Developed Countries

This report contains an error in the section related to consumer spending and fiscal policy. That error somewhat changes the conclusions from the report, and it particularly impacts Chart 3, Table 2 and Table 3. The attached note explains the mistake and includes corrected versions of Chart 3, Table 2 and Table 3. Highlights Duration: A re-rating of Tech stock valuations is likely not a near-term catalyst for significantly lower bond yields. Congress’ continued failure to pass a follow-up to the CARES act is a greater near-term risk for bond bears. We continue to recommend an “at benchmark” portfolio duration stance alongside duration-neutral yield curve steepeners. Fiscal Policy: Without additional household income support from Congress, at least on the order of $500 - $800 billion, consumer spending will massively disappoint expectations during the next 6-12 months. Inflation: Inflation will continue its rapid ascent between now and the end of the year, but it is likely to level-off in 2021. We recommend staying long TIPS versus nominal Treasuries for the time being, but we will be looking to take profits on that position later this year. Feature Bond Implications Of A Tech Stock Sell-Off Risk-off sentiment reigned in equity and credit markets during the past two weeks. The S&P 500 fell 7% between September 2nd and 8th and the average junk spread widened from 471 bps to 499 bps. This represents the largest sell-off since June when the equity market saw a similar 7% decline and the junk spread widened from 536 bps to 620 bps (Chart 1). Chart 1Two Equity Sell-Offs, Two Different Bond Market Reactions A comparison between the September and June episodes is particularly interesting for bond investors because Treasuries behaved very differently in each case. In June, bonds benefited from a flight to quality out of equities and the 10-year Treasury yield fell 22 bps. But this month, Treasuries actually delivered negative returns and the 10-year Treasury yield rose 3 bps (Chart 1, bottom panel). Table 1Selected Asset Class Performance During Last Two Equity Sell-Offs Why would Treasuries perform so well in June but fail in their role as a diversifier of equity risk in September? The answer lies in the underlying drivers of the stock market’s decline, which are easily identified when we look at the performance of different equity sectors. Table 1 shows the performance of different equity sectors in both the June and September sell-offs. In June, it was the cyclical equity sectors – Industrials, Energy and Materials – that led the decline. These sectors tend to be the most sensitive to global economic growth. This month’s equity drawdown was led by Tech stocks, while cyclical and defensive sectors saw much smaller drops. Table 1 also shows that a broad measure of commodity prices – the CRB Raw Industrials index – rose by 0.79% during the September equity sell-off, significantly outpacing gains in the gold price. In June, the CRB index still rose but it lagged gold by a wide margin. The underlying drivers of the stock market’s decline explain why Treasuries performed well in June and underperformed in September. We bring up the performance of different equity sectors, commodity prices and gold because bond yields correlate most strongly with: The performance of cyclical equities over defensive equities (Chart 2, top panel). The ratio of CRB Raw Industrials over gold (Chart 2, bottom panel). Chart 2High-Frequency Bond Indicators These correlations explain why bond yields fell a lot in June but not in September. June’s equity sell-off was more like a traditional risk-off event that saw investors questioning the sustainability of the global economic recovery. The cyclical equity sectors that are most exposed to the global economic cycle experienced the worst losses and demand for safe-haven gold far outpaced the demand for growth-sensitive industrial commodities. In contrast, this month’s sell-off was driven by a re-rating of Tech stock valuations, not so much expectations for a negative economic shock. Technology now makes up such a large portion of the equity index’s market cap that this sort of move can cause the entire stock market to fall, but the pass-through to bonds will be much smaller for any equity sell-off that isn’t prompted by a negative economic shock and led by cyclical equity sectors. Implications For Bond Investors Even after this month’s drop, there remains a legitimate concern about extreme Tech stock valuations. The fact that many of the larger Tech names, like Microsoft and Apple, have benefited from the pandemic only makes it more likely that their stock prices will suffer as the world slowly returns to normal. From a bond investor’s perspective, we doubt that even a large drop in Tech stock prices would lead to significantly lower bond yields, especially if that drop occurs in the context of an economy that continues to recover. Bond yields will only turn down if the market starts to question the sustainability of the economic recovery, an event that would be negative for cyclical equity sectors but much less so for the big Tech names. With that in mind, our base case outlook calls for continued economic recovery during the next 6-12 months, but we do see a significant risk that the failure to pass a follow-up to the CARES act will lead to just such a deflationary shock during the next couple of months. We therefore recommend keeping portfolio duration close to benchmark, while positioning for continued economic recovery via less risky duration-neutral yield curve steepeners. The Outlook For Consumer Spending And The Necessity Of Fiscal Stimulus After plunging during the lock-down months of March and April, consumer spending has rebounded strongly during the past few months. But can this strong rebound continue? Our view is that it cannot. That is, unless Congress delivers more income support to households. Even a large drop in Tech stock prices is unlikely to lead to significantly lower bond yields, especially if that drop occurs in the context of an economy that continues to recover. In this section we consider several different economic scenarios and estimate the amount of further income support that is necessary to sustain an adequate level of consumer spending. First off, to make forecasts for consumer spending we need to consider two main parameters: household income and the personal savings rate (Chart 3). More income leads to more spending in most cases. The only exception would be if cautious households decide to increase the amount they save relative to the amount they spend. Chart 3Consumer Spending Driven By Income & The Savings Rate We’ve actually seen that exception play out somewhat during the past five months. The CARES act provided households with an income windfall, but the savings rate also shot higher. This suggests that households had enough income to spend even more during the past few months but have been much more cautious than usual. We cannot overstate the role the CARES act has played in supporting household incomes since March. Disposable income has grown 7.4% during the past five months compared to the five months prior to COVID, and the CARES act’s provisions pressured income 10.3% higher during that period (Chart 4). The CARES act’s one-time $1200 stimulus checks and expanded $600 weekly unemployment benefits were the two most important provisions in this regard. Together, they pushed disposable income higher by 7.5%. Chart 4Disposable Personal Income Growth And Its Drivers This presents an obvious problem. The income support from the CARES act is now expired and Congress has yet to pass a follow-up stimulus bill. How vital is it that we get a new bill? And how large does it need to be? To answer these questions, we first need to set a target for adequate consumer spending growth. The second panel of Chart 3 shows 12-month over 12-month consumer spending growth. That is, it looks at total consumer spending during the last 12 months and shows how much it has increased (or decreased) compared to the previous 12 months. Notice that the worst 12-month period during the 2008 Great Financial Crisis (GFC) saw 12-month over 12-month consumer spending growth of -3%. During the economic recovery that followed, consumer spending growth fluctuated between +2% and +6%. Exercise 1: The March 2020 To February 2021 Period Chart 5Three Scenarios For Income And Savings In our first exercise, we consider the 12-month period starting at the very beginning of the COVID recession in March 2020 and ending in February 2021. As a bare minimum, we target consumer spending growth of -3% for this 12-month period on the presumption that 12-month spending growth equal to the worst 12 months seen during the GFC is the bare minimum that markets might tolerate. We also consider somewhat rosier scenarios of 0% and 2% spending growth. In addition to consumer spending targets, we also make assumptions for household income and the savings rate. We consider income coming from all sources including automatic government stabilizers, but without assuming any additional fiscal support from the government. We consider three scenarios (Chart 5): A pessimistic scenario where both income and the savings rate hold steady at current levels. An optimistic scenario where both income and the savings rate return to pre-COVID levels by February 2021. A “split the difference” scenario where both income and the savings rate get halfway back to pre-COVID levels by next February. Table 2 shows how much additional income support from the government is needed between now and February to achieve each of our consumer spending growth targets in each of our three scenarios. For example, in the optimistic scenario the government will need to provide $434 billion of additional income support between now and February for consumer spending to hit our minimum -3% threshold. In the more realistic “split the difference” scenario, households will require another $777 billion of stimulus. Table 2 also shows that stimulus on a monthly basis and compares the monthly rate of stimulus to the rate provided by the CARES act. For example, an additional $777 billion of income doled out between August and February works out to $111 billion per month, 61% of the amount of monthly stimulus provided by the CARES act between April and July. Table 2Without More Stimulus COVID's Impact On Consumer Spending Will Be Worse Than The GFC Two main conclusions jump out from this analysis. The first is that more income support from Congress is absolutely required. Otherwise, consumer spending will come in worse during the March 2020 to February 2021 period than it did during the worst 12 months of the GFC. Second, unless we assume a truly dire economic scenario, the follow-up stimulus does not need to be as large as the CARES act. In our most realistic “split the difference” scenario, that $777 billion of required stimulus is only 61% of what the CARES act doled out on a monthly basis. In that same scenario, a follow-up bill that delivered the same monthly stimulus as the CARES act would lead to positive 12-month consumer spending growth. Exercise 2: The August 2020 To July 2021 Period Chart 6One More Scenario One potential problem with our last exercise is that our target was for total consumer spending between March 2020 and February 2021. This period includes five months for which we already have data and the exercise is therefore partially backward-looking. A more relevant analysis might target consumer spending on a purely forward-looking basis from August 2020 to July 2021. We therefore perform our calculations again for the August 2020 to July 2021 period. This time, we consider only one economic scenario where income and the savings rate both return to pre-COVID levels by July 2021 (Chart 6). This scenario works out to be slightly more optimistic than the “split the difference” scenario we considered earlier. Also, since our target 12-month spending growth period no longer contains the downtrodden months of March and April, we require a more ambitious target than -3% growth. A return to the post-GFC range of 2% to 6% represents a target that is likely more representative of market expectations. Table 3 shows the results of this second analysis. Once again, we see that some additional government stimulus is necessary to meet our spending targets. Even to achieve 0% spending growth over the next 12 months will require another $249 billion from the government, and that outcome would almost certainly disappoint markets. We calculate that an additional $534 billion is required to achieve 2% spending growth during the August 2020 to July 2021 timeframe. This result is consistent with the $777 billion we calculated in Table 2, though it has come down a bit because we have made slightly more optimistic economic assumptions. Table 3At Least Half A Trillion More Government Income Support Is Needed Bottom Line: Our analysis suggests that further stimulus is needed to sustain the recovery in consumer spending. A new stimulus package doesn’t need to be as large as the CARES act on a monthly basis, but it should provide at least $500 - $800 billion of additional income support to households. With Congress still dithering on this issue, financial markets appear overly complacent in the near-term. While the economic constraints suggest that a deal should be reached soon, policymakers may need to see a spate of negative economic data and/or poor market performance before being spurred into action. In acknowledgement of this significant near-term risk to the economic outlook, bond investors should refrain from getting too bearish, and keep portfolio duration close to benchmark for the time being. Inflation’s Snapback Phase Chart 7Inflation Coming In Hot The core Consumer Price Index rose 0.4% in August, the third large monthly increase in a row (Chart 7). We see inflation continuing to come in hot between now and the end of the year, before tapering off in 2021. As of now, we would describe inflation as being in a snapback phase. That is, back in March and April, when lock-down measures were widespread across the country, the sectors that were most affected by the shutdowns experienced massive price declines. However, notice that core inflation fell by much more than median or trimmed mean inflation during this period (Chart 7, panels 2 & 3). The median sector’s price didn’t fall that much, but the overall inflation number moved down because of deeply negative prints in a few sectors. Now that the economy is re-opening, many of the sectors that were most beaten down in March and April are coming back to life. As a result, those massive price declines are turning into massive price increases. Once again, the median and trimmed mean inflation figures have been much more stable. This “snapback” dynamic is illustrated very clearly in Chart 8 which shows the distribution of monthly price changes for 41 different sectors in April and in August. Notice that while the middle of the distribution hasn’t changed that much, April’s massive left tail has morphed into August’s massive right tail. Chart 8Distribution Of CPI Expenditure Categories The continued wide divergence between core inflation and the median and trimmed mean measures suggests that this snapback phase has further to run. In other words, we will likely continue to see strong inflation prints for a few more months as the sectors that were most downbeat in March and April continue their rebounds. However, once core catches back up to the median and trimmed mean inflation measures, this snapback phase will come to an end and inflation’s uptrend will probably level-off. The continued wide divergence between core inflation and the median and trimmed mean measures suggests that this inflation’s snapback phase has further to run.  We recommend that bond investors continue to favor TIPS over nominal Treasuries during this snapback phase, but we will be looking for an opportunity to go underweight TIPS versus nominal Treasuries later this year, once core inflation moves closer to the median and trimmed mean measures and the snapback phase ends. Appendix A: Buy What The Fed Is Buying The Fed rolled out a number of aggressive lending facilities on March 23. These facilities focused on different specific sectors of the US bond market. The fact that the Fed has decided to support some parts of the market and not others has caused some traditional bond market correlations to break down. It has also led us to adopt of a strategy of “Buy What The Fed Is Buying”. That is, we favor those sectors that offer attractive spreads and that benefit from Fed support. The below Table tracks the performance of different bond sectors since the March 23 announcement. We will use this to monitor bond market correlations and evaluate our strategy’s success.   Table 4Performance Since March 23 Announcement Of Emergency Fed Facilities Ryan Swift US Bond Strategist rswift@bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification

Special Report

Highlights In a world with low expected returns from various asset classes and still-elevated target returns among largely underfunded pension funds, asset allocators may have to consider the use of leverage to meet liability requirements. Canadian pension funds have been more open to using leverage than their US counterparts, but even the very conservative Japanese Government Pension Investment Fund (GPIF) has an allocation to levered asset classes such as private equity, albeit at a very low weight. Retail investors do not have access to low-cost financing as institutional investors do. Still, they too can add leverage via ETFs and Liquid Alts mutual funds. When leverage is used at the asset-class level such as in alternative asset classes, financing costs play an important role in investment decisions. For pension funds with access to low-cost financing, “direct investing” in alternative assets is more advantageous than indirect investment via alternative funds. When leverage is used at the portfolio level, such as via a risk-parity structure, the financing cost impacts mostly just the return, but the leverage constraints impact both return and volatility. Risk-parity strategy is more advantageous when it’s used as one of the strategies in a total portfolio, rather than at the total-fund level because usually a sub-portfolio can have a much higher leverage ratio than the total fund. Leverage should be managed in a centralized risk-management system at the total-fund level, together with all other risk exposures. 1. Why Should Leverage Be Considered? In a Global Asset Allocation Special Report on long-term return assumptions,1 the key conclusion was that, for the next decade, investors would not be able to achieve the kind of return targets they were used to over the previous two decades because all asset classes would see much lower returns going forward, with the largest reductions coming from fixed-income and alternative assets such as farmland, REITs, and commodities. This is bad news for investors, especially pension fund investors with long-term liabilities to match. For example, according to Wilshire Consulting,2 at the end of 2018, the aggregate funded ratio (defined as the fund assets as a percentage of the fund obligations) of 134 US state retirement systems was 72.2%, which is better than the low at the end of 2016 (Chart 1). However, as shown in Chart 2, there were still about 11% of the funds with assets at less than 50% of liabilities. Chart 1US Pensions' Funded Status* Chart 2US Pensions' Funded Ratio Distribution* Over the past two decades, the risk-return profile of traditional assets like equities and government bonds has already been much less attractive than historical averages, as shown in Chart 3, but investors have diversified into credit and alternative asset classes (which contain embedded leverage) to enhance their portfolios’ risk-return profile. Chart 3Future Risk-Return Profiles Less Attractive Than Historical Averages According to the above-mentioned Global Asset Allocation Special Report, with a conventional 50/30/20 (equities/bonds/alts) allocation, a US investor could comfortably achieve a 7% annual return over the past two decades. Now, alternative asset classes have become mainstream, likely producing a much lower future return. The same 50/30/20 portfolio would currently generate only about 4.9% annually, much less than what’s required to match liabilities. In fact, alternatives’ future return expectations have been cut to 6.1% from their past 20-year average of 15.1% annually, meaning that even if 100% of assets are fully invested in alternatives, the expected return will still be lower than the 7% that’s still assumed by some US state pension funds.3 Not to mention that at the end of 2018, over 34% of US retirement pension funds had long-term rate-of-return expectations higher than 7.5%, as shown in Chart 4. Chart 4Challenging Long-Term Return Expectations Chart 5Why Should Leverage Be Considered? According to Modern Portfolio Theory, to achieve a higher return investors can take higher risk in three different ways, as shown in Chart 5:  1) Allocating more funds to higher-return/higher-risk assets, i.e. moving upwards to the right along the efficient frontier (red line) – for example, a 60-40 equity/bond portfolio is well to the upper-right side of the “optimal” allocation; 2) Levering up one or more assets to alter the shape of the frontier (grey line) – for instance, incorporating private-equity and infrastructure funds that contain embedded leverage; and 3) Levering up the “optimal” (in terms of return per unit of risk) risky portfolio with funds borrowed at the total-fund level (green line). Risk parity is a close proximation. For more detail about the basics of leverage, please see Appendix 1 on pages 21-22. Chart 5 illustrates three different frontiers based on the assumed risk-return forecast for US equities, US Treasurys, and alternative assets.4 We can observe the following: When the target return is low (at target 1), leverage does not provide significant benefit no matter which form is used; As the return target moves up relative to what the underlying assets can provide (target 2), direct leverage produces a better return/risk profile than embedded leverage, which in turn is better than the portfolio without any leverage; When the target return is higher than what any efficient combination from the available asset classes can achieve (target 3), investors must consider the use of direct leverage. In theory, investors should always prefer to use leverage at the total-fund level to lever up the “optimal” portfolio. In reality, however, some investors are constrained from borrowing. In addition, some investors do not have the expertise or infrastructure to manage the additional complexity that results from the use of direct leverage. In fact, direct leverage has typically been considered dangerous by many investors. Misuse of leverage was attributed to some high-profile failures, such as Long-Term Capital Management in 1998 and Lehman Brothers in 2008. So how has leverage been used by asset allocators? What are the key factors that determine if and how leverage should be used? What are the key risks associated with the use of leverage, and how should leverage be managed? In the sections below, we first review how some pension funds and retail investors have been using leverage (we ignore hedge funds, even though they are the most obvious users of leverage, because they are a part of the “alternative” asset class with embedded leverage). From there, we attempt to address, 1) How does financing cost impact leverage at the asset-class level? and 2) How does financing cost impact the decision to use leverage at the portfolio level if investors are constrained by the amount of leverage that can be used? Finally, we suggest a centralized leverage management framework to monitor and manage leverage at the total-portfolio level. 2. Use of Leverage By Pension Funds Leverage can be applied in many different ways. In general, the use of leverage by pension funds can be grouped into four categories: First, with a focus on return-seeking. This is achieved mainly by using alternative asset funds such as private-equity funds, hedge funds, real-estate funds, and infrastructure funds. These funds have embedded leverage, but with much higher costs. They provide diversification and higher risk-adjusted returns, partly because of their embedded leverage and lock-up advantages. Large pension funds, especially the Canadian pension funds which all have excellent credit ratings and strong in-house talent, have also taken advantage of their solid balance sheets to acquire low-cost financing to invest directly in alternatives. For example, the first bond issued by the Ontario Teachers’ Pension Plan (OTPP) in 2001 was $600 million at 5.7%, while the Canada Public Pension Investment Board (CPPIB) even issued euro-denominated bonds in both 2017 (2 billion euros, 7-year, 0.375% coupon) and 2018 (1 billion euros, 15-year, 1.5% coupon).5 Proceeds from these bond issues have mostly been used to invest in alternative assets, which now account for a large proportion of the major Canadian pension funds’ assets under management (Table 1). Table 1“Alternatives” Have Become Mainstream For The Canadians                     Most US state pension funds are more conservative than their Canadian peers. They too have been gradually adding exposure to alternatives with embedded leverage such as private equity, real estate, and hedge funds, as shown in Chart 6. Even Japan’s Government Pension Investment Fund (GPIF), the very conservative Japanese sovereign wealth fund, in its current operating guideline has a 5% allocation to alternatives such as private equity, real estate, and infrastructure.6 That’s an impressive amount considering its first investment in the space was in 2013, as shown in Chart 7. Chart 6The Americans Are Catching Up On Alternatives Chart 7GPIF’s Push Into Alternatives* However, the push into alternative asset classes by large pension funds has made it increasingly difficult to allocate funds to alternative assets. For example, CalPERS has only an 8% allocation to private equity,7  yet its most recent exposure as of June 2019 stood at only 7.1% – because it could not find enough suitable private-equity investments to build the asset class to the desired scale.8 Second, with a focus on liability matching. Pension funds who follow the liability-driven investing (LDI) approach often construct two portfolios. One is the liability-matching portfolio and the other is the active portfolio. The former matches the liabilities, while the latter generates alpha to cover management fees and to provide a cushion for estimation errors. Since most pension liabilities are indexed to inflation, liabilities are often modelled as a combination of nominal bonds and inflation-linked bonds with leverage. The leverage ratio can often be higher than two or three times because of the ultra-long duration of the liabilities versus the available bond instruments. For example, the Healthcare of Ontario Pension Plan (HOOPP) uses an LDI approach, which is why its leverage ratio is much higher than some other pension plans, as shown in Chart 8.  Chart 8Use Of Leverage By Some Pension Funds Third, with a focus on risk diversification. Risk-based strategies such as risk parity generate a more diversified portfolio with lower absolute returns compared to a conventional 60/40 equity/bond portfolio, but in general have a much higher Sharpe ratio, and therefore require leverage to achieve the required return/risk objective.9 Even though most risk-parity believers dedicate a portion of their assets to risk-parity strategies (either internally with direct leverage or externally with embedded leverage), some pension funds have adopted such a risk-diversification approach at the total-fund level. Danish pension fund ATP and the Missouri State Employees Retirement System (MOSERS) are two examples. As shown in Chart 8, as of June 2019, MOSERS’ leverage was about 50%,10 a lot higher than CalPERS’ newly augmented total-fund leverage limit of 20% (from 5% previously),11 because CalPERS does not use the same approach to apply leverage. Fourth, with a focus on more tactical moves, such as tail-risk hedging, revenue generation, and opportunistic strategies to take advantage of short-term dislocations in the marketplace. These tactics are achieved mostly using derivatives such as futures, options, and swaps. For example, equity and bond futures or swaps are often used to tactically adjust asset allocation at the total-fund level without impacting the underlying asset-class management. 3. Use of Leverage By Retail Investors Retail investors do not enjoy low-cost financing as large institutions do. They can use lines of credit or margin accounts to invest, and they can also use derivatives if they are qualified to do so. For those who are not qualified or not comfortable using leverage themselves, there are two types of retail products with embedded leverage: Levered or inverse ETFs and “Liquid Alts” mutual funds or ETFs. Levered or Inverse ETFs: These products are rebalanced daily to a fixed leverage multiple, often -3X, -2X, -1X, 2X, and 3X of the underlying assets. As such, only daily performance matches the intended performance objective. Because of the daily realization of gain and loss, they are not suitable for long-term buy-and-hold investors because the longer the holding period, the larger the drift due to the compounding effect. For example, Chart 9 shows the Nasdaq-100 ETF (QQQ) and the associated levered ETFs. It’s interesting to note that in several annual periods ending in 2011, 2016, 2018 and 2019, QQQ’s one-year return was slightly positive, yet 3X ETF’s corresponding returns were negative! This is due to the “negative diversification return” effect as defined by Qian.12 Chart 9NASDAQ-100 Linked ETF Performances* Liquid Alternative Mutual Funds/ETFs: These are the “liquid” version of alternative investment strategies aimed at retail investors. They are easy to buy and sell. In Canada, since National Instrument 81-102 became effective in January 2019, retail investors who do not have the sophistication to directly invest in alternatives now have access to such investments via mutual funds and ETFs. As shown in Table 2,13 these funds can utilize leverage up to 3X based on gross aggregate exposure by borrowing or short-selling. In the US, liquid alts have been available to retail investors since 2013, and the market has grown rapidly to over US$225 billion.14 Now there are signs emerging that even some institutional investors are starting to look into liquid alts ETFs.15 Table 2Canadian Regulation On Liquid Alts Mutual Funds 4. Leverage At Asset-Class Level Alternative funds, such as funds that invest in private equity, private debt, and infrastructure, typically use leverage. These funds carry a high cost because 1) investors in these funds must pay a premium for not managing leverage, and 2) these funds often have much higher financing costs. For example, the average financing cost for leveraged buyouts in 2014 was more than 5%,16 while the average risk-free rate in 2014 was 0.03%. Research has shown that private-asset performance can be proxied by using leverage and the corresponding public asset. In Table 3, the base case is based on the forecast for US equities and Treasurys without leverage, and a risk-free rate of 2.6%.17 Then equities are leveraged by 1.5 times to proximate private equity. The low-cost case has a financing cost of 1.57% (which is what the average 3-month T-Bill rate was in 2019), while the high-cost case with a financing cost of 3.92%, which is 2.5 times the low-cost rate. Table 3Assumed Returns/Risks* Chart 10Financing-Cost Impact On The Use Of Embedded Leverage Chart 10 shows the corresponding frontiers of the three cases. It’s clear that leverage expands the frontier to the right, meaning that leverage helps to achieve a higher return with better diversification, albeit with higher volatility. However, the financing cost plays a very important role in the feasibility of the leverage decision. When the financing cost is low, leverage is better than the base case at any return-target level. When the financing cost is high, however, leverage is worse – so long as the return target is lower than what the underlying assets can achieve without leverage. This supports the shift to “direct investing” by some institutional investors with access to lower financing when investing in alternative asset classes. 5. Leverage At Portfolio Level Risk parity is an obvious example of using leverage at the portfolio level. As shown in our previous report on risk parity, there are different approaches to implementing risk parity, and they can generate different results – especially when there are more than two assets.18 To analyze the impact of leverage constraints and financing costs, we use a two-asset (US equity/Treasury) risk-parity portfolio as the basis of our analysis. One definitive conclusion we arrived at in our previous report was that risk-parity approach historically always outperforms in recessions. This conclusion has passed the real-time test in the most recent pandemic-induced recession. As shown in Chart 11A, risk-parity portfolios that target the same volatility as a 60/40 US equity/Treasury portfolio have outperformed the latter significantly. The same holds true for the portfolios that target the same volatility as an equity portfolio (Chart 11B).   Chart 11AUS Risk Parity With Same Vol As US 60/40 Chart 11BUS Risk Parity With Same Vol As US Stocks However, as described in the previous Special Report on risk parity, we did not impose any cap on the use ofleverage. As such, some strategies that use a relatively short lookback period to calculate historical statistics required very high leverage ratios at some time periods in our back-tests. What would happen if we set a cap on the leverage ratio? And what if the financing cost is higher than the 3-month T-bill rate assumed in most academic research, and also in our previous report? Chart 12A and Chart 12B show the results when leverage is capped at three times and the financing cost is Libor +25 basis points. It’s clear that Chart 12A looks the same as Chart 11A because the leverage cap is higher than the required leverage employed, while the cost impact is negligible for such a short period. But Chart 12B shows that, even though the risk-parity portfolio still outperformed, the outperformance has been much less so far this year because the required leverage was a lot higher than three times. Chart 12AImpact Of Financing Cost And Leverage Constraint On Low-Vol Target Risk-Parity Chart 12BImpact Of Financing Cost And Leverage Constraint On High-Vol Target Risk-Parity The impact of financing costs in Chart 12A is barely seen because the time period was short and the interest rate was low. What is the long-run impact of leverage restrictions and financing costs then? Chart 13A and Chart 13B show the long-run statistics from April 1945 to July 2020 based on a 180-month look-back period for two portfolios: RPL1, the risk-parity portfolio with the same volatility target as a 60/40 US equity/Treasury portfolio; and RPL2, the risk-parity portfolio with the same volatility as MSCI US equity index. Chart 13C shows the risk-adjusted returns for three portfolios with constant volatility targets at 10%, 12%, and 15%, respectively. Chart 13ALong-Term Impact Of Financing Cost And Leverage Constraint On Risk-Parity With Low Vol Target* Chart 13BLong-Term Impact Of Financing Cost And Leverage Constraint On Risk-Parity With High Vol Target*     Chart 13CLong-Tem Impact Of Financing Cost And Leverage Constraint On Risk-Parity Portfolio* Some observations are worth highlighting: Financing costs mainly impact average return, but have very little impact on volatility. As such, higher financing costs reduce risk-adjusted returns. When there is no financing cost, all risk-parity portfolios with different volatility targets should have the same risk-adjusted return as the underlying unlevered risk-parity portfolio. When financing costs are present, however, this is no longer the case. Leverage constraints impact both returns and volatility in the same direction – i.e., stricter constraints on leverage reduce both return and volatility, and vice versa. The magnitude of the impact from leverage constraints, however, varies because the target volatility of the portfolio plays a key role in the required leverage. For a constant-volatility target, a tighter control on leverage will reduce volatility more than return, resulting in a higher risk-adjusted return (Chart 13C); for a variable-volatility target such as RPL1 and RPL2, however, the same conclusion cannot be drawn (Charts 13A and 13B) Long-run statistics do not tell the full story because they really depend on the period chosen. Chart 14A shows the dynamic impact of financing when there were no constraints on leverage, and Chart 14B shows the dynamic impact of leverage when there were no additional financing costs. Chart 15 shows the combined impact when leverage is capped at three times and the financing cost at Libor+25 basis points. They are for five different risk-parity portfolios with different volatility targets with a lookback window length of 180 months. (For different lookback window, please see Appendix 2 on pages 23-25). Chart 14ADynamic Impact Of Financing On Risk Parity Without Leverage Constraint* Chart14BDynamic Impact Of Leverage Cap On Risk Parity Without Extra Financing * Chart 15Dynamic Impact Of Financing On Risk Parity With Leverage Being Capped At 3X It is interesting to note the following: When there is no restriction on leverage, additional financing cost eats away cumulative total return in a much more significant way when an risk parity portfolio has a higher-volatility target than a lower-volatility target (Chart 14A). This is simply because a higher-volatility target requires higher leverage. When there was no additional cost of financing, constraint on leverage ate away total returns – mostly in the early years of the back-test when required leverage was often very high. In recent years, the impact was significant only when the leverage cap dropped to three times or lower. Also, the higher the volatility target, the more reduction in return risk-parity portfolio would suffer compared to its base case (Chart 14B). When the lookback window length is changed, the impact of leverage and financing cost also changes. The shorter the window length, the larger the impact (Charts in Appendix 2). A 180-month lookback period was the preferred choice in our previous report, and it is still more appropriate to use than 36 months or 360 months. Since additional cost and restriction on leverage eat away total return so much, is it really worthwhile to even consider using a risk-parity approach at all? Charts 16A and 16B show that overall total returns were worse during the entire period from April 1945 to July 2020, when additional cost and leverage constraints are applied. Since the burst of the tech bubble, however, risk-parity portfolios with the same volatility target as US 60/40 and also MSCI US have generated higher total returns than US 60/40 and MSCI US, respectively.  Chart16ADoes Risk-Parity With Same Vol As US 60/40 Outperform US 60/40? Chart 16BDoes Risk-Parity With Same Vol As MSCI USI Outperform MSCI US? We are in a low interest-rate environment, and rates may stay low for a long time to come. In addition, when futures contracts are used to implement leverage, the implied cost is very close to 3-month T-Bills; Libor or Libor + may be present mostly when swaps are used due to factors such as supply/demand and counterparty risk. As such, financing costs will likely play less of a role than leverage constraints until interest rates rise significantly. Given that total-fund leverage is much lower than individual strategy/portfolio leverage, the implication is that risk-parity is more advantageous when it is used as a strategy in a sub-portfolio other than at the total fund level. 6. Suggestion For Leverage Management In a low-return environment, asset allocators face more challenges to meet return targets than in the past. Unless return targets are lowered to what the underlying assets can reasonably provide, asset allocators may have to consider the use of leverage to beef up overall portfolio returns. However, leverage is also a double-edged sword because it also increases portfolio volatility. As such, we suggest a centralized risk-management system to monitor and manage all risks, including risks associated with leverage, in line with our suggestion on currency hedging outlined in our 2017 Special Report. Appendix 1: Leverage Basics Leverage is an investment strategy of using borrowed money – specifically, the use of various financial instruments or borrowed capital – to increase the potential return of an investment. It also amplifies the loss potential if the levered investment does not work out as expected. This is why it is also often called a “double-edged sword.” Leverage has many different forms and is used in many different places. For example, residential home mortgages are a form of leverage that the general public understands very well, yet the leverage embedded in a futures contract may sound alien to some retail investors. For asset allocators, the most important decision on leverage is whether to apply leverage directly at the total-portfolio level or use assets with embedded leverage. For example, issuing bonds to lever up a diversified portfolio (a mean-variance optimal portfolio or a suboptimal risk-parity portfolio) is an example of the former. On the other hand, investing in a private equity fund is an example of the latter. Research has shown that for large pension funds with excellent credit ratings, the latter is less efficient than the former due to the much higher cost of financing.1  For example, in 2014, the average cost of financing for leveraged buyouts was in excess of 5% when the short-term interest rate was close to zero.2 It’s not surprising that pension investors have formed joint ventures to invest in alternative assets directly instead of relying on specialty funds. In terms of financing, there is on-balance-sheet leverage and off-balance-sheet leverage. On-balance-sheet leverage raises liabilities, such as via bond issuance. Off-balance-sheet leverage uses the balance sheet of a counterparty, such as OTC financial derivatives. A repo agreement is a repurchase agreement that involves selling a security (often a government bond) to a counterparty (a lender) with the promise of buying it back after a pre-defined period at a pre-defined price. It’s often used for short-term liquidity management.  Depending on the form of financing, the measurement of leverage differs. Accounting leverage or balance-sheet leverage is calculated as total assets divided by net assets. This measurement is accurate only if on-balance-sheet leverage is used for long-only investments. When off-balance-sheet leverage is used or when shorting is involved, then accounting leverage severely understates the actual leverage. For example, Appendix Table A1 below is a snapshot from the 2018 annual report of Healthcare of Ontario Pension Plan (HOOPP). The notional value of its derivatives was $333 billion, which is over 10 times the fair value of these instruments, and over four times the fund’s net asset value. Appendix Table A1HOOPP's Use Of Derivatives* So, when derivatives are used, accounting leverage is often adjusted for derivative exposure by taking the delta-adjusted notional value of derivative contracts.3 When there are short positions, leverage can be measured as Gross and Net Leverage after derivatives exposure is delta-adjusted. Gross Leverage is defined as the total exposure of long and short positions divided by net assets. This is accurate when the long and short positions are totally separate active bets. Net Leverage, is defined as the net exposure between long and short positions, divided by net assets. This is an accurate measure of leverage when the long and short positions are taken as hedges for one another. 1  Dr. Serguei Zernov, “Leverage to Meet the Pension Promise,” Global Risk Institute, Jan 24, 2019. 2  L’her, J.F., Stoyanova, R., Shaw, K., Scott, W. and Lai, C, “A bottom-up approach to the risk-adjusted performance of the buyout fund market”, Financial Analysts Journal, July/August 2016. 3  Andrew Ang, Sergiy Gorovyy and Gregoty B. van Inwegen, “Hedge Fund Leverage,” Journal of Financial Economics, January 25, 2011. Appendix 2: Impact Of Leverage Caps And Financing Costs With Different Lookback Window Lengths In Section 5, Chart 14A, Chart 14B and Chart 15 were presented using a lookback window of 180 months, a prefered window length based on our previous research on risk parity. However, practioners have been using different lookback windows. Below are the corresponding charts showing lookback windows of 360 months and 36 months, respectively. It’s easy to see that, the shorter the lookback window, the more significant the impact of  both financing costs and leverage constraints. The reason is simple: a shorter lookback window generates much higher leverage compared to a longer lookback window. APPENDIX 2 Chart 1AImpact Of Financing With 360-Month Lookback APPENDIX 2 Chart 1BImpact Of Financing With 36-Month Lookback   APPENDIX 2 Chart 2AImpact Of Leverage Cap With 360-Month Lookback APPENDIX 2 Chart 2BImpact Of Leverage Cap with 36-Month Lookback   APPENDIX 2 Chart 3AImpact Of Financing When Leverage Capped At 3X With 360-Month Lookback APPENDIX 2 Chart 3BImpact Of Financing When Leverage Capped At 3X With 36-Month Lookback   Xiaoli Tang Associate Vice President xiaoliT@bcaresearch.com Footnotes 1 Please see Global Asset Allocation Special Report, “Return Assumptions – Refreshed And Refined,” dated June 25, 2019. 2 Ned McGuire and Brice Shirimbere, "2019 Wilshire Consulting Report on State Retirement Systems: Funding Levels and Asset Allocation," Wilshire Associates, March 2019. 3 “State Pension Funds adjust to ‘New Normal’ of Lower Returns,” Chief Investment Officer, January 2, 2020, 4 Please see Global Asset Allocation Special Report, “Return Assumptions – Refreshed And Refined,” dated June 25, 2019. 5 Martha Porado, “A look at how Canadian pension funds are using leverage,” dated Aug 10, 2018. 6GPIF (Government Pension Investment Fund) 2018 annual report. 7 "2017-18 Comprehensive Annual Financial Report,"CalPERS. p106, 2018 annual report. 8 "CalPERS Falling Short of Private Equity Goals," dated November 18, 2019. 9 Please see Global Asset Allocation Special Report, "Demystifying Risk Parity," dated May 8, 2019. 10 https://www.mosers.org/funding/annual-reports 11 Arleen Jacobius, "CalPERS shifts $150 billion as part of new strategic asset allocation," Pensions And Investments, dated August 20, 2019. 12 Edward Qian, “Rebalance and Diversification Returns of Leveraged Portfolios,” Investment Insight, Panagora, December 2011. 13https://www.mackenzieinvestments.com/content/dam/mackenzie/en/2019/03/mm-investing-in-liquid-alternatives-en.pdf 14 https://perspectives.scotiabank.com/wp-content/uploads/2018/10/Liquid-A… 15 "5 Use Cases for Liquid Alt ETFs," Institutional Investor dated November 18, 2019. 16 L’ her et al, “A bottom-up approach to the risk-adjusted performance of the buyout fund market,” Financial Analysts Journal, 72(4), 2016. 17 Please see Global Asset Allocation Special Report, “Return Assumptions – Refreshed And Refined,” dated June 25, 2019. 18 Please see Global Asset Allocation Special Report, Demystifying Risk Parity," dated May 8, 2019.

Highlights Overweighting the SIFI banks is our highest-conviction call, … : Our enthusiasm for the four banks deemed to be systemically important financial institutions is founded on the view that generous monetary and fiscal policy will lead to considerably smaller credit losses than the SIFIs’ depressed valuations imply. … but investors are none too sure of it, inside and outside of BCA: The SIFIs have underperformed the broad market since we overweighted them in late April, and they will likely run in place until our mild-credit-loss thesis can be borne out. Banks’ fortunes are not tied to the slope of the yield curve … : Banks do not borrow short to lend long and the widespread belief that their stocks are hostage to the yield curve has no empirical support. … and the US banking industry is not in structural decline: US banks have experienced steady growth in real loans, net interest income and net income. Their businesses have yet to be disrupted by new entrants; so far, technology has increased profitability and we expect that the pandemic will point the way to future efficiency improvements. Feature In response to ongoing client questions and a lively internal debate, we are devoting this week’s report to reviewing our highest-conviction call: overweighting the SIFI banks.1 After restating our thesis and what it would take to get us to abandon it, we challenge two arguments that have been cited in support of a bearish view. We hold fast to our underlying rationale, though we concede that it will likely take more time for the call to pan out. We always recommended it for investors with a time frame of at least a year, and it may take until first quarter 2021 earnings to start generating alpha, but we still believe it will. A Feature, Not A Bug Our entire editorial staff gathers every month to define the consensus view on all the major asset classes, which becomes the BCA House View until we revisit it the next month (or sooner, if need be). The House View is not a party line that we all parrot; any individual managing editor is free to express an opposing view, provided s/he clearly states that s/he is departing from the House View and, ideally, explains why. Although this policy does not always lead to neatly packaged views, it affords clients a window on our internal debates, allowing them to evaluate the merits of opposing points of view for themselves. It also helps us attract and retain the informed, opinionated researchers we seek. Banking On Washington The pandemic, and the lockdown measures imposed to limit its spread, tore a huge hole in the economy. Policymakers swiftly mobilized to build a bridge across the hole until the virus could be contained. Before March was out, the Fed had soothed the Treasury market, prized open the corporate bond market and had set bond spreads on a path to tighten. Congress passed measures providing nearly $3 trillion of aid, highlighted by the massive CARES Act. Although another significant round of federal aid is not assured, it would be in the House's, the Senate's and the White House's interest, so we expect it will eventually materialize. Thanks to the CARES Act’s copious household support, personal income reversed its March slide and comfortably exceeded February's pre-pandemic level in April, May, June and July (Chart 1). With much of the economy still in suspended animation, absent another round of direct payments to households, unemployment insurance benefit supplements, support for badly disrupted businesses and aid to state and local governments facing severe revenue shortfalls, potentially dire economic consequences loom. With even run-of-the-mill recessions dooming incumbent administrations’ election prospects, it is in the White House’s best interests to advocate for more spending to hold back the flood. Republican control of the Senate also lies in the balance. Chart 1Fiscal Transfers Have Kept Households Afloat With the Democrats seeking to demonstrate that bigger government is the solution, House, Senate and White House interests all align with the passage of a major new aid package ahead of the election. Despite the worsening climate, we expect that elected officials’ self-interest will carry the day. All creditors stand to benefit, since fiscal transfers have been vital to limiting bankruptcies and defaults, and the SIFIs would get a major boost as we attribute their dreadful year-to-date performance to market fears of credit losses well in excess of the loan loss reserves they’ve already set aside. The key to our pro-SIFIs call is that we see them as the foremost beneficiary of continued fiscal largesse. Just The SIFIs, Please We are not enamored of the entire banking industry. Low rates are likely to undermine net interest margins for an extended period and weakening loan growth, a function of borrower and lender caution, will hurt lending volumes. Banks that principally take deposits and make loans to the households and businesses within their geographic footprint will suffer. Several community banks face stiff headwinds as do some regionals. The SIFIs have quite a few earnings streams, though, and only get around half of their revenues from net interest income. They are hybrids that combine investment banks boasting bulge-bracket underwriting, top-tier sales and trading, and formidable wealth management businesses with a nationwide commercial banking footprint. These companies do not live and die by loan volumes and interest rate spreads, as much of their loan originations are securitized and their loan books are not bound to the intrinsic risk of their local economies. The SIFIs trade slightly below book value and only slightly above tangible book value (Table 1, left panel). This would be cold comfort if their book values were at risk of falling because of optimistic carrying values for their assets or impending reserve builds that would eat away at retained earnings. We are not at all worried about bad marks, however – post-GFC regulation kept the SIFIs from getting out over their skis in the just-concluded expansion – and we think that they are adequately reserved in the aggregate. Assuming that the virus will be contained by the end of the year, we stick to our initial projection that they would need to build sizable loan loss reserves only through this year's first three quarters. Table 1SIFI Book Values On their second quarter earnings calls, the SIFIs were of the view that their reserve building was nearly complete. National infection rates have remained high, however, and the supplemental federal unemployment insurance benefit has since lapsed. We expect that the rollback of re-opening measures and the interruption of CARES Act relief provisions will force the SIFIs to add to their reserves this quarter in amounts approaching first and second quarter levels, but if Congress does provide another round of meaningful aid this month or next, we think that will be the end of the big builds. Equity investors do not seem to have recognized that the SIFIs’ earnings power has allowed them to take their sizable reserve builds in stride. Book values didn’t budge in the first two quarters (Table 1, right panel), and if they continue to hold their ground, the selling in their stocks is way overdone. We are quite happy to find a group that’s so inexpensive against a backdrop in which nearly every public security is trading at elevated levels relative to history, especially when that group will be a clear winner from continuing fiscal support. If further aid on a meaningful scale is not forthcoming, however, we will exit our SIFI overweight. We are not irresolute, but we close out positions when their underlying rationale no longer applies. Psst. The Yield Curve Doesn’t Matter Old superstitions die hard. US Investment Strategy has been presenting evidence for ten years that the yield curve does not drive bank earnings.2 Although the intuition behind the view is logical, it fails to acknowledge that banks do not borrow short to lend long. As the gargantuan interest rate swap market and the FDIC’s Quarterly Banking Profile demonstrate, all but the smallest community banks rigorously match the duration of their assets and liabilities. We typically show line charts overlaying the slope of the yield curve (the 10-year Treasury yield less the 3-month T-bill rate) with aggregate net interest income or net income, showing that there has been no consistent relationship between the two series. We’ve even shown that the yield curve is largely uncorrelated with bank net interest margins. Alas, one may as well try to convince a native New Yorker that s/he is not the most important element of the universe, or an English soccer fan that his/her side is not among the favorites to capture the next World Cup. Fiscal aid has held defaults way below levels that would typically be associated with such a severe economic shock and another hearty round of it would position SIFI credit losses to come in way below the market's worst fears. This time around, we present over 60 years of monthly data in one scatterplot after another that takes the shape of an amorphous blob. They demonstrate that there is no coincident relationship between the level of the slope of the yield curve and bank stocks’ performance relative to the S&P 500 (Chart 2), or the change in the slope of the yield curve and bank stocks’ relative performance (Chart 3). They also show that there is no leading relationship over six- (Chart 4A) or twelve-month periods (Chart 4B) between the level of the slope of the yield curve and bank stocks’ relative performance. The change in the slope of the yield curve also comes a cropper with six- (Chart 5A) and twelve-month lead times (Chart 5B). With every one of the six regressions generating r-squareds below 1%, we conclude that neither the level of the slope of the yield curve, nor its direction, explains any element of relative bank stock performance. Chart 2The Steepness Of The Yield Curve Does Not Influence Bank Stocks' Relative Performance Chart 3The Change In The Steepness Of The Yield Curve Does Not Influence Bank Stocks' Relative Performance Chart 4AThe Steepness Of The Yield Curve Does Not Lead Bank Stocks' Relative Performance Over 6 Months Chart 4BThe Steepness Of The Yield Curve Does Not Lead Bank Stocks' Relative Performance Over 12 Months Chart 5AChanges In Yield Curve Steepness Do Not Lead Bank Stocks' Relative Performance Over 6 Months Chart 5BChanges In Yield Curve Steepness Do Not Lead Bank Stocks' Relative Performance Over 12 Months Rumors Of The Banks’ Structural Decline Have Been Greatly Exaggerated We submit that US banks are not in the throes of a structural decline. Adjusted for inflation, growth in their core lending business has been steady, except during recessions and their aftermath, for 70 years (Chart 6). Despite a persistent trend toward increasing non-bank intermediation that has reduced the industry’s market share, loan volumes continue to expand. Chart 6Real Bank Loan Balances Have Steadily Grown For 70 Years Industry viability is not only about sales volume, however. Participants in a declining industry could retain or even grow volumes, only to see their profits shrink in the face of competition from incumbents or new entrants. Real net interest income has continued to grow, however, more or less in line with real loan growth (Chart 7), demonstrating that margins have not eroded. Real net income, which includes credit costs and fees and other non-interest items that are more sensitive to the business cycle, is much more volatile, but has also followed a broad upward trend (Chart 8). Chart 7Real Net Interest Income Growth Has Decelerated, But It's Still Positive ... Chart 8... While Real Net Income Quickly Surpassed Its Pre-GFC Peak Futurists see fintech and cryptocurrencies as looming disruptive threats to the banking industry, but they have yet to make a significant dent in its volumes or its profits. To this point (Chart 9), technological advances have done more to reduce the industry’s operating costs than they have to undermine its moat. One would expect that a meaningful downward move in the efficiency ratio might be in store, based on what the banks have learned from the pandemic about optimizing human inputs, virtual applications and their costly branch footprints. The data do not support the claim that the industry is in the midst of a structural decline and an efficiency tailwind is likely in the offing once the acute phase of the pandemic passes. Chart 9Banks' Non-Interest Expenses Relative To Revenue Are Structurally Declining Concluding Thoughts Stocks that are oversold can become even more oversold and cheap does not necessarily mean valuable. It is entirely possible that the SIFI banks are a value trap; our call has underperformed since the late May/early June backup in long yields was summarily unwound (Chart 10). Something seems off, however, when the SIFIs are performing nearly as badly year-to-date as office and retail REITs. The latter face a structural shrinking of their businesses while banks are looking at nothing more than a cyclical ebb. Chart 10A Marathon, Not A Sprint Fiscal policymakers demonstrated their ability to counter the cyclical drag over the spring and summer; if they recover their willingness to do so, the SIFIs' outlook is far less grim than markets are currently discounting. Given our view that both the administration’s re-election prospects and Republican control of the Senate depend on staving off severe adverse economic consequences from the pandemic, we think that Congress will rediscover its resolve. If it doesn’t, we will have to close our position and potentially seek a better entry point after the new session of Congress convenes in January. It won't be all hearts and rainbows for the SIFIs over the next year, but concerns about the yield curve and the banking industry's trend earnings and revenue growth are misplaced. They are positioned to climb a wall of worry as soon as the pandemic begins to loosen its grip. Under our base-case policy scenario, the selling in the SIFIs has gone way too far. With policymakers squarely in the SIFIs’ corner, we’re thrilled to have a chance to take a shot at them from the long side below book value. The market is right to recognize that the banks will not have smooth sailing even if Congress eventually comes through, but we think it has failed to consider how much more protected the SIFIs are than their smaller brethren. If it’s holding them down because of yield curve concerns, or the idea that the banking industry is in the midst of a long-run decline, it simply has its facts wrong and we’re confident that they will rise over the next six to nine months. Doug Peta, CFA Chief US Investment Strategist dougp@bcaresearch.com     Footnotes 1     JPM, BAC, C and WFC are the commercial/universal banks that regulators have deemed systemically important. 2     Please see the February 28, 2011 US Investment Strategy Special Report, “Banks And The Yield Curve,” available at usis.bcaresearch.com.