IMPACT OF BIG DATA AND PREDICTIVE ANALYTICS ON FINANCIAL FORECASTING ACCURACY AND DECISION-MAKING IN GLOBAL CAPITAL MARKETS

Abstract

This quantitative study examined how big data intensity and predictive-analytics capability influenced financial forecasting accuracy and decision-making quality in global capital markets. The study was grounded in an extensive review of 112 peer-reviewed empirical and methodological papers spanning econometric forecasting, alternative data applications, machine learning in asset pricing, and decision-value testing across international markets. Using a cross-sectional time-series panel of market–asset–time observations, the analysis integrated developed and emerging markets and covered equities, fixed income, foreign exchange, and derivatives. Big data intensity was operationalized as a composite index reflecting data volume, variety, frequency, latency, dimensionality, and coverage breadth, while predictive-analytics capability was measured as a ranked model-class indicator ranging from classical econometrics to hybrid econometric–machine-learning systems. Forecasting accuracy served as the primary outcome and was evaluated through rolling out-of-sample errors; decision-making quality was assessed through standardized portfolio, trading, and risk-performance indicators.

Descriptive results showed substantial dispersion across markets, with mean big data intensity at 0.56 (SD = 0.19) and predictive-analytics capability at 3.12 (SD = 1.09) on a five-point scale. Developed markets exhibited higher data intensity (0.63) and sophistication (3.54) than emerging markets (0.48 and 2.66, respectively). Correlation analysis indicated that big data intensity and predictive-analytics capability were each inversely associated with forecasting error (r = −0.46 and r = −0.52), and forecasting error was inversely related to decision-making quality (r = −0.48). Fixed-effects panel regressions confirmed significant direct effects: big data intensity reduced forecasting error (β = −0.021, p < .001) and predictive-analytics capability reduced forecasting error (β = −0.018, p < .001). The interaction term was also negative (β = −0.007, p < .01), indicating complementarity such that analytics benefits strengthened under higher data intensity. In the decision model, forecasting error significantly predicted decision quality (β = −28.6, p < .001), demonstrating that statistical accuracy gains translated into improved economic outcomes. Overall, the study provided integrated evidence that richer data environments and advanced predictive analytics jointly enhanced forecasting precision and decision performance across global capital markets.