Independent verification of national greenhouse gas inventories is a vital measure for cross-checking the accuracy of emissions data submitted to the United Nations Framework Convention on Climate Change (UNFCCC). We infer annual UK emissions of HFC-134a from 1995 to 2012 using atmospheric observations and an inverse modeling technique, and compare with the UK's annual UNFCCC submission. By 2010, the inventory is almost twice as large as our estimates, with an "emissions gap" equating to 3.90 (3.20-4.30) Tg CO2e. We evaluate the RAC (Refrigeration and Air-Conditioning) model, a bottom up model used to quantify UK emissions from refrigeration and air-conditioning sectors. Within mobile air-conditioning (MAC), the largest RAC sector and most significant UK source (59%), we find a number of assumptions that may be considered oversimplistic and conservative; most notably the unit refill rate. Finally, a Bayesian approach is used to estimate probable inventory inputs required for minimization of the emissions discrepancy. Our top-down estimates provide only a weak constraint on inventory model parameters and consequently, we are unable to suggest discrete values. However, a significant revision of the MAC servicing rate, coupled with a reassessment of non-RAC aerosol emissions, are required if the discrepancy between methods is to be reduced.