Higher-order statistics, such as skewness and kurtosis, are useful measures to describe the shape of a probability distribution. In hydrology, the assessment of these parameters is an important step for representing precipitation frequencies and understanding rainfall variability. Because of restrictions related to data acquisition and methods of analysis, the global patterns of higher-order precipitation statistics are still poorly described. The objective of this study is to combine global precipitation datasets and state-of-the-art statistical methods in order to build a spatio-temporal assessment of global precipitation frequency. Gridded precipitation records were obtained from NOAA's Precipitation Reconstruction over Land (PREC/L) product. The L-moments approach was applied to describe the shape of the probability distribution and estimate the function parameters in each point of the global grid. Like conventional statistical moments, the L-moments can be used to describe the shape of a probability distribution, but with the advantage of being less susceptible to the presence of outliers and performing better with small sample sizes. The results of this study provide a new perspective on the spatial and temporal patterns of the parameters defining the shape of the precipitation probability distribution at a global scale.