The persistent premium of government debt attributes to two main reasons: absolute nominal safety and liquidity. This paper employs two types of measures of government debt supply to disentangle the safety and liquidity part of the premium. The empirical evidence shows that, after controlling for the opportunity cost of money, the quantitative impact of total government debt-to-GDP ratio is still significant and negative, which is consistent with the theoretical predictions of the CAPM with utility surplus of holding convenience assets. The relative availability measure, the ratio of total government liability to all sector total liability, separates the liquidity premium from the safety premium and has a negative impact too. Both theoretical and empirical results suggest that the substitutability between government debt and private safe assets dictates the quantitative impact of the government debt supply.