Carbon accounting is a key issue in the discussions on global warming/CO$_2$mitigation. This paper applies both the IPCC Approach and the NEAT (Non-Energy use Emission Accounting Tables) model, a bottom-up approach, to estimate the potential CO$_2$ emissions (carbon storage) originating from the non-energy use as to assess the actual CO$_2$ emissions (carbon release) from the use of fossil fuels in Korea. The current Korean carbon accounting seems to overestimate the potential CO$_2$ emissions and with it to underestimate the actual CO$_2$ emissions. The estimation shows that the potential CO$_2$ emissions calculated according to the IPCC Approach are lower than those calculated using the NEAT model. This is because the IPCC default storage fraction for naphtha seems to be low for the Korean petrochemical production structure, on the one hand and because the IPCC Approach does not consider the trade with short life petrochemical products, on the other hand. This paper shows that a bottom-up approach like the NEAT model can contribute to overcome some of limitations of the IPCC guidelines, especially by considering the international trade with short life petrochemical products and by estimating the storage fractions of fossil fuels used as feedstocks for the country in consideration. This paper emphasizes the importance of accurate energy statistics for carbon accounting.