The notional value of non-performing loans of the 15 largest and systemic relevant Italian banks under direct supervision of the Banking Union’s Single Supervisory Mechanism (SSM) summed up to 236 billion Euro by the end of 2015. These problem loans have risen from 9.3 percent of the Italian gross domestic product (gdp) to 14.4 percent. Thereby the share of Italian problem loans of the Eurozone banking sectors’ problem loans rose from 21 percent in 2009 to 30 percent in 2015. This divergence is also due to the non-Italian Eurozone banks’ progress in repairing their balance sheets. While the percentage fraction of Eurozone non-performing loans to the Eurozone’s gdp increased from 7.4 percent in 2009 to 15.5 percent in 2013, this ratio is now back at its 2009 level.

In contrast to Ireland and Spain, which faced housing booms and busts, Italy’s non-performing loans did not result from indebtness. In Italy, corporates’ and households’ indebtness is below their Eurozone averages. The Italian households’ debt-to-disposable-income ratio of 62.6 percent is even lower compared to Germany’s low household debt ratio of 84.5 percent. Spanish households and Irish households have debt ratios of 107.6 and 158.8 percent instead. Italian companies have a debt-to-gdp ratio of 85.0 percent, while Spanish companies’ indebtness fits the Eurozone average of 106.3 percent and Irish companies have a debt ratio of 187.4 per-cent. Instead, Italy’s banking sector problems are in part caused by the severeness of the 2009 recession and the slow recovery thereafter. The Italian recession was not only deeper compared to the Eurozone as a whole, but the Italian economy’s recovery is still below the Euro-zone average.

In addition to that, structural deficiencies in Italy have prevented banks to get rid of their problem loans. One is the inefficient tax treatment of loan-loss provisions and write-offs, the other is the length of insolvency processes, which take on average 1.8 years and which are thereby longer compared to most other European countries. Moreover, Italy has the highest costs of resolving insolvency processes in Europe (Scope, 2015).

Under the Italian government’s plan, a bank bad should be funded with 10 billion Euro of public funds, which should free banks’ balance sheets by 50 billion Euro of non-performing loans (Wallace/Marlow, 2016,). Calculations with balance sheet data from the 15 systemically relevant Italian banks show that by transferring this amount of non-performing loans to a bad bank, Italy’s banking sector’s problem loans would shrink by 21 percent, which would be a reduction from 14.4 percent of the Italian gdp to 11.4 percent.

For this, each of the 15 banks has to transfer on average 21 percent of its problem loans to the bad bank. Because the bad bank is exposed to credit risks, it will buy these problem loans only at a price of 80 percent below their notional value, which would expose banks to losses of 40 billion Euro. Calculations reveal that under this plan 10 of the 15 banks will be in need of additional capital injections, because their capital ratios will erode below the Eurozone average of 6 percent equity capital to total assets. In order to maintain equity capital ratios at, for example here, the Eurozone average, the 10 banks need capital injections ranging from 0.1 billion Euro to 8.0 billion Euro, which sum up to a total of 21.7 billion Euro.

Under the EU’s Bank Restructuring and Resolution Directive (BRRD) and under the EU’s State-Aid Rules, creditors have to bear losses of at least 8 percent of the banks’ liabilities, before any publically funded recapitalisation measures can be applied. Under the above mentioned plan, creditors’ contributions would suffice to cover each bank’s losses. But the problem with creditor bail-in in Italy is, that a significant part of the bail-in-able liabilities are not held by institutional investors with appropriate risk management capabilities, which can bear losses of that amount, but by Italian households. Merler/Minenna (2016) estimate that one third of all issued Italian bank bonds are held by Italian households. Bailing-in retail investors is in line with the BRRD, but politically infeasible. Especially, in light of the upcoming Italian referendum.

In case the creditor bail-in becomes politically infeasible, the Italian government could be tempted to expose the bad bank to more credit risk. If it will transfer the non-performing loans, for example, only at a 50 percent discount to the bad bank’s balance sheet, then only 7 banks would be in need for recapitalization measures, which would sum up to 11.2 billion Euro in total (table). Under this alternative additional 4.5 billion Euro public money are at stake, because shareholders bear less losses, which would not be in line with the BRRD either.

An alternative scenario would be, that the Italian government could be tempted to recapitalise its banks with public money. Under its proposed balance sheet repair measures, it would have to invest 10 billion Euro into the capitalisation of the bad bank and it has to put 21.7 billion Euro into recapitalisation measures. But the government can reduce the usage of public funds by transferring the non-performing loans to the bad bank at a price of zero. Then the government has to fund recapitalisation measures in the amount of 31.7 billion Euros, but it would save 1.7 billion Euro compared to its original plan.

These 1.7 billion Euros must be covered by the stockholders’ capital. But given that Italian bank stocks already lost between 27.4 and 97.0 percent, the banks’ market capitalisation is lower compared to the book value, so stockholders will bear losses of less than 1.7 billion Euro on their portfolios. Given the fact that most creditors are retail investors without any risk management, this option seems to be the best feasible solution for Italy’s bank problems. A similar solution was proposed in the past by Schäfer/Zimmermann (2009).

In order to prevent retail investors to bear losses, which makes a creditor bail-in political infeasible, bail-in-able bank bond holdings should be restricted to institutional investors outside the banking sector, who are actually able to bear losses from a bank restructuring, as proposed by Liikanen et al. (2012). To this end, banks need time to restructure their balance sheets. Moreover, the conditions under which a creditor bail-in takes place have to be pre-specified in the contracts, so that creditors can demand a reasonable risk premium for bearing the additional credit risk.