23 January 2006

Scotia Capital Research on Cdn Insurance

Q4 2005 Earnings Preview

• Despite stretched valuations, the sector’s defensive characteristics provide support in uncertain markets – we continue to recommend market weight. As we approach the Q4/05 earnings season, we find the group’s valuation relative to the market at four-year highs, and, perhaps more importantly, stretched versus other financials, namely Canadian banks and U.S. lifecos. Simply said, financials appear expensive relative to the market, and Canadian lifecos appear to be a little expensive relative to other financials. However, in our opinion, the Canadian lifeco sector’s defensive characteristics, specifically a consistent track record of negligible earnings surprises and significant excess capital positions that allow for share buybacks and dividend increases, do warrant a valuation premium, especially in these uncertain markets. Thus, despite the stretched valuations, we maintain our market weight recommendation.

• Multiple relative to market continues to expand – unlikely to participate in any market rally, but excellent defensive characteristics. The Canadian lifecos’ forward multiple, currently at 13.6x, has climbed 8% over the last three months and has increased nearly 10% over the last 18 months, despite a 5% reduction in forward multiples for both the S&P 500 and the S&P/TSX over the same time. Relative to the S&P 500, the Canadian lifeco forward multiple is now 85% of the S&P 500 forward multiple (up from 83% three months ago), or nearly two standard deviations above its historical 68% mean, the highest premium relative to the market in the last three years. With this sort of valuation premium, we would not expect the Canadian lifecos to significantly participate in a market rally, just as premium valuations were essentially the reason the Canadian lifecos did not broadly participate in the market rally of the fourth quarter of 2004. However, we believe the defensive qualities of the Canadian lifeco group, namely reasonable earnings visibility (we expect 12% EPS growth in 2005 and 11% growth in 2006 and 2007) and exceptional excess capital positions (both to buy back stock and increase dividends), will provide support in uncertain markets. As such, we continue to recommend market weight in the sector, despite the stretched valuations.

• Valuation relative to U.S. lifecos continues to expand – makes sector look expensive relative to U.S. peers but also makes U.S. acquisitions more appealing. The premium to the U.S. lifecos (on a forward P/E basis) has modestly increased over the last three months to 9% from 8%, and Canadian lifecos still trade above the three-year mean of a 5% premium. When you combine these valuation premiums to the U.S. lifecos with increasing levels of excess capital, and a more favourable Canadian currency, we have to believe the Canadian lifecos will look to the highly fragmented U.S. market to make what we consider to be reasonably accretive acquisitions. We expect acquisition activity to increase in 2006, as we believe the big players will continue to take advantage of low debt financing rates to acquire, and sub-scale players will look to rationalize or specialize.

• Valuation relative to Canadian banks – 5% premium versus 2% three-year mean – suggests banks may be slightly more attractive. The Canadian lifecos have modestly outperfomed the Canadian banks over the last three months (by about 3%), as the Canadian lifeco premium to the banks climbed from 1% (below the 2% mean) to 5%. Three months prior to that, the banks outperformed the lifecos by 4%, as the lifeco premium fell from a 7% premium to a 1% premium. We believe a modest premium over historical levels would be justified, assuming long-term interest rates rise and the yield curve remains somewhat flat, which is traditionally an environment less punitive to lifecos than to banks. However, given the uncertainty over the likelihood of a long-term rate rising scenario actually happening (we note consensus has been wrong on this scenario for the past two years), the modest premium is by far not a given. A scenario whereby long-term rates continue to decline will be more punitive to the lifecos than the banks, in our opinion. Given this uncertainty, we believe the current valuation premium may not be entirely justified.

• Fundamentals remain steady – but we point to a couple of items that may cause some concern. Despite the stretched valuations relative to the market, fundamentals remain steady for the group, with ROEs modestly climbing, excess capital positions growing, and targeted dividend payout ratios rising.

• Low long-term interest rates and a flattening yield curve – Industrial-Alliance the most at risk. Long-term rates continue to slide, especially in Canada. With liabilities longer than assets, a declining long-term interest rate scenario can pose significant reinvestment risk to the lifecos. While the reinvestment risk is by far the largest risk in a declining long-term and flattening yield curve environment, a flattening yield curve not only increases the likelihood of increased surrenders on savings-type products (along with associated disintermediation risks), but may also diminish the competitive advantage that the longer investing lifecos have over the banks in attracting sales of short-term asset accumulation-type products. We estimate that a 100 basis point (bp) parallel decline in the yield curve over 2004 year-end levels (and deemed to be permanent for the purposes of reserving) results in a 6% decline in earnings for Sun Life, a 10% decline in earnings for Great-West Lifeco, a 14% decline in earnings for Manulife, and almost a 30% decline in earnings for Industrial-Alliance, all other things equal. Canadian long-term rates have declined 80 bp over 2004 year-end levels, whereas U.S. long-term rates have declined only 20 bp over 2004 year-end levels. Interest rate risk for Great-West Lifeco, Sun Life, and Manulife is much more U.S. related than Canadian related, whereas it is 100% Canadian related for Industrial-Alliance. As such, we remain cautious on Industrial-Alliance as long-term Canadas continue to fall.

• Credit cycle takes a turn for the worse – Manulife the most at risk. Times are good now, but if credit spreads were to widen and the incidence of bond defaults increases, Canadian lifecos’ earnings could suffer. Manulife has the highest exposure of below-investment grade bonds for the group (6% of Manulife’s bond portfolio is below investment grade, as opposed to 3% for Sun Life, 1% for Great-West Lifeco, and 0% for Industrial-Alliance). In terms of exposure to GM and Ford bonds, Sun Life is at $0.80 per share and Great-West Lifeco is at $0.27 per share (as at Q4/04, the latest data we have available), with Manulife at $0.56 per share (as at Q1/05, the latest available). Manulife claims that 95% of its exposure to Ford and GM bonds is secured.

• Equity markets slowing – Sun Life the most at risk. 2005 saw average levels for the S&P 500 up 8% (in line with company and our expectations), and average levels for the TSX up 18% (above expectations). A sizeable portion of the equity market-related earnings for the lifecos is derived from U.S. equity markets (especially in the case of Sun Life with MFS), and if markets were to slide, we could see downward EPS adjustments for 2006. For the purposes of pricing and reserving for segregated fund guarantees, we believe that most lifecos assume 7%-8% annual appreciation in equity markets. While recent trends in equity markets are encouraging, we estimate that if markets declined 10% from current levels over the next two months and then remained flat throughout the remainder of 2006, we could expect EPS estimates for 2006 to decline by 2% in the case of Great-West Lifeco, 3% in the case of Manulife, and 4% in the case of Sun Life, assuming all else remains the same.

• Strengthening Canadian dollar versus the U.S. dollar – Manulife the most at risk. Our EPS estimates assume an average exchange rate of $1.16 (C$/US$) for 2006, with no currency hedging for Manulife and Sun Life and a currency hedge for Great-West Lifeco in line with our average rate. Each 5% movement in our estimate is worth about $0.05 per share for Great-West Lifeco (or 2% of 2006E EPS), $0.17 per share for Manulife (or 4% of 2006E EPS), and $0.06 per share for Sun Life (or 2% of 2006E EPS). As Manulife’s U.S. owners (nearly 50% of the shareholder base) would obviously see their stock benefit from an appreciating Canadian dollar, we believe any perceived impact on Manulife’s share price due to an appreciating Canadian dollar may not be so dramatic.

Canadian P&C insurers – Canadian auto continues to pace ahead of industry norms, but we do not expect impact of hurricanes to result in significant increase in Canadian commercial insurance pricing.

• The profitability of Canadian auto insurance continues to pace well ahead of industry norms due to the sustained effectiveness of automobile reforms and continued low frequency levels. We get the impression from management at ING Canada that this trend will continue through 2006. The U.S. non-standard auto market (a significant portion of Kingsway’s business) remains very competitive and perhaps somewhat irrational, as niche players have gained market share. It remains to be seen as to whether the impact of more expensive reinsurance (in light of the hurricanes) will introduce an element of rationality to the market, an obvious positive for Kingsway. However, should this market become more rational, we would expect large traditional players (State Farm, Geico, Progressive) to no longer “hold back the reins” and re-enter the non-standard market as well, thus increasing competition.

• With the 2005 hurricanes removing US$40 billion to US$60 billion from the balance sheets of insurers worldwide, market conditions for commercial players should improve in 2006, especially for those companies writing U.S. coastal coverages. We are somewhat cautious as to the hurricane impact for our players, especially lower-rated direct commercial players (such as Northbridge) with very limited U.S. coastal business. We believe the magnitude of price increases in the January renewal season is up for debate, and we believe the higher leverage to any significant improving pricing trends lies with the reinsurers and the highrated direct writers. We will continue to revisit our view as January renewals are finalized.

Great-West Lifeco Inc.
2-Sector Perform – $33 one-year target, based on 2.9x 12/31/06E BV and 13.1x 07E EPS
• We are in line with consensus for Q4/05 and $0.03 per share below consensus for 2006. We do not expect a significantly beneficial currency hedge for 2006
• European segment (24% of bottom line), up 27% year-to-date 2005, should continue to show double-digit growth – with further support in 2006 from the recently announced acquisition in the payout annuity market in Britain (closed Q4/05).
• We look for good cost control and good sales growth in Canada (45% of bottom line).
• Possible 8%-10% dividend increase.

Industrial-Alliance Insurance and Financial Services Inc.
2-Sector Perform – $33 one-year target, based on 1.75x 12/31/06E BV and 11.8x 07E EPS
• We are in line with consensus for Q4/05 and $0.04 per share below consensus for 2006.
• Expect to see a modest rebound in individual insurance sales.
• We will carefully monitor interest rate risk.

Manulife Financial Inc.
1-Sector Outperform – $76 one-year target, based on 2.3x 12/31/06E BV and 13.8x 07E EPS
• We are $0.02 per share above consensus for Q4/05 and $0.01 per share above consensus for 2006.
• John Hancock integration is essentially done – U.S. protection segment and U.S. annuity segments should continue to build on recent sales momentum.
• We will pay close attention to the credit risk profile, but expect the improvement in quality we saw in Q3/05 to continue.
• Possible 10%-15% dividend increase.

Sun Life Financial Inc.
2-Sector Perform – $50 one-year target, based on 1.85x 12/31/06E BV and 12.7x 07E EPS
• We are in line with consensus for Q4/05 and $0.06 per share below consensus for 2006.
• A “show me” story – we will closely monitor much-needed progress in organic sales growth.
• Spread improvement in U.S. fixed annuity block should continue in Q4/05 as the company’s “mismatch” assumption is proving to be profitable.
• Possible 8% dividend increase.

Fairfax Financial Holdings Limited
2-Sector Perform – US$184 one-year target, based on 1.0x 12/31/06E BV
• Hurricane Wilma wipes out any gains in the quarter.
• Lots of “noise” surrounding subpoenas but no new news.
• Runoff segment remains under the radar screen.

ING Canada Inc.
2-Sector Perform – $56 one-year target, based on 2.3x 12/31/06E BV
• A good chance the company could exceed consensus with another excellent quarter of underwriting profitability – we forecast a combined ratio of 87%, not as good as the exceptionally strong 83% in Q3/05.

Kingsway Financial Services Inc.
2-Sector Perform – $27 one-year target, based on 1.4x 12/31/06E BV
• We look for a steady quarter in line with consensus, with no significant prior period reserve development and a combined ratio in the 97% range.
• We look for an update regarding the impact of the company’s recent agreement with Robert Plan (a non-standard business partner focusing on New York and New Jersey).
• Last quarter the company reports in Canadian dollars; going forward the company will report in U.S. dollars.

Northbridge Financial Corporation
3-Sector Underperform – $35 one-year target, based on 1.5x 12/31/06E BV
• Hurricane Wilma to negatively impact the quarter by $0.24 per share.
• While underwriting profitability should remain better than the historical average, revenue is expected to continue to decline. • Will the post-hurricane environment be significantly more favourable? We remain sceptical.