We explore the burst energy distribution of fast radio bursts (FRBs) in the low-twist magnetar model of Wadiasingh & Timokhin (2019). Motivated by the power-law fluence distributions of FRB 121102, we propose an elementary model for the FRB luminosity function of individual repeaters with an inversion protocol which directly relates the power-law distribution index of magnetar short burst fluences to that for FRBs. The protocol indicates the FRB energy scales virtually linearly with crust/field dislocation amplitude, if magnetar short bursts prevail in the magnetoelastic regime. Charge starvation in the magnetosphere during bursts (required in WT19) for individual repeaters implies the predicted burst fluence distribution is narrow, < ∼ 3 decades for yielding strains and oscillation frequencies feasible in magnetar crusts. Requiring magnetic confinement and charge starvation, we obtain a death line for FRBs which segregates magnetars from the normal pulsar population, suggesting only the former will host FRBs. We convolve the burst energy distribution for individual magnetars to define the distribution of luminosities in evolved magnetar populations. The broken power-law luminosity function's low energy character depends on the population model, while the high energy index traces that of individual repeaters. Independent of the evolved population, the broken power-law isotropic-equivalent energy/luminosity function peaks at ∼ 10 38 − 10 40 erg with a sharp low-energy cutoff at < ∼ 10 37 erg. Lastly, we consider the local fluence distribution of FRBs, and find that it can constrain the subset of FRB-producing magnetar progenitors. Our model suggests that improvements in sensitivity may reveal flattening of the global FRB fluence distribution and saturation in FRB rates. 1/7 c and ε CR,pp ∝ ρ 1/2 c . There also exist regimes between B ∼ 10 13 − 10 14 G where the splitting length may be shorter than the pair length, owing to the dependence of splitting amplitude integrals M σ on B (for details, see Hu et al. 2019). A more complete study of pair cascades with splitting for FRB models is deferred to the future.