Many scientists are interested in but puzzled by the various inverse power laws with a negative exponent 1 such as the rank-size rule. The rank-size rule is a very simple scaling law followed by many observations of the ubiquitous empirical patterns in physical and social systems. Where there is a rank-size distribution, there will be a hierarchy with cascade structure. However, the equivalence relation between the rank-size rule and the hierarchical scaling law remains to be mathematically demonstrated and empirically testified. In this paper, theoretical derivation, mathematical experiments, and empirical analysis are employed to show that the rank-size rule is equivalent in theory to the hierarchical scaling law (the Nn principle). Abstracting an ordered set of quantities in the form {1, 1/2,..., 1/k,...} from the rank-size rule, I prove a geometric subdivision theorem of the harmonic sequence (k=1, 2, 3,...). By the theorem, the rank-size distribution can be transformed into a self-similar hierarchy, thus a power law can be decomposed as a pair of exponential laws, and further the rank-size power law can be reconstructed as a hierarchical scaling law. A number of ubiquitous empirical observations and rules, including Zipf's law, Pareto's distribution, fractals, allometric scaling, 1/f noise, can be unified into the hierarchical framework. The self-similar hierarchy can provide us with a new perspective of looking at the inverse power law of nature or even how nature works.