Description

Investment recommendation

I recommend purchase of the Bank of Ireland 10.24% perpetual preferred stock trading around 105-106 for a highly likely 8% return for 2 ½ years. (ISIN XS1003011563 or BKIR 10.24 12/29/2049 on Bloomberg)

My premise is that this preferred stock is mispriced because it doesn’t fit within the mandate of most institutional investors. Fixed income investors are turned off by the issuer’s option to pay the dividend in common stock not cash, by the perpetual term, and because the preferred stock ranks pari passu with common stock on a winding up or other return of capital of the bank.  They settle for a paltry 2.5% return on senior notes with a similar maturity (ISIN XS 0940658361 or BKIR 2.75 6/5/16).  Equity investors are turned off by the limited upside. They settle for no earnings today, and pay over 1.3x book on expectations of earnings tomorrow.  Others might be confused by the issue’s quirky structure given that the issuer is a special purpose vehicle with no assets other than the Bank’s 2009 preferred stock.

My main assumption is that management has committed to the bank’s primary regulator, the Central Bank of Ireland, to redeem the preferred stock by 1 August 2016 either with retained earnings or with a new issue of common stock.  Accordingly, I think it is highly likely that investors who pay €105 today will in return receive: (a) €1.35 of interest on 20 February 2014, (b) €10.24 of interest on 20 February 2015, (c) €10.24 of interest on 20 February 2016, (d) €4.57 of interest on 1 August 2016, and (e) €100 of principal on 1 August 2016.

Owing to the constraints of the security's structure, it is difficult for a scenario to unfold that delivers a return much lower than 8%. Of course, it is hard to do much better than 8% either.  Even in the most negative scenario that seems still plausible, the annual return is 4%. The two scenarios are illustrated in the tables below.  

Background

For background on Bank of Ireland, see a post dated 20 January 2012.

Bank of Ireland issued €3.5 billion of preferred stock to the Irish government (National Pension Reserve Fund) in 2009 as part of a mandatory recapitalization during the banking crisis.  During 2010, nearly half was converted to common leaving €1.8 billion of preferred outstanding.  In December 2013, the bank repaid the government using proceeds from two new issues:  €500 million of common stock and €1.3 billion of bonds issued by a special purpose company, Baggot Securities Limited.

This was a carefully negotiated transaction and very nearly, if not, regulatory forbearance. The bank had the right to repurchase the preferred stock at par (€1.00 per unit) until 31 March 2014.  But after that day, the repurchase price increases by 25% (“step up provision”) such that the bank owes preferred stockholders 1.25x par (or €1.25 per unit. Furthermore, the preferred stock qualifies as common equity tier 1 regulatory capital under Basel 3 until 1 January 2018. And without the €1.8 billion of regulatory capital, the bank’s capital adequacy ratio would decline from 14.2% to 10.7% (as of Jun-13) which is very close to the 10% minimum the bank has pledged to remain above.   

The bank had three goals: (1) avoid the 25% step up penalty on the preferred stock, (2) retain the advantageous capital treatment of the preferred stock (i.e. qualification as Common Equity Tier 1 regulatory capital) so regulatory capital ratios are not diminished, and (3) minimize dilution to existing common shareholders from issuance of new shares.  The bank achieved all of these objectives because Baggot agreed to waive the 25% “step up” provision.

The government sought repayment of its €1.8 billion for two reasons: (1) to demonstrate to voters that elected officials were recovering taxpayer funds that the prior government had invested in the banks and (2) to improve the budget and boost Ireland’s financial strength.  The government achieved these objectives.

The transaction required approval from Bank of Ireland’s regulatory, the Central Bank of Ireland (repayment of the 2009 preferred stock requires approval of the Central Bank).  Like most regulators, the Central Bank is focused on quality and quantity of capital.  In this case, the Central Bank is desirous that Bank of Ireland’s preferred equity is replaced with common equity as swiftly as possible.  This can happen either through retained earnings or new issuance.  To require Bank of Ireland to repay the government’s €1.8 billion entirely from new issuance might have threatened a positive feedback loop in Ireland’s financial markets and economy at a precarious moment – just as the government was exiting the bailout program organized by the troika.  My theory is that the Central Bank approved a smaller common stock new issue of €500 million provided that Bank of Ireland committed to repaying the remaining €1.3 billion with either retained earnings or with funds from a new issue no later than 31 July 2016.  Hence the statement from management that the Bank has advised the Central Bank that the Bank does not intend to recognize the preferred stock as common equity tier 1 capital after July 2016.       

Accordingly, repayment is likely between 1 January 2016 and 31 31 July 2016.  Management is on the record stating that they do not expect to redeem the preferred stock prior to 1 January 2016.  Repayment may be delayed in a stress scenario for the economy and credit losses, in which case repayment could be delayed until early 2018.  But it almost certainly will not be long after 31 December 2017 because then the preferred stock will cease to qualify as common equity Tier 1 capital and the 10.24% cost will probably be very expensive for debt unless interest rates rise significantly.  Technically, repayment is allowed at any time and the stock is perpetual.

Losing big seems unlikely; you will probably get an 8% return

This preferred stock was designed with common equity-like features on purpose to qualifying as regulatory capital. To me, this is a sort of legerdemain.  The critical difference between the preferred and the common is underscored by the fact that during the Irish banking-sovereign crisis, nearly half the 2009 preferred issue was converted to common as part of Bank of Ireland’s 2010 recapitalization (see notes 47 and 49 of the 2010 annual report).  If the two securities were really similar, such conversion would have been unnecessary.  Nonetheless, the common-like features are real and could influence the investment return in remote scenarios, so it is important to acknowledge them. 

First, the issuer has the option to pay the 10.24% dividend in common shares rather than in cash. So far, the bank has paid dividends in cash for 2011, 2012, and 2013.  The bank was unable to pay cash dividends on the 2009 preferred stock during 2010 because the European Commission prohibited payment of coupons on more senior securities (Tier 1 and Upper tier 2), and provisions of those securities precluded dividends on common and preferred stock.  

Second, if the issuer elects not to pay the dividend in cash and to issue common shares in lieu of cash, then the date of settlement can be determined by the bank in its sole discretion (provided it is no later than redemption or payment of any cash dividends).  The price used to determine the number of shares issued is 100% of the average daily closing price over the 30 dealing days immediately preceding the original scheduled dividend declaration date if settlement is on the dividend payment date (20 Feb) but 95% if settlement is after the dividend payment date.  

Third, the preferred stock is perpetual although under Basel 3 it will no longer qualify as common equity tier 1 capital beginning 1 January 2018 and management has told the Central Bank of Ireland that the Bank does not intend to recognize the preferred stock as common equity tier 1 capital after July 2016.

So, in theory, the bank could elect to issue common shares in lieu of cash dividends and then defer settlement of the shares in perpetuity.  In reality, it is easy to envision the bank paying the dividend in shares not cash: if credit losses follow a stress case over the next several years and cause the capital ratio to fall sufficiently, then management would elect to suspend cash dividends to preserve capital and to slow or to stop the decline until profitability improved.  But it is difficult to envision the bank deferring settlement and repayment of the preferred for an extended period.  Regulators want the preferred stock replaced with common equity, either from retained earnings or new issuance. 

The preferred must be redeemed in cash, unlike the €1 billion of contingent convertible notes outstanding.  Repaying the preferred stock once it no longer qualifies as regulatory capital (by 2018 at the latest) is a top priority of management.  Capacity to repay should be ample provided earnings recover because the preferred stock balance is only 1.0% of total assets.  Historical returns are presented in the table below.

Total assets stand at €134 billion after declining by 32% from a peak of €197 billion.  The deleveraging plan management agreed to with the regulators (both Irish and European) has been completed.  My work suggests that a further decline of 10% to €120 billion would return the bank to a more sustainable level than at present. 

In the adverse scenario that assets decline to €120 billion and the ROA only improves to 0.50%, each year the bank would generate €600 million to pay €130 million of preferred dividends and to retire €470 million of principal.

A “stress case” scenario that threatens capital adequacy is the only reason Bank of Ireland would pay the dividend in stock and not in cash

The bank would only pay preferred holders a dividend in stock not cash is to avoid the significant dilution caused by conversion of the €1.0 billion Contingent Convertible Notes at €0.05 per share.  Conversion becomes mandatory if common equity tier 1 capital falls below 8.25% of risk weighted assets, so management intends to keep it above 9.0%.  Presently, there is a cushion of €2.5 billion because the ratio is 14.2%.  Tougher new capital rules that restrict what qualifies as capital will phase out contributions from three key sources: (1) deferred tax assets -- €550 million, (2) pension deficit – €535 million, and (3) preferred stock -- €1,300 million.  These reductions total €2.4 billion and erode the entire cushion!  So the bank must generate cumulative capital (net profits) by 2018 of ~€400 million to maintain cash dividends on the preferred stock, and a further €1,300 million to repay the principal. This entails significant improvement over recent results but is below the consensus view.

The key variable for capital generation is credit losses.  Loan loss provisions since 2007 have already totaled €14.5 billion, or 10.7% of gross loans at the beginning of the period.  Amazingly, the balance sheet account still shows €8.1 billion because charge-offs have been much slower to materialize.  Provision expense has declined for three half year reporting periods and was €680 million in the first half of 2013 (excluding a one timer of €100 million).

Scenario #1 – status quo case: loan loss provision expenses continue for years around €1,000 million per annum based on the run-rate in the second half of 2013 so that the bank generates no profit or loss (cumulative breakeven).  This situation isolates the impact from changes in the regulatory framework under Basel 3. The core tier 1 ratio would decline from 14.2% at the end of June 2013 to 11.8% at the end of December 2017 as the deferred tax asset (DTA) and pension deficit are phased out, and then to 8.8% at the end of December 2018 when the preferred stock no longer qualifies as equity.  Because management will not permit the capital ratio to fall below 9.0%, the bank would pay stock dividends from 2018 onwards (this boosts net profit and capital by €130 million per year) until profitability recovered to allow the bank to repay the preferred with earnings or proceeds from new issue of common.

Scenario #2 – consensus base case: loan loss provision expenses of €3.0 billion during the three years 2013-2014-2015. This equals 3.2% of current gross loans, or 1.05% per year (€1,000 million), and reflects strong improvement from recent results based on several factors including: (a) rebound in property prices, (b) decline in early stage delinquencies, and (c) rising unemployment.  If credit recovers as strongly as expected, then the core tier 1 ratio bottoms at 14.1% in 2013 and the bank has ample earnings to pay cash dividends and to repay the principal balance of €1,300 million during 2018. 

Scenario #3 – stress case: loan loss provision expenses of €5.5 billion during the three years 2013-2014-2015, and then a slow recovery toward normal by 2019.  The €5.5 billion equals 5.8% of gross loans and is an annual expense of nearly 200bps.  Note that this level of provisioning is in line with recent results.  In this scenario, management may choose to pay dividends in stock as early as 2016 because the projected ratio of 9.5% is approaching their 9.0% minimum.  Their decision would depend on their outlook for future profitability. 

Stress case losses are unlikely without a macroeconomic shock

No one can forecast loan provisions with accuracy.  But an independent third party can identify negligence, sloppiness, and whether the reserve is significantly inadequate.  Bank of Ireland’s loan portfolio and reserves have been subjected to heightened external scrutiny in each of the past three years by various independent third parties hired by the Central Bank of Ireland.  Moreover, management commissioned Oliver Wyman to perform internal stress tests every year since 2009.  This is all in addition to normal supervision by regulators and external auditor PricewaterhouseCoopers. 

In 2011, BlackRock Solutions performed an independent loan loss assessment exercise for the Central Bank and troika as part of a Prudential Capital Assessment Review (PCAR).  This was a high profile stress test, and the Central Bank also hired Boston Consulting Group to check BlackRock’s work.  Since the Irish government had guaranteed all bank liabilities, a rigorous test was the first step for the sovereign to regain credibility in capital markets.  BlackRock spent several months and reviewed 75% of all commercial mortgages over €50 million as part of its assessment.  They estimated lifetime losses of €10.1 billion in a stress scenario.  By the end of 2013, comparable provisions will have totaled €8.6 billion.  If BlackRock’s 2011 stress case projections are correct, then provisions over the next several years shouldn’t much exceed €1.5 billion.  That is about the consensus view today.

In 2012, BlackRock updated their loan loss estimate by the Central Bank did not publicize the results.  In regard to the rigor of the exam, one gains some insight from a remark by the CEO of an Irish bank who remarked that the bank’s headquarters was flooded with BlackRock employees and the bank was spending thousands of man hours complying with their data requests.

In 2013, the troika organized an asset quality review of Irish banks before the sovereign exited the bailout program. Unlike the prior reviews, this one was rushed in a short time frame due to political expediency.  Resources deployed included about 100 people from KPMG reviewing loans at Bank of Ireland and about 100 people from Ernst & Young doing the same at Allied Irish.  Data has been collected and initial conclusions have been formed.  It seems unlikely the Central Bank of Ireland would approve partial repayment of the preferred stock if Bank of Ireland were deemed to have major capital adequacy problems.  

Catalyst