Imperata grassland soils are widely perceived as having poor physical and chemical properties that render them unproductive and prone to erosion. They are therefore increasingly targeted for reforestation across the tropics. To better understand how reforestation and forest growth affect soil hydrological processes we compared the soil physical characteristics for an Imperata grassland and a 23-year-old ‘reforest’ on Leyte Island, the Philippines. Saturated hydraulic conductivity (Ksat) was determined in the field (Amoozemeter, 20–90 cm depth) and the laboratory (small-core permeametry). Core-based values of Ksat were (much) lower than field-based values, suggesting macropores were not sampled adequately with the small cores. Ksat decreased exponentially with depth in both land-cover types, with a median field-measured Ksat of the grassland of 2.1 mm h−1 at the surface and 2.9 mm h−1 at 20–40 cm depth, declining to ≤1 mm h−1 below 60 cm. Corresponding values for the reforest were 59 (at 20 cm), 37 (at 40 cm) and 7.3 mm h−1 (at 60–100 cm depth). Reforest Ksat-values down to 60 cm depth were significantly higher than corresponding values in the grassland, but the difference disappeared at 90 cm depth. Organic carbon content in the top 40 cm of soil was slightly higher in the reforest than the grassland. Bulk density was higher and porosity marginally lower in the grassland than the reforest at all depths considered. The median five-min rainfall intensity (June 2013–May 2014) was 3.2 mm h−1, suggesting that >50% of the rainfall might generate Hortonian overland flow in the grassland. Overland flow is unlikely in the reforest where lateral flow is likely to be generated around 60 cm depth for ~30% of rain-time, versus 2–3% between 20 and 60 cm. Within the limitations of the space-for-time substitution approach, these results suggest that 23 years of forest development at Manobo had a positive effect on hillslope hydrological functioning.