The impacts of interest rates on banks’ loan portfolio risk-taking
Introduction
The 2008 global financial crisis has shown that controlling excessive banks’ risk-taking is a fundamental factor for the stability of the financial system and the proper functioning of the real economy. In this context, an important issue is to investigate how the definition of interest rates affects the behavior of the banks (Beyer, 2017, Ireland, 2005).
Besides clarifying some of the causes of bank risk-taking, understanding the impacts of interest rates on the composition of banks’ portfolios can help us understand how these effects spill over into the real sector. There are various channels through which interest rates and monetary policy can affect financial stability and the banks’ incentive to take risks (Claessens and Valencia, 2013). Through the risk-taking channel, low interest rates can increase incentives for banks to expand their balance sheets, to over-leverage, and to reduce lending standards by not screening borrowers. This may result in an excessive expansion of credit, amplifying boom-bust cycles, especially when the interest rate is held too low for too long (Adrian, Shin, 2010, Borio, Zhu, 2012). Low interest rates can lead to rises in asset prices and loose collateral constraints through the asset prices channel (Bernanke, Gertler, 1989, Bernanke, Gertler, 1995).
High interest rates may also affect financial stability negatively, once it can hurt the capacity of borrowers to pay their loans, increasing default rates (Allen, Gale, 2000, Illing, 2007). It can induce a shortage of loans that make it unfeasible to finance more ambitious projects, generally associated with high expected returns. Monetary policy can also affect capital inflows, especially for small open economies, which drive credit growth and may contribute to an excessive increase in leverage (Georgiadis and Zhu, 2021).
Our study evaluates how interest rates affect banks’ risk-taking based on their lending profile to real sector firms. We investigate the two different regimes of interest rates, namely low interest rate regime and high interest rate regime. Banks can lend to two types of companies: low-risk and high-risk. We use the fraction between the volume of loans lent to the riskier firms and the total of loans as the banks’ measure of risk. While our main result is that the low interest rate regime produces a higher amount of riskier loans, by varying the model’s parameters, we also find a rich set of banks’ behaviors response to different profit opportunities and risk profiles.
We build our results using an Agent-Based Model (ABM) (Anderson, Anderson, Pines, 1988, Arthur, Durlauf, Lane, 1997, Blume, Durlauf, 2005, Bonabeau, 2002, Chakraborti, Toke, Patriarca, Abergel, 2011, Dawid, Fagiolo, 2008, Farmer, Foley, 2009, Gallegati, Kirman, 2012, Gatti, Fagiolo, Gallegati, Richiardi, Russo, 2018, Holland, Miller, 1991, Steinbacher, Raddant, Karimi, Camacho Cuena, Alfarano, Iori, Lux, 2021, Tesfatsion, 2006). We consider five types of agents: banks, depositors, the Central Bank, firms, and the clearinghouse. Banks are bounded-rational agents with adaptive strategies. In different setups, depositors are either noisy agents or bounded-rational agents that withdraw their deposits when they have concerns over their banks’ solvency. The other players’ behaviors are used as references to understand how these main agents respond strategically to different incentives and situations. More precisely, we bring in new features to the model introduced by Barroso et al. (2016). Its main methodological contribution is to discuss bank regulatory policies in a more flexible environment than considered in Diamond et al. (1983), Gale and Douglas (1998) and Gale and Douglas (2000).1 Our extension grants banks the possibility to lend to firms with different risk profiles. As in Barroso et al. (2016), our model assumes a repeated game framework in which banks choose their liquidity and risk-taking levels. In this scenario, the greater the strategy’s payoff, the more likely it will be adopted again in the future. Unlike the traditional ABMs, where the choice of agents follows if-else structures, an important aspect of our model is that our learning scheme is based on Experience-weighted Attraction Learning (Camerer et al., 1999). In this configuration, banks can adapt their behavior.
We use two different model setups for the simulations. Firstly, we set banks as the only intelligent agents. Secondly, the banks and depositors are both intelligent agents. We run the simulations considering the following (four) scenarios: developed and emerging countries under low and high interest rate regimes. The parameters are chosen based on real-world values obtained from the World Bank, the International Monetary Fund, and the Federal Reserve Economic Data.
We can summarize our findings in four main results that converge with the literature claims about interest rates and monetary policy impacts on banks’ behavior. First, when interest rates are low, there is an increase in real sector loans, particularly for riskier clients. This evidence is in line with the risk-taking channel, which states that low interest rates can increase incentives for banks to take risks and may result in an excessive expansion of credit (Adrian, Shin, 2010, Agur, Demertzis, Maria, 2012, Borio, Zhu, 2012, Chen, Wu, Jeon, Wang, Rui, 2017).
Second, we show that the interbank market plays a fundamental role in banks’ liquidity management. Higher interbank rates also affect banks lending decisions in the interbank market when they face an excess of liquidity. As typically observed under high interest rate regimes, higher interbank rates increase lending in the interbank market (Lucchetta, 2007). In this context, our paper adds to the evidence of several works that have been studying the interbank market, other networks of banks, systemic risk, and, in particular, how a network of banks responds to shocks using ABMs (Gurgone, Iori, Jafarey, 2018, Iori, Jafarey, Padilla, 2006, Ladley, 2013, Lenzu, Tedeschi, 2012, Lux, 2016, Nier, Yang, Yorulmazer, Alentorn, 2007, Poledna, Thurner, 2016).
Third, we find that banks avoid borrowing resources from the Central Bank. This result is related to the findings of Armantier et al. (2015), which explain that it happens not only because of the punitive interest rate but also due to a discount window stigma, since depositors and other banks may perceive the borrower institutions as being in a weakened financial condition.
Fourth, when interest rates are high, banks increase capital level buffers and the Capital Adequacy Ratio (CAR), maintaining capital buffers on top of minimal capital requirements to enhance banks resilience against shocks. This result is in line with the evidence that monetary policy affects capital regulation through the risk-taking channel since banks react to monetary policy by changing the volume of provisions (de Moraes et al., 2016).
Our findings offer new insights regarding the relationship between interest rates and bank risk-taking, opening an avenue to investigate the banks’ learning process dynamics. In particular, we find that banks learn to manage their balance sheets in different interest rate regimes to deal with noisy and earlier depositor withdraws and risky firms defaults. Furthermore, this learning process is essential to avoid banks insolvencies and contagions.
Our paper also connects to interesting ABM approaches to address relevant questions related to the interaction among banks and the real side of the economy. We emphasize the resilience of financial networks (Battiston et al., 2012a), the role of the interbank market (Gabbi, Iori, Jafarey, Porter, 2015, Ladley, 2013, Popoyan, Napoletano, Roventini, 2020), the impact of macro-prudential regulations (Popoyan et al., 2017), and the resolution of failing institutions (Klimek et al., 2015). Also, we share many framework settings with several of these mentioned works - especially with Gabbi et al. (2015) - such as the banks’ balance sheets explicit definitions, the roles of the central bank, the interbank market, and depositors. Despite these similarities, it is worth mentioning that our paper aims at answering different questions, and our banks’ choices follow distinct mechanisms regarding those considered by Gabbi et al. (2015).
Besides this introduction, we organize the manuscript as follows: Section 2 presents a literature review about the contributions of ABMs in banking in order to make clearer how our paper connects to this literature. Section 3 details all the features of our model and we divide this section into three subsections: Section 3.1 describes the behavior of the agents that populate our model; Section 3.2 reports how the agents learn their strategies; Section 3.3 shows how we design the timeline of our model. Section 4 shows our main results. We divide this section into two subsections. Section 4.1 presents the findings for the setup where banks are the only intelligent agents (that is, they choose their actions strategically to maximize their payoff). Section 4.2 introduces the results for the case in which banks and depositors are intelligent agents. Finally, Section 5 brings the main conclusions of our paper. There is also a supplemental material associated with this paper that presents additional simulations using a different set of parameters than those used here.
Section snippets
Literature review about ABMs in banking
In the ABM approach, we build a model of the system using a bottom-up approach: (1) We define the agents and the collection of simple rules they have to follow. (2) We define the interaction between the agents, which is constrained by the institutional arrangements; (3) We explore the emergent behavior of the system (Anderson, Anderson, Pines, 1988, Arthur, Durlauf, Lane, 1997, Blume, Durlauf, 2005, Bonabeau, 2002, Chakraborti, Toke, Patriarca, Abergel, 2011, Dawid, Fagiolo, 2008, Farmer,
The model
Our model is a sequential non-collaborative game of imperfect and incomplete information. Players do not know either each others’ moves or payoffs. They only have information about themselves and, thus, cannot cooperate. As we present in Section 3.1, five types of agents populate our stochastic game model: banks, depositors, the Central Bank, firms, and clearinghouse. Both banks and depositors (in some of the setups) have bounded-rationality and strategically play the game according to a scheme
Results
In this section, we present and discuss the results of our paper. We use two different model setups for the simulations. While in the first banks are the only intelligent agents, in the second, banks and depositors are both intelligent agents. For each model setup, we consider two different scenarios, one where we use parameters based on developed countries and one where the parameters are based in emerging economies. We assume that Central Banks have the potential to affect the interbank
Conclusions
This study has investigated the relationship between interest rates and bank risk-taking using an agent-based model (ABM) approach. The model counts with different-sized banks and uses the volume of banks’ loans to risky real sector clients as a proxy for risk. To carry on such analysis, we introduce firms with different risk types in the agent-based model of Barroso et al. (2016). We consider two different setups. In one setup, banks are the only intelligent agents. In the other setup, the
Acknowledgement
LFSA, RAE and DOC thank to cnpq for financial support.
References (175)
- et al.
Financial intermediaries and monetary economics
Handbook of Monetary Economics
(2010) - et al.
Bank networks: contagion, systemic risk and prudential policy
J. Econ. Behav. Organ.
(2017) - et al.
Multiplex interbank networks and systemic importance: an application to european data
J. Financ. Stab.
(2018) - et al.
Rollover risk, network structure and systemic financial crises
J. Econ. Dyn. Control
(2012) - et al.
Discount window stigma during the 2007–2008 financial crisis
J. Financ. Econ.
(2015) - et al.
Banks, market organization, and macroeconomic performance: an agent-based computational analysis
J. Econ. Behav. Organ.
(2017) - et al.
Disentangling bipartite and core-periphery structure in financial networks
Chaos Solitons Fractals
(2016) - et al.
Calibrating limits for large interbank exposures from a system-wide perspective
J. Financ. Stab.
(2016) - et al.
Liaisons dangereuses: increasing connectivity, risk sharing, and systemic risk
J. Econ. Dyn. Control
(2012) - et al.
Complex networks: structure and dynamics
Phys. Rep.
(2006)
Capital regulation, risk-taking and monetary policy: a missing link in the transmission mechanism?
J. Financ. Stab.
The role of banks in the brazilian interbank market: does bank type matter?
Physica A
The use and effectiveness of macroprudential policies: new evidence
J. Financ. Stab.
Shedding light on el farol
Physica A
Emergence of cooperation and organization in an evolutionary game
Physica A
Monetary policy and bank risk-taking: evidence from emerging economies
Emerg. Mark. Rev.
Evolving traders and the business school with genetic programming: a new architecture of the agent-based artificial stock market
J. Econ. Dyn. Control
On the emergent properties of artificial stock markets: the efficient market hypothesis and the rational expectations hypothesis
J. Econ. Behav. Organ.
Interbank tiering and money center banks
J. Financ. Intermed.
Interbank tiering and money center banks
J. Financ. Intermed.
Agent-based models for economic policy design: introduction to the special issue
J. Econ. Behav. Organ.
Bank runs, deposit insurance and liquidity
J. Polit. Economy
Systemic risk in financial systems
Manag. Sci.
Network analysis of the e-MID overnight money market: the informational value of different aggregation levels for intrinsic dynamic processes
Comput. Manag. Sci.
Macroprudential policy – a literature review
J. Econ. Surv.
Banking panics and business cycles
Oxf. Econ. Pap.
Tipping points in macroeconomic agent-based models
J. Econ. Dyn. Control
Where do interorganizational networks come from?
Am. J. Sociol.
Excessive Bank Risk Taking and Monetary Policy
ECB Working Paper 1457
Statistical mechanics of complex networks
Rev. Mod. Phys.
Combining monetary policy and prudential regulation: an agent-based modeling approach
J. Econ. Interact. Coord.
A noise trader model as a generator of apparent financial power laws and long memory
Macroecon. Dyn.
Bubbles and crises
Econ. J.
Understanding Financial Crises
Does Monetary Policy Affect Bank Risk?
ECB Working Paper 1166
The Economy as an Evolving Complex System
Size and complexity in model financial systems
Proc. Natl. Acad. Sci.
Inductive reasoning and bounded rationality
Am. Econ. Rev.
The Economy as an Evolving Complex System II
Policy experiments in an agent-based model with credit networks
Economics
Understanding the consequences of diversification on financial stability
J. Econ. Interact. Coord.
The physics of financial networks
Nat. Rev. Phys.
The multiplex structure of interbank networks
Quant. Finance
Interbank network and regulation policies: an analysis through agent-based simulations with adaptive learning
J. Netw. Theory Finance
The structure of financial networks
Network Science
Debtrank: too central to fail? Financial networks, the fed and systemic risk
Sci. Rep.
Decision spread in the corporate board network
Adv. Complex Syst.
Agency costs, net worth, and business fluctuations
Am. Econ. Rev.
Inside the black box: the credit channel of monetary policy transmission
J. Econ. Perspect.
The Transmission Channels of Monetary, Macro- and Microprudential Policies and Their Interrelations. Occasional paper series/European Central Bank
Cited by (4)
Impact of higher federal funds rates on bank risk during higher inflation in the U.S.
2024, Finance Research LettersHousehold loan portfolios and financial characteristics of Korean banks
2023, Asian-Pacific Economic LiteratureInflation Targeting Regimes in Emerging Market Economies: To Invest or Not to Invest?
2023, Computational Economics