The popularity of macroprudential policies (MaPs) has greatly increased in recent years, especially after the Global Crisis. Yet our understanding of these measures and their effects is still open to debate. Macroprudential interventions should be judged primarily by whether they reduce the occurrence and magnitude of financial crises – their ultimate goals. However, measuring this remains an elusive task, given the relative infrequency of crises, and the difficulty of precisely attributing them to fundamental factors. One potential way of assessing the effectiveness of macroprudential policies is to evaluate their impact on some intermediate targets, such as credit growth and the evolution of bank risk.
So far, the literature has focused on the effects of MaPs on credit growth, mainly by using country-wide data or aggregating them at the bank level (Cerutti et al. 2017). In two recent studies, we complement the existing evidence by:
- Analysing the effects of MaPs on lending growth using very granular credit registry data (Gambacorta and Murcia 2017);
- Evaluating the impact of MaPs directly on measures of bank risk (Altunbas et al. 2017).
Macroprudential policies have been used in different ways
In order to correctly evaluate MaPs, we need to first classify them according to the goals for which they might best be suited. Following Claessens et al. (2013), we can split MaPs into instruments whose main objective can be regarded to be to enhance the resilience of the financial sector (capital and liquidity based requirements) or to mitigate a credit boom or credit crunch (LTVs, currency-based instruments, reserve requirements).
Interestingly, only one quarter of the measures used appear to primarily target the resilience of the financial sector (Figure 1, left-hand panel). By contrast, the vast majority appear to have smoothing the cycle as their primary goal. About half have involved changes in reserve requirements. Overall, 60% of the interventions were meant to tighten financial conditions (Figure 1, centre panel). Of all the macroprudential measures adopted, 80% were taken by emerging market economies (Figure 1, right-hand panel).
Figure 1 Use of macroprudential instruments (in percentages)1
Notes: 1 The sample covers 64 countries over the period 1990–2014. Macroprudential tools for resilience include (a) capital-based instruments (countercyclical capital requirements, leverage restrictions, general or dynamic provisioning) and (b) liquidity requirements. Cyclical macroprudential tools include: (c) asset-side instruments (credit growth limits, maximum debt service-to-income ratio, limits to banks’ exposures to the housing sector as a maximum loan-to-value ratio); (d) changes in reserve requirements; and (e) currency instruments (variations in limits on foreign currency exchange mismatches and net open positions).
Source: Authors’ calculations.
MaPs help stabilise credit cycles: Evidence from credit registry data
Among emerging market economies, Latin America countries are a particularly good laboratory for the evaluation of the effectiveness of macroprudential tools. This is for two reasons:
- The use of MaPs has a relatively long history there (Tovar et al. 2012);
- Many countries have developed granular credit registry data that help to disentangle loan demand from loan supply effects without making strong assumptions.
However, the confidentiality of credit registry data does not allow pooling them into a unique dataset. This is why studies of the impact of MaPs using credit registry data have been scarce so far. It also explains why they have focused on country-level experiences (exceptions are Jimenez et al. 2016 for Spain, and Camors et al. 2016 for Uruguay).
To study the impact of macroprudential measures and their interaction with monetary policy, the Bank for International Settlements initiated (under the auspices of the Consultative Council for the Americas) a joint project covering eight of their shareholding central banks. Five of these – Argentina, Brazil, Colombia, Mexico, and Peru – made use of credit registry data and ran regressions based on a common protocol (i.e., with similar modelling strategies and data definitions). Gambacorta and Murcia (2017) summarised the results for a total of 15 episodes of macroprudential interventions (six primarily aimed at enhancing resilience, and nine at dampening the cycle, according to our classification).
Using meta-analysis techniques, we find that macroprudential tools help stabilise credit cycles (see Table 1). In particular, a tightening of macroprudential measures is associated with an average reduction of annual credit growth of 4.6% after three months, and 7.2% after one year. However, the propagation of the effects is quite heterogeneous across macroprudential tools – it is more rapid for prudential measures that aim at dampening the cycle (the effects are significant after one quarter) than for those aimed at fostering resilience (the effects materialise within a year). Finally, the effectiveness of macroprudential measures on credit growth is affected by monetary policy conditions. In particular, macroprudential tools that were adjusted to reinforce monetary policy (i.e. pushed in the same direction, when both were tightened or eased) were relatively more effective.
Table 1 Effects of macroprudential policies on credit growth
Meta-analysis of estimated coefficient of MaPs on credit growth
Notes: (1) The Q Measure evaluates the level of homogeneity/heterogeneity among studies. It is calculated as the weighted squared difference of the estimated effects with respect to the mean. The statistical distribution of this measure follows a χ2 distribution. The null hypothesis of the test assumes homogeneity in the effect sizes. (2) It corresponds to the weighted average of coefficients reported in different estimations. ***,** and * denote significance at the 1%,5% and 10%, respectively.
Source: Gambacorta and Murcia (2017).
MaPs help contain bank risk: Evidence from bank-level data
The evidence presented so far is based on the impact of MaPs on credit growth. However, as the ultimate goal of MaPs is to reduce the probability of a financial crisis, we also analysed their impact on bank risk. In particular, we performed the test on a panel of more than 3,000 banks operating in 61 advanced and emerging market economies and, as measures of risk, we used the expected default frequency and the Z-score. The calculation of the expected default frequency indicator requires bank issuance of equity on the stock market while the Z-score relies on balance sheet variables to determine the probability of default.
There are three main results, as presented in Altunbas et al. (2017). First, the evidence suggests that macroprudential tools have a significant impact on bank risk. And this applies to both of the types discussed so far (see Figure 2, left-hand panel). Second, MaPs are more effective in tightening than in easing episodes, which is in line with Cerutti et al. (2017) who analysed the asymmetric impact of MaPs on bank credit. Third, the responses to changes in macroprudential tools depends on bank-specific characteristics. In particular, banks that are small, weakly capitalised, and with a higher share of wholesale funding react more strongly to changes in macroprudential tools.
To get a sense of the quantitative effects of such a heterogeneity, the right-hand panel of Figure 2 summarises the effects of macroprudential tools for banks with different levels of capital. The estimates roughly imply that a tightening of macroprudential measures leads to a decline in the expected default probability of around 0.7% for the average bank. The effect is higher for weakly capitalised banks (−0.9%) than for strongly capitalised ones (−0.4%), probably because they have better access to markets. Similar results are detected when taking the Z-score as an indicator of bank risk. These results are in line with Gambacorta and Shin (2016) – well-capitalised banks are considered to be less risky by the market and pay less on their debt funding.
Figure 2 Effects of macroprudential policies on bank risk
Notes: The figures reports the effect on bank risk of a tightening in macroprudential tool. The left part of each graph indicates the effects on banks’ expected default frequency (left-hand axis), the right part the effects on the Z-score (right-hand axis). A higher Z-score corresponds to a lower upper bound of insolvency risk, therefore it implies a lower probability of insolvency risk. To compare the signs of the coefficients in the regressions, we have therefore multiplied the Z-score by -1.
Source: Altunbas et al. (2017).
Macroprudential measures are designed to make financial crises less likely and severe. At the same time, most of the existing literature studies their impact on credit growth, focusing on country-wide data or bank-level information. We present new evidence on the effectiveness of macroprudential measures using credit registry data at the bank-firm level and analysing also their link with bank risk measures. An open issue remains the effects of MaPs on long-run economic performance (such as output growth and its volatility; see Boar et al. 2017 for some preliminary results). Moreover, structural features, such as a country’s financial development (i.e. the relative importance of the shadow banking system) and its degree of openness, could alter the effects of prudential measures through possible leakage effects. Further research is needed to shed light on these aspects.
Authors’ disclaimer: The views expressed here are those of the authors and should not be attributed to the Bank for International Settlements, Banco de la República or the IMF, their Executive Boards, or management.
Altunbas, Y, M Binici and L Gambacorta (2017), “Macroprudential policies and bank risk”, Journal of International Money and Finance, forthcoming (also published as CEPR Discussion Paper No. 12138).
Boar, C, L Gambacorta, G Lombardo and L Pereira da Silva (2017), “What are the effects of macroprudential policies on macroeconomic performance?”, BIS Quarterly Review, September.
Camors C, Dassatti, J Peydró and F.R Tous (2016), “Macroprudential and monetary policy: loan level evidence from reserve requirements”, mimeo, Universitat Pompeu Fabra, Spain.
Cerutti, E, S Claessens and L Laeven (2017), “The use and effectiveness of macroprudential policies: new evidence”, Journal of Financial Stability, 28: 203–224.
Claessens, S, S R Ghosh, and R Mihet (2013), “Macro-prudential policies to mitigate financial system vulnerabilities”, Journal of International Money and Finance 39: 153-185.
Gambacorta, L and A Murcia (2017), “The impact of macroprudential policies and their interaction with monetary policy: an empirical analysis using credit registry data”, BIS Working Papers, 636 (also published as CEPR Discussion Paper No. 12027).
Gambacorta, L and H S Shin (2016), “Why bank capital matters for monetary policy”, Journal of Financial Intermediation, forthcoming.
Jimenez, G, S Ongena, J-L Peydro and J Saurina (2016), “Macroprudential policy, countercyclical bank capital buffers and credit supply: Evidence from the Spanish dynamic provisioning experiments”, Journal of Political Economy, forthcoming.
Tovar, C, M García-Escribano and M Vera Martin (2012), “Credit growth and the effectiveness of reserve requirements and other macroprudential instruments in Latin America”, IMF Working Papers, no 142.
 Three other central banks – Canada, Chile, and the US – complemented the analysis by studying specific prudential measures using information on credit origination and borrower characteristics. The country studies and the summary paper can be found at: https://www.bis.org/am_office/wgfinstab/teompatiwmp.htm.
 The expected default frequency used in the study estimates the probability that a bank will default within one year. Its value, expressed as a percentage, is calculated by combining banks’ financial statements with stock market information and Moody’s proprietary default database. The Z-score is an alternative measure for risk that measures the number of standard deviations that a returns realisation has to fall in order to deplete equity, under the assumption of normality of bank returns.