Canada

Special Report

Highlights The Fed: The Fed will formally discuss tapering plans over the course of this summer and fall and announce the slowing of asset purchases before the end of 2021. Its labor market objectives will also be achieved in time to lift rates in 2022. Non-US Developed Markets: The central banks outside the US most likely to deliver tapering and/or outright rate hikes over the next 1-2 years are those facing housing bubbles – the Bank of Canada and Reserve Bank of New Zealand. The ECB will do nothing on rates while adjusting asset purchase programs to preserve the size of its balance sheet, while the Reserve Bank of Australia will also sit on their hands for longer. Bond Strategy Recommendations: Investors should maintain below-benchmark portfolio duration in US-only and global fixed income portfolios. Global bond investors should also favor exposure in markets where central banks will be more dovish than expected (core Europe, Australia), while limiting exposure to markets where hawkish surprises are more likely (the US, Canada, New Zealand). Feature The recovery from the 2020 COVID recession is now well underway and many investors are getting antsy about when central bankers might respond by removing monetary policy accommodation. Some central banks appear more eager than others. Both the Bank of Canada and Bank of England, for instance, have already started to reduce their rates of bond buying. Meanwhile, the US Federal Reserve is only just now starting to talk about the timing of its own tapering. This Special Report lays out a timeline for what central bank actions we should expect during the next two years. The first section focuses exclusively on the US Federal Reserve and the second section incorporates likely announcements from other central banks. Based on a comparison of our expected central bank timeline with current market prices, we conclude that investors should maintain below-benchmark portfolio duration in US-only and global fixed income portfolios. Global bond investors should also favor government bonds in countries where central banks are likely to be less hawkish than markets expect (core Europe, Australia) versus bonds from countries where hawkish surprises are more likely (US, Canada, New Zealand and, potentially, the UK and Sweden).   The Federal Reserve’s Timeline Chart 1 shows our anticipated timeline for when the Federal Reserve will make specific policy announcements between now and the start of 2024. Chart 1The Federal Reserve’s Timeline First, over the course of this summer, the Fed will initiate discussions about when to taper its asset purchases. Then, asset purchase tapering will be announced at the December 2021 FOMC meeting with purchases set to decline as of the beginning of 2022. We expect that net Fed purchases will fall to zero by the end of Q3 2022. That is, by that time the Fed will no longer be adding to its securities holdings. Rather, it will keep the size of its balance sheet constant. Then, with its balance sheet no longer growing, the Fed will begin the process of lifting interest rates. We expect the first rate hike to occur at the December 2022 FOMC meeting. Finally, some time after the fed funds rate is well above the zero bound, the Fed will try to reduce the size of its securities portfolio. How do we arrive at this timeline? Table 1A Checklist For Liftoff We start with the Fed’s forward guidance about the timing of the first rate hike (Table 1). The Fed has told us that it will lift rates off the zero bound once (i) PCE inflation is above 2%, (ii) the labor market is at “maximum employment” and (iii) inflation is expected to remain above 2% for some time. The first item on the Fed’s liftoff checklist has already been met and the third item logically follows from the other two. That is, if inflation is above 2% and the labor market is at “maximum employment” then the Fed will certainly expect inflation to remain high. This means that the second item on the Fed’s checklist is the most critical for assessing the timing of liftoff. In assessing the US labor market’s progress toward “maximum employment” we first have to define what “maximum employment” means. Based on the Fed’s communications, we infer that “maximum employment” means an unemployment rate between 3.5% and 4.5% - a range consistent with the Fed’s NAIRU estimates – and a labor force participation rate that has recovered back to pre-pandemic levels (Chart 2). Table 2 presents the average monthly growth in nonfarm payrolls that is required to reach that definition of maximum employment by specific future dates. For example, we calculate that average monthly payroll growth of 698k to 830k will cause the labor market to reach maximum employment by the end of this year. Average monthly payroll growth of 412k to 493k is required to hit the Fed’s target by the end of 2022. Chart 2Defining "Maximum Employment" Table 2Average Monthly Nonfarm Payroll Growth Required To Reach Maximum Employment By The Given Date The most recent issue of the Bank Credit Analyst posits several reasons why US employment growth will pick up steam in the coming months.1 We agree with this view and note that indicators of labor demand such as job openings, the NFIB “jobs hard to get” survey and the Conference Board’s “jobs plentiful” survey also point to accelerating employment gains.2 All told, we think that average monthly payroll growth of 412k to 493k is eminently achievable (Chart 3). This means that the Fed will hit its three liftoff criteria in time to hike rates before the end of 2022. Chart 3Max Employment By The End of 2022 Working backwards from the expected liftoff date, the Fed has said that it needs to see “substantial progress” toward the criteria listed in Table 1 before it will taper its pace of asset purchases. The definition of “substantial progress” remains somewhat unclear, but a few recent Fed communications provide some clues. First, Fed Chair Jay Powell said that he wants to see a “string of months” like the strong March employment report before it will be appropriate to reduce the pace of asset purchases. The question of how many months constitutes a “string” remains unclear, but it certainly seems plausible that we could see two or three more strong employment reports over the course of the summer. Other Fed Governors appear to agree with this timeline. Governor Randal Quarles: If my expectations about economic growth, employment, and inflation over the coming months are borne out, however, and especially if they come in stronger than I expect, then, as noted in the minutes of the last FOMC meeting, it will become important for the FOMC to begin discussing our plans to adjust the pace of asset purchases at upcoming meetings.3 Fed Vice-Chair Richard Clarida: I myself think that the pace of labor market improvement will pick up. […] It may well be the time that – there will come a time in upcoming meetings we’ll be at the point where we can begin to discuss scaling back the pace of asset purchases …4 Fed Governor Christopher Waller: The May and June jobs report[s] may reveal that April was an outlier, but we need to see that first before we start thinking about adjusting our policy stance.5 Our takeaway from these comments is that two or three more strong employment reports, say 500k or higher, would be sufficient for the Fed to more formally discuss tapering plans. Further, several Fed Governors seem to agree with our forecast that nonfarm payroll growth will accelerate in the coming months. With that in mind, it seems reasonable to expect that the Fed will discuss tapering plans over the course of the summer and fall, and that it will have seen sufficient labor market gains to announce a formal plan before the end of this year. Assuming that a tapering announcement occurs before the end of this year and that asset purchases actually start declining as of Jan 1st 2022, we estimate that the tapering process will conclude by the end of Q3 2022. That is, the Fed will hold the size of its balance sheet constant as of that date. Chart 4Balance Sheet Growth Will End Before The First Rate Hike At the very least, the Fed will certainly bring its net purchases to zero before it lifts rates. This is because it would be incoherent for the Fed to be tightening policy through its interest rate actions while it eases policy with its balance sheet strategy. Indeed, this is the roadmap that the Fed followed leading up to the 2015 rate hike cycle (Chart 4). Finally, we note that the Fed will try to reduce the size of its balance sheet only after the process of rate hikes is well underway. This will be consistent with the last tightening cycle when the Fed waited until the funds rate was 1.5% before it pared the size of its securities portfolio (Chart 4). We also want to stress that the Fed will only try to reduce the size of its balance sheet. In fact, we doubt that this process will get very far. The main reason for our skepticism is that there is an ongoing structural issue in the Treasury market where the supply of securities keeps growing while stricter regulations make it more costly for primary dealers to intermediate trades.6 In this environment, there are strong odds that Treasury market liquidity will evaporate whenever there is a significant shock to financial markets. When that happens, the Fed will be forced to support Treasury market liquidity through large-scale purchases, as was the case during last March’s market turmoil (Chart 5). In essence, the likelihood of future shocks that will necessitate Fed intervention in the Treasury market makes it unlikely that the Fed will make much progress reducing the size of its balance sheet. Chart 5Fed Had To Support Treasury Market In March 2020 Market Expectations And Investment Implications We can get a sense of how our Fed timeline compares to consensus expectations by looking at the New York Fed’s Surveys of Market Participants and Primary Dealers (Tables 3A & 3B). Respondents to these surveys expect tapering to start in early 2022, in line with our expectations, though they generally see it taking longer for net purchases to fall to zero. Respondents also expect a later Fed liftoff date than we do and don’t see the Fed trying to reduce the size of its balance sheet until well after rate hikes have begun. Table 3ASurvey of Market Participants Expected Fed Timeline Table 3BSurvey Of Primary Dealers Expected Fed Timeline But more important for investors than survey results is what is currently priced into the yield curve. In that regard, the overnight index swap curve is priced for Fed liftoff in February 2023 and a total of 75 bps of rate hikes by the end of 2023 (Chart 6). We expect rate hikes to start earlier and proceed more quickly than that, and therefore recommend running below-benchmark duration in US bond portfolios. Chart 6Market Rate Expectations The Timelines For Other Central Banks Policymakers outside the US are facing many of the same issues that the Fed is – rapidly recovering economies coming out of the pandemic, inflation overshoots, and surging asset prices. However, not every central bank will respond at the same time, or same pace, as the Fed. In Charts 7a and 7b, we show additional timelines for two of the most important non-Fed central banks: the European Central Bank (ECB) and the BoE. We see the likely dates and policy decisions playing out as follows. Chart 7AThe ECB’s Timeline Chart 7BThe Bank Of England’s Timeline European Central Bank For the ECB, the timing of its upcoming inflation strategy review is the most critical element. That report is due to be delivered in the latter half of this year, most likely in September or October (no firm release date has been announced by the ECB). It is highly unlikely that any meaningful policy changes will be implemented before that strategic review is completed. Some ECB officials have hinted that a move to a Fed-like interpretation of the ECB inflation target, tolerating overshoots of the target to make up for past undershoots, could result from the strategy review. The more likely option will be a move to an inflation target range, perhaps a 1-3% tolerance band, that offers more policy flexibility than the current target of just below 2%. This will potentially “move the goalposts” for the ECB in a way that will make monetary tightening even less likely compared to previous cycles. Looking at past ECB tightening episodes dating back to the central bank’s inception in 1998, it is clear that a majority of countries within the euro area must be seeing inflation that is high enough, with unemployment low enough, before any policy tightening can take place. Chart 8 illustrates this point, by showing “breadth” measures for unemployment and inflation across the euro area.7 Chart 8The ECB Usually Tightens When Growth AND Inflation Are Broad Based Specifically, the chart shows the percentage of euro area countries with an unemployment rate below the OECD’s estimate of full employment (second panel), the percentage of euro area countries with headline inflation higher than one year earlier (third panel) and the percentage of euro area countries with headline inflation above the ECB’s 2% target (bottom panel). We compare those breadth measures to the actual path of policy interest rates and the size of the ECB’s balance sheet (top panel). The conclusion from the chart is that the euro area is still a long way from having the sort of broad-based rise in inflation or fall in unemployment necessary to trigger a reduction in the size of its balance sheet or actual interest rate hikes. Chart 9The ECB Is Under No Pressure To Tighten Pre-Emptively Nonetheless, our expectation is that the ECB will want to begin preparing the markets for the end of the Pandemic Emergency Purchase Program (PEPP) - which has been buying government bonds since March 2020 in a less constrained fashion than previous asset purchase programs - shortly after the inflation strategy review is concluded. Much of the euro area economy is already showing signs of rapid recovery from pandemic induced lockdowns, amid an accelerating pace of vaccinations. On top of that, the Next Generation European Union (NGEU) recovery fund is set to begin distributing funds in the final quarter of 2021, providing a meaningful lift to government investment and expected growth in 2022. It will be difficult for the ECB to justify the need for an “emergency” program like the PEPP to continue against such a growth backdrop, especially with euro area inflation no longer at the depressed levels seen in 2020. We expect the ECB to begin preparing the market for the end of PEPP heading into the December 2021 ECB policy meeting, when it will be announced that the program will not be renewed when it expires in March 2022 (Chart 9). As always for such major policy announcements, the ECB will wish to do so when there is a new set of economic forecasts used to justify any changes. This is why December – the first meeting after the strategic review is completed that will also have new forecasts – is the earliest realistic date for an announcement on the PEPP. The communication around the PEPP announcement will need to be delicate, as the PEPP has significantly increased the ECB’s footprint in European bond markets. The share of government bonds owned by the ECB has increased by anywhere from five to ten percentage points since the PEPP began (Chart 10). We expect the ECB will be forced to expand its existing Public Sector Purchase Program (PSPP) to make up for the eventual disappearance of the PEPP. This means that the PEPP will be effectively “rolled into” the PSPP, to limit the damage from a likely post-PEPP surge in bond yields in the more fragile markets like Italy, Spain and even Greece – especially with the euro now trading close to pre-2008 highs on a trade-weighted basis (Chart 11). Chart 10The PEPP Can Expire, But Cannot Disappear Chart 11ECB Must Avoid A 'PEPP Taper Tantrum' There is a chance that the ECB will want to avoid any “PEPP taper tantrum” in Peripheral European yields (and spreads versus Germany) by making an announcement on PEPP expiry and PSPP expansion at the same meeting. If that happens, we suspect it would happen in December of this year rather than sometime in the first quarter of 2022. Beyond that, the ECB will likely seek to keep financial conditions as accommodative as possible by keeping policy interest rates unchanged well into 2023, with an actual rate hike not likely until mid-2024 at the earliest. The ECB could deliver a more modest form of “tightening” before then by letting some of the cheap bank funding programs (TLTROs) expire. Although we suspect that even those programs will need to be renewed, perhaps at less attractive financing terms, to prevent an unwanted tightening of credit conditions in the euro area banking system. Bank Of England Chart 12BoE Forecasts Are Conservative Having already announced a tapering of the pace of its bond buying in early May, the BoE is likely to continue along that path over the next year. We expect the BoE, like the ECB, to make any future taper announcements when new sets of economic forecasts are published in Monetary Policy Reports. Thus, the next taper announcements are expected in August 2021, November 2021 and February 2022, with a full tapering down to zero net purchases (new buying only replacing maturing bonds) by May 2022 at the latest. The first rate hike will occur between 6-12 months after the end of tapering, possibly as early as November 2022 but, more likely in our view, sometime closer to mid-2023. The most recent set of BoE economic forecasts calls for headline UK CPI inflation to rise to 2.3% in 2022 before settling down to 2% in 2023 and 1.9% in 2024 (Chart 12). This would be a mild inflation outcome by recent UK standards during what will certainly be a period of strong post-pandemic growth over the next 12-18 months. Longer-term inflation expectations, both survey-based and extracted from CPI swaps and inflation-linked Gilts, are priced for a bigger inflation upturn above 3%. The BoE has been one of the least active central banks in the developed world since the 2008 financial crisis. The BoE main policy rate, the Bank Rate, has been no higher than 0.75% since then, even with the BoE threatening to lift rates to higher levels many times under the leadership of former Governor Mark Carney when inflation was overshooting the bank’s 2% target. Of course, the Brexit uncertainty since mid-2016 effectively tied the hands of the central bank and prevented any possible policy tightening. Now that Brexit has actually happened, however, the BoE has more flexibility to respond to developments with UK economic growth and inflation, as needed. A possible path for the UK Cash Rate was laid out in a recent speech by BoE Monetary Policy Committee (MPC) member Gertjan Vlieghe.8 He triggered a selloff across the Gilt market with his comment that a BoE rate hike could occur as early as Q2 2022 – with the Bank Rate rising to 1.25% from the current 0.1% by 2024 - under more optimistic scenarios for UK growth and employment. His base case, however, was that the coming uptick in UK inflation will prove to be temporary, but that a move towards full employment will make the first hike more likely toward the end of 2022 with modest rate increases in 2023 and 2024 that will take the Bank Rate to 0.75% (Chart 13). Chart 13Gilts Are Vulnerable To A Hawkish Surprise Vlighe’s base case scenario on growth and interest rates is in line with the BoE’s current forecasts that call for spare capacity in the UK economy to be fully eliminated by mid-2022, with rate hikes to begin in mid-2023. That is broadly in line with our projected BoE timeline and with current pricing in the UK OIS curve, although we see risks tilted towards faster growth and inflation – and the BoE moving more aggressively than projected – over the next 12-18 months. Other Major Developed Market Central Banks Looking beyond the “Big Three” of the Fed, ECB and BoE, central bank timelines have become increasingly dependent on a single factor – the strength of domestic housing markets. House prices are booming in Canada, New Zealand and Sweden, with valuation measures like the ratio of median house prices to median incomes soaring to historical extremes according to the OECD (Chart 14). House prices are also climbing fast in the US and UK, but the valuation measures have not surpassed the peaks seen during the mid-2000s housing bubble. The housing boom has already motivated some central banks to respond by turning less dovish sooner than expected, even with unemployment rates still above pre-pandemic peaks (Chart 15).9 The BoC noted that soaring Canadian housing values motivated the taper announcement in April. The Reserve Bank of New Zealand (RBNZ) has come under political pressure over the growing unaffordability of New Zealand homes, with the government changing the central bank’s remit earlier this year to force the RBNZ to explicitly consider house price inflation when setting monetary policy. Chart 14Surging House Prices Can Turn Doves Into Hawks Chart 15These CBs Could Turn More Hawkish Before Reaching Full Employment We expect more tapering announcements from the BoC over the latter half of 2021, with a first rate hike likely sometime in the first quarter of 2022. We see the RBNZ moving aggressively, as well, tapering over the remainder of 2021 before lifting rates by the spring of 2022 at the latest. Sweden’s Riksbank will be the next central bank to turn more hawkish because of surging home values, although they will lag the pace of the BoC and RBNZ with Sweden only now beginning to emerge from lockdowns associated with a third wave of COVID-19 cases. Importantly, Australia – a country that has dealt with house price surges in the past – has seen house price valuations retreat over the past few years, even with the Reserve Bank of Australia (RBA) slashing policy rates to historic lows. The RBA also introduced yield curve control in 2020 to anchor the level of short-term bond yields, while also engaging in outright bond purchases to mitigate the rise in longer-term bond yields. With Australian inflation still remaining well below target in a year of rising global inflation, and with subdued labor costs likely to keep price pressures moderate over the next 12-18 months, we expect the RBA to move very slowly on both tapering and rate hikes. Finally, for completeness, we should note that we do not expect any policy changes from the Bank of Japan (BoJ) over the next two years, with inflation likely to remain far below the central bank’s 2% target. Non-US Investment Implications In Table 4, we show the timing of the first rate hike (i.e. “liftoff”), and the subsequent amount of total rate hikes to the end of 2024, as currently discounted in the OIS curves of the eight countries discussed in this report. We rank the countries in the table in order of liftoff dates, starting with the closest to today. Table 4The “Pecking Order” Of Central Bank Rate Hikes The RBNZ is expected to hike first in May 2022, followed by the BoC (September 2022), the Fed (February 2023), the RBA (April 2023), the Riksbank (May 2023), the BoE (May 2023), the ECB (June 2023) and the BoJ (October 2025). The cumulative amount of rate hikes discounted to the end of 2024 rank similarly: more rate increases are expected in New Zealand (167bps), Canada (150bps), the US (137bps) and Australia (113bps); while fewer rate increases are expected in the Sweden (63bps), the UK (61bps), the euro area (31bps) and Japan (7bps). According to our various central bank timelines discussed in this report, we see the risks of a rate hike coming sooner than discounted by markets in the US, Canada and New Zealand. We see central banks moving slower than markets expect in the euro area and Australia, while we see Sweden and UK priced in line with our base case views (although we see risks tilted towards a more hawkish turn faster than expected in the latter two). The story is the same in terms of cumulative rate hikes discounted in OIS curves, with markets not pricing in enough rate hikes in New Zealand, Canada and the US – and, possibly, Sweden and the UK – while pricing too many hikes in Australia and the euro area. This leads us to recommend the following country allocations in a global government bond portfolio: Underweight the US, Canada and New Zealand Overweight Australia and core Europe (and Japan) Neutral Sweden and the UK, but with a bias to downgrade. Ryan Swift US Bond Strategist rswift@bcaresearch.com Robert Robis, CFA Chief Fixed Income Strategist rrobis@bcaresearch.com Footnotes 1 Please see The Bank Credit Analyst June 2021 Monthly Report, "Global House Prices: A New Threat For Policymakers", dated May 27, 2021. 2 Please see US Bond Strategy Weekly Report, “Lower For Longer, Then Faster Than You Think”, dated May 25, 2021. 3 https://www.federalreserve.gov/newsevents/speech/quarles20210526b.htm 4 https://ca.news.yahoo.com/federal-reserve-vice-chair-richard-clarida-yahoo-finance-transcript-may-2021-173007192.html 5 https://www.federalreserve.gov/newsevents/speech/waller20210513a.htm 6 For a longer discussion of Treasury market liquidity issues please see US Investment Strategy / US Bond Strategy Special Report, “Alphabet Soup 2: Shocked And Awed”, dated July 28, 2020. 7 For more details, please see Global Fixed Income Strategy Report, “ECB Outlook: Walking On Eggshells”, dated May 19, 2021. 8 The full speech can be found here: https://www.bankofengland.co.uk/speech/2021/may/gertjan-vlieghe-speech-hosted-by-the-department-of-economics-and-the-ipr 9 For more details on the global housing boom, see Global Fixed Income Strategy Special Report, “Global House Prices: A New Threat For Policymakers”, dated May 28, 2021. Fixed Income Sector Performance Recommended Portfolio Specification

Special Report

Highlights House prices are rising rapidly across the developed markets, in response to the extraordinary monetary and fiscal policy stimulus implemented to fight the pandemic. Evidence points to the house price surge being driven by monetary policy that has left real interest rates far below equilibrium levels. Supply factors are a secondary cause of the house price boom. Financial stability risks stemming from rising house prices are less acute than the pre-2008 experience, as overall household leverage has grown more slowly during the pandemic and global banks are better capitalized. Rapidly rising house prices are forcing some central banks to turn less accommodative earlier than expected. The recent hawkish turns by the Bank of Canada and Reserve Bank of New Zealand may be canaries in the coal mine for other central banks – perhaps even the Fed – if house prices and household leverage start rising together. Feature The COVID-19 pandemic led to the sharpest economic recession since World War II, alongside an enormous rise in unemployment. Consensus expectations call for the output gap to be closed (or mostly closed) in most advanced economies by the end of this year, but it remains an open question how quickly these economies will be able to return to full employment amid potentially permanent shifts in demand for office space and goods sold at physical, “brick and mortar” retail locations. Despite this sizeable and swift economic shock, house price appreciation accelerated last year in the developed world. Chart 1 highlights that US house prices rose at an 18% annualized pace in the second half of 2020, whereas they accelerated at a high-single digit pace in developed markets ex-US (on a GDP-weighted basis). This, in conjunction with a sharp rise in the household sector credit-to-GDP ratio (Chart 2), has unnerved some investors while raising questions about the implications for monetary policy. Chart 1House Prices Are Surging Around The World Chart 2Rising Fears About Deteriorating Household Balance Sheets Before we discuss the investment implications of the global housing boom, however, we must first accurately determine the reasons why it is happening. The Work-From-Home Effect: Less Than Meets The Eye When analyzing the surprising behavior of the housing market last year, the working-from-home effect brought upon by the pandemic emerges as an obvious factor potentially explaining house price gains. Last year, following recommended or mandatory stay-at-home orders from governments, most office-based businesses rapidly shifted to work-from-home arrangements as an emergency response. However, in the month or two following the beginning of stay-at-home orders, several national US surveys found many office workers preferred the flexibility afforded by work-from-home arrangements. Many employers, correspondingly, found that the productivity of their employees did not suffer while working from home, or that it even improved. Several prominent corporations in the US have subsequently made some work-from-home options permanent, or even allowed employees to work from offices in a different city than they did prior to the pandemic. Newfound work-from-home options have undoubtedly created new demand for housing, and thus explained the surge in house prices seen over the past year in the minds of some investors. However, in our view, evidence from the US, the UK, and France suggests that the work-from-home effect better explains differences in price gains across housing types and within large metropolitan areas, rather than aggregate or national-level changes in house prices. Chart 3 provides some quantification of the impact of work-from-home policies by plotting US resident migration patterns by city. This data has been compiled by CBRE, and the impact of COVID is shown as the change in net move-ins from 2019 to 2020 per 1000 people. This helps control for the underlying migration pattern that existed in US cities prior to the pandemic. Chart 3Work From Home Policies Have Impacted Migration Trends… The chart highlights that the negative migration impact from COVID has been mostly concentrated in New York City and the three most populous cities on the West Coast (by metro area): Los Angeles, San Francisco, and Seattle. And yet, Chart 4 highlights that house price inflation in these four cities has accelerated to a double-digit pace, only modestly below the national average. Chart 4...But Cities With Outward Migration Still Have Very Strong House Price Gains The house price indexes shown in Chart 4 represent aggregate, metro area trends, and clearly some regions within these metro areas have experienced house price deceleration or outright deflation versus gains in areas outside the urban core. But Chart 5 highlights that house prices have declined in Manhattan basically in line with the change in net move-ins as a share of the population, underscoring that double-digit metro area-wide house price gains appear to be vastly disproportionate to changes in net migration. Similarly, Chart 6 highlights that rents decelerated in the US over the past year but remained in positive territory and grew at a 3.5% annualized rate from February to April. Chart 5In Manhattan, House Prices Have Tracked Net Migration Chart 6Rent Costs Have Decelerated, But Have Not Contracted Evidence from Paris and London also suggests that a work-from-home effect is insufficient to explain broad house price gains. Panel 1 of Chart 7 highlights that house prices in France have accelerated significantly, but that apartment prices have decelerated only fractionally in lockstep. Panel 2 shows that the acceleration in house prices does reflect a work-from-home effect, as prices have risen faster in inner Parisian suburbs. Panel 3, however, highlights that Parisian apartment prices, the dominant property type in the urban core, have decelerated modestly. Chart 8 highlights that house price gains have not even decelerated in greater London; they have been merely been modestly outstripped by gains in Outer South East (outside of the Outer Metropolitan Area). Chart 7In France, Parisian Apartment Prices Are Simply Lagging, Not Falling Chart 8In The UK, Greater London Property Prices Are Accelerating     The Policy Effect: The Fundamental Driver Of The Housing Market Despite the broader location flexibility that work-from-home policies now provide to potential homeowners, it seems inconceivable that the housing market would have responded in the manner that it has over the past year given the size of the economic shock brought on by the pandemic without significant support from policy. Above-the-line fiscal measures to the pandemic have totaled in the double-digits in advanced economies (Chart 9), and monetary policy has contributed to easier financial conditions via rate cuts, asset purchases, and sizeable programs to support financial market liquidity. Chart 9There Has Been A Massive Fiscal Policy Response To The Crisis In fact, Charts 10-13 present compelling evidence that fiscal and monetary policy have been the core drivers of significant house price gains over the past year. Charts 10 and 11 plot the above-the-line fiscal response of advanced economies against the year-over-year growth rate in house prices as well as its acceleration (the change in the year-over-year growth rate). The charts show a clearly positive relationship, with a stronger link between the pandemic fiscal response and the acceleration in house prices. Chart 10Differences In Last Year’s Fiscal Response… Chart 11…Help Explain Differences In House Price Gains Chart 12Pre-Pandemic Differences In The Monetary Policy Stance… Chart 13…Do An Even Better Job Of Explaining 2020 House Price Gains   Charts 12 and 13 highlight the even stronger link between house prices and the pre-pandemic monetary policy stance in advanced economies, defined as the difference between each country’s 2-year government bond yield and its Taylor Rule-implied policy interest rate as of Q4 2019. We construct each country’s Taylor Rule using the original specification, with core consumer price inflation, a 2% inflation target, and real potential GDP growth as the definition of the real equilibrium interest rate. The charts make it clear that easy monetary policy strongly explains house price gains in 2020, particularly the year-over-year percent change rather than its acceleration. This makes sense, given that monetary policy was already quite easy in many countries at the onset of the pandemic – meaning that changes were less pronounced than they would have been had interest rates been higher. The explanation that emerges from Charts 10-13 is that historic fiscal easing, combined with an easy starting point for monetary policy – that became even easier last year – enabled demand from work-from-home policies to manifest during an extremely severe recession. We agree that work-from-home policies have shifted the geographic preferences of some home buyers and likely provided a new source of net demand from renters in urban cores purchasing homes in outlying areas. But we strongly doubt that the net effect of work-from-home policies in the midst of an extreme shock to economic activity would have caused the rise in house prices that we have observed, certainly not to this level, without major support from policy. This underscores that policy, and not the work-from-home effect, has and will likely remain the core driver of the global housing market. The Supply Effect: Mostly A Red Herring Chart 14Countries Fall Into Two Groups In Terms Of The Relative Trend In Real Residential Investment One perennial question that emerges when analyzing the housing market, particularly in markets with outsized house price gains, is the impact of constrained supply. It is frequently argued that constrained supply is squeezing prices higher in many markets, and that the appropriate policy solution to extreme house price gains is to enable widespread housing construction – not to raise interest rates. We do not rule out the potential impact of constrained supply in certain cities or regional housing markets, and we have highlighted in previous research that a positive relationship does exist between population density in urban regions and median house price-to-income ratios.1 But as a broad explanation for supercharged house price gains, the supply argument appears to fall flat. Chart 14 presents the most standardized measure of cross-country housing supply available for several advanced economies, the trend in real residential investment relative to real GDP over time. These series are all rebased to 100 as of 1997, prior to the 2002-2007 US housing market boom. The chart makes it clear that advanced economies generally fall into two groups based on this metric: those that have seen declines in real residential investment relative to GDP, especially after the global financial crisis (panel 1), and those that have experienced either an uptrend in housing construction relative to output or have seen a flat trend (panel 2). If scarce housing supply was the core driver of outsized house price gains, then we would expect to see stronger gains in the countries shown in panel 1 and smaller gains in the countries shown in panel 2. In fact, mostly the opposite is true: Charts 15 and 16 highlight that the relationship between the level of these indexes today relative to their 1997 or 2005 levels is positively related to the magnitude of house price gains last year, suggesting that housing market supply has generally been responding to demand over the past decade. The US and possibly New Zealand stand as possible exceptions to the trend, suggesting that relatively scarce supply may be boosting prices even further in these markets beyond what fiscal and monetary policy would suggest. Chart 15Countries That Have Seen A Stronger Pace Of Residential Investment… Chart 16…Have Experienced Stronger House Price Gains   Chart 17Is This Not Enough Supply, Or Too Much Demand? As a final point about the inclination of investors to gravitate towards supply-side arguments related to the housing market, Chart 17 presents a simple thought experiment. The chart shows a simple housing supply-demand curve diagram, in a scenario where the demand curve for housing has shifted out more than the supply curve has (thus raising house prices). Is this a scenario in which supply is too tight? Or is it a case in which demand is too strong? In our view, the tight supply answer is reasonable in circumstances where the increase in demand is normal or otherwise sustainable. But Charts 10-13 clearly showed that housing demand is being boosted by easy policy, which in the case of some countries has occurred for years: interest rates have remained well below levels that macroeconomic theory would traditionally consider to be in equilibrium, and this has occurred alongside significant household sector leveraging (Chart 18). As such, in our view, investors should be more inclined to view the global housing market as generally being driven by demand-side rather than supply-side factors. This Is Not 2007/08 … Yet We highlighted in Chart 2 above that the household sector debt-to-GDP ratio increased sharply last year, which has raised some questions about debt sustainability among investors. For the most part, the rise in this ratio actually reflects denominator effects (namely a sharp contraction in nominal GDP) rather than a huge surge in household debt. Chart 19 shows BIS data for the annual growth in total household debt in developed economies was roughly stable last year, at least until Q3 (the most recent datapoint available from the BIS). Chart 18Low Interest Rtaes Have Fueled Household Leveraging Chart 19Total Credit Growth Has Been Stable, But Mortgage Credit Growth Is Accelerating Chart 20US Mortgage Growth Is Picking Up, As Repayments Slow Consumer Credit Growth But Chart 19 shows the recent trend in total household debt, which masks diverging mortgage and non-mortgage debt trends. In the US, euro area, Canada, and Sweden, household mortgage debt has accelerated to varying degrees, underscoring that households have likely paid down non-mortgage debt with some of the savings that they have accumulated from a significant reduction in spending on services. Chart 20 shows this effect directly in the case of the US; mortgage debt growth accelerated by roughly 1.5 percentage points in the second half of the year, whereas consumer credit growth (made up of student loans, auto loans, credit cards, and other revolving credit) decelerated significantly. This aligns with data showing that US households have used some of their savings windfall to pay down their credit card balances. This changing mix within household debt - less higher-interest-rate consumer credit, more lower-interest-rate collateralized mortgage debt – could, on the margin, help mitigate financial stability risks from the housing boom by moderating overall debt service burdens. The starting point for the latter matters, though, in accurately assessing the risks from rising house prices and increased mortgage debt, particularly in countries where household debt levels are already high. According to data from the BIS, the US already has one of the lowest household debt service ratios (7.6%) among the developed economies (Chart 21).2 This compares favorably to the double-digit debt service ratios in the “higher-risk” countries like Canada (12.6%), Sweden (12.1%) and Norway (16.2%). On top of that, US commercial banks have become far more prudent with mortgage loan underwriting standards since the 2008 financial crisis. The New York Fed’s Household Debt and Credit report shows that an increasing majority of mortgage lending made by US banks since the 2008 crisis has been to those with very high FICO credit scores (Chart 22). This is in sharp contrast to the steady lending to “subprime” borrowers with poor credit scores that preceded the 2008 financial crisis. The median FICO score for new mortgage originations as of Q1 2021 was 788, compared to 707 in Q4 2006 at the peak of the mid-2000s US housing boom. Chart 21Diverging Trends In Global Household Debt Servicing Costs Chart 22US Banks Have Become More Prudent With Mortgage Lending   US bank balance sheets are also now less directly exposed to a fall in housing values. Residential loans now represent only 10% of the assets on US bank balance sheets, compared to 20% at the peak of the last housing bubble (Chart 23). This puts the US in the “lower-risk” group of countries in Europe, the UK and Japan where mortgages are less than 20% of bank balance sheets. This compares favorably to the “higher risk” group of countries where residential loans are a far larger share of bank assets (Chart 24), like Canada (32%), New Zealand (49%), Sweden (45%) and Australia (40%). Chart 23Banks Have Limited Direct Exposure To Housing Here Chart 24Banks Are Far More Exposed To Housing Here   Like nature, however, the financial ecosystem abhors a vacuum. “Non-bank” mortgage lenders have filled the void from traditional US banks reducing their lending to lower-quality borrowers, and they now represent around two-thirds of all US mortgage origination, a big leap from the 20% origination share in 2007. Non-bank lenders have also taken on growing shares of new mortgage origination in other countries like the UK, Canada and Australia. Chart 25Global Banks Can Withstand A Housing Shock Non-bank lenders do not take deposits and typically fund themselves via shorter-term borrowings, which raises the potential for future instability if credit markets seize up. These lenders also, on average, service mortgages with a higher probability of default, so they are exposed to greater credit losses when house prices decline. However, the risk of a full-blown 2008-style commercial banking crisis, with individual depositors’ funds at risk from a bank failure, are reduced with a greater share of riskier mortgage lending conducted by non-bank entities. This is especially true with global commercial banks far better capitalized today, with double-digit Tier 1 capital ratios (Chart 25), thanks to regulatory changes made after the Global Financial Crisis. Net-net, we conclude that the overall financial stability implications of the current surge in house prices in the developed economies are relatively modest on average. The acceleration in mortgage growth has occurred alongside reductions in non-mortgage growth, at a time when banks are better able to withstand a shock from any sustained future downturn in house prices. However, if house prices continue to accelerate and new homebuyers are forced to take on ever increasing amounts of mortgage debt, financial stability issues could intensify in some countries. Services spending will recover in a vaccinated post-COVID world, as economies reopen and consumer confidence improves, which will likely end the trend of falling non-residential consumer debt offsetting rising mortgage debt in countries like the US and Canada. Overall levels of household debt could begin to rise again relative to incomes, building up future financial stability risks when central banks begin to normalize pandemic-related monetary policies – a process that has already started in some countries because of the housing boom. The Monetary Policy Implications Of Surging House Prices Rapidly appreciating house prices are becoming an area of concern for policymakers in countries like Canada and New Zealand, where the affordability of housing is becoming a political, as well as an economic, issue. In the case of New Zealand, the government has actually altered the remit of the Reserve Bank of New Zealand (RBNZ) to more explicitly factor in the impact of monetary policy on housing costs. The Bank of Canada announced in April that it would taper its pace of government debt purchases and signaled that its decision was based, at least in small part, on signs of speculative behavior in Canada’s housing market. Macroprudential measures like limiting loan-to-value ratios of new mortgage loans are a policy option that governments in those countries have already implemented to try and cool off housing demand. Yet while such measures can help alleviate demand-supply mismatches in certain cities and regions, the efficacy of such measures in sustainably slowing the ascent of house prices on a national scale is unclear. In the April 2021 IMF Global Financial Stability Report, researchers estimated that, for a broad group of countries, the implementation of a new macro-prudential measure designed to cool loan demand reduced national household debt/GDP ratios by a mere one percentage point, on average, over a period encompassing four years.3 If macroprudential measures are that ineffective in sustainably reducing demand for mortgage loans, then the burden of slowing house price appreciation will have to fall on the more blunt instruments of monetary policy. Importantly, surging house price inflation is not likely to give a boost to realized inflation measures – an important issue given the current backdrop of rapidly rising realized inflation rates in many countries. Housing costs do represent a significant portion of consumer price indices in many developed countries, ranging from 19% in New Zealand to 33% in the US (Chart 26), with the euro area being the outlier with housing having a mere 2% weighting in the headline inflation index. Chart 26A Limited Impact On Actual Inflation From Housing Yet those so-called “housing” categories overwhelmingly measure only housing rental costs and not actual house prices. This is an important distinction because rents – which are often imputed measures like in the US and not even actual rental costs - are rising at a far slower pace than actual house prices in most countries, so the housing contribution to realized inflation is relatively modest. So the good news is that booming house prices will not worsen the acceleration of realized global inflation that has concerned investors and policymakers in 2021. Yet that does not mean that central bankers will not be forced to tighten policy to cool off red-hot housing demand that is clearly being fueled by persistently negative real interest rates. In Chart 27 and Chart 28, we show both nominal and real policy interest rates for the “lower risk” and “higher risk” country groupings that we described earlier. The real policy rates are nominal policy rates versus realized headline CPI inflation. The dotted lines in the charts represent the future path of rates discounted by markets. Specifically, the projection for nominal rates is taken from overnight index swap (OIS) forward curves, while the projection for real rates is calculated by subtracting the discounted path of inflation expectations extracted from CPI swap forwards. Chart 27Markets Discounting Negative Real Rates For The Next Decade Chart 28Negative Real Rates Are Unsustainable During A Housing Bubble   There are two key takeaways from these charts: Real policy interest rates are at or very close to the most deeply negative levels seen since the 2008 financial crisis. Markets are discounting that real rates will be at or below 0% for most of the next decade. Admittedly, there is room for debate over what the equilibrium level of real interest rates (a.k.a. “r-star”) should be in the coming years. However, we deem it a major stretch to believe that real rates need to be persistently low or negative for the next ten years to support even trend growth across the developed economies. In our view, the current boom in housing demand and mortgage borrowing provides clear evidence that negative real rates are below equilibrium and, thus, are stimulating credit demand. Thus, the only way for a central bank to cool off housing demand will be to raise both nominal and, more importantly, real interest rates. Canada and New Zealand will be the “canaries in the coal mine” among developed market central banks for such a move. According to the latest Bank of Canada Financial Stability Review, nearly 22% of Canadian mortgages are highly levered, with a loan-to-value ratio greater than 450%, a greater share of such mortgages than during the 2016/17 housing boom (Chart 29). Canadian house prices have risen to such an extent that home prices in major cities like Toronto, Vancouver and Montreal are among the most expensive in North America.4  Stunningly, a recent Bloomberg Nanos opinion poll revealed that nearly 50% of Canadians would support Bank of Canada rate hikes to cool off the red-hot housing market (Chart 30). The central bank will be unable to resist the pressure to use monetary policy to slam on the brakes of the housing market – investors should expect more tapering and, eventually, rate hikes from the Bank of Canada over at least the next couple of years. Chart 29Canadians Are Leveraging Up To Buy Expensive Homes Chart 3050% Of Canadians Want A Rate Hike To Cool Housing   In New Zealand, worsening housing affordability has reached a point where a 20% down payment on the median national house price is equal to 223% of median disposable income (Chart 31). This is forcing more first-time home buyers to take on levels of mortgage debt that the RBNZ deems highly risky (top panel). Like the Bank of Canada, the RBNZ will prove to be one of the most hawkish central banks in the developed world over the next couple of years as the central bank follows their newly-revised remit to try and cool off housing demand in New Zealand. Who is next? Housing values, measured by the ratio of median national house prices to median national household incomes, are rising in the US and UK but are still below the peaks of the mid-2000s housing bubble (Chart 32). Meanwhile, housing is becoming more expensive across the euro area, but not in a consistent manner, with valuations in Germany and Spain having increased far more than in France or Italy. Housing valuations have actually improved in Australia over the past couple of years on a price-to-income basis. The most likely candidates for a housing-related hawkish turn are in Scandinavia, with housing valuations in Sweden and Norway closing in on Canada/New Zealand levels. Chart 31New Zealand Housing Is Wildly Unaffordable Chart 32Global House Price/Income Ratios Are Trending Higher   Investment Conclusions The current acceleration in global house prices is an inevitable outcome of the extraordinary monetary and fiscal easing implemented during the pandemic. Higher realized inflation is pushing real rates deeper into negative territory in many countries, fueling the demand for housing. Central banks in countries with more stretched housing valuations will be forced to turn more hawkish sooner than expected, leading to tapering and, eventually, rate hikes to cool housing demand. This has negative implications for government bond markets in countries where housing is more expensive and real yields remain too low, like Canada, New Zealand and Sweden (Chart 33). Investors should limit exposure to government bonds in those markets over the next 6-12 months. Chart 33Negative Real Yields & Expensive Housing Valuations – An Unsustainable Mix Bond markets in countries where house prices are not rising rapidly enough to force policymakers to turn more hawkish more quickly – like core Europe, Australia and even Japan - are likely to be relative outperformers. The US and UK are “cuspy” bond markets, as housing valuations are becoming more expensive in those two countries but the Fed and Bank of England are not facing the same domestic political pressure to use monetary policy tools to fight the growing unaffordability of housing. That could change, though, if overall household leverage begins to rise alongside house price inflation as the US and UK economies emerge from the pandemic. Current pricing in OIS curves shows that markets expect the RBNZ and Bank of Canada to begin hiking rates in May 2022 and September 2022, respectively (Table 1). This is well ahead of expectations for “liftoff” from other developed markets central banks, including the Fed in April 2023. The cumulative amount of rate hikes following liftoff to the end of 2024 is highest in Canada, New Zealand, the US and Australia. Those are also countries with currencies that are trading at or above the purchasing power parity levels derived from our currency strategists’ valuation models. This highlights the difficult choice that central bankers facing housing bubbles must confront, as the rate hikes that will help cool off housing demand will lead to currency appreciation that could impact other parts of their economies like exports and manufacturing. Table 1Hawkish Central Banks Must Live With Currency Strength Tracking the second-round economic consequences of eventual monetary policy actions to control excessive house price inflation, particularly in “higher risk” countries, is likely to be the subject of future Bank Credit Analyst / Global Fixed Income Strategy reports. Robert Robis, CFA Chief Fixed Income Strategist Jonathan LaBerge, CFA Vice President The Bank Credit Analyst   Footnotes 1 Please see Global Investment Strategy "Canada: A (Probably) Happy Moment In An Otherwise Sad Story," dated July 14, 2017, available at gis.bcaresearch.com 2 Importantly, the BIS debt service ratios include the payment of both principal and interest, thus making it a true measure of debt service costs that includes repayment of borrowed funds – a critical issue in countries with high loan-to-value ratios for home mortgages. 3 Please see page 46 of Chapter 2 of the April 2021 IMF Global Financial Stability Report, which can be found here: https://www.imf.org/en/Publications/GFSR/Issues/2021/04/06/global-financial-stability-report-april-2021 4 “Vancouver, Toronto and Hamilton are the least affordable cities in North America: report”, CBC News, May 20, 2021

Special Report

Highlights House prices are rising rapidly across the developed markets, in response to the extraordinary monetary and fiscal policy stimulus implemented to fight the pandemic. Evidence points to the house price surge being driven by monetary policy that has left real interest rates far below equilibrium levels. Supply factors are a secondary cause of the house price boom. Financial stability risks stemming from rising house prices are less acute than the pre-2008 experience, as overall household leverage has grown more slowly during the pandemic and global banks are better capitalized. Rapidly rising house prices are forcing some central banks to turn less accommodative earlier than expected. The recent hawkish turns by the Bank of Canada and Reserve Bank of New Zealand may be canaries in the coal mine for other central banks – perhaps even the Fed – if house prices and household leverage start rising together. Feature The COVID-19 pandemic led to the sharpest economic recession since World War II, alongside an enormous rise in unemployment. Consensus expectations call for the output gap to be closed (or mostly closed) in most advanced economies by the end of this year, but it remains an open question how quickly these economies will be able to return to full employment amid potentially permanent shifts in demand for office space and goods sold at physical, “brick and mortar” retail locations. Despite this sizeable and swift economic shock, house price appreciation accelerated last year in the developed world. Chart II-1 highlights that US house prices rose at an 18% annualized pace in the second half of 2020, whereas they accelerated at a high-single digit pace in developed markets ex-US (on a GDP-weighted basis). This, in conjunction with a sharp rise in the household sector credit-to-GDP ratio (Chart II-2), has unnerved some investors while raising questions about the implications for monetary policy. Chart II-1House Prices Are Surging Around The World Chart II-2Rising Fears About Deteriorating Household Balance Sheets Before we discuss the investment implications of the global housing boom, however, we must first accurately determine the reasons why it is happening. The Work-From-Home Effect: Less Than Meets The Eye When analyzing the surprising behavior of the housing market last year, the working-from-home effect brought upon by the pandemic emerges as an obvious factor potentially explaining house price gains. Last year, following recommended or mandatory stay-at-home orders from governments, most office-based businesses rapidly shifted to work-from-home arrangements as an emergency response. However, in the month or two following the beginning of stay-at-home orders, several national US surveys found many office workers preferred the flexibility afforded by work-from-home arrangements. Many employers, correspondingly, found that the productivity of their employees did not suffer while working from home, or that it even improved. Several prominent corporations in the US have subsequently made some work-from-home options permanent, or even allowed employees to work from offices in a different city than they did prior to the pandemic. Newfound work-from-home options have undoubtedly created new demand for housing, and thus explained the surge in house prices seen over the past year in the minds of some investors. However, in our view, evidence from the US, the UK, and France suggests that the work-from-home effect better explains differences in price gains across housing types and within large metropolitan areas, rather than aggregate or national-level changes in house prices. Chart II-3 provides some quantification of the impact of work-from-home policies by plotting US resident migration patterns by city. This data has been compiled by CBRE, and the impact of COVID is shown as the change in net move-ins from 2019 to 2020 per 1000 people. This helps control for the underlying migration pattern that existed in US cities prior to the pandemic. Chart II-3Work From Home Policies Have Impacted Migration Trends… The chart highlights that the negative migration impact from COVID has been mostly concentrated in New York City and the three most populous cities on the West Coast (by metro area): Los Angeles, San Francisco, and Seattle. And yet, Chart II-4 highlights that house price inflation in these four cities has accelerated to a double-digit pace, only modestly below the national average. Chart II-4...But Cities With Outward Migration Still Have Very Strong House Price Gains The house price indexes shown in Chart II-4 represent aggregate, metro area trends, and clearly some regions within these metro areas have experienced house price deceleration or outright deflation versus gains in areas outside the urban core. But Chart II-5 highlights that house prices have declined in Manhattan basically in line with the change in net move-ins as a share of the population, underscoring that double-digit metro area-wide house price gains appear to be vastly disproportionate to changes in net migration. Similarly, Chart II-6 highlights that rents decelerated in the US over the past year but remained in positive territory and grew at a 3.5% annualized rate from February to April. Chart II-5In Manhattan, House Prices Have Tracked Net Migration Chart II-6Rent Costs Have Decelerated, But Have Not Contracted Evidence from Paris and London also suggests that a work-from-home effect is insufficient to explain broad house price gains. Panel 1 of Chart II-7 highlights that house prices in France have accelerated significantly, but that apartment prices have decelerated only fractionally in lockstep. Panel 2 shows that the acceleration in house prices does reflect a work-from-home effect, as prices have risen faster in inner Parisian suburbs. Panel 3, however, highlights that Parisian apartment prices, the dominant property type in the urban core, have decelerated modestly. Chart II-8 highlights that house price gains have not even decelerated in greater London; they have been merely been modestly outstripped by gains in Outer South East (outside of the Outer Metropolitan Area). Chart II-7In France, Parisian Apartment Prices Are Simply Lagging, Not Falling Chart II-8In The UK, Greater London Property Prices Are Accelerating     The Policy Effect: The Fundamental Driver Of The Housing Market Despite the broader location flexibility that work-from-home policies now provide to potential homeowners, it seems inconceivable that the housing market would have responded in the manner that it has over the past year given the size of the economic shock brought on by the pandemic without significant support from policy. Above-the-line fiscal measures to the pandemic have totaled in the double-digits in advanced economies (Chart II-9), and monetary policy has contributed to easier financial conditions via rate cuts, asset purchases, and sizeable programs to support financial market liquidity. Chart II-9There Has Been A Massive Fiscal Policy Response To The Crisis In fact, Charts II-10-II-13 present compelling evidence that fiscal and monetary policy have been the core drivers of significant house price gains over the past year. Charts II-10 and II-11 plot the above-the-line fiscal response of advanced economies against the year-over-year growth rate in house prices as well as its acceleration (the change in the year-over-year growth rate). The charts show a clearly positive relationship, with a stronger link between the pandemic fiscal response and the acceleration in house prices. Chart II-10Differences In Last Year’s Fiscal Response… Chart II-11…Help Explain Differences In House Price Gains Chart II-12Pre-Pandemic Differences In The Monetary Policy Stance… Chart II-13…Do An Even Better Job Of Explaining 2020 House Price Gains   Charts II-12 and II-13 highlight the even stronger link between house prices and the pre-pandemic monetary policy stance in advanced economies, defined as the difference between each country’s 2-year government bond yield and its Taylor Rule-implied policy interest rate as of Q4 2019. We construct each country’s Taylor Rule using the original specification, with core consumer price inflation, a 2% inflation target, and real potential GDP growth as the definition of the real equilibrium interest rate. The charts make it clear that easy monetary policy strongly explains house price gains in 2020, particularly the year-over-year percent change rather than its acceleration. This makes sense, given that monetary policy was already quite easy in many countries at the onset of the pandemic – meaning that changes were less pronounced than they would have been had interest rates been higher. The explanation that emerges from Charts II-10-II-13 is that historic fiscal easing, combined with an easy starting point for monetary policy – that became even easier last year – enabled demand from work-from-home policies to manifest during an extremely severe recession. We agree that work-from-home policies have shifted the geographic preferences of some home buyers and likely provided a new source of net demand from renters in urban cores purchasing homes in outlying areas. But we strongly doubt that the net effect of work-from-home policies in the midst of an extreme shock to economic activity would have caused the rise in house prices that we have observed, certainly not to this level, without major support from policy. This underscores that policy, and not the work-from-home effect, has and will likely remain the core driver of the global housing market. The Supply Effect: Mostly A Red Herring Chart II-14Countries Fall Into Two Groups In Terms Of The Relative Trend In Real Residential Investment One perennial question that emerges when analyzing the housing market, particularly in markets with outsized house price gains, is the impact of constrained supply. It is frequently argued that constrained supply is squeezing prices higher in many markets, and that the appropriate policy solution to extreme house price gains is to enable widespread housing construction – not to raise interest rates. We do not rule out the potential impact of constrained supply in certain cities or regional housing markets, and we have highlighted in previous research that a positive relationship does exist between population density in urban regions and median house price-to-income ratios.1 But as a broad explanation for supercharged house price gains, the supply argument appears to fall flat. Chart II-14 presents the most standardized measure of cross-country housing supply available for several advanced economies, the trend in real residential investment relative to real GDP over time. These series are all rebased to 100 as of 1997, prior to the 2002-2007 US housing market boom. The chart makes it clear that advanced economies generally fall into two groups based on this metric: those that have seen declines in real residential investment relative to GDP, especially after the global financial crisis (panel 1), and those that have experienced either an uptrend in housing construction relative to output or have seen a flat trend (panel 2). If scarce housing supply was the core driver of outsized house price gains, then we would expect to see stronger gains in the countries shown in panel 1 and smaller gains in the countries shown in panel 2. In fact, mostly the opposite is true: Charts II-15 and II-16 highlight that the relationship between the level of these indexes today relative to their 1997 or 2005 levels is positively related to the magnitude of house price gains last year, suggesting that housing market supply has generally been responding to demand over the past decade. The US and possibly New Zealand stand as possible exceptions to the trend, suggesting that relatively scarce supply may be boosting prices even further in these markets beyond what fiscal and monetary policy would suggest. Chart II-15Countries That Have Seen A Stronger Pace Of Residential Investment… Chart II-16…Have Experienced Stronger House Price Gains   Chart II-17Is This Not Enough Supply, Or Too Much Demand? As a final point about the inclination of investors to gravitate towards supply-side arguments related to the housing market, Chart II-17 presents a simple thought experiment. The chart shows a simple housing supply-demand curve diagram, in a scenario where the demand curve for housing has shifted out more than the supply curve has (thus raising house prices). Is this a scenario in which supply is too tight? Or is it a case in which demand is too strong? In our view, the tight supply answer is reasonable in circumstances where the increase in demand is normal or otherwise sustainable. But Charts II-10-II-13 clearly showed that housing demand is being boosted by easy policy, which in the case of some countries has occurred for years: interest rates have remained well below levels that macroeconomic theory would traditionally consider to be in equilibrium, and this has occurred alongside significant household sector leveraging (Chart II-18). As such, in our view, investors should be more inclined to view the global housing market as generally being driven by demand-side rather than supply-side factors. This Is Not 2007/08 … Yet We highlighted in Chart II-2 above that the household sector debt-to-GDP ratio increased sharply last year, which has raised some questions about debt sustainability among investors. For the most part, the rise in this ratio actually reflects denominator effects (namely a sharp contraction in nominal GDP) rather than a huge surge in household debt. Chart II-19 shows BIS data for the annual growth in total household debt in developed economies was roughly stable last year, at least until Q3 (the most recent datapoint available from the BIS). Chart II-18Low Interest Rates Have Fueled Household Leveraging Chart II-19Total Credit Growth Has Been Stable, But Mortgage Credit Growth Is Accelerating Chart II-20US Mortgage Growth Is Picking Up, As Repayments Slow Consumer Credit Growth But Chart II-19 shows the recent trend in total household debt, which masks diverging mortgage and non-mortgage debt trends. In the US, euro area, Canada, and Sweden, household mortgage debt has accelerated to varying degrees, underscoring that households have likely paid down non-mortgage debt with some of the savings that they have accumulated from a significant reduction in spending on services. Chart II-20 shows this effect directly in the case of the US; mortgage debt growth accelerated by roughly 1.5 percentage points in the second half of the year, whereas consumer credit growth (made up of student loans, auto loans, credit cards, and other revolving credit) decelerated significantly. This aligns with data showing that US households have used some of their savings windfall to pay down their credit card balances. This changing mix within household debt - less higher-interest-rate consumer credit, more lower-interest-rate collateralized mortgage debt – could, on the margin, help mitigate financial stability risks from the housing boom by moderating overall debt service burdens. The starting point for the latter matters, though, in accurately assessing the risks from rising house prices and increased mortgage debt, particularly in countries where household debt levels are already high. According to data from the BIS, the US already has one of the lowest household debt service ratios (7.6%) among the developed economies (Chart II-21).2 This compares favorably to the double-digit debt service ratios in the “higher-risk” countries like Canada (12.6%), Sweden (12.1%) and Norway (16.2%). On top of that, US commercial banks have become far more prudent with mortgage loan underwriting standards since the 2008 financial crisis. The New York Fed’s Household Debt and Credit report shows that an increasing majority of mortgage lending made by US banks since the 2008 crisis has been to those with very high FICO credit scores (Chart II-22). This is in sharp contrast to the steady lending to “subprime” borrowers with poor credit scores that preceded the 2008 financial crisis. The median FICO score for new mortgage originations as of Q1 2021 was 788, compared to 707 in Q4 2006 at the peak of the mid-2000s US housing boom. Chart II-21Diverging Trends In Global Household Debt Servicing Costs Chart II-22US Banks Have Become More Prudent With Mortgage Lending   US bank balance sheets are also now less directly exposed to a fall in housing values. Residential loans now represent only 10% of the assets on US bank balance sheets, compared to 20% at the peak of the last housing bubble (Chart II-23). This puts the US in the “lower-risk” group of countries in Europe, the UK and Japan where mortgages are less than 20% of bank balance sheets. This compares favorably to the “higher risk” group of countries where residential loans are a far larger share of bank assets (Chart II-24), like Canada (32%), New Zealand (49%), Sweden (45%) and Australia (40%). Chart II-23Banks Have Limited Direct Exposure To Housing Here Chart II-24Banks Are Far More Exposed To Housing Here   Like nature, however, the financial ecosystem abhors a vacuum. “Non-bank” mortgage lenders have filled the void from traditional US banks reducing their lending to lower-quality borrowers, and they now represent around two-thirds of all US mortgage origination, a big leap from the 20% origination share in 2007. Non-bank lenders have also taken on growing shares of new mortgage origination in other countries like the UK, Canada and Australia. Chart II-25Global Banks Can Withstand A Housing Shock Non-bank lenders do not take deposits and typically fund themselves via shorter-term borrowings, which raises the potential for future instability if credit markets seize up. These lenders also, on average, service mortgages with a higher probability of default, so they are exposed to greater credit losses when house prices decline. However, the risk of a full-blown 2008-style commercial banking crisis, with individual depositors’ funds at risk from a bank failure, are reduced with a greater share of riskier mortgage lending conducted by non-bank entities. This is especially true with global commercial banks far better capitalized today, with double-digit Tier 1 capital ratios (Chart II-25), thanks to regulatory changes made after the Global Financial Crisis. Net-net, we conclude that the overall financial stability implications of the current surge in house prices in the developed economies are relatively modest on average. The acceleration in mortgage growth has occurred alongside reductions in non-mortgage growth, at a time when banks are better able to withstand a shock from any sustained future downturn in house prices. However, if house prices continue to accelerate and new homebuyers are forced to take on ever increasing amounts of mortgage debt, financial stability issues could intensify in some countries. Services spending will recover in a vaccinated post-COVID world, as economies reopen and consumer confidence improves, which will likely end the trend of falling non-residential consumer debt offsetting rising mortgage debt in countries like the US and Canada. Overall levels of household debt could begin to rise again relative to incomes, building up future financial stability risks when central banks begin to normalize pandemic-related monetary policies – a process that has already started in some countries because of the housing boom. The Monetary Policy Implications Of Surging House Prices Rapidly appreciating house prices are becoming an area of concern for policymakers in countries like Canada and New Zealand, where the affordability of housing is becoming a political, as well as an economic, issue. In the case of New Zealand, the government has actually altered the remit of the Reserve Bank of New Zealand (RBNZ) to more explicitly factor in the impact of monetary policy on housing costs. The Bank of Canada announced in April that it would taper its pace of government debt purchases and signaled that its decision was based, at least in small part, on signs of speculative behavior in Canada’s housing market. Macroprudential measures like limiting loan-to-value ratios of new mortgage loans are a policy option that governments in those countries have already implemented to try and cool off housing demand. Yet while such measures can help alleviate demand-supply mismatches in certain cities and regions, the efficacy of such measures in sustainably slowing the ascent of house prices on a national scale is unclear. In the April 2021 IMF Global Financial Stability Report, researchers estimated that, for a broad group of countries, the implementation of a new macro-prudential measure designed to cool loan demand reduced national household debt/GDP ratios by a mere one percentage point, on average, over a period encompassing four years.3 If macroprudential measures are that ineffective in sustainably reducing demand for mortgage loans, then the burden of slowing house price appreciation will have to fall on the more blunt instruments of monetary policy. Importantly, surging house price inflation is not likely to give a boost to realized inflation measures – an important issue given the current backdrop of rapidly rising realized inflation rates in many countries. Housing costs do represent a significant portion of consumer price indices in many developed countries, ranging from 19% in New Zealand to 33% in the US (Chart II-26), with the euro area being the outlier with housing having a mere 2% weighting in the headline inflation index. Chart II-26A Limited Impact On Actual Inflation From Housing Yet those so-called “housing” categories overwhelmingly measure only housing rental costs and not actual house prices. This is an important distinction because rents – which are often imputed measures like in the US and not even actual rental costs - are rising at a far slower pace than actual house prices in most countries, so the housing contribution to realized inflation is relatively modest. So the good news is that booming house prices will not worsen the acceleration of realized global inflation that has concerned investors and policymakers in 2021. Yet that does not mean that central bankers will not be forced to tighten policy to cool off red-hot housing demand that is clearly being fueled by persistently negative real interest rates. In Chart II-27 and Chart II-28, we show both nominal and real policy interest rates for the “lower risk” and “higher risk” country groupings that we described earlier. The real policy rates are nominal policy rates versus realized headline CPI inflation. The dotted lines in the charts represent the future path of rates discounted by markets. Specifically, the projection for nominal rates is taken from overnight index swap (OIS) forward curves, while the projection for real rates is calculated by subtracting the discounted path of inflation expectations extracted from CPI swap forwards. Chart II-27Markets Discounting Negative Real Rates For The Next Decade Chart II-28Negative Real Rates Are Unsustainable During A Housing Bubble   There are two key takeaways from these charts: Real policy interest rates are at or very close to the most deeply negative levels seen since the 2008 financial crisis. Markets are discounting that real rates will be at or below 0% for most of the next decade. Admittedly, there is room for debate over what the equilibrium level of real interest rates (a.k.a. “r-star”) should be in the coming years. However, we deem it a major stretch to believe that real rates need to be persistently low or negative for the next ten years to support even trend growth across the developed economies. In our view, the current boom in housing demand and mortgage borrowing provides clear evidence that negative real rates are below equilibrium and, thus, are stimulating credit demand. Thus, the only way for a central bank to cool off housing demand will be to raise both nominal and, more importantly, real interest rates. Canada and New Zealand will be the “canaries in the coal mine” among developed market central banks for such a move. According to the latest Bank of Canada Financial Stability Review, nearly 22% of Canadian mortgages are highly levered, with a loan-to-value ratio greater than 450%, a greater share of such mortgages than during the 2016/17 housing boom (Chart II-29). Canadian house prices have risen to such an extent that home prices in major cities like Toronto, Vancouver and Montreal are among the most expensive in North America.4  Stunningly, a recent Bloomberg Nanos opinion poll revealed that nearly 50% of Canadians would support Bank of Canada rate hikes to cool off the red-hot housing market (Chart II-30). The central bank will be unable to resist the pressure to use monetary policy to slam on the brakes of the housing market – investors should expect more tapering and, eventually, rate hikes from the Bank of Canada over at least the next couple of years. Chart II-29Canadians Are Leveraging Up To Buy Expensive Homes Chart II-3050% Of Canadians Want A Rate Hike To Cool Housing   In New Zealand, worsening housing affordability has reached a point where a 20% down payment on the median national house price is equal to 223% of median disposable income (Chart II-31). This is forcing more first-time home buyers to take on levels of mortgage debt that the RBNZ deems highly risky (top panel). Like the Bank of Canada, the RBNZ will prove to be one of the most hawkish central banks in the developed world over the next couple of years as the central bank follows their newly-revised remit to try and cool off housing demand in New Zealand. Who is next? Housing values, measured by the ratio of median national house prices to median national household incomes, are rising in the US and UK but are still below the peaks of the mid-2000s housing bubble (Chart II-32). Meanwhile, housing is becoming more expensive across the euro area, but not in a consistent manner, with valuations in Germany and Spain having increased far more than in France or Italy. Housing valuations have actually improved in Australia over the past couple of years on a price-to-income basis. The most likely candidates for a housing-related hawkish turn are in Scandinavia, with housing valuations in Sweden and Norway closing in on Canada/New Zealand levels. Chart II-31New Zealand Housing Is Wildly Unaffordable Chart II-32Global House Price/Income Ratios Are Trending Higher   Investment Conclusions The current acceleration in global house prices is an inevitable outcome of the extraordinary monetary and fiscal easing implemented during the pandemic. Higher realized inflation is pushing real rates deeper into negative territory in many countries, fueling the demand for housing. Central banks in countries with more stretched housing valuations will be forced to turn more hawkish sooner than expected, leading to tapering and, eventually, rate hikes to cool housing demand. This has negative implications for government bond markets in countries where housing is more expensive and real yields remain too low, like Canada, New Zealand and Sweden (Chart II-33). Investors should limit exposure to government bonds in those markets over the next 6-12 months. Chart II-33Negative Real Yields & Expensive Housing Valuations – An Unsustainable Mix Bond markets in countries where house prices are not rising rapidly enough to force policymakers to turn more hawkish more quickly – like core Europe, Australia and even Japan - are likely to be relative outperformers. The US and UK are “cuspy” bond markets, as housing valuations are becoming more expensive in those two countries but the Fed and Bank of England are not facing the same domestic political pressure to use monetary policy tools to fight the growing unaffordability of housing. That could change, though, if overall household leverage begins to rise alongside house price inflation as the US and UK economies emerge from the pandemic. Current pricing in OIS curves shows that markets expect the RBNZ and Bank of Canada to begin hiking rates in May 2022 and September 2022, respectively (Table II-1). This is well ahead of expectations for “liftoff” from other developed markets central banks, including the Fed in April 2023. The cumulative amount of rate hikes following liftoff to the end of 2024 is highest in Canada, New Zealand, the US and Australia. Those are also countries with currencies that are trading at or above the purchasing power parity levels derived from our currency strategists’ valuation models. This highlights the difficult choice that central bankers facing housing bubbles must confront, as the rate hikes that will help cool off housing demand will lead to currency appreciation that could impact other parts of their economies like exports and manufacturing. Table II-1Hawkish Central Banks Must Live With Currency Strength Tracking the second-round economic consequences of eventual monetary policy actions to control excessive house price inflation, particularly in “higher risk” countries, is likely to be the subject of future Bank Credit Analyst / Global Fixed Income Strategy reports. Jonathan LaBerge, CFA Vice President The Bank Credit Analyst Robert Robis, CFA Chief Fixed Income Strategist Footnotes 1 Please see Global Investment Strategy "Canada: A (Probably) Happy Moment In An Otherwise Sad Story," dated July 14, 2017, available at gis.bcaresearch.com 2 Importantly, the BIS debt service ratios include the payment of both principal and interest, thus making it a true measure of debt service costs that includes repayment of borrowed funds – a critical issue in countries with high loan-to-value ratios for home mortgages. 3 Please see page 46 of Chapter 2 of the April 2021 IMF Global Financial Stability Report, which can be found here: https://www.imf.org/en/Publications/GFSR/Issues/2021/04/06/global-finan… 4 “Vancouver, Toronto and Hamilton are the least affordable cities in North America: report”, CBC News, May 20, 2021

Highlights Global Tapering: The Bank of England has joined the Bank of Canada as central banks tapering the pace of bond buying. Markets are now trying to sort out who is next and concluding that it will not be the Federal Reserve, with US employment still well below the pre-pandemic peak. US Treasury yields will continue trading sideways until there is greater clarity on the pace of US labor market improvement, especially after the big downside miss in the April jobs report. US Treasury Curve: We are adding a new recommended US butterfly trade to our Tactical Overlay portfolio, going long the 5-year bullet and short the 2/30 barbell using US Treasury futures. This trade should benefit with US Treasury curve steepening overshooting the pace of past cycles, while offering attractive carry if persistent Fed dovishness slows the cyclical transition to a bear-flattening curve regime. Feature Heading into 2021, one of our key investment themes for the year was that no major central bank would shift to a less dovish monetary policy stance before the Fed. Not even five months into the year, our theme has already been proven incorrect. Last week, the Bank of England (BoE) announced a slower pace of its asset purchases, following a similar tapering decision by the Bank of Canada (BoC) last month. Chart of the WeekUS Jobs Recovery Lagging, Despite Vaccine Success We had assumed that no central bank could tolerate the currency strength that would inevitably occur by tapering ahead of the Fed. That was clearly not the case in Canada, and the Canadian dollar has already appreciated 4.6% versus the greenback since the BoC taper announcement April 21. The British pound also rallied solidly against both the US dollar and euro immediately after the BoE taper announcement last week. Markets are beginning to speculate on future taper candidates, like the Reserve Bank of New Zealand (RBNZ), with the New Zealand dollar being one of the strongest currencies in the G10 versus the US dollar since the end of March (+4.4%). Investors had been debating the possibility that the Fed could begin tapering sometime in the second half of 2020, largely based on what has to date been a successful US vaccination campaign. Yet while that led to optimism that the US economy can quickly reopen and return to normal, the fact remains that the recovery in US employment from the COVID shock has lagged other major economies (Chart of the Week). The big downside miss on the April US payrolls report highlights how the Fed can be patient before joining the tapering club. US Treasury yields are likely to continue trading sideways, and the US dollar will trade soft, until markets can sort out the true state of US labor demand versus supply. Which Central Bank Could Follow The BoC And BoE? Back in March, we published a report that discussed what we called the “pecking order of global liftoff”.1 We looked at how interest rate markets were pricing in an increasingly diverse path out of the coordinated global monetary easing enacted last year during the COVID recession (Chart 2). We looked at both the timing of “liftoff” (the first rate hike) and the pace of hikes afterward to the end of 2024. We then ranked the countries by the market-implied timing of liftoff. Chart 2Sorting Out The Relative Hawks & Doves Among Global CBs At the time, overnight index swap (OIS) curves were discounting the earliest liftoff from the RBNZ (June 2022) and BoC (August 2022). The Fed was expected to hike in January 2023, followed by the BoE in June 2023 and Reserve Bank of Australia (RBA) in July 2023. The European Central Bank (ECB) and Bank of Japan (BoJ) were the laggards, with no rate hiked discounted until September 2023 and February 2025, respectively. In terms of the pace of rate hikes after liftoff through 2024, our list was broken into two groups. The more aggressive central banks were expected to be the BoC (+175bps), RBA (+156bps), RBNZ (+140bps) and the Fed (+139bps). Much smaller amounts of rate hikes were anticipated from the BoE (+63bps), ECB (+25bps) and BoJ (+9bps). In the two months since our March report, the market timing of liftoff, and the pace of subsequent hikes, has shifted for all those countries (Table 1). The BoC is now expected to move in September 2022, ahead of the RBNZ (October 2022). In 2023, the Fed is now priced for liftoff in March 2023, followed by the BoE and RBA (both in July 2023). The ECB liftoff date is little changed (now August 2023), while the market has dramatically pushed out the timing of any BoJ hike (now November 2025). The cumulative rate hikes through 2024 are moderately lower for all countries except Australia (a reduction in total tightening of 56bps). Table 1The Fed Is Sliding Down The “Pecking Order Of Liftoff” List What is interesting about these changes is that the market has pulled forward the timing of liftoff for the BoE and RBA, while pushing it out for the BoC, RBNZ, BoJ and, most importantly, the Fed. The Fed is now drifting down the “pecking order” for liftoff, expected to lift rates only a couple of months before the BoE or RBA. This is a major change from previous monetary policy cycles, when the Fed would typically be a first mover when it comes to tightening policy. Chart 3The Momentum Of Global QE Has Already Been Slowing While the BoC and BoE decisions to taper quantitative easing (QE) have garnered the headlines, the pace of global central bank balance sheet expansion had already peaked at the start of 2021 (Chart 3). The pace has slowed most dramatically in Canada and the US, but this was a result of certain emergency programs expiring – most notably the Fed’s corporate bond buying vehicles late last year and the BoC’s short-term repo facilities more recently. Greater financial market stability was the reason cited to end those programs, while still leaving government bond QE buying in place unchanged. The year-over-year pace of global QE was set to slow, simply from less favorable comparisons to 2020 after the surge in central bank balance sheet expansion last year. Yet now we are starting to see actual tapering of government bond purchases from some central banks. Is such “early tightening” warranted? Back in that same March report where we discussed the order of global liftoff, we gave our assessment of the most important factors that could drive central banks to consider a shift to a less dovish stance (like tapering). For the BoC, we cited booming house prices and robust business confidence as reasons the BoC could turn less dovish sooner (Chart 4). For the BoE, we noted a sharper-than-expected recovery in domestic investment and consumer spending, as the locked-down UK economy reopens, as reasons why the BoE could begin to tweak its policy settings. For both central banks, all those indicators were mentioned as factors leading to their decision to taper. For the Fed, we determined that rising inflation expectations and increasing labor market tightness would both be required for the Fed to turn less dovish. Only inflation expectations have reached that goal, with the US Employment/Population ratio still well below the pre-pandemic peak (Chart 5). For the RBA, we looked solely at realized inflation measures, as the RBA has explicitly noted that Australian wage growth must rise sustainably towards 3% - nearly double current levels - before realized CPI inflation could return to the 2-3% target range. For both the Fed and RBA, the necessary conditions for a change in current policy settings have not yet been met. Chart 4What The More Hawkish CBs Are Watching Chart 5What The More Dovish CBs Are Watching For the ECB, we noted that realized inflation (and the ECB’s inflation forecasts), along with the Italy-Germany government bond spread as a measure of financial conditions, were the most important indicators to watch before the ECB could consider any move to taper its QE programs (Chart 6). Italian spreads have widened a bit in recent months, while the latest set of ECB economic forecasts still call for headline euro area inflation to remain well south of the 2% target out to 2023. For the BoJ, we simply cited a rise in realized inflation as the only possible development that could lead to a BoJ taper. The BoJ now forecasts that Japanese inflation will not reach the 2% central bank target until at least 2024. So for both the ECB and BoJ, the conditions do not warrant any imminent tapering of bond buying. Chart 6What The Most Dovish CBs Are Watching As another way to determine who could taper next, we turn to our Central Bank Monitors, which are designed to measure the pressure on policymakers to ease or tighten monetary setting. All the Monitors have responded to the recovery in global growth and inflation, along with the easing of financial conditions implied by booming markets, over the past year. Yet only the RBA Monitor is calling for tightening (Chart 7), indicating that the RBA’s current focus on only wages and realized inflation is a departure from their behavior in the past. The Fed and BoE Monitors have risen to the zero line, suggesting no further pressure to ease policy but no tightening is needed either. The ECB, BoJ and RBNZ Monitors are all close, but just below, the zero line, suggesting diminishing need for more monetary stimulus (Chart 8). Chart 7Bond Yields Have Moved Ahead Of Our CB Monitors Chart 8Yields Overshooting Tightening Pressures Here Too Based on our assessment of the above indicators, we judge the RBNZ to be the next central bank most likely to taper, sometime in the 2nd half of 2021. We still see the Fed starting to signal tapering later this year, but with actual slowing of US Treasury (and Agency MBS) purchases not occurring until early 2022. The year-over-year momentum of bond yields correlates strongly with the Central Bank Monitors. The rise in global bond yields seen over the past year has exceeded the pace implied by the Monitors. This is unsurprising given how rapidly the global economy has recovered from pandemic-fueled recession in 2020. Supply chain disruptions and surging commodity prices have also given a lift to bond yields via rising inflation expectations, even as central banks have promised to keep rates on hold for at least the next couple of years. Yet purely from a monetary policy perspective, the surge in global bond yields looks to have gone a bit too far, too fast. Bottom Line: Markets are now trying to sort out who will taper next after the BoC and BoE, and have concluded that it will not be the Federal Reserve, with US employment still well below the pre-pandemic peak. US Treasury yields will continue trading sideways until there is greater clarity on the pace of US labor market improvement, especially after the big downside miss in the April jobs report. Bond yields in other developed markets appear to have overshot economic momentum, and a period of consolidation is needed before yields can begin moving higher again. US Treasury Curve: How Much Steepening Left? Chart 9A Pause In The UST Bear-Steepening Trend For most of the past year, the primary trend in the US Treasury curve has been one of bear steepening. Longer maturity yields have borne the brunt of the upward pressure stemming from the rapid recovery in US (and global) economic growth from the depths of the 2020 COVID-19 recession. In recent weeks, however, the surge in longer-maturity Treasury yields has stalled, as have the immediate steepening pressures (Chart 9). Purely from a fundamental economic perspective, a steepening Treasury curve is an expected result of the reflationary mix of growth, inflation and monetary policy currently at work in the US. For example, since the 2020 lows, 5-year/5-year forward inflation expectations from the TIPS market have risen 143bps while the ISM manufacturing index surged from a low of 41 to a high of 65 in March of this year (Chart 10). Combine that with the Fed cutting rates to 0% last year, while promising to keep rates unchanged through 2023 and reinforcing that commitment through QE, and it is no surprise to see a steeper US Treasury curve. Chart 10UST Curve Steepening Has Been Driven By Reflation Yet even despite these obvious steepening pressures, the pace of the Treasury curve steepening does seem to be a bit rapid compared to history. In Chart 11, we show a “cycle-on-cycle” analysis, comparing the slope of various US Treasury curve segments (2-year versus 5-year, 5-year versus 10-year, 10-year versus 30-year) to the average of the previous five US business cycles, dating back to the 1970s. The curves are lined up to the start date of the previous recession, with the vertical line in the chart representing that date. Thus, this chart allows us to see how the Treasury curve evolved heading into, and coming out of, economic downturns. Chart 11 shows that the current 2-year/5-year curve, with a steepness of 63bps, is in line with past steepening moves coming out of recession. For the curve segments at longer maturities, the pace of steepening has been much more rapid than in the past. In fact, the current 5-year/10-year slope of 82bps is already above the average past peak level, as is the 10-year/30-year curve of 72bps. If we do the same cycle-on-cycle analysis for the three previous US recessions dating back to 1990, the current curve slopes are more in line with levels seen one year into the economic expansion (Chart 12). During those previous cycles, the curve steepening trend ended around two years into the expansion. This suggests that the current curve steepening could continue into 2022, except for one major difference – the Fed cut rates to 0% very rapidly last year, far faster than in the previous easing cycles. This suggests that additional curve steepening from current levels can only occur through a surge in US inflation. Chart 11Current UST Steepening Has Moved Fast Compared To Past Cycles Chart 12Can More UST Curve Steepening Occur With A 0% Funds Rate? The slope of the Treasury curve is typically correlated to the level of the nominal fed funds rate, but is even more strongly correlated to the funds rate minus actual inflation, or the real fed funds rate. When the real funds rate is below the natural real rate of interest, a.k.a. r-star, the Treasury curve has historically exhibited its strongest steepening trend. That can be seen in Chart 13, where we show the real fed funds rate (adjusted by US core CPI inflation) compared to the New York Fed’s estimate of r-star. The gap between the two series is shown in the bottom panel, correlating very strongly to the 2-year/30-year Treasury curve slope. Chart 13Curve Steepening Results When Real Rates Are Below R* With the nominal funds rate at zero, that gap between r-star and the real fed funds rate can only widen in a fashion that would support more curve steepening if a) realized US inflation moves higher or b) r-star moves higher. Both outcomes are possible as the US economic recovery, fueled by expanding vaccinations and fiscal stimulus. Both real rates and r-star are much lower in the current cycle than in previous economic recoveries, although the r-star/real funds rate gap appears to be following a more typical path that suggests potential additional steepening pressure (Chart 14). The wild card in this analysis is the Fed itself. If US economic growth and inflation evolve in way that makes it more likely the Fed would have to begin tapering QE and, eventually, signal future rate hikes, the Treasury curve may shift to a more typical bear-flattening trend seen during tightening cycles. We saw an example of that after the release of the March US employment report, where over a million jobs were created in a single month, causing 5-year Treasury yields to jump higher than longer-maturity Treasuries (i.e. curve flattening). Looking ahead, it appears that the US yield curve is more likely to slowly transition to a bear-flattening/bull-steepening regime than continue the bear-steepening/bull-flattening: trend of the past twelve months. One way to position for this is to enter into butterfly curve trades that offer attractive carry or valuation. For that, we turn to our Treasury curve valuation models. We have been recommending a Treasury yield curve trade in our Tactical Overlay portfolio on page 19, going long a 7-year bullet versus going short a 5-year/10-year barbell (Chart 15). This barbell is now very cheap on our models, which measure value by regressing the butterfly spread on the underlying slope of the curve. In this case, the spread between the 5/7/10 butterfly is unusually wide compared to the slope of the 5/10 Treasury curve. According to our model, this butterfly spread discounts nearly 100bps of additional 5/10 steepening, an excessive amount compared to past cycles. Chart 14R* - Real Funds Rate Gap Below Previous Cyclical Peaks Chart 15Maintain Our Current 5/7/10 UST Butterfly Trade While the valuation is attractive on the 5/7/10 butterfly (Table 2), the carry on this position is a modest 12bps. A butterfly with more attractive carry is the 2/5/30 butterfly. Table 2US Butterfly Strategy Valuation: Standardized Residuals Table 3US Butterfly Strategies: Carry Chart 16Enter A New 2/5/30 UST Butterfly Trade This butterfly has a neutral valuation (Chart 16) on our model, but offers 35bps of carry - the most attractive among all butterflies involving a 5-year bullet (Table 3). With US Treasury yields, and the Treasury curve slope, likely to remain rangebound for the next few months, going for higher carry trades is an attractive strategy – particularly if used in conjunction with a below-benchmark duration stance, which we still advocate. The 2/5/30 butterfly represents an attractive near-term hedge to that more defensive duration posture. Bottom Line: We are adding a new recommended US Treasury butterfly trade to our Tactical Overlay portfolio, going long the 5-year bullet and short the 2/30 barbell. This trade should benefit with US Treasury curve steepening overshooting the pace of past cycles, while offering attractive carry if persistent Fed dovishness slows the cyclical transition to a bear-flattening curve regime.   Robert Robis, CFA Chief Fixed Income Strategist rrobis@bcaresearch.com Footnotes 1 Please see BCA Research Global Fixed Income Strategy Report, "Harder, Better, Faster, Stronger", dated March 16, 2021, available at gfis.bcaresearch.com. Recommendations The GFIS Recommended Portfolio Vs. The Custom Benchmark Index ​​​​​​​ Duration Regional Allocation Spread Product Tactical Trades Yields & Returns Global Bond Yields Historical Returns

Special Report

Highlights The US fiscal outlook has deteriorated substantially over the past two decades, as a consequence of the fiscal response to both the global financial crisis and the COVID-19 pandemic. US government debt-to-GDP is now nearly as high as it was at the end of the Second World War, and is projected by the US Congressional Budget Office (CBO) to explode higher over the coming 30 years. Some investors argue that extreme levels of government debt now virtually guarantee that interest rates will remain structurally low, and we test this claim alongside a scenario that limits the projected rise in the primary deficit. We find that US fiscal reform, when it eventually occurs, will likely be negative for health care stocks. We also note that even in a scenario where the US limits the size of its future primary budget deficit, net interest outlays will likely rise to elevated levels compared to history. A comparison with the Canadian experience in the 1990s suggests a structurally negative outlook for the US dollar, from an overvalued starting point. Finally, we note that the US fiscal outlook does not necessarily prevent an increase in interest rates over the coming few years in a scenario where investors raise their expectations for the neutral rate of interest, a possibility that we discussed in last month’s report. This scenario is not our base case view, but it is plausible and should actively be monitored by investors over the coming one to two years. For now, we do not expect that rising interest rates pose a risk to stocks over the coming 6-12 months. Investors should remain cyclically overweight equities within a multi-asset portfolio, and should maintain a below-benchmark level of duration on a risk-adjusted basis. In 2001, US government debt held by the public as a share of GDP stood at 31.5%, after having fallen roughly 16 percentage points from early 1993 levels. Today, as a result of both the global financial crisis and the COVID-19 pandemic, the debt to GDP ratio has risen to a whopping 100%, and is projected to rise meaningfully higher over the coming decades. Feature In this report we review the long-term US fiscal outlook in the wake of the pandemic, with a focus on the implications for interest rates. Some investors argue that extreme levels of government debt now virtually guarantee that interest rates will remain structurally low, and we test this claim alongside a scenario that limits the projected rise in the primary deficit. We find that US fiscal reform, when it eventually occurs, will likely be negative for health care stocks, whose fundamental performance has outstripped that of the broad equity market since the mid-1990s (reflecting pricing power that stands to be curtailed through regulation). We also note that even in a scenario where the US limits the size of its future primary budget deficit, net interest outlays will likely rise to elevated levels compared to history. A comparison with the Canadian experience in the 1990s suggests a structurally negative outlook for the US dollar, from an overvalued starting point. Finally, we note that the US fiscal outlook does not necessarily prevent an increase in interest rates over the coming few years in the hypothetical scenario that we described in last month’s report,1 i.e., an environment where the narrative of secular stagnation is challenged and investor expectations for the neutral rate rise closer to trend rates of economic growth. This scenario is not our base case view, but it is plausible and should actively be monitored by investors over the coming one to two years. For now, investors should remain cyclically overweight equities within a multi-asset portfolio, and should maintain a below-benchmark level of duration on a risk-adjusted basis. Debt Sustainability, And The CBO’s Baseline Projection When analyzing the US fiscal outlook, the Congressional Budget Office’s Long-Term Budget Outlook report is typically the reference point for investors. The report provides annual projections for the budget deficit and the debt-to-GDP ratio for the next three decades, as well as a breakdown of the projected deficit into its primary (i.e., non-interest) and net interest components. Charts II-1 and II-2 present the most recent baseline projections from the CBO, which clearly present a dire long-term outlook. The deficit and debt-to-GDP ratio are projected to be relatively stable over the next decade, but explode higher over the subsequent 20 years. In 2051, the CBO’s baseline projects that the budget deficit will be roughly 13% of GDP, with net interest costs accounting for approximately two-thirds of the deficit. Chart II-1The CBO’s Fiscal Outlook Is Extremely Negative Chart II-2In 2051, The CBO Projects A 13% Annual Budget Deficit In order to understand what is driving the CBO’s dire long-term budget and debt forecast, it is important to review the government debt sustainability equation shown below. The equation highlights that the change in a government’s debt-to-GDP ratio is approximately equal to 1) the primary deficit plus 2) net interest costs as a share of GDP, the latter being defined as the product of last year’s debt-to-GDP ratio and the difference between the average interest rate on the debt and the rate of GDP growth. Δ Debt-To-GDP Ratio ≈ Primary Deficit As A % Of GDP2 + (r-g)*(Prior Period Debt-To-GDP Ratio) Where: r = Average interest rate on government debt and g = Nominal GDP growth The equation highlights that expectations of a persistently rising debt-to-GDP ratio must occur either because of expectations of a persistent primary deficit, or expectations that interest rates will persistently exceed the rate of economic growth (or some combination of the two). This underscores why debt sustainability analysis often focuses on the primary budget balance, as a country’s debt-to-GDP ratio will be stable if no primary deficit exists and interest costs are at or below the prevailing rate of economic growth. Chart II-3 illustrates the source of the CBO’s projected rise in debt-to-GDP beyond 2031, by presenting the two components of the debt sustainability equation alongside the projected annual change in the debt-to-GDP ratio. The chart makes it clear that while the CBO is forecasting a sizeable primary deficit to continue, it is projected to grow at a slower pace than the debt-to-GDP ratio itself. The increasing rate at which the debt-to-GDP ratio is projected to grow in the latter years of the CBO’s forecast period is clearly driven by the interest rate component, meaning that “r” is projected to be greater than “g”. Chart II-4 presents this point directly, by highlighting that the CBO is forecasting the average interest rate on government debt to exceed that of nominal GDP growth in 2038, and to continue to exceed growth (by an increasing amount) thereafter. Chart II-3Decomposing The CBO's Projected Change In The Debt-To-GDP Ratio Chart II-4The CBO's Projections Rest, In Part, On Rates Eventually Exceeding Growth   Three Adjustments To The CBO’s Baseline We make three adjustments to the CBO’s baseline in order to assess how the US fiscal outlook shifts under an interest rate path that is different than that projected by the CBO. First, we adjust the CBO’s projected budget deficit over the coming few years based on deficit forecasts from our US Political Strategy service following the passage of the American Recovery Plan act.3 Chart II-5We Test The Effect Of An Initially Higher, But More Sustainable, Rate Path Next, we adjust the interest component of the total budget deficit based on a new path for short- and long-term interest rates that models a scenario in which the neutral rate of interest rises to, but not above, GDP growth (Chart II-5). In last month’s report we outlined a scenario in which this could feasibly occur,1 and the hypothetical path for interest rates shown in Chart II-5 thus incorporates both the negative budgetary impact of an earlier rise in interest rates and the positive budgetary impact of “r” never rising above “g”. We explicitly exclude any crowding out effect on long-term interest rates, based on the view that term premia are likely to remain muted in a world of low potential economic growth, unless a fiscal crisis appears to be imminent (see Box II-1). Box II-1 Arguing Against The CBO’s Crowding Out Assumption The CBO’s projection that interest rates will ultimately rise above the rate of economic growth rests on the view that increased government spending will absorb savings that would otherwise finance private investment (a “crowding out” effect). We agree that crowding out can occur over the course of the business cycle, especially in a scenario where increased government spending pushes output above its potential (creating a cyclical acceleration in inflation and eventually an increase in interest rates). But the CBO is assuming that high government debt-to-GDP ratios will crowd out private investment on a structural basis, and on this basis we disagree. First, Chart Box II-1 highlights that there is essentially no empirical relationship across countries between a country’s debt-to-GDP ratio and its long-term government bond yield. Japan is a clear outlier in the chart, but including Japan implies that the relationship is negative, not positive. Chart Box II-1There Is No Empirical Relationship Between Debt-To-GDP And Interest Rates In addition, given that central banks directly control interest rates at the short-end of the curve, a structural crowding out effect can only manifest itself in the form of an elevated term premium embedded in longer-term government bond yields. Our bet is that term premia are likely to stay low in a world of low falling nominal growth, as evidenced by the experience of the past decade.4 Finally, we model the impact of two changes, beginning in 2031, that would work towards reducing the primary deficit: an increase in average government revenue to 20% of GDP (its peak level reached in 2000), and a slower pace of increase on major health care program spending. Despite the fact that population aging will increase mandatory spending on social security and health care over the coming three decades, the CBO has highlighted that the majority of the increase in spending towards these programs is projected to occur due to rising health care costs per person (Chart II-6). We thus model the impact of medical care cost control by limiting the rise in net mandatory outlays on health care programs between 2021 and 2051 to roughly half of what the CBO baseline projects. This adjustment does not prevent mandatory spending on health care programs from rising, given the strong political challenges involved in limiting spending increases that are caused by an aging population. Chart II-6The US Structural Primary Balance Is Heavily Impacted By Medical Costs Charts II-7 and II-8 illustrate how these three adjustments impact the long-term US fiscal outlook. Relative to the CBO’s baseline projections, the American Recovery Plan (ARP) budget deficit forecasts from our US Political Strategy service imply that the debt-to-GDP ratio will be approximately three to four percentage points higher over the very near term, and roughly ten points higher over the long term. Chart II-7Even With Higher Rates, The Fiscal Outlook Is Meaningfully Less Bad… Relative to this new baseline, an increase in interest rates to, but not above, the projected rate of nominal economic growth increases the debt-to-GDP ratio by an additional ten percentage points (20 points higher versus the CBO’s baseline) in the middle of the forecast period, but it lowers the debt-to-GDP ratio over the longer run by eliminating the effect of outsized interest rates magnifying a persistent primary deficit. Still, the debt-to-GDP ratio is projected to rise to a whopping 207% of GDP by 2051 in this scenario, with a budget deficit in excess of 10% of GDP. The third adjustment shown in Charts II-7 and II-8 underscores the impact on the US fiscal outlook of actions aimed at reducing the primary deficit. Increases in government revenue and the prevention of rising health care costs per person results in the debt-to-GDP ratio that is 64 percentage points lower in 2051 than in our normalized interest rate scenario. The budget deficit in this scenario still increases to approximately 6% of GDP thirty years from today, but in this case most of the deficit is due to the net interest component rather than the primary deficit, meaning that the debt-to-GDP ratio would be increasing at a much slower rate if interest rates were no higher than the rate of economic growth. Chart II-8 highlights that net interest spending in this scenario would rise to 4.5% of GDP, which would be meaningfully higher than the prior high of roughly 3% in the late 1980s and early 1990s. Chart II-8...With Higher Taxes And Medical Cost Control Chart II-9A Meaningful, But Not Unprecedented, Rise In Net Interest Outlays But that is far from unprecedented or necessarily consistent with a fiscal crisis. Chart II-9 also shows that Canada’s public debt charges rose to 6.5% of GDP in the early 1990s without triggering a public debt crisis. It is true that Canada subsequently embarked on a painful fiscal consolidation program in order to reduce its public debt burden, but this, in part, occurred because of a cyclically-adjusted primary deficit of approximately 3% - twice as large as that projected for the US in 2051 in our adjusted scenario shown in Charts II-7 and II-8. Revenue And Health Care Cost Reform Our third adjustment to the CBO’s long-term budget outlook involved changes to revenue and health care cost control to reduce the US’ projected primary deficit. Are these adjustments achievable? In our view, the answer is yes: As noted above, our scenario modeled these changes taking place a decade from today, which allows for policymakers and stakeholders to have a substantial amount of time to act and adjust to these changes. On the revenue front, we noted above that US government revenue has reached 20% of GDP in the past, in the year 2000. Chart II-10 highlights that while raising taxes will likely reduce US competitiveness, the US maintains a sizeable tax advantage relative to other advanced economies, and that this was true prior to the tax cuts that took place under the Trump administration. On the health care cost front, Chart II-11 highlights that US healthcare expenditure is much larger as a share of GDP than other countries, which was not the case prior to the 1980s. Chart II-12 highlights that this cost difference is entirely due to inpatient (i.e., hospital) and outpatient (i.e., drug) costs. While it is not clear what form it will take, it seems likely that future reforms by policymakers to eliminate rising health care costs per person will occur and can be achieved. Chart II-10The US Government Can Afford To Raise Revenue Chart II-11The US Spends Much More On Health Care Than Other Countries   Chart II-12The US Significantly Outspends The World On Hospital And Drug Costs The key point for investors is not whether these changes should or should not occur, but whether there are any feasible scenarios in which spiraling government debt and interest payments are avoided without the Fed purposely maintaining monetary policy at levels persistently below the rate of economic growth – and thus risking major inflationary pressure. Our analysis above highlights that there are; the question is when policymakers will choose to act and in what form. A potential tipping point may be when US government spending on net interest as a % of GDP exceeds its prior high, which occurs in 2026 in the scenario modeled in Chart II-8. In a scenario where reforms fail to materialize or where financial markets force policymakers to act, a fiscal risk premium could certainly emerge in longer-term government bond yields, which could lead the Fed to maintain lower short-term interest rates than it otherwise would. But this scenario is only likely to emerge after interest rates converge towards rates of economic growth, as US government debt will remain highly serviceable for some time if "r" remains meaningfully lower than "g". Investment Conclusions There are three potential investment implications of our research. First, the fact that rising medical costs have such a significant impact on the CBO’s projections of the primary deficit implies that fiscal reform, when it eventually occurs, will be negative for US health care stocks. Chart II-13 highlights that US health care sector earnings have outperformed broad market earnings since the mid-1990s, and that the sector has consistently delivered an above-average return on equity. This historical performance likely reflects the sector’s pricing power, which stand to be curtailed through regulatory efforts in a world where rising health care costs per person collide with fiscal belt-tightening. Interestingly, Chart II-12 highlighted that US per capita spending on medical goods is not significantly higher than in other developed markets, suggesting that the health care equipment & supplies industry may fare better over a very long term time horizon than overall health care. Second, Charts II-7 and II-8 highlighted that even if the US does raise revenue as a share of GDP and limits excessive growth in medical costs, a primary deficit will still exist and net interest outlays will still rise to elevated levels compared to what has historically been the case. We noted that Canada experienced a higher public debt burden in the 1990s and did not suffer from a fiscal crisis, but Chart II-14 highlights that the fiscal situation did weigh on the Canadian dollar, which progressively traded 10-20% below its PPP-implied fair value level over the course of the 1990s. Thus, the implication is that eventual fiscal reform in the US may be structurally negative for the US dollar, from an overvalued starting point (panels 3 and 4 of Chart II-14). Chart II-13Eventual Fiscal Reform Will Likely Be Negative For Health Care Stocks Chart II-14The US Fiscal Outlook, Even With Some Reforms, Is Dollar-Negative   Finally, our scenario analysis highlights that very elevated levels of government debt do not guarantee that interest rates will remain structurally low, especially over the next decade when the US primary deficit is projected to remain relatively stable. For investors focused on forecasting the direction of 10-year Treasury yields from the perspective of valuation, it should be noted that the next decade is the relevant projection period for the Fed funds rate, not what occurs to net interest outlays in the two decades that follow. Over the very long run, it is true that there may ultimately be very strong political pressure on the Fed to keep interest rates below the prevailing rate of economic growth, as policymakers in 2030 will be able to avoid a structural adjustment to the primary deficit of roughly 1.1-1.3% of GDP for every percentage point that average interest rates on government debt are below nominal GDP growth. However, we noted above that this pressure is unlikely to build before the second half of this decade even in a scenario where interest rates rise significantly over the coming few years, and it remains an open questions whether the Fed will acquiesce to this pressure given its strong potential to fuel excess private sector leveraging. Over the coming one to two years, the key conclusion is that the US fiscal outlook is not likely to prevent an increase in interest rates over the coming few years in the hypothetical scenario that we described in last month’s report, i.e., an environment where the narrative of secular stagnation is challenged and investor expectations for the neutral rate rise closer to trend rates of economic growth. This remains a risk to our overweight stance towards risky assets and is not our base case view. But it does highlight the importance of monitoring long-dated rate expectations over the coming year, and argues, on a risk-adjusted basis, for a below-neutral duration stance within a fixed-income portfolio. Jonathan LaBerge, CFA Vice President The Bank Credit Analyst Footnotes 1 Please see The Bank Credit Analyst Special Report "R-star, And The Structural Risk To Stocks," dated March 31, 2021, available at bca.bcaresearch.com 2 Presented in this fashion, a budget deficit (surplus) is recorded with a positive (negative) sign. 3 For more information, please see US Political Strategy report “Biden’s Pittsburgh Speech And Legislative Agenda,” dated April 1, 2021, available at usp.bcaresearch.com 4 Please see “Term premia: models and some stylised facts”, by Cohen, Hördahl, and Xia, BIS Quarterly Review, September 2008.