Tuesday, May 12, 2009

The Credit Cycle - Banks Positioned to Outperform

Scotia Capital, 11 May 2009

• Canadian banks continue to be impacted by the credit cycle as loan loss provisions rise and fall with the economic tide. The major question is how vulnerable are Canadian banks to the current credit cycle from an earnings and capital perspective and how are bank stocks likely to perform on an absolute and relative basis. In this report, we attempt to compare and contrast the current credit cycle with past cycles, taking into account macroeconomic factors and micro-factors such as loan mix and concentration.

• However, before looking in detail at the current credit cycle, we attempt to put the Canadian banking industry in an historical context. Canada has been known, we believe, from a global perspective as a small and concentrated banking market with undue loan concentration, especially from a corporate lending perspective. The one blemish in Canada’s now recognized enviable banking system has been undue loan concentration, which has historically overshadowed the underlying strength of the bank system.

• Canada has traditionally had significantly higher exposure to high-risk assets/loans than its global peers, with the exception of the current cycle. As a multiple of common equity, Canadian banks’ exposure to lesser developed countries (LDC) loans was 10.9x the exposure level of U.S. banks in 1982, with commercial real estate (CRE) exposure being 2.5x its U.S. counterparts in the early 1990s and Telco exposure 6.7x larger. The high exposure to high-risk assets was further compounded by single name concentration within these high-risk sectors, especially CRE in the early 1990s. Canadian banks had single name exposure as high as 35% of common equity in 1985 and 20% in 1992 with single name exposure now estimated at only 2% of common equity, with very few exceptions.

• The Canadian banking system’s historical leverage to credit and concentration can be reflected by the fact that the Canadian banking system absorbed $40 billion in loan loss provisions from 1984 to 1994 on a common equity base of only $14 billion. The 1984 to 1994 period was the darkest days for credit in Canadian banking, which included massive LDC provisions in the 1980s, particularly 1987, and the equally devastating commercial real estate provisions of the early 1990s.

• Interestingly, despite major credit blowups, particularly from the 1984 to 1994 period, Canadian banks have managed to grow their earnings and dividends by an impressive 10% CAGR over the past 50 years, which speaks to the core strength of the banking system. Bank share prices have also substantially outperformed over the long term.

• The Canadian banking system was transformed with the acquisition of the major investment dealers in the late 1980s (no Glass-Steagall Act) and subsequent acquisitions of the major trust companies in the early 1990s, which assisted in the significant reduction in corporate credit dependency. Thus the evolution of our current strong banking system with a diversified revenue mix, balanced operating platforms, and sound effective regulation.

• This cycle, Canada has less than half the exposure to high-risk assets such as sub-prime and structured products versus their global peers, which is a significant reversal of fortunes. As a consequence of the relatively low exposure to high risk-assets this cycle, Canadian banks have comfortably maintained their dividend levels versus global banks that have cut or essentially eliminated their dividends to survive, in addition to accepting government aid.

• The strength of the Canadian system was tested in 2008 with the market crisis as structured products and trading securities values plummeted and mark-to-market losses skyrocketed globally. Canadian banks’ relatively low exposure to toxic securities allowed the banks to report a fully loaded return on equity of 12% in 2008 after all writedowns, and allowed banks to grow their book values by double digits and pay their dividends, which reached yields never seen before versus both government and corporate bond yields.

• The Canadian banks’ operating performance in 2008 resulted in the banks outperforming the TSX, a rarity in global markets, although this is of little consolation given the poor absolute returns driven by global valuation contagion. Canadian banks are also comfortably outperforming the TSX thus far in 2009.

• The current credit cycle began in earnest in late 2008 and early 2009, and the major focus now is what level will loan losses peak at and the banks’ ability to absorb these losses from an earnings and capital perspective, and bank share price performance through the credit cycle. The key issues in deriving peak loan losses are what will be the severity and duration of the economic slowdown and the risks and mix of the banks’ loan portfolios.

• In analyzing the credit cycle, we look at both macroeconomic variables and micro-factors such as loan mix, sector and name concentration, and risk management including credit or underwriting standards. We also compare and contrast the current credit cycle with the past three cycles, particularly the 1983 and 1992 cycles, as the 2002 cycle was probably the only one in history not initiated by a decline in the economy as measured by two consecutive quarters of negative GDP.

• The macroeconomic variables we focus on are the unemployment rate, debt levels for households and non-financial corporations, pre-tax corporate profits, housing affordability, equity in real estate, house price declines, as well as depth and breadth of the economic retreat (GDP) and Canada’s fiscal position. The macroeconomic variables are used in conjunction with historical loss ratios, impaired loan levels, and micro-factors such as loan mix and concentration to derive expected loss ratio by loan type bank by bank in order to compute an overall expected peak loss ratio.

• We believe the positive impact of Canadian banks’ lower loan concentration and lower leverage to credit will be evident this credit cycle. We believe that banks’ leverage to credit has declined successively over the past 20 or 30 years and, combined with lower loan concentration and higher underwriting standards in Corporate, should more than offset higher consumer loan losses and perhaps the risk of a more severe recession than 1982 and 1991. Canadian banks, we believe, are positioned to outperform through this credit cycle.

• Our analysis has loan loss provisions peaking this cycle in 2010 at $10 billion, or 76 basis points (bp) of loans (82 bp on Q1/09 loan balances), for return on equity of 16%. We believe that fiscal 2008 ROE of 12% (14% excluding CIBC) fully loaded after large mark-to-market writedowns will be the bottom in profitability this cycle. The 12% ROE included $13 billion in mark-to-market losses and $5 billion in credit losses. Thus, banks have an $18 billion or so cushion to maintain a 12% ROE. So with peak loan losses expected at $10 billion, we expect positive ROE momentum off the 2008 level. If unemployment were to spike into uncharted territory in the 14%-16% range, loan losses could increase to the extremes of $17 billion, or 135 bp level, for a modest return on equity of 11%.

• The three previous loan loss provision peaks were 1983 at 100 bp, 1992 at 161 bp (80 bp excluding CRE/O&Y), and 2002 at 106 bp (87 bp excluding TD’s sectoral), with trough LLPs in the 22-25 bp range in 1988, 1997, and 2006.

• The credit cycle has tended to be five years from trough to peak and another five years from peak to trough, resulting in loan losses peaking approximately every 10 years. Loan loss provisions also typically peak one year after the bottoming of the economy (negative GDP).

• Historical patterns would place the next peak in loan losses in the 2010-2012 time period. The fallout from the dramatic implosion of Wall Street (sub-prime/structured products/CDS) has no doubt accelerated the global economic decline.

• Our $10 billion expected peak is based on retail loan losses (personal loans and credit cards) of $4.4 billion-$6.1 billion representing 137-191 bp of loans, significantly higher than past cycle peaks of approximately 140 bp. We expect corporate and commercial loan losses to be in the $3.3 billion-$5.6 billion range, or 65-111 bp below previous peaks of 200 bp due to lower corporate debt levels, higher credit standards, and lower industry and single name concentration. We expect residential mortgage loss ratios to increase to the 11-16 bp range versus the previous peak of 7 bp in the early 1990s.

• The major macroeconomic factors that should result in overall loan losses being contained well within the context of historical provision levels are expected unemployment rates, non-financial corporate debt levels, corporate pre-tax profit decline rate, and equity levels in household real estate.

• Our peak loan loss ratio for this cycle, we believe, is consistent with past peaks’ loss ratio factoring in loan mix, loan concentration, underwriting standards, and macro-factors. Our loss ratio of 76 bp for 2010 compares to 100 bp in 1983, 80 bp in 1992 (excluding CRE/O&Y), and 87 bp in 2002 (excluding TD’s sectoral). In terms of loan mix, residential mortgages now represents 35% of the loan portfolio versus 29% in 1992 and 13% in 1983, and given the very low loss ratio on these loans, it lowers the bank overall loss ratio substantially. Thus, excluding residential mortgages, our loss ratio peak this cycle is estimated at 125 bp versus 114 bp in 1992 (excluding CRE/O&Y concentration) and 114 bp in 1983. In addition, retail loans have grown to 60% of total loans versus 23% in 1982, thus much lower leverage to corporate lending. Also, on the corporate loan loss side, we see the lower sector concentration, lower single name exposure, lower corporate debt levels, and lower decline in pre-tax corporate profits versus past cycles as supportive of lower loss ratios from corporate and commercial.

• Bank pre-tax, pre-provision earnings are estimated in the $36 billion range and have increased from the $3 billion level in 1982, representing an impressive CAGR of 9% (exhibit 4). A $10 billion peak in loan losses would equate to a return on equity of 15% to 16%. Loan loss provisions of $10 billion are estimated to peak at 25% of pre-tax, pre-provision earnings and 11% of revenue and 9% of common equity, significantly below previous cycles. In terms of absorbability of loan losses, the pre-tax pre-provision earnings would allow the banks to take a charge of 340 bp (excluding insured mortgages), which we believe is three to four times higher than past peaks and the current plausible peak provision levels.

• We conclude that Canadian banks are in the best shape in 50 years to absorb credit losses based on broad fundamentals. Bank profitability (exhibits 2, 3) on a return on equity or return on risk-weighted assets basis is near historical highs, and the revenue base (exhibits 6) is more diversified. In addition, leverage to credit has declined materially and the loan portfolio is lower risk with less sector and single name concentration. Capital levels are at the highest levels in history despite the market hysteria and panic and stampede to deleverage.

• We also conclude that bank shares can outperform during the credit cycle and that the banking system has never been this well positioned on an absolute basis and relative basis to U.S. and U.K. banks. Bank stocks have outperformed in the past three credit cycles, managing to record positive absolute and relative returns in the face of rising credit losses, which is counterintuitive and contrary to popular belief. The current credit cycle is thus far the exception, which we attribute to valuation contagion mainly from the banking crisis in the U.S. and U.K. and the fact that the cycle is not over. Interestingly, bank stocks’ major underperformance is typically not credit cycle related but commodity/asset bubble related. The three years where banks recorded the greatest underperformance, and by a wide margin, are 1979 (commodities), 1999 (Nortel/Tech), and 2007 (commodities).

• In summary, we expect banks to outperform over the next several years despite the credit cycle, recovering some of its underperformance from the commodity spike in 2007 and valuation contagion of 2008. In the absence of valuation contagion, we would have expected bank P/E multiples to have bottomed at 9.0x, similar to the 1998 Asia crisis as opposed to 6.5x at the March 9, 2009, bottom. We expect bank P/E multiple expansion through 2012 similar to that experienced post the 2002 credit cycle. We expect bank P/E multiples to expand back to 14x in the next few years and eventually 16x. The sustainability of bank dividends and the resumption of superior dividend growth will be a catalyst for significantly higher bank share prices. We remain overweight the bank group based on strong fundamentals and depressed valuations.

Increasing Bank Target Prices - Remain Overweight - RY Top Pick

• We are increasing our target prices for the Canadian banks by 18% (exhibit 26) based on a
higher target P/E multiple of 12.7x versus our previous 10.7x.

• We expect bank P/E multiples to continue to recover as the market focuses on fundamentals as opposed to fear and hysteria and valuation contagion is expected to be less of a factor as investors refocus on earnings power and P/E multiples as opposed to capital, solvency, and MV/BV ratios.

• We expect significant P/E multiple expansion over the next few years.

• We remain overweight the bank group with RY (1-Sector Outperform) remaining our top pick as we believe that the bank is the best positioned for growth.