EU’s deep renovation rates have not been sufficiently fast enough to significantly reduce building stocks’ emissions. As a reaction against these facts, the newly published policy packages “Fit to 55” aim to boost renovation activities and to decarbonise the building stock until 2050. Nevertheless, building owners’ affordability to pay for renovation has been recognised as a significant barrier. Therefore, a high amount of investments and adequate financing schemes are important instruments to enable the achievement of these goals. The present paper contributes to this context by classifying households natural gas expenditure and their budget restrictions, relevant information to develop more user-targeted financing instruments. Budget restrictions are expressed through the household’s savings (income minus expenditures). In this paper, the authors develop an approach to statistically match and test it using HBS and SILC data for 2015 Spain. Following research questions will be answered: what is the replicability of a method to merge HBS and SILC datasets? What can we learn about household annual natural gas expenditures, savings and incomes of four different household types? To carry out this analysis, mainly two databases of EU-households were used: EU-SILC (European Union Statistics on Income and Living Conditions) and HBS (Household Budget Surveys). The method consists of two steps. First, the application of a logistic regression model to perform a statistical match of both datasets. Then, statistical describing the income, savings and natural gas expenditure for four household types: single-family house owner- occupied, single-family house rented, multi-family house rented and multi-family house owner- occupied. The whole approach was carried out, tested and validated for data from Spain. 16% of the total households spend annually more than 600 euro for natural gas. Rented single family houses were identified as the most vulnerable household type, due to their low income and saving. Next steps are replicating the workflow to other countries. And, using the estimated budget restrictions as input data in building stock models.