This work documented that investor narratives are net originators of information that propagates into systematic risk factors, reframing Fama-French factors as endogenous downstream constructs rather than exogenous risk primitives. Using daily data from 2015 to 2024 for 503 S&P 500 constituents, variance-weighted indices were constructed from analyst recommendations, social media, and stock prices, combined with Fama-French factors and Global Database of Events, Language and Tone (GDELT) media-attention measures. Quantile vector autoregression (VAR) network analysis showed that the market, size, and value factors are net information receivers, whereas growth and value narratives are net transmitters. Granger causality tests indicated that analyst recommendations predict factor movements, while social-media sentiment follows rather than leads prices. Structural impulse responses identified a transmission mechanism: growth narrative shocks significantly reduce the size premium, consistent with narrative-driven capital reallocation toward large-cap stocks.
Citation: José Gabriel Astaíza-Gómez. Narratives as risk transmitters[J]. Data Science in Finance and Economics, 2026, 6(4): 567-594. doi: 10.3934/DSFE.2026020
This work documented that investor narratives are net originators of information that propagates into systematic risk factors, reframing Fama-French factors as endogenous downstream constructs rather than exogenous risk primitives. Using daily data from 2015 to 2024 for 503 S&P 500 constituents, variance-weighted indices were constructed from analyst recommendations, social media, and stock prices, combined with Fama-French factors and Global Database of Events, Language and Tone (GDELT) media-attention measures. Quantile vector autoregression (VAR) network analysis showed that the market, size, and value factors are net information receivers, whereas growth and value narratives are net transmitters. Granger causality tests indicated that analyst recommendations predict factor movements, while social-media sentiment follows rather than leads prices. Structural impulse responses identified a transmission mechanism: growth narrative shocks significantly reduce the size premium, consistent with narrative-driven capital reallocation toward large-cap stocks.
| [1] |
Afzal T, Afridi MA, Jan MN (2025) Integrating LSTM with Fama-French six factor model for predicting portfolio returns: Evidence from Shenzhen stock market China. Data Sci Financ Econ 5: 177–204. https://doi.org/10.3934/DSFE.2025009 doi: 10.3934/DSFE.2025009
|
| [2] |
Bybee L, Kelly B, Su Y (2023) Narrative asset pricing: Interpretable systematic risk factors from news text. Rev Financ Stud 36: 4759–4787. https://doi.org/10.1093/rfs/hhad042 doi: 10.1093/rfs/hhad042
|
| [3] |
Chatziantoniou I, Gabauer D, Stenfors A (2021) Interest rate swaps and the transmission mechanism of monetary policy: A quantile connectedness approach. Econ Lett 204: 109891. https://doi.org/10.1016/j.econlet.2021.109891 doi: 10.1016/j.econlet.2021.109891
|
| [4] |
Dabara KA, Tahir IM (2025) Leveraging hybrid ensemble models in stock market prediction: A data-driven approach. Data Sci Financ Econ 5: 355–386. https://doi.org/10.3934/DSFE.2025015 doi: 10.3934/DSFE.2025015
|
| [5] |
Dash SR, Gabauer D, Goel G, Subramaniam S (2024) Investigating dynamic connectedness of global equity markets: the role of investor attention. Appl Econ 56: 7222–7243. https://doi.org/10.1080/00036846.2023.2277705 doi: 10.1080/00036846.2023.2277705
|
| [6] |
Diebold FX, Yilmaz K (2012) Better to give than to receive: Predictive directional measurement of volatility spillovers. Int J Forecast 28: 57–66. https://doi.org/10.1016/j.ijforecast.2011.02.006 doi: 10.1016/j.ijforecast.2011.02.006
|
| [7] |
Diebold FX, Yilmaz K (2014) On the network topology of variance decompositions: Measuring the connectedness of financial firms. J Econom 182: 119–134. https://doi.org/10.1016/j.jeconom.2014.04.012 doi: 10.1016/j.jeconom.2014.04.012
|
| [8] |
Fama EF, French KR (1996) Multifactor explanations of asset pricing anomalies. J Finance 51: 55–84. https://doi.org/10.1111/j.1540-6261.1996.tb05202.x doi: 10.1111/j.1540-6261.1996.tb05202.x
|
| [9] |
Fama EF, French KR (2016) Dissecting anomalies with a five-factor model. Rev Financ Stud 29: 69–103. https://doi.org/10.1093/rfs/hhv043 doi: 10.1093/rfs/hhv043
|
| [10] |
Gabauer D (2021) Dynamic measures of asymmetric and pairwise connectedness within an optimal currency area: Evidence from the ERM I system. J Multinatl Financ Manag 60: 100680. https://doi.org/10.1016/j.mulfin.2021.100680 doi: 10.1016/j.mulfin.2021.100680
|
| [11] |
Goetzmann WN, Kim D, Shiller RJ (2022) Crash narratives. NBER Working Paper No. 30195. https://doi.org/10.3386/w30195 |
| [12] |
Goldfarb D, Idnani A (1983) A numerically stable dual method for solving strictly convex quadratic programs. Math Program 27: 1–33. https://doi.org/10.1007/BF02591962 doi: 10.1007/BF02591962
|
| [13] |
Granger CW (1969) Investigating causal relations by econometric models and cross-spectral methods. Econometrica 37: 424–438. https://doi.org/10.2307/1912791 doi: 10.2307/1912791
|
| [14] |
Hilary G, Hsu C (2013) Analyst forecast consistency. J Finance 68: 271–297. https://doi.org/10.1111/j.1540-6261.2012.01800.x doi: 10.1111/j.1540-6261.2012.01800.x
|
| [15] |
Hirshleifer D (2020) Presidential address: Social transmission bias in economics and finance. J Finance 75: 1779–1831. https://doi.org/10.1111/jofi.12906 doi: 10.1111/jofi.12906
|
| [16] |
Johnson SG, Tuckett D (2022) Narrative expectations in financial forecasting. J Behav Decis Mak 35: e2245. https://doi.org/10.1002/bdm.2245 doi: 10.1002/bdm.2245
|
| [17] |
Khalfaoui R, Jabeur SB, Dogan B (2022) The spillover effects and connectedness among green commodities, Bitcoins, and US stock markets: Evidence from the quantile VAR network. J Environ Manag 306: 114493. https://doi.org/10.1016/j.jenvman.2022.114493 doi: 10.1016/j.jenvman.2022.114493
|
| [18] |
Li Z, Chen Z, Huang Z (2026) Modelling the data-generating mechanism of China's commodity market by identifying hidden information flow regimes. Financ Innov 12: Article 28. https://doi.org/10.1186/s40854-025-00804-w doi: 10.1186/s40854-025-00804-w
|
| [19] |
Long S, Chatziantoniou I, Gabauer D, Lucey B (2024) Do social media sentiments drive cryptocurrency intraday price volatility? New evidence from asymmetric TVP-VAR frequency connectedness measures. Eur J Finance 30(: 1470–1489. https://doi.org/10.1080/1351847X.2024.2314085 doi: 10.1080/1351847X.2024.2314085
|
| [20] |
Merton RC (1973) An intertemporal capital asset pricing model. Econometrica 41: 867–887. https://doi.org/10.2307/1913811 doi: 10.2307/1913811
|
| [21] |
Sharif T, Bouteska A, Abedin MZ, Cotturone S (2025) An enquiry into the monetary policy and stock market shocks in the US. Int Rev Econ Finance 98: 103925. https://doi.org/10.1016/j.iref.2025.103925 doi: 10.1016/j.iref.2025.103925
|
| [22] |
Shiller RJ (2019) Narrative Economics: How Stories Go Viral and Drive Major Economic Events. Princeton University Press. |
| [23] |
Shiller RJ (2015) Irrational Exuberance, 3rd ed. Princeton University Press. |
| [24] |
Sims CA (1980) Macroeconomics and reality. Econometrica 48: 1–48. https://doi.org/10.2307/1912017 doi: 10.2307/1912017
|
Figures(11) / Tables(9)
José Gabriel Astaíza-Gómez. Narratives as risk transmitters[J]. Data Science in Finance and Economics, 2026, 6(4): 567-594. doi: 10.3934/DSFE.2026020
Analyst coverage
Histograms of analyst price target gaps and gap volatilities
Price target gap and gap volatility for Apple stock
Informativeness weights
Market indices: Price, analyst activity, and social sentiment
Fama-French factors and GDELT article counts
Stationarity. All eigenvalues lie within the unit circle, indicating the stability of the VAR model. The variables included in (a) are: projected returns, buy recommendations, sell recommendations, number of Twitter publications, Twitter sentiment, growth, value and Fama-French factors. The variables included in (b) are: buy recommendations, sell recommendations, number of Twitter publications, Twitter sentiment, growth, value and Fama-French factors
(a) Heatmap visualization of the cross-quantilogram matrix. The X- axis represents the quantile values of the price target index and the Y-axis represents the quantile values of the stock price index. The critical values are obtained by stationary bootstrap. (b) Time series plot of the projected stock return
Net pairwise directional connectedness (NPDC) based on the quantile VAR framework with a lag length as suggested by BIC and 6-step-ahead projection with a 150-day rolling window
(a) Response of HML to an impulse from the growth narrative; (b) response of SMB to an impulse from the growth narrative; (c) response of HML to an impulse from the value narrative; (d) response of SMB to an impulse from the value narrative; (e) response of SMB to an impulse from the growth narrative placing Mkt-RF first in the SVAR ordering
(a) Response of buy recommendations to an impulse from projected returns; (b) response of buy recommendations to an impulse from Twitter sentiment
DownLoad: