Climate change impacts assessment involves downscaling of coarse-resolution climate variables simulated by general circulation models (GCMs) using dynamic (physics-based) or statistical (data-driven) approaches. Here we use a statistical downscaling technique for projections of all-India monsoon rainfall at a resolution of 0.5° in latitude/longitude. The present statistical downscaling model utilizes classification and regression tree, and kernel regression and develops a statistical relationship between large-scale climate variables from reanalysis data and fine-resolution observed rainfall, and then applies the relationship to coarse-resolution GCM outputs. A GCM developed by the Canadian Centre for Climate Modeling and Analysis is employed for this study with its five ensemble runs for capturing intramodel uncertainty. The model appears to effectively capture individual station means, the spatial patterns of the standard deviations, and the cross correlation between station rainfalls. Computationally expensive dynamic downscaling models have been applied for India. However, our study is the first to attempt statistical downscaling for the entire country at a resolution of 0.5°. The downscaling model seems to capture the orographic effect on rainfall in mountainous areas of the Western Ghats and northeast India. The model also reveals spatially nonuniform changes in rainfall, with a possible increase for the western coastline and northeastern India (rainfall surplus areas) and a decrease in northern India, western India (rainfall deficit areas), and on the southeastern coastline, highlighting the need for a detailed hydrologic study that includes future projections regarding water availability which may be useful for water resource policy decisions.