This code simulates the propagation of negatives shocks through a supply chain network and a bank-firm credit network. The modelling of negatives shocks spread across the supply chain is based on the simulator of Inoue and Todo (2018) available on Github.
Moreover, the model simulates the effects of initial exoegouns negatives shocks on the supply chain, on the bank-firm credit network. The model simulates the non-performing loans and the liquidity crisis of banks following negatives shocks on the supply chain.
Based on real Japanese data, two negatives shocks were simulated: the 2008 Lehman brothers bankruptcy and the 2011 Great earthquake and tsunami. The model reproduces the real 1-year dynamics of the Index of Industrial Production (IIP) of the Japanese economy as shown in these figures from Krichene et al. (2019).
The SNSE simulator is a ready to use code. However, the example given in this Github repo is for windows 64bit users.
- Download the release of this link.
- Uncompress the file .rar
- Open windows command line and execute ABM_Disasters.exe.
The user will obtain the same results shown in "Results" directory of this repo. The simulation is based on toy data. You can use the model on any data; replace the toy data by your real data.
The parameters should reflect the properties of your economy: How firms and banks behave? In our work in Krichene et al. (2019) we used a Latin hypercube sampling to calibrate our parameters and reproduced the IIP dynamics shown in Figures 1&2. The user may define any different approach. In this section we explain the meaning and roles of each parameter in the SNSE.
Two types of parameters are in the SNSE: behavioral parameters and simulation parameters. The first are used to define the strategies of agents. The latters are used to define different scenarios of the simulation.
- n: the number of days of the inventory.
- GammaMin, GammaMax: the recovery speed of damaged firms.
- LimitSolvencyRatio: the acceptable solvency by banks to supply loans.
- LoanMaturity, LTLoanMaturity: maturity of short and long term loans.
- DamageMagnitude: the magnitude of the initial negative shock to the supply chain.
- NumberDamagedFirms: the number of initial damaged firms.
- t: simulation time, assumed to be one day: daily simulation.
- disaster: if 0 no negative shock; if 1 simulation of negative shock at t = 1.
- WithPayment: if 0, the SNSE simulates Inoue and Todo (2018); if 1, the SNSE considers the bank-firm network.
- HelpFirms: if 0, firms can get funding only as loans from banks; if 1, firms may have an exogenous funding.
- BankRiskManager: if 0, banks supply loans to all received demand; if 1, banks monitor their risk based on the LimitSolvencyRatio parameter.
Example of outputs are given in the Results directory of this repo.
- DamagedFirms.txt: list of initially damaged firms.
- FinalGvtSupport.txt: the amount of exogenous funding.
- FinalLiquidity.txt: the liquidity ratio of banks.
- FinalNPL.txt: the generated non-performing loans.
- GDP.txt: the dynamics of the simulated VA, used as a proxy of IIP.
Krichene, H., Inoue, H., Isogai, T., Chakraborty, A. A model for indirect losses of negatives shocks in production and finance, (2019). SSRN.