By Erin Davis
As the European Central Bank prepares to complete yet another stress test of the European banks, we question whether it is focusing on the right metrics. Given the significant increase in the sector's stock prices over the past 16 months, we think investors might be too optimistic as well. We acknowledge that the banks are in better shape than they were, as the average core Tier 1 ratio has improved to 13.5% from 10% in 2010. However, we caution that European banks' unweighted capital ratios have not improved nearly as much as headline regulatory capital levels would lead one to believe. For example, one more relevant measure (tangible common equity/tangible common assets) shows that the majority of European banks fail to meet a reasonable 5% standard. Shell-shocked investors who think brighter days are here again for European banks should be more wary.
Has European Bank Health Really Improved?
The ECB is preparing to take over supervision of the majority of the euro-area banking system from local regulators in 2014 and perform a related "comprehensive assessment" of the banks it will supervise. We think the ECB will find most of the euro-area banks that we cover to be in fairly good health. European banks have been working diligently to prepare for the tests, with several announcing major balance sheet reductions or capital raisings in the back half of 2013. By most measures, European banks are much less risky than they were even a few years ago--Greek sovereign debt has been written off, peripheral European sovereign yields have fallen, more stable sources of funding have been found, and, most notably, regulatory capital ratios have improved materially. The average core Tier 1 ratio of the 22 European banks that we cover was an impressive 13.5% as of June 30, up 250 basis points from year-end 2010 despite the significant headwinds created by changes in how regulatory capital is measured. At the same time, the economic outlook for Europe has brightened. The United Kingdom's economic growth is accelerating, and the euro area is expected to emerge from recession in 2014, according to November figures from the Organisation for Economic Co-operation and Development. This improving outlook has not gone unnoticed by investors; a recent study by the Financial Times found that the number of European bank shares owned by U.S. investment funds rose 10% between June and November, helping to reverse the funds' sharp pullback from the sector in 2011.
Impressively, all of the European banks that we cover have achieved, or are close to achieving, a 10% core Tier 1 ratio on a fully loaded Basel III basis, a much tougher standard and one to which regulators are not scheduled to hold banks until 2019.
While the short-term outlook for European banks looks increasingly benign, we think their overleveraged capital structures continue to pose material risks for long-term investors. Although we now expect that the ECB comprehensive assessment will center on regulatory capital ratios, rather than on the balanced consideration of weighted and unweighted capital ratios that we had once hoped for, an increasing number of regulators and scholars are voicing concerns about the dangers of the current reliance on risk weightings.
In our opinion, leverage will continue to make its way to the forefront of European bank regulation and, in doing so, will negatively affect the returns of the most leveraged banks. Long-term investors in these banks are likely to see their returns pressured by dilutive capital raises and/or delays in returning capital to shareholders as banks are forced to retain funds instead. At a more basic level, the likely higher future capital requirements mean that current earnings, which leverage boosts, will prove to be an overly optimistic indicator of future returns.
We have been calling investors' attention to this issue since 2011, most recently in our spring 2013 report, "European Banks: Should Investors Continue to Dance While the Music is Playing?" We caution investors that European banks' unweighted capital ratios have not improved nearly as much as headline regulatory capital levels would lead one to believe.
Regulators' Focus on Risk-Weighted Ratios Is Misguided
We think European regulators' focus on risk-weighted ratios rather than unweighted ratios is dangerous for several reasons. The lack of correlation between risk-weighted capital ratios and unweighted leverage ratios is not as unusual as one might imagine. Rather, risk-based capital measures and leverage have been entirely uncorrelated for the past two decades. We find it particularly noteworthy that average risk weightings continued to fall between 2005 and 2008, as leverage was spiking.
We are not altogether opposed to risk-based capital measures, unlike many scholars who argue that such measures offer too little benefit for the large and growing cost involved in generating them. We merely argue that unweighted leverage measures should be used in tandem with risk-weighted measures; we think current regulatory requirements give too much weight to risk-based capital measures. However, as bank executives argue fervently and unrelentingly against leverage ratios and for risk weightings, it is important for investors to understand the failings of risk-based measures.
We note first that risk weightings are not measuring risks that capital needs to address. Risk weightings measure known unknowns. They are inherently backward-looking and measure the historical volatility of assets -- the known potential losses. Capital, however, is necessary to guard against unknown unknowns -- losses that have not been foreseen and are not in the samples used to calculate risk weightings.
Risk-Based Measures Have Become Too Complex and Are Measured Inconsistently
We argue that risk-based measures have become too complex to be fully understood by any market participant, whether a banker, regulator, or investor. The original Basel agreement, Basel I, was only 30 pages long. The Basel II accord, which was agreed upon in 2004, was 347 pages long. Basel III, which is non-salutatory and must be implemented separately by each participating country, runs 509 pages and includes 78 calculus equations. European directives aimed at implementing the accord run more than 2,000 pages and are likely to cover tens of thousands more as rules are finalized. As Bank of England executive director Andrew Haldane has pointed out, a single representative large bank now uses more than 200 million calculations to measure its capital.
While executives may imply that regulators are too lazy to check the calculations, we argue that doing so is impossible. Moreover, small modeling errors can result in material misstatements, as J.P. Morgan confirmed in 2012 when a modeling error undervalued the bank's value at risk (a measure of potential losses) by 50%.
As a consequence of this complexity, risk weightings have become impossible for investors or analysts to double-check or compare across banks. Also, risk weightings of the same asset can vary considerably across institutions.
The risk density (risk-weighted assets/accounting assets) of European banks varies significantly across institutions, but much of this can be explained by differences in business model. However, as bank risk-weighting calculations are opaque and complex, exactly how much is due to structural factors and how much is due to modeling differences is difficult to ascertain.
As a way to help answer this question, the Bank for International Settlements recently asked a sample of 15 global banks to run a hypothetical model portfolio through their risk-rating systems. The BIS found significant variation in outputs, some of which it attributed to local supervisory decisions (such as a restriction on modeling options). However, the bank also noted that modeling choices are an important source of variation and found a standard deviation of 31% for the implied capital requirement for the test portfolio. In the same report, the BIS measured banks' actual risk weights to average industry risk weights across asset classes and found considerable variability.
Which Ratios Should Investors Focus On?
Since the financial crisis, which caught many regulators off guard, banks have been under pressure to increase their capital and reduce their exposure to toxic assets. One of the key changes this has brought has been the gradual transition from Basel II to Basel III capital rules, which is increasing capital requirements through changes both in the way in which capital is calculated (the numerator in the equation) and in the way that risk-weighted assets are calculated (the denominator). For example, under Basel III, most hybrid debt, deferred tax assets, and minority investments without control must be deducted from capital. At the same time, many assets like derivatives carry greater risk weightings. Altogether, we estimate that these changes knock 100-200 basis points off the core Tier 1 capital ratios of the European banks that we cover, with banks with larger investment banking operations experiencing decreases near the higher end of the range. The Basel committee has also proposed a minimum 3.0% unweighted leverage ratio, which nets derivatives (which U.S. GAAP does but IFRS--the accounting standard of most European banks -- does not) and takes into account some off-balance-sheet items. In the U.K., regulators have called upon banks to meet this standard by year-end 2013. While we appreciate the gesture, we believe 3.0% is far too low of a bar.
We have long argued that risk-weighted capital ratios need to be paired with unweighted leverage ratios in order for investors, and other market participants, to have a full view of a bank's capital and thereby its ability to withstand unexpected shocks. Our preferred measure is the ratio of tangible common equity/tangible assets, which is a measure of a bank's book value, excluding its intangible assets, divided by its total assets excluding intangible assets. For European banks, which typically report under IFRS standards, we net the banks' derivative positions, which creates ratios more comparable with those calculated for banks that report under U.S. GAAP.
We argue that banks should maintain TCE levels of at least 5%–7%, with some variation allowed for business models. Predominantly retail banks could get by at the lower end of this range, while banks with large investment banks, which have much more variable results and whose risks are more opaque to investors, should tend toward the higher end. This is in line with the recent U.S. leverage ratio proposals for the largest banks, which would require a minimum 5% level at bank holding companies and 6% at their bank subsidiaries. The U.S. proposed leverage calculation includes off-balance-sheet items in the denominator, which makes the standard even tougher than it would first appear. It is also in line with OECD recommendations and with International Monetary Fund estimates of losses during the most recent financial crisis. IMF research shows that between 2007 and 2010, U.S. bank losses and write-downs were approximately 7% of assets and euro area bank losses were well over 5%. While these estimates are prone to considerable uncertainty, the lower loss rate of euro area banks is probably related to, at least in part, differences in accounting rather than differences in realized losses.
We consider 5%-7% to be a conservative but reasonable expectation and note that scholars have suggested that significantly higher levels of capital would be necessary to remove implicit "too big to fail" subsidies. The IMF has noted that U.S. and U.K. bank leverage ratios were 10%-12% in the early 20th century and around 6.5% in the 1950s to the 1970s.
However, European banks have made much less progress on leverage than on regulatory capital since the crisis, and only 6 of the 22 European banks that we cover had a ratio of common tangible equity/tangible assets of more than 5% as we calculate it.
European Banks Remain Highly Leveraged
Source: Company filings, Morningstar.
This looks even worse with a stricter definition of capital, which deducts assets that may disappear during a systemic crisis. The following chart shows TCE ratios after deducting deferred tax assets from capital (deferred tax assets can only be realized by profitable companies) and applying a 20% haircut to Level 3 assets (assets whose value cannot be determined with observable measures).
Common Tangible Equity Looks Even Worse Under a Stricter Definition
Source: Company filings, Morningstar.
Capital Will Be a Major Driver of Long-Term Shareholder Returns
Influential regulators and academics are paying increasing attention to the leverage of European banks. Given our concerns about the health of the European banks, we are concerned that undercapitalized banks may be forced to hold significantly more capital, pushing long-term returns on equity lower. The upcoming ECB comprehensive assessment is unlikely to be catalyst for change, in our opinion -- recent statements suggest to us that it will be designed so that most banks pass, as Europe's stress tests have been in the past. Notably, until the Single Resolution Mechanism (cross-country bailout fund) is finalized, it would be difficult to "fail" banks in peripheral Europe, as these banks would probably have to rely on deeply indebted sovereigns for funding, which could trigger another default crisis. Despite this, we think the calls for reduced leverage are growing stronger and will not go unheeded. In fact, we worry that if the comprehensive assessment is viewed as an easy test by markets, as it probably will be, it may trigger a renewed round of doubts about the strength of Europe's banks.
Although the near-term outlook for European banks has improved, we advise long-term investors to continue to exercise caution. The recent rally in share prices has been fairly indiscriminate across companies and provides an attractive exit point for more troubled banks, or an opportunity to trade into higher-quality names. We recommend that investors shift into well-capitalized, moaty banks such as HSBC (NYSE:HSBC) and Standard Chartered (OTC:SCDRF), both of which are trading at a discount to our fair value estimates. We also increasingly like Lloyds Banking Group (NYSE:LYG) and UBS (NYSE:UBS), both of which are trading at a discount to our fair value estimates and have been strengthening their moats by focusing on their moaty core businesses and cutting risk and scale in other divisions. Investors seeking a greater risk/reward profile could look to Royal Bank of Scotland (NYSE:RBS) and BNP Paribas (OTCQX:BNPQF), which are trading at sizable discounts to book value despite above-average capital levels. Both have very attractive core retail banking businesses and have been reducing their exposure to less attractive markets and products.