Data of the Credit Market of Low-Income Households in a Semi-Peripheral Country

The data is the results of the network based ABM model of the financial system of low income families. The complex financial system of low-income families manage the gaps between the families expenditure and incomes, financial liquidity shocks and permanent financial crises. The code for the multi-agent model and simulation is available on CoMSES Network-Computational Model Library as: Credit and debt market of low-income families (version 1.0.0),”https://www.comses.net/codebase-release/74832082-f455-4d58-88f3-\ \7efeb29b1966” The economic actions of the actors within the system are embedded in a dynamic network of formal and informal credit institutions and connections. The total number of households in the database is 159, that is the village’s entire population. The monthly income of each household is calculated using evidence-based data and the possible connections of different credit institutions or interpersonal credit connections are also based on measured network data. The entire measured credit network of the households in a year it includes. The households appear as decision-making agents in the model, while the seven formal ind informal credit institutions appear as quasi-agents. The model assumes that the agents maintain credit partners in the market to balance the differences between their incomes and expenditures. If the budget of the household becomes negative, they can apply for loans or credits provided by their network connections. The agents may have several credits simultaneously from different institutions to remedy their liquidity shocks but can have only one pending credit relation from one particular credit source. Further, they are free to exit their credit connections to some extent, but the existence of a network connection is defined by an agent’s previous transactions. An interval indicates one day in the dataset. The total run of the model was 365 days, starting from the hypothetical Spring month. Each day, our model calculates the local socioeconomic strata and many other model indicators. The local socioeconomic strata are calculated based on a complex set of variables that take into account the incomes, consumption structure, size of the family and the basis of the available liquid capital. The model classifies the households into three local social segments: Poor households (PoorHH or l_ ) that are below the poverty line, middle households characterised by medium incomes at the local level (MiddleHH or m_ ) and finally high-income households (HighHH or h_ ). A thousand simulations were made on each parameterised model, thus, the analysis is based on a total of 54.000 simulated models, including 40.000 models used for validation and 14.000 for analysis.