Abstract: The most pristine remnants of the Solar system's planet formation epoch orbit the Sun beyond Neptune, the small bodies of the trans-Neptunian object populations. The bulk of the mass is in ~100 km objects, but objects at smaller sizes have undergone minimal collisional processing, with New Horizons recently revealing that ~20 km effective diameter body (486958) Arrokoth appears to be a primordial body, not a collisional fragment. This indicates bodies at these sizes (and perhaps smaller) retain a record of how they were formed, and are the most numerous record of that epoch. However, such bodies are impractical to find by optical surveys due to their very low brightnesses. Their presence can be inferred from the observed cratering record of Pluto and Charon, and directly measured by serendipitous stellar occultations. These two methods produce conflicting results, with occultations measuring roughly ten times the number of ~km bodies inferred from the cratering record. We use numerical models to explore how these observations can be reconciled with evolutionary models of the outer Solar system. We find that models where the initial size of bodies decreases with increasing semimajor axis of formation, and models where the surface density of bodies increases beyond the 2:1 mean-motion resonance with Neptune can produce both sets of observations, though comparison to various observational tests favours the former mechanism. We discuss how to evaluate the astrophysical plausibility of these solutions, and conclude extended serendipitous occultation surveys with broad sky coverage are the most practical approach.