As the final exercise, the optimal level of loan supply when holding different bundles of government bond portfolios is examined. Loan weights take the val-ues of 50% to 100% and government bond bundles weights vary from 0% to 50%.
Banks hold either their normal government bond compositions, the fully diver-sified government bond portfolio or SBBS assets. The results for the five countries that are being analyzed are reported in figure 20 for Greek banks, figure 21 for Italian banks, figure 22 for Spanish banks, figure 23 for Portuguese banks and figure 24 for German banks. For 100% loans-to-total-assets a bank is indifferent to which composition to choose. However, as the weight of loans decreases, and the weight of government bonds increases, the gap between the three lines in-creases. For all the countries the most efficient government bond composition is the bundle that contains SBBS assets, because it returns higher supply of loans for every given level of C-VaR. The second most efficient composition of govern-ment bonds is the diversified portfolio for Greek, Italian, Spanish and Portuguese banks. For German banks the diversified portfolio is the second most efficient up to a certain point. The diversified bundle is the second most efficient when expo-sure to government bonds is 6.1% or lower. For expoexpo-sure higher than 6.1%, the line of normal bond holdings of German banks produces higher loan supply for every given C-VaR level.
If the risk management of each country’s representative bank behaves in a way that has a certain level of credit risk that has to satisfy and decides the levels of loans supplied and bonds bought with a certain constraint for the level of C-VaR, the improvement from adopting a different bundle instead of the nor-mal composition of government bonds can be examined. Assuming a C-VaR con-straint of 41.5% for the risk manager of the Greek bank, the improvement from adopting a fully diversified government bond bundle is very high. The loan sup-ply when Greek banks hold their usual government bond compositions is around 50% of total assets. The diversified bundle returns a level of loan supply around 90.35% of total assets and the SBBS portfolio returns 90.5% of loans supplied.
LGD Benchmark Adverse
VaR(%) # wiped out VaR(%) # wiped out
35% 13.281 0 61.127 0
50% 18.972 0 87.324 6
75% 28.458 0 98.771 175
Figure 20. Greek bank’s credit risk for different compositions of government bonds
Assuming a constraint C-VaR level of 31.3% for the Italian banks, for the normal bundle of government bonds the loans supplied equals the 89.85% of the banks’ total asset value. For a diversified bundle, the supply of loans increases a little to 90.4% of total asset value and when the banks hold SBBS assets, supply of loans is around 91.3%.
Figure 21. Italian bank’s credit risk for different compositions of government bonds
For a constraint set to 26.5% of C-VaR for Spanish banks, the supply of loans for the normal composition is around 89.4%, for the diversified portfolio around 90.05% and for SBBS assets around 91.1%.
50 60 70 80 90 100
20 25 30 35 40 45 50
Loans-to-total-assets (%)
C-VaR (%)
GR GRSBBS GRdiversified
50 60 70 80 90 100
15 20 25 30 35 40
Loans-to-total-assets (%)
C-VaR (%)
IT ITSBBS ITdiversified
Figure 22. Spanish bank’s credit risk for different compositions of government bonds
For Portuguese banks, the normal composition of government bonds is very risky compared to the given riskiness of loans in the country and hence, the improvement from the diversified bundle is very high. For a C-VaR constraint equal to 25.76% and the normal bond composition, the loans to assets ratio is at 94%. With the diversified bundle the level of loans supplied is around 97% and with SBBS the loans to assets ratio reaches 97.2%.
Figure 23. Portuguese bank’s credit risk for different compositions of government bonds 50
60 70 80 90 100
10 15 20 25 30 35
Loans-to-total-assets (%)
C-VaR (%)
PT PTSBBS PTdiversified 50
60 70 80 90 100
10 15 20 25 30 35
Loans-to-total-assets (%)
C-VaR (%)
ES ESSBBS ESdiversified
Setting a C-VaR constraint of 11.8% for German banks, the normal bundle of bonds leads to 90.65% of loan supply to total assets. The diversified portfolio supplies lower amount of loans, around 90%. The portfolio that contains SBBS assets is again the best portfolio and increases the supply of loans to 92.4% of the total amount of assets that German banks own.
Figure 24. German bank’s credit risk for different compositions of government bonds
Lastly, taking into account the increase in loan supply from the analysis above, as well as, the values of consolidated banking assets in each country, the change in supply of loans in each country from the different government bond bundles is reported in table 8 below. The change from government bond compo-sition that banks normally hold in these five countries to a fully diversified port-folio that consists of 19 government bonds from EMU members leads to a 1.58%
increase in the loan supply to their economies. Greece and Portugal are the two economies that receive the highest increase in loan supply due to the wide spread between loan supply with normal government bond composition and the diver-sified portfolio. For Germany the loan supply decreases when banks hold the di-versified portfolio by 55 billion euros, a 0.72% decrease of loan supply for the German economy. However, the increase in loan supply from the other 4 coun-tries is high enough to create a 110.2 billion euro increase in the overall loan sup-ply, which translates to a 0.9% increase. For SBBS, the increase in loan supply is more than three times the increase from holding the diversified portfolio, an in-crease of 336.93 billion euros, a 2.75% inin-crease in the supply of loans. When Ger-man bank hold SBBS, their loan supply increases as well which is the reason for the higher increase compared to the diversified approach.
50 60 70 80 90 100
6 7 8 9 10 11 12 13 14
Loans-to-total-assets (%)
C-VaR (%)
DEU DEUSBBS DEUdiversified
Table 8. Changes in loan supply for three government bond bundles
C-VaR constraint Normal Diversified SBBS
Greece 41.50% 131.25 237.17 237.56
Italy 31.30% 2365.93 2380.41 2404.11
Spain 26.50% 3161.54 3184.53 3221.66
Portugal 25.76% 358.33 369.76 370.53
Germany 11.80% 6222.31 6177.69 6342.43
Sum 12239.36 12349.56 12576.29
Increase in
loan supply 0.90% 2.75%
6 POLICY IMPLICATIONS
One of the main recommendations in the recent literature has been the imposition of a new policy to increase diversification of governments bonds inside the euro area in order to improve bank stability and mitigate the bank-sovereign loop (Ar-nold, 2012; Bénassy-Quéré et al., 2018). The simulation model of this paper shows evidence that diversification of government bonds can possibly reduce banks’
credit risk in countries with high NPL ratios. However, banks that operate in countries within sovereigns that issue government bonds with very low proba-bility of default (e.g. Germany) might experience increase in credit risk when di-versification of government bonds is higher. More specifically, banks which al-ready have low-risk government bonds might experience a 1-5% increase in their credit risk if they increase their diversification to government bonds. In the ex-treme case where, German banks increase their exposure to government bonds up to 50%, the simulation shows that the increase in VaR could be up to 30%. In addition, it is shown that for exposures higher than 6.1% to government bonds the normal bundle of German government bonds results into safer investments compared to the fully diversified bundle.
Furthermore, various scholars (Lenarcic, Mevis & Siklos, 2016; Arnold, 2012) have suggested that government bonds should not be treated differently compared to other assets and that the large exposure framework should be ap-plied to government bonds in a similar manner. While increased exposure to risky government bonds increases credit risk for banks, increased exposure to low risk government bonds might have a neutral effect or even reduce credit risk.
This finding suggests that in an environment where there is high freedom for applying zero risk weights for government bonds, a 25% exposure limit to gov-ernment bonds in the banks’ balance sheets would have a positive impact on the reduction of credit risk that occurs from high percentage holdings of risky gov-ernment bonds.
Consequently, the high exposure framework to government bonds can positively impact banks of countries that suffer from high exposure to risky gov-ernment bonds and have a very small negative impact, or no impact at all, on banks that hold safe government bond portfolios. This proposition would pro-hibit banks from increasing exposure to risky government bonds during adverse times, a bank behavior that has been observed previously. However, it should be pointed out that such an approach may increase flight to quality as also argued by Schneider & Steffen (2017).
Lenarcic, Mevis & Siklos (2016) propose that positive risk weights for gov-ernment bonds diminishes the doom loop. Simulation results of this paper show that applying zero risk-weights to government bonds that have a non-zero prob-ability of default leads to insufficient capital requirements. As it is expected from higher exposure to risky government bonds, the importance of applying proper risk weights also rises. However, one of the important lessons learned from the global financial crisis and the European sovereign debt crisis was that liquidity
squeeze can cause major problems to banks, as well as their sovereign ments. It is thus, essential to determine whether positive risk weights on govern-ment bonds would have a negative impact on the ability of governgovern-ments to re-ceive funding, so they can implement their fiscal policy without issues.
The introduction of LRR is not sufficient to compensate for zero-risk weights. Minimum capital requirement of 8% together with LRR equal to 3% re-sults in a minimum risk weight level for all assets equal to 37.5%. Being as it may, the findings of Kiema & Jokivuolle (2014) state that an LLR of 3% is extremely low and it would not increase bank stability. In the case of a negative shock to low risk loans, bank stability might actually reduce, due to contamination effects.
As it is suggested only a higher level of LRR, close to the average level of risk-based capital requirements, would be more successful in increasing bank stability, and as a result reduce the occurrence of twin crises.
The creation of a safe asset in the euro area as proposed by Brunnermeier et al. (2016, 2017) is a feasible solution for the mitigation of doom loop phenom-ena. Their findings suggest that the senior tranch of SBBS will bear at least the same risks as most low-risk government bonds in the euro area, leading to an increase in the supply of safe assets, so that all EMU members can have access to safe assets and their benefits. Simulation results in this paper support the pro-spect of removing the sovereign risk from banks’ balance sheets, which originates from their government bond holdings. However, this proposal is only going to steer the majority of risk, if not all, to other recipients, the holders of lower sen-iority tranches of SBBS. Some scholars have raised their concerns that this ap-proach could result in the same adverse effects of the sub-mortgage and sover-eign debt crisis. In addition, it has been pointed out that the supply of a safe asset in the euro area can only be increased either through ECB’s statement about guar-anteed safety for government debt in the region via the shadow banking system (Gabor & Vestergaard, 2018) or through risk sharing between nations (De Grauwe, & Ji, 2018).
Banks are obligated to follow exogenous constraints related to the credit risk of their assets that they hold, such as the propositions of the Basel accord.
For this reason, the level of loans supplied to the economy depends on banks’
credit risk. If a bank holds riskier assets their loan supply is going to be lower.
This paper finds support for the findings of Gennaioli, Martin, and Rossi (2018) that government bond holdings negatively affect the loan supply of banks. In addition, higher exposure to risky government bonds and NPLs further increases credit risk of banks’ balance sheets and thus, loan supply reduces further. The introduction of diversification and/or SBBS assets effectively diminishes the neg-ative effect of exposure to government bonds on the supply of loans.
Even though diversification of government bonds can possibly reduce credit risk that emerges from high exposure to government bonds, special atten-tion is needed for those banks that operate in safe countries. The government bond portfolio composition of such banks might be much safer compared to the diversified bundle, so regulators need to take that into account when imposing laws that affect the levels of diversification and exposure of banks to government bonds. Furthermore, the introduction of a safe asset, such as the SBBS has shown
positive results in its ability to reduce credit risk of banks’ balance sheets and break the direct linkages that create doom loops, increase home bias and increase loan supply for households and enterprises. Nonetheless, the feasibility of SBBS is subject to many variables that are not examined in this paper thoroughly, for example demand for and regulation of junior and senior SBBS, credit rating given to SBBS and the structure of tranching. Therefore, an initiation of a test phase for such an asset might be the next step to determine whether it will be a feasible solution for eliminating the doom loop and moving forward the market integra-tion in the euro area without mutualizing sovereign risks amongst the members.
7 CONCLUSION
This study examines the impact of banks’ exposure to government bonds for five countries in the euro area via a Monte Carlo simulation. The relevance of this connection relates to major issues inside the euro area’s financial system, like the vicious cycle of default incidents between banks and their governments, home bias and general instability in the financial system, due to low supply of a safe asset in the area. The simulation study has shown firstly, that diversification of government bonds is effective in reducing credit risk in banks that operate in countries which issue risky government bonds, but can be countereffective in countries whose governments issue safe assets. Secondly, high exposure limits to government bonds can positively affect the composition of banks that operate in risky environments. However, flight to quality, is most likely going to increase in this case which can put further pressure to governments that are already in a bad financial situation. Thirdly, levying zero risk weights to government bonds that are not risk-free makes banks hold lower capital than what would be required for them to be in safe against sovereign defaults. Moreover, the introduction of SBBS in the euro area would effectively remove the sovereign risk from banks’ balance sheets and thus, break one of the main linkages that create the vicious cycle of defaults between banks and their sovereign. Finally, the exogenous credit risk constraints that banks have to follow, like the minimum capital requirement of the Basel accord, can possibly impact the level of loan supply to the economy. A government bond composition that has lower credit risk can reduce the overall credit risk of banks’ credit risk and provide the opportunity for banks to increase their loan supply. Consequetively, a government bond composition that contains all 19 issuers of government bonds in the euro area, manages to increase the loan supply in the euro area overall.
Germany is the only country out of the five countries examined, that experiences a loan supply decrease when banks hold the diversified composition of government bonds. Additionally, the introduction of SBBS leads to an increase in loan supply for all five countries, a further positive development that the creation of SBBS has to offer in the euro area.
The research method of this paper is a simulation and therefore the results are indicative. A technique such as a Monte Carlo simulation can provide the means for the researcher to understand issues that otherwise would be extremely difficult to examine and find answers for, due to lack of data or because of the complicated connections between different factors. However, calibration of the model is a difficult task. There can be occasions in which deciding the right value for an input is not a straight forward decision. For this reason, empirical evidence that would further support the findings of this paper would increase the trustworthiness of this study.
Furthermore, due to the complicative nature of this paper’s point of focus , some of the assumptions that have been introduced simplify things. Asset values follow the standard normal distribution which causes the expected returns to be
equal to zero. In addition, LGD and PD values are appointed separately.
Although, Altman et al. (2004) show that expected and unexpected losses are largely undervalued if PD and LGD are uncorrelated.
Future research can take many different paths. One of those can be testing the role of exposure in government bonds in a dynamic model. A further understanding of the causes of loan supply shocks and their implications can also be examined via a dynamic model to a greater extend. Further empirical evidence that supports the indications of theoretical models has always been a necessity in the literature as well.
REFERENCES
Acharya, V., & Steffen, S. (2015). The “greatest” carry trade ever? Understand-ing eurozone bank risks. Journal of Financial Economics, 115(2), 215-236.
Acharya, V., Drechsler, I., & Schnabl, P. (2014). A pyrrhic victory? Bank bailouts and sovereign credit risk. The Journal of Finance, 69(6), 2689-2739.
Albertazzi, U., Ropele, T., Sene, G., & Signoretti, F. M. (2014). The impact of the sovereign debt crisis on the activity of Italian banks. Journal of Banking &
Finance, 46, 387-402.
Altman, E., Resti, A., & Sironi, A. (2004). Default recovery rates in credit risk modelling: a review of the literature and empirical evidence. Economic Notes, 33(2), 183-208.
Andritzky, J., Gadatsch, N., Körner, T., Schäfer, A., & Schnabel, I. (2016). Re-moving privileges for banks’ sovereign exposures–a proposal. European Economy, (1), 139.
Arellano, C., & Kocherlakota, N. (2014). Internal debt crises and sovereign de-faults. Journal of Monetary Economics, 68, S68-S80.
Arnold, I. J. (2012). Sovereign debt exposures and banking risks in the current EU financial crisis. Journal of policy modeling, 34(6), 906-920.
Arslanalp, S., & Liao, Y. (2014). Banking sector contingent liabilities and sover-eign risk. Journal of Empirical Finance, 29, 316-330.
Balteanu, I., & Erce, A. (2018). Linking Bank Crises and Sovereign Defaults: Evi-dence from Emerging Markets. IMF Economic Review, 66(4), 617-664.
Basel Committee on Banking Supervision (BCBS) (2018, April 30). Liquidity Cov-erage Ratio (LCR) - Executive Summary. Retrieved from
https://www.bis.org/fsi/fsisummaries/lcr.htm
BCBS (2018, February 13). Countercyclical capital buffer (CCyB). Retrieved from https://www.bis.org/bcbs/ccyb/
Bénassy-Quéré, A., Brunnermeier, M., Enderlein, H., Farhi, E., Fratzscher, M., Fuest, C., ... & Schnabel, I. (2018). Reconciling risk sharing with market discipline: A constructive approach to euro area reform. Centre for Eco-nomic Policy Research Policy Insight, No.91
Benczur, P., Cannas, G., Cariboni, J., Di Girolamo, F., Maccaferri, S., & Giudici, M. P. (2017). Evaluating the effectiveness of the new EU bank regulatory framework: a farewell to bail-out?. Journal of Financial Stability, 33, 207-223.
Boone, P., & Johnson, S. (2014). Forty years of leverage: what have we learned about sovereign debt?. American Economic Review, 104(5), 266-71.
Brunnermeier, M. K., Garicano, L., Lane, P. R., Pagano, M., Reis, R., Santos, T., ... & Vayanos, D. (2016). The sovereign-bank diabolic loop and
ES-Bies. American Economic Review, 106(5), 508-12.
Brunnermeier, M. K., Langfield, S., Pagano, M., Reis, R., Van Nieuwerburgh, S.,
& Vayanos, D. (2017). ESBies: Safety in the tranches. Economic Pol-icy, 32(90), 175-219.
Caballero, R. J., & Farhi, E. (2013). A model of the safe asset mechanism (sam):
Safety traps and economic policy (No. w18737). National Bureau of Eco-nomic Research.
Caballero, R. J., Farhi, E., & Gourinchas, P. O. (2017). The safe assets shortage
Caballero, R. J., Farhi, E., & Gourinchas, P. O. (2017). The safe assets shortage