The September 2008 Guarantee

The media have decided that one of the headline conclusions of the Honohan report was that “Honohan supported the September 2008 guarantee” and the government have been very keen to consistently repeat that line.

I think the truth is that the report says that some kind of guarantee was required but that it asks very serious questions about the nature of the guarantee put in place. The only issue I have heard discussed in relation to the report’s questioning of the guarantee is the inclusion of subordinated debt. Indeed, the Taoiseach’s speech in the Dail this afternoon also mentioned the inclusion of subordinated debt as the only quibble that the report had with the guarantee.

In fact, my reading of the relevant section is that the report questions the essential nature of the type of guarantee that was introduced here. Here, I want to highlight the relevant discussion in the report and to flesh out its implications.

Here’s the relevant section (8.39 in the report):

The scope of the Irish guarantee was exceptionally broad. Not only did it cover all deposits, including corporate and even interbank deposits, as well as certain asset backed bonds (covered bonds) and senior debt it also included, as noted already, certain subordinated debt. The inclusion of existing long-term bonds and some subordinated debt (which, as part of the capital structure of a bank is intended to act as a buffer against losses) was not necessary in order to protect the immediate liquidity position. These investments were in effect locked-in. Their inclusion complicated eventual loss allocation and resolution options. Arguments voiced in favour of this decision, namely, that many holders of these instruments were also holders of Irish bonds and that a guarantee in respect of them would help banks raise new bonds are open to question: after all, extending a Government guarantee to non-Government bonds has the effect of stressing the sovereign to the disadvantage of existing holders of Government bonds; besides, new bonds could have been guaranteed separately. The argument for simplicity also is weakened significantly by the fact that an actual dividing line between covered and non-covered liabilities was drawn at as least an equally arbitrary point; moreover, such instruments were held only by sophisticated investors.

The key point here is not that subordinated bonds were included (this decision is singled out in the short paragraph 8.40 that followed and later in 8.50 as “not easy to defend”). Rather the point was that the guarantee applied to existing bonds rather than new debt issues. While most of the discussion of the guarantee in Ireland has focused on the very broad range of instruments covered, this distinction between existing and new debt is at least as important.

Irish banks, like many banks during this period, hit a liquidity crisis (based on concerns about their long-run solvency). Government guarantees were required if they were to be able to go into the market and raise new funds to pay off existing debts. So a government guarantee of some sort was required if the liquidity problems were not to trigger widespread defaults. However, the guarantee of almost the entire stock of existing debt was, as the report notes, “not necessary in order to protect the immediate liquidity position.”

We have heard time and again that guarantees were provided in lots of countries. However, the Irish decision to guarantee the full stock of existing debt, rather than new issues, was followed by hardly anyone. Here’s a link to an ECB document from last summer that outlined the various measures taken by European governments: You won’t find many guarantees of existing debt, you’ll find plenty of guarantees of new issues.

This decision to underwrite the full stock of debt of the Irish banks rather than assist them with new borrowing had profound implications. As the above paragraph noted, it “complicated eventual loss allocation and resolution options.” Of course, there’s a sense in which it made things pretty simple: It put the taxpayer on the hook for almost all losses.

The decision was motivated by a view that any failure of Irish banks to repay debts could not be countenanced. As Sections 8.48 and 8.49 of the report notes, this decision:

was an oversimplification which short-circuited decisions that deserved closer scrutiny. Under the circumstances of the extraordinary international financial market environment of those weeks, it was an understandable position. But it could not be a permanent policy if severe moral hazard was to be avoided. And it was also conducive to downplaying the importance of developing an appropriate legal framework for a special bank resolution regime scheme.

The “no failure” policy also took the question of optimal loss-sharing off the table. In contrast to most of the interventions by other countries, in which more or less complicated risk-sharing mechanisms of one sort or another were introduced, the blanket cover offered by the Irish guarantee pre-judged that all losses in any bank becoming insolvent during the guarantee period – beyond those absorbed by some of the providers of capital – would fall on the State.

To be concrete, consider an alternative decision in September 2008 to fully guarantee deposits and guarantee new issues. These new issues could be guaranteed out to some maximum maturity with the agreement of the Department of Finance and the policy of guaranteeing new issues could be reviewed in three or six months.

Such a policy would have allowed for a review that could have deemed Anglo Irish Bank to be no longer covered by the guarantee. This would have given real teeth to the later plan to split Anglo into a good bank and bad bank because the bad bank could have ended up with debt that has since been paid off by an Irish government unwilling to dishonour its guarantee.

The failure to focus on new issues, and to instead guarantee all debt up to September 2010, has also lead to the potentially serious problem we now face in rolling over bank debt over the next few months.

The government was slow to introduce a new scheme that focused on guaranteeing new issues out to maturity. The ELG scheme was passed in December 2009 and did not start operating until earlier this year. For this reason, the banks issued very large amounts of debt after October 2008 that mature prior to October 2010: €57.8 billion in senior debt and €16.4 billion in interbank deposits. The introduction of a scheme that focused only on new issues could have been designed in a way that avoided the single “wall of cash” moment that we have descending upon us.

Governor Honohan is more than capable of speaking for himself on these issues, so I don’t want to get into the question of the importance he places on the qualifications he has aired (i.e. whether it is the case, as I heard on the radio last Friday from a political commentator, that he was 97% in favour of the guarantee!)

However, speaking for myself, I think these flaws in the original formulation of the guarantee are extremely serious and may prove to be very costly.