 2011,
        "posts": 307364,
        "negative": 41413,
        "neutral": 209416,
        "positive": 56535,
        "all_posts": 311875
      },
      {
        "year": 2012,
        "posts": 195806,
        "negative": 45306,
        "neutral": 91165,
        "positive": 59335,
        "all_posts": 196250
      },
      {
        "year": 2013,
        "posts": 1222826,
        "negative": 151499,
        "neutral": 887275,
        "positive": 184052,
        "all_posts": 1247520
      },
      {
        "year": 2014,
        "posts": 2505559,
        "negative": 607057,
        "neutral": 1319875,
        "positive": 578627,
        "all_posts": 2517529
      },
      {
        "year": 2015,
        "posts": 6570314,
        "negative": 1925860,
        "neutral": 2970860,
        "positive": 1673594,
        "all_posts": 6580074
      },
      {
        "year": 2016,
        "posts": 16767574,
        "negative": 5523164,
        "neutral": 7242726,
        "positive": 4001684,
        "all_posts": 16778500
      },
      {
        "year": 2017,
        "posts": 25101583,
        "negative": 7823695,
        "neutral": 12080444,
        "positive": 5197444,
        "all_posts": 25116047
      },
      {
        "year": 2018,
        "posts": 23174433,
        "negative": 8102147,
        "neutral": 10251651,
        "positive": 4820635,
        "all_posts": 23195321
      },
      {
        "year": 2019,
        "posts": 31720576,
        "negative": 11171244,
        "neutral": 13107445,
        "positive": 7441887,
        "all_posts": 31734353
      },
      {
        "year": 2020,
        "posts": 47670467,
        "negative": 17265522,
        "neutral": 19186948,
        "positive": 11217997,
        "all_posts": 47684359
      },
      {
        "year": 2021,
        "posts": 55865998,
        "negative": 20079217,
        "neutral": 22965865,
        "positive": 12820916,
        "all_posts": 55879053
      },
      {
        "year": 2022,
        "posts": 48241198,
        "negative": 18011895,
        "neutral": 19975940,
        "positive": 10253363,
        "all_posts": 48260584
      },
      {
        "year": 2023,
        "posts": 36501573,
        "negative": 12019035,
        "neutral": 17017927,
        "positive": 7464611,
        "all_posts": 38916383
      },
      {
        "year": 2024,
        "posts": 36177807,
        "negative": 11752370,
        "neutral": 16858004,
        "positive": 7567433,
        "all_posts": 39291313
      },
      {
        "year": 2025,
        "posts": 26142181,
        "negative": 8481325,
        "neutral": 12124093,
        "positive": 5536763,
        "all_posts": 27788196
      }
    ],
    "summary": {
      "posts": 358369728,
      "negative": 123022321,
      "neutral": 156437348,
      "positive": 78910059,
      "all_posts": 365701929
    },
    "method": "Eastmoney Guba discussion archive. Annual counts across stock boards. Shares use only valid −1/0/+1 labels. Posts with missing or unresolved labels are excluded from sentiment shares. The 2025 sample ends on 7 August. Publication dates include non-trading days."
  },
  "validation": {
    "training_posts": 9000,
    "benchmark_posts": 600,
    "benchmark_accuracy": 89.17,
    "benchmark_f1": {
      "neutral": 82.35,
      "positive": 90.26,
      "negative": 93.84
    },
    "external": [
      {
        "sample": "English translation",
        "n": 600,
        "accuracy": 86.83
      },
      {
        "sample": "StockTwits",
        "n": 210,
        "accuracy": 90.0
      },
      {
        "sample": "Long-form financial texts",
        "n": 135,
        "accuracy": 90.37
      }
    ]
  }
}