Because the economic crisis, stress tests have grown to be an essential supervisory and financial stability tool. From this background, an issue is whether or not stress tests lead to financial stability your clients’ needs risk decrease in the banking sector as recent evidence suggests. Stress tests offer deep insights into banks’ vulnerabilities to supervisors and also the public with an intense supervisory process. Inside a recent paper, we reveal that greater supervisory scrutiny brought to some disciplining effect for banks which were area of the 2016 EU-wide stress test, coordinated through the European Banking Authority (EBA) and conducted through the European Central Bank (ECB).
How supervisory scrutiny is exerted in stress tests
In Europe, stress tests involve interactions between banks and supervisors on banks’ risk management practices in addition to private communications about best stress-testing practices and methods. We use data on these private interactions to approximate just how much scrutiny was exerted on every bank underneath the direct supervision of Single Supervisory Mechanism (SSM) throughout the 2016 EU-wide stress test. These interactions arise included in the restricted bottom-up approach went after within the EBA-coordinated exercises (see figure 1). Within this context, banks use their very own internal models to create projections, for instance, for credit losses. Meanwhile, banks’ projections are challenged through the competent supervisory government bodies typically by making use of top-lower models along with other challenger tools. In the existence of material deviations between both of these teams of projections, “flags” are triggered and then discussed between supervisors and banks. Banks have to adhere to or explain the problems elevated within the interactions using the ECB. We create a scrutiny measure by counting the flags associated with credit risk projections. Without effort, banks that received more flags needed to continue to work harder on their own resubmissions coupled with lengthier and most likely more serious interactions with supervisors, while banks that received no flags in principle didn’t have further interactions with supervisors.
Figure 1: Simplified instance of one quality assurance cycle underneath the restricted bottom-up approach.
A diagram (text boxes) showing sign of 1 quality assurance cycle underneath the restricted bottom-up approach.
Source: Own illustration according to Mirza, H. and Zochowski, D. (Macroprudential Bulletin Issue 3, Ch.2, 2017).
Scrutiny measures the concentration of stress tests
We use a variations-in-variations approach where we make use of the stress test like a treatment and also the involved scrutiny like a way of measuring the concentration of the therapy. In the initial step, we compare the loan perils of banks which were area of the stress make sure banks which were not area of the stress test four quarters before and 4 quarters following the 2016 stress test. Within the next step, we compare the loan chance of banks which were subject to some more intense supervisory scrutiny and banks that received less or none.1 The 2016 EU-wide stress test was performed on significant institutions (SIs). Decreased institutions (LSIs) weren’t tested so we therefore rely on them because the control group.2
Aftereffect of supervisory scrutiny on credit risk
We focus our analysis on credit risk that makes up about most from the stress testing projections and typically for 86 percent of risk exposure amounts in bank balance sheets. To determine credit risk in the bank level, we make use of the risk-weight density (RWD), that’s, the mixture risk weight allotted to total credit risk exposures based on regulatory standards.
Figure 2: Estimate and 90 % confidence interval from the differential impact on RWD between tested and non-tested banks pre and post the reported quarter.
A regular chart showing Figure 2: Estimate and 90 % confidence interval from the differential impact on RWD between tested and non-tested banks pre and post the reported quarter.
First, we discover no factor in RWD between your treatment and control groups prior to the stress test (2015q1 and 2015q4, see figure 2) but significant negative variations for that period following the test (2017q1 to 2017q4). The decrease in RWD of tested banks following the stress test was typically 4.2 percentage points less than the decrease in not-tested banks. This effect is economically material because it comes down to a big change of approximately 20 % from the standard deviation of RWD.3 These results read the findings according to US data that “treating” banks with stress tests can impact their risk.
Figure 3: Marginal effect and average effect estimates with 90 % confidence times from the differential aftereffect of scrutiny intensity on RWD.
A line (stock) chart showing Figure 3: Marginal effect and average effect estimates with 90 % confidence times from the differential aftereffect of scrutiny intensity on RWD.
Second, we reveal that the greater interactions banks had with supervisors, the greater was their decrease in RWD following the stress test exercise (see figure 3). We discover that individuals banks that received more scrutiny (the half with scrutiny intensity over the median) exhibit 5.6 percentage points greater reduction in credit risk compared to half that received less scrutiny.4 Overall, these bits of information provide novel evidence the tighter and much more intrusive supervisory scrutiny connected using the EU-wide stress-test can enhance banks’ risk management practices and induce lower bank risk.
Policy implications
We lead towards the emerging evidence on the potency of supervisory scrutiny. Our results claim that stress tests which are conducted by making use of a strong quality assurance of banks’ projections and designs include disciplining effects on stress tested banks’ risk. However, among the stress tests’ primary objectives would be to assess banks’ risk profiles properly. Our findings don’t showcase how good this objective is met. The potential proper underreporting of banks’ vulnerabilities within bottom-up approach could undermine the longevity of the strain test outcomes out of this perspective. Going after a far more impartial top-lower approach while retaining supervisory interactions with banks after and during the strain test is much more appropriate to do this goal. Therefore, with this analysis, we simply deliver one insight among many who could serve the insurance policy discussion on the style of future stress tests in Europe.