<![CDATA[Google Trends search-interest data are widely reported to improve tourism-demand forecasts, but most evidence relies on monthly or weekly series. This study examines whether an annual composite Google Trends Index (GTI), constructed from four Malaysia-related English-language queries, is associated with and adds incremental predictive information for Malaysian international tourist arrivals from 2014 to 2024. The full-sample contemporaneous correlation is very high (r = 0.952; r² ≈ 0.906), but the association reverses in the small pre-pandemic window (2015–2019; r = −0.902), while the one-year-lagged correlation is near zero (r = −0.127, N = 4). In nested annual regressions, lagged GTI is not statistically significant and is affected by substantial multicollinearity. A leave-one-year-out sensitivity analysis also yields higher RMSE and MAE for the GTI-augmented model than for the AR(1) benchmark. Because this validation is not temporally ordered and the sample is extremely small, it should not be interpreted as definitive prospective forecasting evidence. Instead, the combined findings indicate that the strong pandemic-inclusive correlation is unstable and likely reflects shared collapse-and-recovery dynamics. Claims about routine forecasting value for Malaysia therefore require replication with a longer monthly series, source-market-sensitive search measures, and rolling-origin validation.]]>
Copyrights © 2026
