ABSTRACT The post common-envelope eclipsing binary HW Virginis (HW Vir) has had many circumbinary companions proposed based on eclipse timing variations. Each proposed solution has lacked in predictability and orbital stability, leaving the origin of the eclipse timing variations an active area of research. Leveraging the catalogue of Hipparcos and Gaia proper motion anomalies, we show there is slight evidence for a circumbinary companion orbiting HW Vir. We place an upper limit in mass for such a companion which excludes some previously claimed companions. We also apply this method to V471 Tauri and confirm the non-detection of a previously claimed brown dwarf. We adapt the kima nested sampling code to analyse eclipse timing variations and re-analyse archival data on HW Vir, varying the order of the ephemeris that we fit for and the amount of the data that we use. Although signals are clearly present, we find two signals around 2500 and 4000-d periods that are not coherent between different chunks of the data, so are likely to not be of planetary origin. We analyse the whole data set and find the best solution to contain four signals. Of these four we argue the outermost is the most compatible with astrometry and thus the most likely to be of planetary nature. We posit the other three pseudo-periodic signals are caused by physical processes on the white dwarf. The eventual release of the full Gaia epoch astrometry is a promising way to confirm whether circumbinary planets exist around HW Vir (and other similar systems), and explore white dwarf physics.