For a financial crisis of such a magnitude that it threatens both the solvency of the Spanish state and effectively destroyed the credibility of the Spanish supervisors, it is surprising how slowly it has developed. With a similar combination of a large real estate bubble and a complicated financial crisis the Irish bank recapitalisation (and the nationalisation of Anglo Irish) in January 2009 took place a full three years before the entire Spanish financial system had to confront the reality of its losses with the collapse of Bankia in May 20121. By this time, the Eurozone crisis was in full swing, and Spain did not have access to the markets, forcing Spain to seek European help. This would have been unnecessary had the crisis been forced in the open before the euro doubts came to the fore and while Spain had a low debt to GDP ratio and easy access to the market. An additional casualty of the crisis, apart from Spain’s solvency, was the Banco de España’s reputation, as evidenced by the fact that Europe forced Spain to have an external private sector firm (Oliver Wyman) undertake the stress tests.
Why was the crisis response so slow? What can we learn from the Spanish supervisory debacle for the future European Single Supervisory Mechanism?
A hint of the answer comes from the knowledge that, with a few relatively small exceptions, the Spanish financial crisis has been a crisis of the caja (savings and loan) sector. Adding up the previous injections by the Spanish bank rescue funds (FROB) to the estimates of the recent official Oliver Wyman (September 2012) report, just the three most problematic Spanish cajas (Bankia, CatalunyaCaixa and Novagalicia) have had capital deficits (to be covered partly or fully by the taxpayer) of €54 billion – over 5% of Spanish GDP, a larger amount than what Spain will have to request from the European rescue funds.
The Oliver Wyman report, the fourth evaluation of the solvency of the financial system in three years, makes very clear that the problematic cajas were busy reclassifying, refinancing, and extending loans to cover up their losses in the previous four years. Indeed, the evidence of a cover up on the part of the worst cajas’ management during 2008, 2009, and 2010 was overwhelming2. And yet, the Banco de España did not confront it.
In fact, it kept being surprised when in each caja that failed the holes uncovered where larger than expected. Already the first entity that was intervened (CCM) as far back as March 2009, showed that the real NPL levels post intervention (17.6%) were more than twice as large as the reported ones. This should have been the point for the Banco de España to get ahead of the curve by ordering an audit of the whole sector (which eventually did happen, three and a half years later). Instead, no one went back to the other cajas to try to correct the numbers. Each further intervention (CajaSur, CAM) resulted in similar jumps, and each time the reaction was circumscribed to the fallen entity. More evidence of a cover up was uncovered, in a widely reported analysis, by Santiago López Díaz from Credit Suisse. He showed (Coterill 2010) that there was a clear cyclical pattern in non performing loan (NPL) recognition: NPLs increased sharply in the first two months of each quarter, and then became systematically negative in the third one, in the third one, when the numbers had to be reported. Finally, all market observers were shocked that the stock of real estate developer loans (32% of Spain’s GDP) was still growing through that period (see Figure 1), in spite of large bankruptcies in the sector and few new developments being started. This suggests many loans were being informally restructured or refinanced.
Figure 1. The fast buildup and slow drop-off of real estate developer loans (€bn)
Source: Banco de España. Series BE_4_19.10.
If the evidence was in plain sight, why did the Banco de España not react? There is no intimation by anyone of outright corruption in the Banco de España supervisory role, and given the professionalism of the institution it is unlikely that there was any. There are four likely reasons for this failure and they suggest lessons for other supervisors.
All supervisors are reluctant to uncover their own previous mistakes – an instance of career concerns (i.e. jamming the signals to improve the perception of performance, as in Holmstrom 1982) of those responsible for the failure. Thus not surprisingly, Banco de España supervisors had little interest in discovering that Spain’s vaunted regulator had in fact missed the largest financial crisis in the history of the country. Unfortunately, often supervisors in charge of the failing entity in the years of the debt run up were the ones charged with uncovering the problems. Indeed, they were not too eager to put in question their own previous work.
But all supervisors have career concerns, and they would not have mattered as much had the crisis developed more suddenly. The dynamics were partly dulled by the existence of dynamic provisions, which were in that sense working as desired. Spain was the leader in the introduction of a dynamic provision – a provisioning tool that forces banks to increase provisions without reference to any specific loan. The intention of this tool was twofold: to mitigate the bad times, and to cool the booms in the good times (Holmstrom and Tirole 1997). Dynamic provisions were endorsed as part of the Basel III standards in December 2010, in part on the strength of Spain’s experience. And indeed the existing evidence (Jiménez et al. 2012) shows that the tool worked as intended, dampening the credit boom and softening somewhat the credit crunch. However, it is clear by now that their level was not nearly enough, as their size – 3% of GDP at their highest point (2004) – was simply not of a magnitude commensurate with the credit losses.
A more insidious consequence of the existence of this buffer is that it allowed the reality to be hidden in plain sight for longer than it would have been otherwise possible. Without the provisions, the reality of the cajas' accounts would have become much faster a concern, and would have imposed itself on the regulator. While this is not an argument to abandon dynamic provisions, it is an argument to make sure to take into account the impact they may have in dulling supervisor´s incentives.
Another reason for the lack of action by the Banco de España was the lack of an appropriate resolution framework at the time. Had the Banco de España ordered an audit of the system after uncovering numerous irregularities in CCM, it would have not been able to deal with the capital shortfalls uncovered as there was no appropriate resolution regime in Spain at the time (indeed there has not been one until the Summer of 2012).
But the main explanation for the supervisory failure of the Banco de España has to do with the political control of the cajas. As in many previous financial crises (most recently, the Savings and Loans in the USA, the Asian crisis and the current Irish crisis), governance played a critical role in the development of the Spanish crisis. In the Spanish case, the supervisor, confronted with powerful and well connected ex-politicians decided to look the other way in the face of obvious building trouble.
Indeed the political connection of the managers of the entities was a good predictor of brewing trouble. Anecdotally, the worst in terms of the losses it will impose on taxpayers, Caja Madrid/Bankia, was the most politicised. For just one nugget, the appointment of the CEO that led it in its out-of-control-bubble years was the result of a formal but secret pact between the Madrid conservative Popular Party and the main hard left trade union in 1986, CCOO, which involved among other things access by the union to the executive commission and participation in restructuring decisions behind the back of the caja board. More systematic evidence of the role played by these governance issues is provided in a 2009 paper (Cuñat and Garicano 2009b) where we showed that cajas with chief executives who had no previous banking experience (!), no graduate education, and were politically connected did substantially worse in the run up to the crisis (granting more real estate developer loan, up to half of the entire loan book in some instances) and during the crisis (with higher NPLs).
Even more important was the role of these political connections in diluting the role of the supervisor after the crisis started, in what was meant to be the crisis resolution stage but which was in fact a crisis cover up stage. The mergers that were decided followed political and regional criteria, rather than economic rationales. The Popular Party cajas (those controlled by Popular Party regional politicians) merged together, which meant that two of the most problematic cajas in the crisis (Bancaja from Valencia and Caja Madrid) ended up as part of the same undercapitalised (now nationalised) entity. The two Galician cajas also merged, because Galicia 'must have' a local credit institution, creating a monster that is also nationalised. Same with the medium-sized Catalan cajas, also now nationalised. In all of these instances, the regulator was mindful of the figures, and understood the costs of these politicised entities, but refused to impose its will in the face of concerted actions by the politicians.
What are the takeaways of the Spanish supervisory debacle for the new European supervisor? I would suggest five.
First, in systemic crisis, the problems do not necessarily have to affect large institutions, but may instead impact a lot of small institutions. Moreover, the small institutions may play the role of the canary in the mine in anticipating the systemic problems. The supervisor should have all the relevant information and that requires covering (essentially) the entire banking system, as in the European Commission 12 of September SSM proposal (COM 2012 511).
Second, career concerns of supervisors are crucial. Like auditors post Enron, supervisors must be rotated from accounts with a certain frequency, and certainly when problems are building up in the system.
Third, dynamic provisioning is a good idea, but the supervisor must be mindful it may delay decision making in problem cases and as a result, and contrary to their intent, make the crisis larger than it should have otherwise been.
Fourth, the supervisor must be able and willing to stand up to politicians. Tackling corporate governance issues in semi-public entities is delicate, as it involves challenging regional power centres, political parties and unions. However, the supervisor must have the courage to be as intrusive as necessary to ensure the necessary professionalism and knowledge in these institutions in the good times, even in the absence of obvious problems, in the knowledge that when they fall, they are likely to take the taxpayers money with them.
Fifth, supervision and an appropriately tough resolution regime must go hand in hand. It is impossible for the supervisor to act sufficiently aggressively to build confidence and get in front of the market if there is no appropriate resolution regime that it can use once irregularities and capital shortfalls are uncovered in the supervised entities.
Coterill, J (2010), "What is up with Spanish mortgage lending?", FT Alphaville, Jul 20.
Cuñat, V and L Garicano (2009a), "¿Para cuándo la reestructuración del sistema financiero español?", El Pais, September 13.
Cuñat, V and L Garicano (2009b), "Did Good Cajas Extend Bad Loans? The Role of Governance and Human Capital in Cajas’ Portfolio Decisions", FEDEA monograph.
Holmstrom, B (1982/99), "Managerial incentive problems: A dynamic perspective", Review of Economic Studies 66, 169–182; originally published in Essays in Economics and Management in Honour of Lars Wahlbeck, Helsinki, Finland.
Holmstrom, B and J Tirole (1997), "Financial Intermediation, Loanable Funds, and the Real Sector", The Quarterly Journal of Economics 112( 3), 663-691.
Jiménez, G, S Ongena, J L Peydró and J Saurina (2012), "Macroprudential Policy, Countercyclical Bank Capital Buffers and Credit Supply: Evidence from the Spanish Dynamic Provisioning Experiments", Banco de España. Mimeo.
1 To be sure, Irish banks (particularly Anglo Irish) unlike the Spanish cajas, were exposed to the US real estate bubble as well, and have been able to generate lower pre-provisioning profits than the Spanish financial system. This, however, does not justify the length of the delay.
2 For a view at the time, see Cuñat and Garicano (2009a).