EQUITY RESEARCH. A profound accounting change is coming but you might not notice it (at first)

Transcription

1 EQUITY RESEARCH November 13, 2017 Canadian Banks RBC Dominion Securities Inc. Darko Mihelic, CFA (Analyst) Sean McQuade, CFA (416) (Associate) (416) Vanessa Wan, CA, CFA, CPA (Senior Associate) (416) A profound accounting change is coming but you might not notice it (at first) We believe IFRS 9 is one of the most profound accounting changes the Canadian banks will have faced in a very long time. The first communication of impact will be Q4/17. This note focuses on the mechanics of IFRS 9 and each bank's starting position. IFRS 9 starts in Q1/18 but the large Canadian banks will provide IFRS 9 opening balance sheet and capital impact guidance during Q4/17 reporting. Here are some things we believe investors should think about. 1. Great timing - small impact. The timing of adopting IFRS 9 is relatively ideal and thus, we do not expect IFRS 9 opening balance sheet impacts on book value and capital to be overly material for the large Canadian banks. Nevertheless, investors should be aware that some banks could record a larger reserve increase and thus, the accounting change may hit book value per share for some banks and perhaps not at all for others. We do not expect a significant hit (perhaps no hit at all) to common equity tier 1 (CET1) capital as the capital regime in Canada adjusts CET1 capital for greater conservatism in allowance for credit loss (ACL) reserves. 2. Which bank might be more or less conservative with existing ACL reserves? IFRS 9 incorporates significant judgement and forward-looking information when setting allowances. In this report, we offer up an initial view on existing ACL conservatism by comparing reserves against recent actual losses and against regulatory requirements. For our coverage universe; we offer the following opinion, on a spectrum from most conservative to least conservative; NA, BMO, BNS, TD and lastly CM. 3. We expect opening balance sheet impacts on book value and capital to vary across the large Canadian banks; however, we believe investors should focus more on where each bank has "positioned" its ACL reserves. Each of the large Canadian banks currently has very different ACLs and each bank will record different opening balance sheet impacts. Differences amongst the Canadian banks will likely continue after the adoption of IFRS 9. However, in our view, each bank's starting point may be a key determinant of how provisions for credit losses (PCL) will behave under IFRS 9 - theoretically (but not necessarily always the case in practice) banks with more conservative ACL reserves may experience less volatile PCLs when the credit cycle turns (and vice versa). 4. Post-Q4/17, we expect a period of "discovery" in which analysts and investors will rethink how they view PCLs and alternative measures and views of PCLs will likely surface - although we believe PCLs remain a fundamental measure of risk. There could be views gravitating towards net charge-offs or perhaps an emphasis on pre-tax pre-provision earnings; however, we believe PCLs under current accounting standards (and even more so under IFRS 9) remain as leading indicators of when credit quality turns for a bank. 5. We have been "caveating" our PCL forecasts for some time and reiterate our concern that our PCL forecasts are probably inaccurate as the new IFRS 9 regime kicks in; however, we also believe our PCL forecasts are probably conservative. Our average PCL forecast for 2018 is 38 bps - this would compare to an estimated 30 bps in 2017 and 38 bps in Under IFRS 9, we anticipate our PCL forecasts to become more volatile and far more "linked" to economic parameters than in the past - perhaps valuations will follow suit (a topic for another day). Disseminated: Nov 13, :30ET; Produced: Nov 13, :00ET Priced as of prior trading day's market close, EST (unless otherwise noted). All values in CAD unless otherwise noted. For Required Non-U.S. Analyst and Conflicts Disclosures, see page 24.

2 Canadian Banks Exploring IFRS 9 opening balance sheet and capital impacts The large Canadian banks will provide IFRS 9 opening balance sheet and capital impacts during Q4/17 reporting this report explores these day 1 impacts in further detail. This report does not intend to quantify opening balance sheet and provision for credit loss (PCL) impacts. In our view, it is not possible to accurately quantify these impacts at this time given the significant level of judgement that can be applied by each bank and the incorporation of forward-looking information. However, we do review the mechanics of IFRS 9, the Canadian bank capital regime, and point out some interesting early observations. For a more pure educational review of IFRS 9 concepts, please refer to our previous note Preparing for IFRS 9, highlights from Canadian banks' educational event. It is a relatively ideal time to adopt IFRS 9 The current benign credit environment and improving economic outlook makes it a relatively ideal time to adopt IFRS 9, in our view. In a stable economic environment, IFRS 9 generally introduces higher allowance for credit loss (ACL) reserves as it requires 12-month expected credit losses (ECL) to be recognized on loans classified as Stage 1 (i.e. performing loans) and lifetime ECLs to be recognized on loans classified as Stage 2 (i.e. loans that are not impaired but have experienced a significant increase in credit risk). However, IFRS 9 overlays forward-looking information into ACL reserves and thus, we expect an improving economic environment in both Canada and the U.S. to help mute the opening balance sheet impact from adopting IFRS 9. Exhibit 1: RBC Economics is forecasting an improving economy in Canada and the U.S. YoY CHG E 2018E E 2018E RBC Economics Forecasts: Canada: Real GDP Growth 1.50% 3.10% 2.10% 0.60% 1.60% -1.00% Overnight Rate 0.50% 1.00% 1.75% 0.00% 0.50% 0.75% Unemployment 7.00% 6.40% 6.20% 0.10% -0.60% -0.20% United States: Real GDP Growth 1.50% 2.20% 2.50% -1.10% 0.70% 0.30% Fed Funds Rate 0.75% 1.50% 2.50% 0.25% 0.75% 1.00% Unemployment 4.90% 4.40% 4.20% -0.40% -0.50% -0.20% Source: RBC Economics forecasts, RBC Capital Markets November 13,

3 Specific Allowance Collective Allowance Canadian Banks Exhibit 2: Comparison of the allowance for credit loss model under IAS 39 and IFRS 9 IAS 39 IFRS 9 Collective allowance for incurred but not identified losses Stage 1 allowance - non-impaired (performing) loans - 12-month ECL using point-in-time PD, LGD, and EAD parameters - large Canadian banks use a PD x LGD x EAD methodology - allowance is intended to cover losses on loans that are impaired but have yet to been identified as impaired Stage 2 allowance - non-impaired (performing) loans with significant increase in credit risk - lifetime ECL using point-in-time PD, LGD, and EAD parameters - 30 days past due rebuttable assumption incorporates forward looking information Allowance for impaired loans - established upon impairment of loan Stage 3 allowance - impaired loans - lifetime ECL (similar to stage 2; however PD is likely much higher under this stage) - 90 day past due rebuttable assumption Note: Probability of default (PD), loss given default (LGD), and exposure at default (EAD) Source: RBC Capital Markets While we believe it is a relatively ideal time to adopt IFRS 9, lower beginning ACL reserves could result in potentially higher PCL volatility for the large Canadian banks in the future (but not necessarily ). All else equal, the incorporation of forward-looking information under IFRS 9 will result in declining ACLs during periods of improving credit quality. When the economic outlook and underlying credit quality deteriorates, this may result in higher PCL volatility, similar to that observed for U.S. banks. Exhibit 3 provides a comparison of Canadian banks historical PCLs relative to U.S. banks PCLs. Except for when Canadian banks have recorded sectoral provisions in the past (which in effect, meant TD was incorporating forward-looking assumptions into its reserves), U.S. banks PCLs have generally occurred earlier than the Canadian banks. November 13,

4 Exhibit 3: U.S. banks have recorded PCLs earlier in the cycle than the Canadian banks 4.00% Canadian vs. U.S. Banks PCL Ratio 3.50% 3.00% 2.50% 2.00% 1.50% 1.00% 0.50% 0.00% -0.50% TD records sectoral provisions for telecom and other loan portfolios Oil and gas credit deterioration Cdn Banks PCL Ratio U.S. (FDIC Insured) Banks PCL Ratio Source: Company reports, FDIC, Haver, RBC Capital Markets We are expecting muted opening balance sheet and capital impacts OSFI has not explicitly provided transitional relief for the adoption of IFRS 9; however, we are not expecting the adoption of IFRS 9 to have an overly material impact on book value and capital for the large Canadian banks. In August 2017, OSFI released draft capital adequacy guidelines to be adopted for fiscal 2018, which largely reiterated how ACLs are currently treated for regulatory capital purposes. In its cover letter, OSFI requested institutions that projected material IFRS 9 impacts to provide estimated impacts to OSFI (during the consultation period that ended in September 2017) and suggested it would consider appropriate mitigation if any. Three of the large Canadian banks (CM, RY, and TD) have meaningful shortfall deductions that can help mitigate the potential capital impact upon adopting IFRS 9 (Exhibit 4). While the other large Canadian banks (BMO, BNS, and NA) do not have any (or meaningful) shortfall deductions, as we detail later on in this report, these banks may have relatively conservative ACL reserves and in our view, may not necessarily need to increase their ACL reserves significantly (some could possibly reduce their ACL reserves). Each bank will be subject to different IFRS 9 opening balance sheet capital implications, depending on their current provisioning shortfall (or excess) and their opening balance sheet increase (if any) in ACLs. A shortfall (excess) provision occurs for regulatory capital purposes when a bank s accounting ACL is less (more) than regulatory expected losses calculated on advanced internal ratings based (AIRB) exposures. Shortfall provisions are deducted against common equity tier 1 (CET1) capital banks currently with a shortfall deduction have room to mitigate the capital impact upon adopting IFRS 9, as a higher ACL will reduce the bank s current shortfall deduction. Only November 13,

5 provisioning shortfalls on AIRB exposures have a direct impact on CET1 capital, as there is no similar deduction for standardized exposures. Excess provisioning on AIRB exposures and collective provisions allocated to standardized exposures are added to Tier 2 capital (Exhibit 5), up to a limit. It is possible for a bank to have a shortfall provision deducted from CET1 capital but a collective provision recognized in Tier 2 capital as collective allowances allocated between AIRB and standardized approaches are treated differently for regulatory capital purposes. We discuss the mechanics of provisioningrelated capital adjustments in further detail later on in this report. Exhibit 4: Three of the large Canadian banks currently have meaningful shortfall deductions CET1 Ratio and Shortfall Deduction (Q3/17) 0.3% 0.2% 11.2% 11.3% 0.2% 11.2% 10.9% 11.0% 10.4% BMO BNS CM NA RY TD CET1 Ratio Shortfall Deduction Shortfall Deduction (Q3/17, $MM) BMO BNS CM NA RY TD , Source: Company reports, RBC Capital Markets November 13,

6 Exhibit 5: All of the large Canadian banks have collective allowance additions included in their total capital ratios Total Capital Ratio and Collective Allowance Additions to Tier 2 Capital (Q3/17) 15.5% 15.6% 15.2% 0.1% 0.2% 0.4% 0.1% 0.1% 14.8% 0.2% 14.4% 0.1% 15.1% 14.6% 13.7% 0.1% 15.2% 14.3% 15.2% 13.6% BMO BNS CM NA RY TD Total Capital Ratio (ex-collective additions) Excess Provisions (AIRB) Collective on Standardized Q3/17, $MM BMO BNS CM NA RY TD Excess provisions - AIRB Collective allowance - Standardized ,571 Collective recognized in Tier 2 capital ,571 Source: Company reports, RBC Capital Markets In this section, we highlight each of our covered bank s current shortfall deduction (if it exists) and mitigation each bank has against capital impacts from adopting IFRS 9; however, each bank will likely have different changes to their ACL reserves and thus, different resulting capital impacts. We explore each bank s provisioning conservatism in the next section of this report. Currently, of the large Canadian banks that we cover, CM and TD have meaningful provisioning shortfalls deducted from their calculation of CET1 capital, which could help mitigate the majority (if not all) of the possible capital impact from adopting IFRS 9. Alternatively, one can view banks with meaningful shortfall deductions against CET1 capital as already being penalized for having ACLs less conservative than what the regulatory capital regime requires (for AIRB exposures). As a result, banks with meaningful shortfall deductions have some room to increase their ACLs before experiencing a hit to their CET1 capital ratio. Exhibit 6 shows each of the large Canadian bank s maximum potential increase to their ACL reserves prior to affecting their CET1 ratios. Using each of the large Canadian bank s effective tax rate on a taxable equivalent basis, we calculate that the banks currently with a shortfall deduction can absorb the capital impact from a significant potential increase in their ACL reserves upon adoption of IFRS 9. November 13,

7 Exhibit 6: CM and TD s current shortfall deduction can absorb a material increase in ACLs Maximum Increase in ACLs Prior to CET1 Impact Q3/17 $MM (1) As a % of Current ACLs BMO 0 0% BNS 3 0% CM % NA 0 0% TD % (1) Pre-tax figure calculated based on each bank's effective tax rate (teb) in Source: Company reports, RBC Capital Markets estimates While IFRS 9-related capital impacts may be mitigated to some extent by each bank s shortfall (if it exists), there is little mitigation for impacts on book value. An increase to a bank s ACL will result in a reduction in retained earnings. The impact may be mitigated partly by related deferred tax impacts but will likely result in a direct reduction in book value. More specifically, CM and TD could theoretically reduce their book value by approximately 1-2% upon adopting IFRS 9 without affecting their CET1 capital ratio. Assuming IFRS 9 was adopted at the end of Q3/17, CM could reduce its book value per share (BVPS) by approximately $1.00 and TD could reduce its BVPS by approximately $0.40 and could still leave its CET1 capital ratio unaffected. Exhibit 7: CM and TD could reduce its BVPS by 1-2% before affecting its CET1 ratio Maximum Impact on BVPS Prior to CET1 Impact Q3/17 Per Share Current BVPS As a % of Current BVPS BMO $0.00 $ % BNS $0.00 $ % CM -$1.03 $ % NA $0.00 $ % TD -$0.40 $ % Source: Company reports, RBC Capital Markets estimates While banks currently with no shortfall deduction (i.e. BMO and NA) would have a capital impact if their ACLs were to increase, in our view, this could suggest that their current ACL reserves are conservative and it may be possible that these banks could actually reduce their reserves upon adopting IFRS 9. BMO and NA hold higher ACLs than imputed for regulatory capital purposes, while BNS has a small shortfall but also has an ACL slightly higher than that imputed for regulatory capital purposes. We detail how we arrive at a regulatory view of ACLs in the next section of this report. In our view, this may suggest that BMO and NA perhaps have more conservative ACL reserves relative to their peers and that these banks could record a smaller increase (or even a reduction) to their ACL than their peers upon adopting IFRS 9. Regulatory requirements can help provide a guiderail on conservatism in ACLs Detailing provisioning-related calculations in the regulatory capital regime In our view, expected losses calculated for regulatory purposes can help serve as a general guiderail or benchmark for each bank s ACL reserves in other words, regulatory November 13,

8 expected losses can be viewed as a conservative assessment of accounting based ACLs (with some limitations). Regulatory capital requirements generally impute a more conservative view on expected losses than current IAS 39 requirements, which are more reflective of the current (benign) credit environment. Given that the economy is expected to improve in Canada, we continue to expect expected losses for regulatory capital purposes to generally be more conservative than ACLs under IFRS 9. Exhibit 8 provides a comparison of probability of default (PD), loss given default (LGD), and exposure at default (EAD) modelling parameters used for regulatory capital purposes and IFRS 9. Exhibit 8: Comparison of regulatory and IFRS 9 modelling parameters Regulatory Capital IFRS 9 PD Through the cycle 12-month loss Point-in-time 12-month or lifetime horizon Definition of default is generally 90 days past due except Based on past experience, current conditions, and forward for credit cards, which uses 180 days past due looking information Definition of default consistent with regulatory capital LGD Downturn LGD based on a severe economic downturn Expected LGD based on 12-month or lifetime horizon Certain regulatory floors apply Adjusted for forward looking information where Includes direct and indirect costs associated with appropriate collection No regulatory floors Only direct costs included EAD Includes expected draws prior to default and cannot be Expected exposure across a 12-month or lifetime horizon lower than current outstanding amount Adjusted for economic conditions Can be lower than the current outstanding amount Other Lifetime losses are discounted back from point of default to the balance sheet date Source: Company reports, RBC Capital Markets There are some nuances in the regulatory capital calculations relating to ACLs. Exhibit 9 illustrates how collective allowances are reflected for regulatory capital purposes under both AIRB and standardized approaches, based on the following steps: (1) Calculate expected loss: Expected losses are calculated for exposures on internal ratings based (IRB) approaches and is calculated as a multiple of the PD, LGD, and EAD. Exhibit 8 provides a more detailed description of each of these model parameters. (2) Attribute collective allowances: Collective allowances are split proportionately between IRB and standardized approaches, based on credit risk-weighted assets (RWA) calculated under each approach. (3) Calculate shortfall or excess: A provisioning shortfall or excess is calculated for IRB exposures as: collective allowances plus all other allowances for credit losses, minus the expected loss amount. There is no equivalent calculation for standardized exposures. (4) Determine capital adjustments: Provisioning shortfalls are deducted from CET1 capital while provisioning excess and collective allowances on standardized approaches are included in Tier 2 capital, up to a limit. Essentially, collective allowances allocated to AIRB exposures are compared to a more conservative regulatory view of losses, which gives rise to a shortfall or excess provision. Collective allowances allocated to standardized exposures are not compared against any regulatory benchmark and banks receive credit for the allowance in Tier 2 capital, up to a limit. November 13,

9 Exhibit 9: Illustration of how ACL-related capital adjustments are determined 1 Calculate expected loss 2 Attribute collective allowances 4 Determine Capital Adjustments 3 Calculate shortfall or excess Expected Loss Shortfall or Excess Specific Allowance (IRB) Collective Allowance (IRB) Collective Allowance (Standardized) CET1 Capital Adjustment: Deduct shortfall (exists when expected loss is greater than collective allowance plus all other ACLs allocated to IRB exposures) Tier 2 Adjustments: Add excess provisioning up to the lower of 0.6% of IRB credit RWA or the amount of collective allowances allocated to IRB exposures (exists when expected loss is less than collective allowance plus all other ACLs allocated to IRB exposures) Add collective allowances allocated to Standardized exposures in Tier 2 capital up to a limit of 1.25% of standardized credit RWA Specific Allowance (Standardized) Source: OSFI, RBC Capital Markets As a result, each bank s shortfall or excess provisioning is a function of how conservative each bank s ACLs are relative to regulatory scenarios but also the bank s proportion of AIRB versus standardized exposures. In our view, this results in the following outcomes: (1) There is no equivalent regulatory concept of expected losses for standardized exposures and thus, using regulatory expected losses as a benchmark/guiderail is useful only to the extent each bank relies on the AIRB approach. While some of the large Canadian banks have a relatively higher proportion of exposures calculated on the standardized approach, the majority of exposures are on the AIRB approach. Thus, we still view regulatory requirements as a good benchmark for each bank s overall ACL conservatism but acknowledge there are some limitations. (2) It is possible that a bank may be conservatively provisioned but may also currently have a shortfall deduction against CET1 capital due to its proportion of AIRB versus standardized exposures. We believe the best example of this scenario may be TD. As shown in Exhibit 9, for regulatory capital purposes, the collective allowance is split proportionately for AIRB and standardized exposures based on credit risk RWA. Standardized exposures have higher RWA intensity than AIRB exposures this inherently results in a higher proportion of the collective allowance allocated to standardized exposures over AIRB exposures, which may result in a provisioning shortfall. Exhibit 10 shows the proportion of each bank s collective allowance allocated to AIRB and November 13,

10 standardized exposures for regulatory capital purposes note this is likely not reflective of the true allocation of the collective allowance. Exhibit 10: For regulatory capital purposes, BNS and TD have a higher proportion of their collective allowance allocated to standardized approaches relative to peers Collective Allowance Allocation* Q3/17 AIRB Standardized BMO 85% 15% BNS 62% 38% CM 78% 22% NA 89% 11% TD 54% 46% Average 73% 27% *Regulatory capital guidelines prescribe the allocation of the collective allowance between AIRB and standardized exposures the split is determined based on the credit risk RWA associated with each approach. The proportion of the collective allowance allocated to standardized exposures shown above is calculated as the disclosed collective allowance allocated to standardized exposures included in Tier 2 capital divided by each bank s total collective allowance. Based on available disclosure to us, none of the large Canadian banks appear to have hit the cap for including collective allowances on standardized exposures in Tier 2 capital. The proportion of the collective allowance allocated to AIRB exposures shown above is implied. Source: Company reports, RBC Capital Markets Comparing each bank s relative conservatism in its ACL reserves Each of the large Canadian banks current ACL reserves are very different but we view BMO and NA as the most conservative with its ACL reserving of the large Canadian banks. The large Canadian banks ACL for accounting and regulatory capital purposes are shown as a multiple of last twelve month (LTM) net charge-offs in Exhibit 11, as a measure of ACL coverage. We show each bank s implied minimum ACL for regulatory capital purposes in two ways one for CET1 purposes and another for total capital purposes. As discussed above, the regulatory view of ACLs for CET1 capital purposes only captures AIRB exposures capital adjustments relating to standardized exposures only consist of additions to Tier 2 capital (i.e. there is no penalization for standardized exposures). For CET1 purposes, we calculate the regulatory view of ACLs for banks with a shortfall deduction, by adding the shortfall deduction to each bank s current level of ACLs. For banks that do not have a shortfall deduction, we deduct each bank s excess provisioning from their current level of ACLs to arrive at a minimum ACL for regulatory capital purposes. For total capital purposes, we also deduct collective allowances on Standardized exposures included in Tier 2 capital to arrive at a minimum ACL for regulatory capital purposes. November 13,

11 Exhibit 11: BMO and NA have the highest proportion of ACLs against LTM net charge-offs 3.5x 3.0x Accounting (1) and CET1 Regulatory Coverage (2) of LTM Net Charge-Offs 3.5x 3.0x Accounting (1) and Regulatory Coverage (Total Capital) (3) of LTM Net Charge- Offs 2.5x 2.5x 2.0x 2.0x 1.5x 1.5x 1.0x 0.5x 0.0x BMO BNS CM NA TD CET1 Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL / LTM Net Charge-Offs 1.0x 0.5x 0.0x BMO BNS CM NA TD Total Capital Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL / LTM Net Charge-Offs (1) Accounting coverage is equal to each bank s current ACL divided by its LTM net charge-offs. (2) CET1 Regulatory coverage is equal to each bank s implied minimum ACL for CET1 capital purposes (calculated as the sum of current ACLs and current shortfall (if it exists), less provision excess (if it exists)) divided by its LTM net charge-offs. (3) Regulatory coverage on a total capital basis is equal to each bank s implied minimum ACL for total capital purposes (calculated as the sum of current ACLs and current shortfall (if it exists), less collective provisioning recognized in Tier 2 capital) divided by its LTM net charge-offs. Source: Company reports, RBC Capital Markets Exhibit 12 measures each of the large Canadian bank s accounting ACL coverage relative to the implied regulatory ACL coverage for both CET1 and total capital purposes ( coverage measured as a multiple of LTM net charge-offs). In our view, whether we look at accounting and regulatory ACL coverage on a CET1 or total capital basis, we see BMO and NA as the two banks that have the most conservative ACL reserves and CM as the bank with the least conservative ACL reserve, amongst the large Canadian banks in our coverage. We view BNS and TD as being in the middle of the pack, but it is a little less clear how the two banks compare to each other. Both banks rely relatively more on standardized approaches versus peers and as a result, comparing their accounting ACL coverage relative to their regulatory ACL coverage is slightly more difficult. As mentioned earlier, this phenomenon is particularly pronounced for TD TD has a sizeable shortfall deduction against CET1 capital but also has a significant collective allowance added to Tier 2 capital relating to standardized exposures. From a CET1 perspective, it would appear that TD is not as conservative as regulatory requirements, given TD s shortfall deduction arising from AIRB exposures. From a total capital perspective, TD s large addition to its Tier 2 capital is indicative of the bank s higher proportion of standardized exposures that are not compared against any regulatory benchmark. In either case, from either a CET1 or total capital perspective, TD still falls in the middle of the pack amongst the large Canadian banks that we cover. November 13,

12 Exhibit 12: BMO and NA appear to have more conservative ACL reserves than peers Accounting Coverage (1) less CET1 Regulatory Coverage (2) of LTM Net Charge- Offs 0.6x 0.3x 0.0x -0.5x -0.4x CM TD BNS BMO NA less conservative more conservative Accounting Coverage (1) less Regulatory Coverage (Total Capital) (3) of LTM Net Charge-Offs 0.9x 0.2x 0.4x 0.6x -0.2x CM BNS TD BMO NA less conservative more conservative (1) Accounting coverage is equal to each bank s current ACL divided by its LTM net charge-offs. (2) CET1 Regulatory coverage is equal to each bank s implied minimum ACL for CET1 capital purposes (calculated as the sum of current ACLs and current shortfall (if it exists), less provision excess (if it exists)) divided by its LTM net charge-offs. (3) Regulatory coverage on a total capital basis is equal to each bank s implied minimum ACL for total capital purposes (calculated as the sum of current ACLs and current shortfall (if it exists), less collective provisioning recognized in Tier 2 capital) divided by its LTM net charge-offs. Source: Company reports, RBC Capital Markets estimates Considerations when comparing provisioning conservatism across the banks Each bank s current starting point (and its starting point upon adopting IFRS 9 in Q1/18) theoretically should be the key determinant of its day 1 and day 2 impacts from adopting IFRS 9. For example, consider a bank with a more conservative ACL reserve (based on a more conservative economic outlook) and a bank with a less conservative ACL reserve (based on a more optimistic economic outlook). As IFRS 9 incorporates forward-looking information into ACL reserves, changes in each bank s economic outlook will result in volatility in each bank s PCLs. The bank with the more conservative ACL reserve theoretically will record less PCLs when the credit cycle turns, compared to a bank with a less conservative ACL. November 13,

13 There could exist a view that banks with a provisioning shortfall may be more inclined to strengthen their ACL reserves and use up some of their current shortfall deduction. A stronger beginning ACL reserve upon transitioning to IFRS 9 will theoretically result in lower PCLs going forward, while a less conservative ACL reserve will theoretically result in higher PCLs when the credit cycle turns. However, our covered large Canadian banks ACL reserves will likely continue to be very different after adopting IFRS 9 as a result, while we may view some banks as being more conservative than others, it is possible that they remain conservative even when transitioning over to IFRS 9. While the banks ACL positioning will be reset at the beginning of Q1/18 when IFRS 9 is adopted, each bank s ACL position will likely continue to vary given significant judgement required in the application of IFRS 9. Exhibit 13 shows the large Canadian banks that we cover and their ACL reserve as a multiple of LTM net charge-offs historically this measure of coverage has varied for each bank and has varied over time. For example, BMO and NA had relatively more conservative ACL reserves relative to peers prior to deterioration in oil and gas credits and both banks currently still appear to be more conservative than peers (NA having accomplished this from establishing a sectoral allowance). On the other hand, BNS s ACL relative to LTM charge-offs has slowly declined over time, whereas CM s ACL reserves has always remained one of the lowest relative to peers. Exhibit 13: Each bank s accounting ACL coverage has varied relative to peers and over time 8.0x Total ACLs / LTM Net Charge-Offs 7.0x 6.0x 5.0x 4.0x 3.0x 2.0x 1.0x 0.0x Source: Company reports, RBC Capital Markets BMO BNS CM NA TD Thus far, we have sized each bank s provisioning against LTM net charge-offs we believe this is a good way to size each bank s ACL against one another but there may be some limitations with this approach. Each bank s underlying rate of change in credit quality could be very different as each bank s product mix and geographical presence differs. As a result, sizing each bank s ACL coverage against LTM net charge-offs, particularly in a relatively November 13,

14 Q1/01 Q3/01 Q1/02 Q3/02 Q1/03 Q3/03 Q1/04 Q3/04 Q1/05 Q3/05 Q1/06 Q3/06 Q1/07 Q3/07 Q1/08 Q3/08 Q1/09 Q3/09 Q1/10 Q3/10 Q1/11 Q3/11 Q1/12 Q3/12 Q1/13 Q3/13 Q1/14 Q3/14 Q1/15 Q3/15 Q1/16 Q3/16 Q1/17 Q3/17 Canadian Banks benign credit environment, may not result in the most accurate comparison across the banks. For example, a bank with a riskier loan portfolio may need to hold higher ACLs relative to their current net charge-offs if the credit conditions are expected to deteriorate or normalize. Alternative risk measures likely to form when IFRS 9 is adopted We will revisit and explore day 2 IFRS 9 impacts post-q4/17 reporting; however, we believe there will likely be a period of discovery where analysts and investors will rethink of how they view PCLs. While we expect to use alternative measures to help ground an initial view of PCLs, we continue to see PCLs as a core measure of risk. In particular, during the first few quarters under IFRS 9, we could see alternative measures such as net chargeoffs or pre-tax pre-provision earnings surfacing as a smart way of explaining results. Pre-tax pre-provision earnings can be used as a measure of a bank s ability to absorb credit losses but we are cautious on looking through PCLs during the initial period of discovery under IFRS 9, as it would ignore each bank s cost of risk. Net charge-offs will likely be a helpful measure to ground PCLs but remains a lagging indicator of credit losses in our view. We view net charge-offs as a helpful measure to ground PCLs as they represent economic losses (which will be unchanged under IFRS 9) but in our view tends to be a lagging indicator of credit losses and changes in credit risk. Under current accounting standards (IAS 39), PCLs are recorded on an incurred loss basis (i.e. recorded when impairment occurs) but under IFRS 9, PCLs are recorded on an expected loss basis (i.e. recorded when there is a reasonable expectation of impairment). In other words, PCLs under current accounting standards already lead net charge-offs and IFRS 9 will further accelerate the timing of when PCLs are recorded. Exhibit 14: Net charge-offs is a lagging measure of credit losses Large Canadian Banks - Net Charge-Offs and PCLs ($MM) 3,500 3,000 2,500 2,000 1,500 1, PCLs Net charge-offs Source: Company reports, RBC Capital Markets November 13,

15 Valuations may become more sensitive to the economic outlook While it is still early days of transitioning to IFRS 9, we believe bank valuations will become more sensitive to economic indicators. PCLs will be expedited under IFRS 9 when the credit environment turns and as IFRS 9 overlays forward-looking information into provisioning, we expect to focus more on changes in the economic outlook going under IFRS 9. While Canadian bank performance has historically been correlated to economic indicators, in our view, this may increase after adoption IFRS 9. During periods of significant credit deterioration such as the significant decline in oil prices observed relatively recently, the performance of the Canadian bank index has been highly correlated with macroeconomic factors such as the WTI and was very lowly correlated with PCL ratios. However, over the longer-term, Canadian banks have still been less sensitive than the U.S. banks on other indicators such as unemployment rates (Exhibit 15) and less correlated to PCLs (Exhibit 16). November 13,

16 Exhibit 15: Canadian banks have historically been less correlated to unemployment rates than U.S. banks Canadian Bank Index vs. Canadian Unemployment Rate 4,000 3,500 3,000 2,500 2,000 1,500 1, R-squared 6% Correlation -24% 0.0% 1.0% 2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0% Canadian Bank Index (LHS) Canadian Unemployment Rate (Inverted, RHS) U.S. Bank Index vs. U.S. Unemployment Rate R-squared 62% Correlation -78% 0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00% U.S. Bank Index (LHS, Indexed) U.S. Unemployment Rate (Inverted, RHS) Source: Company reports, FDIC, Haver, Bloomberg, RBC Capital Markets November 13,

17 Exhibit 16: Canadian banks have historically been less correlated to PCLs than U.S. banks Canadian Bank Index vs. PCL Ratio and WTI Weak Oil Prices* Since 2001 Canadian Bank Index R-Squared vs.: WTI 70% 9% PCL ratio 5% 18% 0.50% 0.45% 0.40% 0.35% 0.30% 0.25% Canadian Banks PCL Ratio (RHS) Canadian Bank Index (LHS, Indexed) WTI (LHS, Indexed) U.S. Bank Index vs. PCL Ratio and WTI Weak Oil Prices* Since 2001 U.S. Bank Index R-Squared vs.: WTI 3% 21% PCL ratio 40% 36% 1.15% 1.05% 0.95% 0.85% % 0.65% 0.55% % 0.35% 0.25% U.S. Banks PCL Ratio (RHS) U.S. Bank Index (LHS, Indexed) WTI (LHS, Indexed) *Period of weak oil prices shown above is the period from July 2014 to March Source: Company reports, FDIC, Haver, Bloomberg, RBC Capital Markets Our PCL forecasts likely still have significant forecast error We have not revised our 2018/2019 PCL forecasts and our PCL forecasts likely continue to have significant forecast error; however, our PCL forecasts are probably also conservative. We are currently assuming an increase in PCLs of 29% in 2018 and an increase of 14% in We will likely gain further insight on PCL behaviour under IFRS 9 when we receive day 2 disclosures in Q1/18. November 13,

18 Exhibit 17: We forecast a significant rise in PCLs in 2018/2019 but our forecasts likely continue to have significant forecast error 1.40% Large Canadian Banks - PCL Ratio 1.20% 1.00% 0.80% 0.60% 0.40% 0.20% 0.00% E 18E 19E % CHG E 2018E 2019E 2017E 2018E 2019E Total Large Canadian Banks in Our Coverage ($MM) 6,943 6,393 8,250 9,432-8% 29% 14% Average Large Canadian Banks in Our Coverage 0.38% 0.30% 0.38% 0.41% -0.08% 0.07% 0.04% Source: Company reports, RBC Capital Markets estimates What should we be considering post-q4/17 reporting? During Q4/17 reporting, we expect the large Canadian banks to disclose IFRS 9-related impacts on: (1) opening retained earnings (and accumulated other comprehensive income) and (2) initial day 1 impact on capital ratios. Any significant balance sheet reclassification adjustments due to changes to IFRS 9 classification and measurement standards will also be disclosed. With these disclosures, post-q4/17 reporting, we expect to have a better idea on where each bank has positioned their starting point under IFRS 9, which should help provide some initial insights on day 2 IFRS 9 impacts. The large Canadian banks will be providing further disclosure upon adoption of IFRS 9 in fiscal Q1/18, which will be much more helpful in shaping our view of PCLs under IFRS 9; however, in the meantime, some questions and considerations we will be focusing on post-q4/17 reporting include: How do opening ACLs under IFRS 9 compare to current ACLs and current implied ACLs for regulatory capital purposes? Currently, the large Canadian banks ACL reserves vary significantly. With the disclosure of opening balance sheet impacts for all of the large Canadian banks, we may be able to better size ACLs relative to the regulatory view of ACLs. Theoretically, banks with a relatively more conservative ACL will experience less volatile PCLs when the credit cycle turns, while banks with less conservative ACLs will experience more volatile PCLs (though in practice this may not necessarily be the case). How do the banks ACL reserves compare against each other? As we have discussed earlier, the large Canadian banks ACL reserves vary significantly relative to regulatory capital requirements and across the banks. Current differences in ACL reserving are partly attributable to differences in product mix, credit quality, assumptions, and the November 13,

19 application of judgement. IFRS 9 requires significant application of judgement and thus, we expect continued differences across each bank s ACL positioning. However, initial day 1 impacts can help provide some insights on each bank s ACL coverage as we have shown earlier in Exhibit 11, each bank s current ACL reserves represents roughly times of LTM net charge-offs. While the adoption of IFRS 9 will change the timing and magnitude of PCLs recorded, net charge-offs will be unchanged and this can be a useful tool to size each bank s ACL reserves post-ifrs 9 adoption. What is the economic outlook incorporated into each bank s IFRS 9 ACL reserves and how does this compare across the banks? As the bulk of IFRS 9 related disclosures will be provided when IFRS 9 is adopted in Q1/18, outside of management commentary, it will likely be more difficult to infer each bank s economic outlook embedded in IFRS 9 ACL reserves. Looking at each bank s change in ACL coverage could perhaps provide some insight for example, if a bank reduces its ACL coverage (relative to LTM net charge-offs), it could suggest that the bank is expecting the economic environment to improve. We are generally expecting the economic environment to improve and thus, we expect this to be reflected in each bank s ACL reserves under IFRS 9. Which banks will still have shortfall deduction? Currently three of the large Canadian banks (CM, RY, and TD) have meaningful shortfall reductions against CET1 capital. Their current shortfall deduction will be reduced upon adoption of IFRS 9 if ACLs increase under IFRS 9 versus current IAS 39 standards. As PCLs are expected to be more volatile under IFRS 9, banks may be able to mitigate some of the volatility if they have a shortfall deduction. November 13,

20 Appendix 1 Valuation Exhibit 18: Bank comp table Stock Price Mkt Cap ROE Operating EPS EPS Growth P/E Ratio Indicated Dividend Ticker per share (Bln) BVPS P/BV TBVPS P/TBV E 2018E 2019E E 2018E 2019E 16/15 17/16 18/17 19/ E 2018E 2019E $/share Yield Canadian Banks (CAD) - RBCCM estimates 1 Bank of Montreal BMO $ $ x x 13.0% 13.7% 13.0% 12.7% % 10% 1% 5% 11.9x 11.8x 11.2x % Bank of Nova Scotia BNS $ $ x x 14.6% 14.7% 14.1% 14.0% % 9% 3% 7% 12.8x 12.4x 11.6x % CIBC CM $ $ x x 19.0% 17.7% 16.3% 16.1% % 5% -1% 7% 10.6x 10.7x 10.0x % 2 National Bank NA $ $ x x 15.5% 18.1% 17.6% 17.4% % 24% 6% 7% 11.6x 10.9x 10.2x % 3 Royal Bank of Canada RY $ $ x x 16.3% n/a n/a n/a 6.78 n/a n/a n/a 1% n/a n/a n/a n/a n/a n/a % TD Bank TD $ $ x x 13.9% 15.0% 14.6% 14.5% % 15% 4% 8% 13.0x 12.5x 11.6x % Canadian Western CWB $ $ x x 9.9% 10.4% 11.3% 11.6% % 11% 15% 10% 13.5x 11.7x 10.6x % Laurentian Bank LB $ $ x x 11.8% 12.5% 12.6% 12.4% % 9% 8% 5% 9.4x 8.7x 8.2x % Median Big Six 1.9x 2.5x 15.1% 15.0% 14.6% 14.5% 6% 10% 3% 7% 11.9x 11.8x 11.2x 3.7% U.S. Money Centre Banks and Investment Dealers (USD) - Consensus estimates Bank of America BAC $ $ x x 7.9% 7.8% 8.7% 9.2% % 25% 18% 12% 14.6x 12.3x 11.0x % Citigroup C $ $ x x 6.9% 6.9% 7.2% 7.8% % 12% 11% 16% 13.7x 12.3x 10.6x % Goldman Sachs GS $ $ x x 10.8% 10.5% 10.2% 11.2% % 13% 5% 11% 12.5x 11.9x 10.7x % JP Morgan Chase JPM $ $ x x 11.0% 10.6% 11.3% 11.8% % 13% 11% 10% 14.1x 12.8x 11.6x % Morgan Stanley MS $ $ x x 9.6% 9.3% 9.9% 10.6% % 26% 11% 9% 13.5x 12.1x 11.2x % Median 1.2x 1.4x 9.6% 9.3% 9.9% 10.6% -5% 13% 11% 11% 13.7x 12.3x 11.0x 1.8% U.S. Large Cap Regionals (USD) - Consensus estimates BB&T Corp BBT $ $ x x 9.4% 9.4% 9.6% 9.9% % 7% 10% 6% 14.9x 13.6x 12.7x % Fifth Third FITB $ $ x x 8.7% 9.1% 9.4% 9.8% % 9% 12% 9% 15.1x 13.5x 12.4x % KeyCorp KEY $ $ x x 10.8% 10.1% 10.7% 11.6% % 17% 10% 9% 13.0x 11.8x 10.8x % SunTrust STI $ $ x x 8.5% 8.7% 9.1% 9.5% % 14% 9% 8% 14.0x 12.9x 11.9x % US Bancorp USB $ $ x x 13.4% 13.3% 13.5% 13.8% % 6% 9% 7% 15.1x 13.8x 12.9x % Wells Fargo WFC $ $ x x 11.4% 11.0% 11.5% 11.8% % -2% 10% 9% 13.6x 12.4x 11.4x % PNC Financial PNC $ $ x x 9.5% 9.7% 10.3% 10.4% % 14% 11% 9% 15.8x 14.2x 13.1x % M&T Bancorp MTB $ $ x x 8.8% 9.3% 9.7% 10.0% % 16% 9% 7% 17.2x 15.8x 14.7x % Median 1.4x 1.9x 9.5% 9.5% 10.0% 10.2% 2% 11% 10% 9% 15.0x 13.5x 12.6x 2.3% U.S. Custodian Banks (USD) - Consensus estimates State Street STT $ $ x x 12.2% 12.6% 12.9% 13.3% % 19% 11% 8% 14.5x 13.1x 12.1x % Bank of New York Mellon BK $ $ x x 9.8% 10.1% 10.3% 10.6% % 12% 10% 9% 14.4x 13.1x 12.0x % Northern Trust NTRS $ $ x x 11.6% 12.0% 13.1% 14.0% % 11% 13% 7% 19.3x 17.1x 16.0x % Median 1.7x 2.9x 11.6% 12.0% 12.9% 13.3% 11% 12% 11% 8% 14.5x 13.1x 12.1x 1.9% 1) RBC Dominion Securities Inc. is a direct wholly owned subsidiary of Royal Bank of Canada and, as such, is a related issuer of Royal Bank of Canada. 2) BMO's definition of core EPS includes a net gain of $0.20 in 2017 relating to a net gain on the sale of Moneris U.S. and a portion of BMO's U.S. indirect auto loan portfolio. 3) NA's definition of core EPS includes a gain of approximately $0.12 per share in 2016 relating to the revaluation of ABA. Note: RBC Capital Markets estimates for Canadian banks only including ROE and EPS. Other banks estimates sourced from Bloomberg. Source: RBC Capital Markets estimates, Bloomberg estimates (U.S. Banks), Reuters, Company reports. In all jurisdictions where RBC conducts business, we do not offer investment advice on Royal Bank. Certain regulations prohibit member firms from soliciting orders and offering investment advice or opinions on their own stock. References to Royal Bank are for informational purposes only and are not intended as a direct or implied recommendation for investing in Royal Bank and all related securities. November 13,

21 Appendix 2 Historical regulatory capital view of ACLs Exhibit 19: Shortfall deduction against CET1 capital has increased for some banks over time 30% Shortfall Deduction as a % of Total ACLs 25% 20% 15% 10% 5% 0% Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 BMO BNS CM NA TD Source: Company reports, RBC Capital Markets November 13,

22 Exhibit 20: Accounting (1) and CET1 Regulatory Coverage (2) of LTM net charge-offs over time BMO BNS 5.0x 3.6x 4.5x 4.0x 3.5x 3.4x 3.2x 3.0x 2.8x 3.0x 2.6x 2.5x 2.0x 1.5x 2.4x 2.2x 2.0x 1.8x 1.0x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 1.6x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 CET1 Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs CET1 Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs CM NA 2.6x 4.0x 2.4x 3.5x 2.2x 3.0x 2.0x 1.8x 2.5x 1.6x 2.0x 1.4x 1.5x 1.2x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 1.0x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 CET1 Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs CET1 Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs TD 3.0x 2.8x 2.6x 2.4x 2.2x 2.0x 1.8x 1.6x 1.4x 1.2x 1.0x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 CET1 Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs (1) Accounting coverage is equal to each bank s current ACL divided by its LTM net charge-offs. (2) CET1 Regulatory coverage is equal to each bank s implied minimum ACL for CET1 capital purposes (calculated as the sum of current ACLs and current shortfall (if it exists), less provision excess (if it exists)) divided by its LTM net charge-offs. Source: Company reports, RBC Capital Markets November 13,

23 Exhibit 21: Accounting (1) and Regulatory Coverage (Total Capital) (2) of LTM net charge-offs over time BMO BNS 5.0x 3.5x 4.5x 4.0x 3.0x 3.5x 2.5x 3.0x 2.5x 2.0x 2.0x 1.5x 1.5x 1.0x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 1.0x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 Total Capital Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs Total Capital Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs CM NA 2.6x 4.0x 2.4x 3.5x 2.2x 3.0x 2.0x 1.8x 2.5x 1.6x 2.0x 1.4x 1.5x 1.2x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 1.0x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 Total Capital Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs Total Capital Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs 2.6x TD 2.4x 2.2x 2.0x 1.8x 1.6x 1.4x 1.2x 1.0x Q3/13 Q4/13 Q1/14 Q2/14 Q3/14 Q4/14 Q1/15 Q2/15 Q3/15 Q4/15 Q1/16 Q2/16 Q3/16 Q4/16 Q1/17 Q2/17 Q3/17 Total Capital Regulatory "Min" ACL / LTM Net Charge-Offs Current Accounting ACL/ LTM Net Charge-Offs (1) Accounting coverage is equal to each bank s current ACL divided by its LTM net charge-offs. (2) Regulatory coverage on a total capital basis is equal to each bank s implied minimum ACL for total capital purposes (calculated as the sum of current ACLs and current shortfall (if it exists), less collective Source: Company reports, RBC Capital Markets November 13,