This project aims to study, from an empirical perspective, the time-evolving dynamics of credit market fragilities. The investigation will focus on private borrowing dynamics over the cycle, disentangling between secured and unsecured credit contracts. The empirical analysis will possibly be carried out employing estimated Bayesian time-varying parameters structural vector autoregression models. These models will be structuralized with mixed point/set restrictions, able to separately identify demand, supply, monetary and financial shocks. On this basis, the possibly nonlinear effects of the different sources of variability will highlight both the emerging smooth nonlinearities characterizing slow-building credit cycles and episodes of major structural changes observed in financial time series.
The first aim of this project is to create reliable and comprehensive measures of US secured and unsecured credit. Up to the present, published papers mainly used annual firms¿ budget data from Compustat and Capital IQ databases. I argue that these measures could be improved in two directions: on the one hand, searching for data with higher frequency allows for a better performance of the model, while on the other hand using the flow of funds FRED data allows the consideration of the households' sector, reaching a more comprehensive picture. This requires collecting data and to reorganise them following an original, collateral-based approach. I began the task of gathering data for an ongoing research paper.
Therefore, I will use quarterly US data, spanning the 1975-present period, and covering credit accounts to both the private non-financial business sector, namely corporate and non-corporate businesses, and the households and nonprofit organizations sector. I will then estimate a structural Bayesian Time-varying VAR model, identified with mixed exclusion and sign restrictions, consistently with the recent literature (Furlanetto et al. [2017], Peersman and Wagner [2014]). This has three main purposes: i) uncover the GDP responses to secured and unsecured credit; ii) simulate a monetary policy tightening; iii) quantify the importance of both credit supply and credit demand shocks. The estimation of a time-varying coefficient model allows me to explore to what degree the model dynamics is affected by relevant nonlinearities characterizing slow-building credit dynamics. Modeling data by allowing for time-varying coefficients to the contemporaneous and dynamic relations of the endogenous variable allows me to pinpoint some episodes of structural change.
More specifically, I will assess whether the explanatory power of unsecured and secured credit shocks changed over time and, if so, whether its time variation hinges on either a change in the size of the shocks or a change in the transmission mechanism of credit shocks to GDP. As for the former, following Prieto et al. [2016], I will look at the time-varying standard deviations of the orthogonalized shocks to have a picture of their changing magnitude over time. As for the latter, I will assess the possible changing dynamics of the shocks¿ transmission mechanism by computing the median Impulse Response Functions of the GDP to unit credit shocks at all points in time.
To the best of my knowledge, the literature contains no attempts to uncover the cyclical stance of secured and unsecured credit contracts within a conditional analysis based on a structural Bayesian VAR model. Hence, the qualifying contribution of this proposal is twofold. On the one hand, I will identify the cyclical behaviour of these two credit contracts (including also the households' sector) and analyse their relative major contribution in explaining GDP fluctuations over time. On the other hand, I will search for relevant nonlinearities of their cyclical pattern over time.
References:
Furlanetto, F., Ravazzolo, F., & Sarferaz, S. (2017). Identification of financial factors in economic fluctuations. The Economic Journal, 129(617):311¿337.
Peersman, G. & Wagner, W. (2014). Shocks to bank lending, risk-taking, securitization, and their role for us business cycle fluctuations.
Prieto, E., Eickmeier, S., & Marcellino, M. (2016). Time variation in macro-financial linkages. Journal of Applied Econometrics, 31(7):1215¿1233.