Developed Countries

Highlights Portfolio Strategy Vibrant and broad-based bank credit growth, pristine credit quality, pent up bank buyback demand and a V-shaped recovery in bank ROE more than offset the risk of 10/2 yield curve inversion, and suggest that the path of least resistance is higher for the S&P banks index. Rising residential construction versus stalling residential investment, easing interest rates, cheapened lumber prices, and alluring valuations and technicals all signal that more gains are in store for homebuilders at the expense of home improvement retailers. Recent Changes Initiate a long S&P homebuilding/short S&P home improvement retail pair trade today. Table 1 Feature Equities have retraced 50% of the peak-to-trough losses, and are still consolidating the post December Fed meeting tremor. Chart 1 shows that the VIX has been cut in half and the high-yield corporate bond option-adjusted spread has dropped 105bps. Retrenching volatility and deflating junk spreads suggest that the equity risk premium (ERP) remains uncharacteristically high. The path of least resistance is for the ERP to narrow in the coming months as we do not foresee recession in 2019. As a reminder, the ERP and the economy are inversely correlated. Chart 1Risk Premia Renormalization Nevertheless, in order for the reflex rebound since the late-December lows to morph into a durable rally, the macro/policy backdrop has to turn from a headwind to a tailwind. We are closely monitoring three potential positive catalysts: A definitively more dovish Fed, which would help restrain the greenback A positive U.S./China trade resolution A continuation of the earnings juggernaut With regard to the macro related catalysts, an update to our reflation gauge (RG) is in order. The trade-weighted U.S. dollar has been depreciating since early November, the 10-year U.S. Treasury yield has come undone since the early November peak and oil prices are 33% lower than the early-October peak. These three variables comprise our RG and the signal is unambiguously bullish. In other words, a reflationary impulse looms in the months ahead which should pave the way for a rebound in both plunging investor sentiment and the gloomy economic surprise index (RG shown advanced, Chart 2). Chart 2Reflating Away On the earnings front, last week we trimmed our end-2020 SPX EPS forecast to $181 while we sustained the multiple at 16.5 times which resulted in a 3,000 SPX target.1 Drilling beneath the surface and analyzing the composition of SPX profits is revealing. Table 2 highlights sell side analysts’ profit levels and growth projections on a per GICS1 sector basis and also their contribution to overall earnings along with each sector’s projected earnings weight and most recent market capitalization weight. Table 2S&P 500 Earnings Analysis Chart 3 shows that financials, health care and industrials are responsible for 61% of the SPX’s profit growth in 2019. Interestingly, technology’s contribution has fallen to a mere 7.2% and even if we add the new communication services sector’s 9.6% contribution it still falls well shy of the tech sector’s market cap and earnings weight. Another worthwhile observation is that energy profits are no longer off the charts, as base effects since the early-2016 $25/bbl oil trough have filtered out of the dataset. While the risk of disappointment surrounds financials, health care and industrials, there are high odds that tech surprises to the upside as it has borne the brunt of recent negative earnings revisions (Charts 4 & 5). In addition, if our Commodity & Energy Strategy service’s bullish oil forecast pans out this year, the negative energy sector contribution to SPX profit growth will get a sizable upward revision (please look forward to our GICS1 sector EPS growth models updates and profit margin analysis in next week’s report). Chart 4Earnings Revisions... Chart 5...Really Weigh On Tech​​​​​​​ In sum, if the Fed pauses its hiking cycle through at least the first half of the year, we see a positive U.S./China trade resolution and SPX profits sustain their upward trajectory, then the SPX budding recovery will morph into a durable rally. This week we are updating an interest rate sensitive index that is highly levered to the surging U.S. credit impulse (Chart 6) and are initiating an early cyclical intra-sector and intra-industry pair trade. Chart 6Heed The U.S. Credit Impulse Signal Stick With Banks While our overweight call in the S&P banks index suffered a setback last month, since inception it has moved laterally, and we continue to recommend an above benchmark allocation to this key financials sub group. Not only are the odds of recession low for this year, but narrowing credit spreads and a reversal in financial conditions are also waving the green flag (junk spread shown inverted & advanced, bottom panel, Chart 7). Chart 7Bank On Banks Unlike the previous three reporting seasons when banks revealed blowout numbers and stocks subsequently fell, this season some profit and top line growth misses have been greeted with rising bank stocks prices. Such a reaction suggests that the worst is behind this sector and a sustainable recovery looms. Importantly, on the loan growth front, our credit impulse diffusion index is reaccelerating (Chart 6) and the overall credit impulse is expanding (middle panel, Chart 7). Our total loans & leases growth model and BCA’s C&I loan growth model both corroborate this encouraging credit backdrop (second & bottom panels, Chart 8). The latter is significant given that C&I loans are the single biggest credit category in bank loan books (Chart 9). Importantly, C&I loans have gone vertical recently topping the 10.5% growth mark despite softening capex intentions and CEO confidence. Chart 8Credit Models Flashing Green Chart 9Credit Models Flashing Green Multi-decade highs in consumer confidence are offsetting the Fed’s tightening cycle and suggest that consumer loans, another key lending category, will also gain traction (third panel, Chart 8). The outlook for the second largest credit category, residential real estate, remains upbeat in spite of last quarter’s soft housing related data releases. The recent easing in monetary conditions has breathed life back into the mortgage purchase applications index and also house prices continue to expand at a healthy pace (Chart 10). The upshot is that first-time home buyers will show up this spring selling season. Chart 10Residential Loans Also On Solid Footing Beyond positive credit growth prospects, credit quality remains pristine. BCA’s no recession in 2019 view remains intact, thus NPLs and chargeoffs should stay muted. As a reminder, U.S. banks are the best capitalized banks in the world,2 and their reserve coverage ratio has returned to 124%, a level last seen in 2007 (Chart 11). Chart 11Pristine Credit Quality Another important source of support is equity retirement. Banks have been late to the buyback game as the GFC along with the new strict bank regulatory body, the Fed, really tied their hands with regard to shareholder friendly activities. In fact, according to flow of funds data, the financial sector is still a net equity issuer, albeit at a steeply decelerating pace especially relative to the non-financial corporate sector (Chart 12). Pent up financial sector buyback demand is a boon for bank EPS growth. Chart 12Pent Up Buyback Demand Getting Unleashed This is significant at a time when analysts have been swiftly downgrading EPS growth figures for the SPX. Encouragingly, our bank EPS growth model captures all these positive forces and while it is decelerating it still suggests that profit growth will be stellar in 2019 and easily outpace the overall market (Chart 13). Chart 13Banks EPS Growth Will Outpace The Market Despite all this enticing news, bank valuations remain anchored near rock bottom levels and a resurgent ROE is signaling that a re-rating phase looms (Chart 14). Chart 14Rerating In Still In The Early Innings Nevertheless, there is one headwind banks face as the business cycle is long in the tooth and on track to become the longest expansion on record: the price of credit. One reason for the deflating relative stock price ratio since the January 2018 peak has been the yield curve slope flattening (Chart 15), as it suppresses bank net interest margins. Banks have been fighting this off partly by keeping their source of funding ultra-low judging by still anemic CD rates, according to Bankrate’s national average (bottom panel, Chart 15). Chart 15One Minor Headwind While yield curve inversions have widened all the way out to the 7/1 slope, the key 10/2 slope has yet to invert. Were the 10-year U.S. treasury to resume its selloff, even a mild yield curve steepening will go a long way, as BCA’s bond strategists expect. Clearly a flattening curve is a risk to our sanguine bank view, but the rest of the positives we outlined above more than offset the yield curve blues. Adding it all up, vibrant and broad-based bank credit growth, pristine credit quality, pent up bank buyback demand and a V-shaped recovery in bank ROE more than offset the risk of the 10/2 yield curve inversion, and suggest that the path of least resistance is higher for the S&P banks index. Bottom Line: Maintain the overweight stance in the S&P banks index. The ticker symbols for the stocks in this index are: BLBG: S5BANKX – WFC, JPM, BAC, C, USB, PNC, BBT, STI, MTB, FITB, CFG, RF, KEY, HBAN, CMA, ZION, PBCT, SIVB, FRC, . Buy Homebuilders/Sell Home Improvement Retailers While we reiterate our recent overweight call on the S&P homebuilding index3 and the high-conviction underweight call on the S&P home improvement retail (HIR) group,4 it also makes sense to initiate a market neutral trade: long homebuilders/short HIR. This pair trade is levered on the swings of residential construction compared with residential investment. Currently the former is significantly outpacing the latter and suggests that relative share prices have ample room to run (top panel, Chart 16). Chart 16A Play On Residential Construction Vs. Investment Put differently, this share price ratio moves in tandem with homebuilders breaking new ground versus home owners renovating their existing house. Chart 17 shows the NAHB’s homebuilder sales expectations survey compared with the remodeling expectations survey. This relative sentiment gauge has ticked up recently, confirming the message from national accounts that residential construction has the upper hand over residential investment. The upshot is that the bull market in relative share prices is in the early innings. Chart 17Relative Survey Expectations... Keep in mind that housing starts and building permits are extremely sensitive to interest rates, depend on first time home buyers and move in lockstep with the homeownership rate. Currently, interest rates are easing, the homeownership rate is coming out of its GFC funk and first time home buyers are slated to make a comeback this spring selling season. This is a boon for homebuilders at the expense of HIR (middle & bottom panels, Chart 16). More specifically on the interest rate front, while both groups move with the oscillation of lending rates, new home sales are more sensitive than HIR sales to the price of credit. Our proxy of mortgage application purchase to refinance index does an excellent job in capturing this relative interest rate sensitivity and the recent jump signals that a catch up phase looms in the relative share price ratio (top panel, Chart 18). Chart 18...Easing Interest Rates... Relative loan growth activity also corroborates that demand for residential real estate is outpacing demand for home renovation (bottom panel, Chart 18). Beyond these macro tailwinds for this intra-sector trade, the price of lumber is a key determinant of relative profitability: lumber represents an input cost to homebuilders whereas it is an important selling item in Big Box building & supply retailers that make a set margin on it. In other words, rising lumber prices are a boon for HIR and a bane to homebuilders and vice versa. The recent drubbing in lumber prices should ease margin pressures on homebuilders but eat into HIR profits (Chart 19). Chart 19...And Cheapened Lumber Prices Favor Homebuilders Over HIR Finally, oversold relative technicals, depressed valuations and extreme sell side analysts’ relative profit pessimism, offer a very compelling entry point in the pair trade for fresh capital (Chart 20). Chart 20Oversold And Unloved Netting it all out, rising residential construction versus stalling residential investment, easing interest rates, cheapened lumber prices, and relative alluring valuations and technicals all signal that more gains are in store for homebuilders at the expense of home improvement retailers. Bottom Line: Initiate a new long S&P homebuilding/short S&P home improvement retail pair trade today. The ticker symbols for the stocks in these indexes are: BLBG: S5HOME – DHI, LEN and PHM, and BLBG: S5HOMI – HD and LOW, respectively.   Anastasios Avgeriou, Vice President U.S. Equity Strategy anastasios@bcaresearch.com footnotes 1 Please see BCA U.S. Equity Strategy Report, “Catharsis” dated January 14, 2019, available at uses.bcaresearch.com. 2 Please see BCA U.S. Equity Strategy Special Report, “Top 10 Reasons We Still Like Banks” dated March 5, 2018, available at uses.bcaresearch.com. 3 Please see BCA U.S. Equity Strategy Report, “Indurated” dated September 24, 2018, available at uses.bcaresearch.com. 4 Please see BCA U.S. Equity Strategy Report, “2019 Key Views: High-Conviction Calls” dated December 3, 2018, available at uses.bcaresearch.com. Current Recommendations Current Trades Size And Style Views Favor value over growth Favor large over small caps

Highlights 2018 Performance Breakdown: Our recommended model bond portfolio generated a modest outperformance versus the custom benchmark index of +6bps for all of 2018. Winners & Losers: The outperformance of our model bond portfolio in 2018 mostly came from country selection on our government bond portfolio (underweight U.S. Treasuries, overweight the U.K. and Australia). However, our below-benchmark overall duration stance, as well as our bias favoring U.S. credit over non-U.S. corporates, were drags on performance during the risk-off moves at the end of 2018. Scenario Analysis For 2019: The tactical upgrade to global corporates that we initiated last week is projected to generate outperformance versus the model portfolio benchmark index in the next six months - both from below-benchmark duration positioning and higher exposure to U.S. corporates. Feature 2019 has gotten off to a very busy start, with significant news and market moves forcing us to devote our first two Weekly Reports of the year to analysis and even changes to our views. This week, we belatedly take care of one final piece of housekeeping for 2018 – reporting the performance of the BCA Global Fixed Income Strategy (GFIS) model bond portfolio for the fourth quarter and for the entire calendar year. We also present an updated scenario return analysis for the next six months after the tactical upgrade to global corporate bonds that we initiated last week.1 As a reminder to existing readers (and to new clients), the model portfolio is a part of our service that complements the usual macro analysis of global fixed income markets. The portfolio is how we communicate our opinion on the relative attractiveness between government bond and spread product sectors. This is done by applying actual percentage weightings to each of our recommendations within a fully invested hypothetical bond portfolio. A Quick Summary Of The Full Year Performance For 2018 The 2018 performance of the model portfolio can really be broken up into two periods: the first ten months of the year and November/December. This is an unsurprising consequence of the severe market moves around year-end that went contrary to our two most significant recommendations – maintaining a below-benchmark stance on overall portfolio duration and overweighting U.S. investment grade corporate debt versus non-U.S. equivalents in Europe and emerging markets (EM). The overall portfolio return in 2018 was +1.10% (hedged into USD), which outperformed our custom benchmark index by +6bps (Chart of the Week).2 That outperformance was considerably higher before the year-end plunge in global bond yields, reaching a peak of +32bps on November 20. In terms of the breakdown of outperformance, our recommended positioning on government bonds (duration and country allocation) contributed +22bps, while our credit tilts (by country and broadly defined credit sectors) were a drag on performance to the tune of -16bps. Chart of the WeekA Small Gain For 2018 After A Q4 Round-Trip The full breakdown of the full-year 2018 performance can be found in the Appendix tables and charts on Pages 14-16. For the government bond portion of the portfolio the full-year outperformers by country were the U.S. (+18bps), Germany (+10bps), Australia (+4bps) and the U.K. (+3bps). These are in line with our long-standing underweight position on the U.S. versus Germany, and our recommended overweights on Australia and the U.K. The laggards were relatively modest, led by our overweight stance on Japan (-4bps) and underweights on France (-3bps) and Italy (-3bps). For the credit portion of the portfolio, the winners were EM USD-denominated corporates (+7bps), U.S. B-rated high-yield (HY) corporates (+3bps) and U.S. Caa-rated high-yield (HY) corporates (+2bps). This was in line with our long-standing bias to favor U.S. junk bonds over EM credit. The losers were our overweights on U.S. investment grade (IG) financials (-15bps), U.S. IG industrials (-8bps), U.S. Ba-rated HY (-4bps), and euro area IG corporates (-2bps). Our overweight tilts on U.S. IG were the issue here. Q4/2018 Model Portfolio Performance Breakdown: A “Risk-Off” Hit To Our Core Recommendations The detailed data on our model bond performance for Q4/2018 only can be found in Table 1. Table 1GFIS Model Bond Portfolio Q4/2018 Overall Return Attribution The total return of the GFIS model bond portfolio was +1.5% (hedged into USD) in Q4, which underperformed the custom benchmark index by a mere -1bp. The main cause for the slight underperformance is from our below-benchmark duration positioning with the Bloomberg Barclays Global Treasury Index yield falling by 20bps over the full quarter. In terms of the specific breakdown between the government bond and spread product allocations in our model portfolio, the former generated -2bps of underperformance versus our custom benchmark index while the latter outperformed by just +1bp. The bar charts showing the total and relative returns for each individual government bond market and spread product sector are presented in Charts 2 and 3.   The main individual sectors of the portfolio that drove the excess returns were the following: Biggest outperformers Underweight U.S. government bonds with maturities between 5-7 years (+12bps) Overweight Japanese government bonds (JGBs) with maturities between 7-10 years (+5bps) Underweight Germany government bonds with maturities between 7-10 years (+3bps) Overweight U.K. government bonds with maturities between 5-7 years (+2bps) Biggest underperformers Overweight Japanese government bonds (JGBs) with maturities beyond 10 years (-15bps) Underweight U.S. government bonds with maturities beyond 10 years (-8bps) Underweight Italy government bonds with maturities beyond 10 years (-3bps) Underweight France government bonds with maturities beyond 10 years (-2bps) Chart 4 presents the ranked benchmark index returns of the individual countries and spread product sectors in the GFIS model bond portfolio for Q4/2018. The returns are hedged into U.S. dollars (we do not take active currency risk in this portfolio) and are adjusted to reflect duration differences between each country/sector and the overall custom benchmark index for the model portfolio. We have also color-coded the bars in each chart to reflect our recommended investment stance for each market during Q4/2018 (red for underweight, blue for overweight, gray for neutral). Government bonds are dominating the left half of the chart, as yields declined in the final months of 2018. This was a drag on our model portfolio performance. However, the best performing sector was U.K. government bonds, generating a total return of 4.7% in Q4/2018 (on a currency-hedged and duration-matched basis). The GFIS model portfolio benefited from this move, given our long-standing overweight bias for U.K. Gilts. The right side of Chart 4 is occupied by global spread product, where currency-hedged returns were flat-to-negative in Q4. This was due to credit spread widening as investors feared both slower global growth and additional Fed tightening. The riskier parts of the corporate bond universe – high-yield, EM corporates – suffered the largest losses. The total return of Bloomberg Barclays U.S. High-Yield Index (currency-hedged into USD) for Q4 was -2.7%, as the option-adjusted spread (OAS) widened by +206bps. Unfortunately for our model portfolio, our preference for U.S. corporate bonds over European and EM credit hurt performance, although not by as much as the below-benchmark duration stance. We are disappointed by the final result for the year, although we are still pleased to generate even a small positive outperformance given the ferocity of the market moves seen at the end of 2018. We can attribute that to lingering gains from good calls made earlier in 2018, but also from our recommended cautious stance on overall portfolio risk (i.e. tracking error) in a more-volatile investment environment. Bottom Line: The outperformance of our model bond portfolio in 2018 mostly came from country selection on our government bond portfolio (underweight U.S. Treasuries, overweight the U.K. and Australia). However, our below-benchmark overall duration stance, as well as our bias favoring U.S. credit over non-U.S. corporates, were drags on performance during the risk-off moves at the end of 2018. Future Drivers Of Portfolio Returns Looking ahead, the performance of the model bond portfolio will benefit from two main factors: our below-benchmark duration bias and our underweight stance on global government bonds versus corporate debt. In terms of specific weighting in the GFIS model bond portfolio, we now have a tactical bias favoring global corporate debt over government debt coming on top of our below-benchmark duration stance (Chart 5). We are sticking with the latter position, about one full year short of the duration of our benchmark index, with global yield curves priced for inflation expectations that are too low and with no rate hikes discounted for 2019 in all major developed markets. Chart 5Portfolio Duration: Staying Below-Benchmark However, we are also keeping our current country allocations on the government bond side of the model portfolio, even after our tactical credit upgrade. That means staying underweight countries where policymakers are only pausing on rate hiking cycles (U.S. and Canada), while overweighting countries that are likely to keep rates on hold for all of 2019 (Japan, U.K., Australia). Our decision to upgrade global credit exposure helps boost the yield of our model portfolio (Chart 6). However, the portfolio is still yielding less than the benchmark thanks to our bear-steepening bias on government yield curves that involves underweights to longer-maturity bonds with higher yields. Chart 6Portfolio Yield: Credit Upgrade Helps Offset Defensive Duration Tilt Importantly, all the changes that were made to our portfolio allocations last week – raising weights on all global corporate bond markets, cutting exposure to government debt in the U.S., Germany and France – did not materially change the tracking error (relative volatility versus the benchmark) of the model portfolio. We do not see the current backdrop as being conducive to taking high levels of overall portfolio risk, even given our tactical view that the U.S. monetary policy will be on hold for the next 3-6 months. We prefer to recommend more relative value positioning via country and credit allocations that help dampen overall portfolio risk and reduce exposure to the kind of volatility spikes that became more frequent in 2018. Thus, we will continue to target a tracking error for the model portfolio of 40-60bps, well below our self-imposed 100bps ceiling (Chart 7). Chart 7Portfolio Risk Budget Usage: Staying Cautious Scenario Analysis & Return Forecasts In April 2018, we introduced a framework for estimating total returns for all government bond markets and spread product sectors, based on common risk factors.3 For credit, returns are estimated as a function of changes in the U.S. dollar, the Fed funds rate, oil prices and market volatility as proxied by the VIX index (Table 2A). For government bonds, non-U.S. yield changes are estimated using historical betas to changes in U.S. Treasury yields (Table 2B). This framework allows us to conduct scenario analysis of projected returns for each asset class in the model bond portfolio by making assumptions on those individual risk factors.   In Tables 3A & 3B, we present three differing scenarios, with all the following changes occurring over a six-month horizon. Note that this differs from how we have typically presented these scenario analyses, with projections over the subsequent twelve months. Given that the changes to our recommended allocations introduced last week were tactical in nature (i.e. up to six months), we are shortening our forecast window for this particular scenario analysis to line up with that shorter investment horizon.   The scenarios are all driven by what we believe will be the most important driver of market returns in 2019 – the path of U.S. monetary policy. Our Base Case: the Fed stays on hold, the U.S. dollar remain flat, oil prices rise by +10%, the VIX index falls to 15, and there is a bear-steepening of the U.S. Treasury curve. This scenario is the one we laid out in last week’s report, with the Fed taking a pause through at least the March FOMC meeting, allowing market volatility to drift lower as U.S. monetary and financial conditions ease. A Very Hawkish Fed: the Fed does a surprise +25bps rate hike in March, the U.S. dollar rises by +5%, oil prices increase +20%, the VIX index climbs to 25 and there is a sharp bear-flattening of the U.S. Treasury curve. This would be the case if the U.S. economy maintains firm growth, the global growth downturn stabilizes, U.S. inflation expectations increase and market volatility increases from a surprisingly hawkish Fed. A Very Dovish Fed: the Fed cuts the funds rate by -25bps, the U.S. dollar falls by -5%, oil prices decline -20%, the VIX index increases to 35 and there is a sharp bull steepening of the U.S. Treasury curve. This is a scenario where U.S./global growth slows rapidly from the current pace and the Fed has no choice but to ease monetary policy as market volatility surges alongside elevated recession risks. The model bond portfolio is expected to outperform the custom benchmark index by +19bps in our Base Case scenario. This comes from the relative outperformance of credit versus government bonds in an environment of rising bond yields (which also benefits our below-benchmark duration stance), and tighter credit spreads. In the Very Hawkish Fed scenario, our model portfolio is expected to outperform the benchmark by +29bps. This is not only due to our duration tilt and our corporates-versus-governments bias. As in the Base Case, the relative stance favoring U.S. corporates over EM credit would benefit from a backdrop of tightening U.S. monetary conditions and rising market volatility (Chart 8) – both of which are worse for EM credit. Chart 8Risk Factors Assumptions For The Scenario Analysis In the Very Dovish Fed scenario, our portfolio is projected to underperform the benchmark index by -26bps, with falling bond yields (Chart 9) hitting both our defensive duration bias and the overweight on corporates relative to governments. Chart 9UST Yield Assumptions For The Scenario Analysis In all three scenarios, there is a drag on expected performance from the relative carry of the model portfolio versus the benchmark (-17bps). This comes mostly from the below-benchmark overall duration stance that involves reduced exposure to longer-maturity government bonds with higher yields. The drag on carry also comes from the underweight positioning on high-yielding EM debt. We are maintaining that tilt given our concerns that China’s policymakers will be unable to provide enough stimulus to benefit EM economies through greater Chinese demand. Importantly, our recommended allocations win in scenarios that do not involve Fed rate cuts, a very low probability outcome in 2019, in our view. Thus, we expect our current allocations to generate alpha in the first half of the year, even if the Fed returns to a hawkish bias faster than we currently anticipate. Bottom Line: The tactical upgrade to global corporates that we initiated last week is projected to generate outperformance versus the model portfolio benchmark index in the next six months - both from below-benchmark duration positioning and higher exposure to U.S. corporates.   Robert Robis, CFA, Senior Vice President Global Fixed Income Strategy rrobis@bcaresearch.com Ray Park, CFA, Research Analyst ray@bcaresearch.com Footnotes 1 Please see BCA Global Fixed Income Strategy Weekly Report, “Enough With The Gloom: Upgrade Global Corporates On A Tactical Basis”, dated January 15th 2019, available at gfis.bcareseach.com. 2 The GFIS model bond portfolio custom benchmark index is the Bloomberg Barclays Global Aggregate Index, but with allocations to global high-yield corporate debt replacing very high quality spread product (i.e. AA-rated). We believe this to be more indicative of the typical internal benchmark used by global multi-sector fixed income managers. 3 Please see BCA Global Fixed Income Strategy Weekly Report, “GFIS Model Bond Portfolio Q1/2018 Performance Review: A Rough Start”, dated April 10th 2018, available at gfis.bcareseach.com. Appendix     Recommendations The GFIS Recommended Portfolio Vs. The Custom Benchmark Index Duration Regional Allocation Spread Product Tactical Trades Yields & Returns Global Bond Yields Historical Returns

Highlights Yield Curve Drivers: A rebound in rate hike expectations will cause the curve to steepen somewhat during the next few months, though accelerating wages limit the upside. The yield curve will not invert until after long-dated inflation expectations are fully re-anchored, probably not until late in the year. Yield Curve Positioning: Correlations that have been in place since the financial crisis show that the 5-year and 7-year maturities are most sensitive to changes in near-term rate hike expectations. With the discounter likely to move higher in the coming months, investors should favor yield curve trades that are short that portion of the curve. Investment Recommendation: Close our recommended long 2-year short 1-year/5-year trade for a profit of 2 bps. Replace it with a position short the 7-year bullet and long a duration-matched 2-year/30-year barbell. Feature The yield curve flattened throughout most of 2018, and actually fell enough that talk of curve inversion hit a fever pitch last November, around the same time that the market started to doubt the Fed’s ability to lift rates (Chart 1). As of today, the 2/10 Treasury slope sits at a mere 17 basis points, but we don’t see it falling below zero any time soon.1 Chart 1Too Soon For Curve Inversion In this week’s report we consider the factors that will determine how the slope of the curve evolves over the next few months, and also recommend an investment strategy to take advantage of those movements. Yield Curve: Macro Drivers Driver 1: Rate Hike Expectations The number one factor that will influence the slope of the yield curve in the coming months is the market’s assessment of the near-term path for Fed rate hikes. Chart 2 shows the 5-year rolling correlation between monthly changes in the 2/10 slope and monthly changes in our 12-month Fed Funds Discounter. A positive correlation means that the 2/10 slope steepens when the market prices in more rate hikes and flattens when it prices in fewer hikes. A negative correlation means that the slope flattens when the market prices in more hikes and steepens when it prices in fewer hikes. Chart 2Rising Rate Expectations = Steeper 2/10 Slope The correlation was consistently negative throughout the pre-crisis period because the 2-year yield reacted more to changes in near-term rate hike expectations than the 10-year yield. In other words, a given increase (decrease) in the discounter would lead to a larger increase (decrease) in the 2-year yield than in the 10-year yield, and the curve flattened (steepened) as a result. But this correlation flipped following the Great Recession. Zero-bound interest rates and Fed forward guidance were an important reason for the switch. But even during the past few months, as the 12-month discounter fell from 66 bps in early November to -1 bp currently, the 10-year yield fell by 45 bps and the 2-year yield by only 36 bps. Even with interest rates off zero and the Fed scaling back its forward guidance, the positive correlation between the 2/10 slope and the 12-month discounter persists. We think that the 12-month discounter is close to its near-term bottom. Our Fed Monitor has fallen somewhat in recent months but it remains above zero, suggesting that the economy requires further monetary tightening (Chart 3). A look at the three components of our Monitor gives us even more confidence that the discounter is near its trough. The economic growth component of the Monitor is nicely above zero (Chart 3, panel 3), and the inflation component continues to trend up (Chart 3, panel 4). All of the Fed Monitor’s recent weakness can be attributed to tighter financial conditions (Chart 3, bottom panel). As we discussed in last week’s report, now that the market views Fed policy as much more accommodative, it is only a matter of time before financial conditions ease.2 Chart 3Fed Monitor Still Suggests Tightening In fact, some easing has already begun (Chart 4): Chart 4Financial Conditions Starting To Ease The stock-to-bond total return ratio has bottomed (Chart 4, top panel) High-Yield spreads have peaked (Chart 4, panel 2) The VIX has moderated (Chart 4, panel 3) The trade-weighted dollar has started to depreciate (Chart 4, bottom panel) Ironically, easier financial conditions will give the Fed the green light to re-start rate hikes, probably by June, and this could re-test risk assets in the second half of the year. But between now and then, a move higher in 12-month rate expectations will apply some steepening pressure to the 2/10 slope. Driver 2: Inflation Expectations Instead of looking at nominal yields and rate hike expectations, another approach is to split yields into their real and inflation components. This is potentially revealing in the current environment since a large portion of the recent drop in yields was driven by the cost of inflation compensation. Since the November 8 peak in the discounter, the cost of 10-year inflation protection fell 26 bps and the real 10-year yield fell 19 bps. The cost of 2-year inflation protection declined 46 bps while the real 2-year yield actually rose 10 bps. Based on those numbers, it is evident that when the cost of inflation compensation fell alongside the oil price, it exerted a steepening pressure on the yield curve that was offset by a flattening in the real yield curve. One might conclude that a rebound in inflation will cause the curve to flatten going forward. That is probably true in the event of a pure inflation shock that does not impact global growth. But such a shock is highly unlikely. Oil (and other commodity) prices fell during the past few months because of a slowdown in global growth. A rebound in commodity prices that drives inflation higher will almost certainly occur alongside stronger global growth. In other words, splitting nominal yields into the real and inflation components probably doesn’t get us any closer to figuring out the near-term path for the yield curve. A better way to incorporate the cost of inflation compensation into our thinking about the yield curve is to focus on the 5-year/5-year forward TIPS breakeven inflation rate. That rate is currently 1.99%, well below the range of 2.3%-2.5% that has historically been consistent with well-anchored inflation expectations (Chart 5). Chart 5Inflation Expectations Are Too Low For The Fed It is difficult to believe that the Fed would allow the yield curve to invert with the 5-year/5-year breakeven rate so low. The combination of an inverted yield curve and below-target inflation expectations would signal that the Fed wants to run a restrictive monetary policy before inflation has fully recovered. That would be completely contrary to the Fed’s mandate. From this argument, we reason that the 2/10 slope is unlikely to sustainably fall below zero until the 5-year/5-year forward TIPS breakeven rate is at least above 2.3%. With the 2/10 slope already at 17 bps, this means it is much more likely to stay near its current level or steepen somewhat during the next few months. Driver 3: Wage Growth The third factor driving our yield curve view is the pace of wage growth. Stronger wage growth is tightly correlated with a flatter yield curve, though the yield curve tends to lead wage growth by 6-12 months (Chart 6). Chart 6A Flatter Curve Leads Faster Wage Growth Higher Wage Growth = Flatter Curve In fact, a typical cyclical pattern is for the 2/10 slope to flatten rapidly and then stay at a low (but positive) level for some time as wage growth catches up. In that sense, this cycle is playing out just like every other. The yield curve has already undergone its large flattening and wage growth is now accelerating to catch up. Bottom Line: The three factors discussed above lead us to expect a small amount of curve steepening during the next few months. A rebound in rate hike expectations due to easier financial conditions will cause the curve to steepen, though accelerating wages limit the upside. The yield curve will not invert until after long-dated inflation expectations are fully re-anchored, probably not until late in the year. Yield Curve Positioning In the first section of this report we noted that the 10-year yield fell by more than the 2-year yield between the early-November peak in the 12-month discounter and today. But Table 1 shows that the 5-year and 7-year yields fell by even more. This is the expected result. Table 1Treasury Curve From Peak In 12-Month Discounter To Present Turning once again to the correlations between different segments of the yield curve and our 12-month discounter, we see that yield curve segments out to the 5-year maturity point are all positively correlated with the 12-month discounter. Also, curve segments beyond the 7-year maturity point are all negatively correlated with the discounter. The 5/7 slope has virtually no correlation (Chart 7). Chart 75-Year & 7-Year Are Most Sensitive To Rate Expectations These correlations tell us that we should expect the 5-year and 7-year yields to move the most in response to changes in the 12-month discounter. In other words, if we expect the discounter to move higher in the coming months we should maintain short exposure to this part of the curve. This short exposure should be offset by long exposure at either the very short-end or the very long-end of the curve, where yields will see less upside when the discounter rebounds. To figure out where to focus this long exposure we can turn to our butterfly spread models.3 Table 2 presents the raw residuals from our butterfly spread models. These models are based on regressions of different butterfly spreads versus the slope of the yield curve segment that spans the two wings of the barbell portion of the trade. For example, Table 2 shows a residual of -9 bps for the 5-year bullet relative to the 2/10 barbell. This means that the 5-year appears 9 bps expensive versus the 2/10 barbell, given where the slope of the 2/10 curve is today. Table 3 shows the standardized residuals from the different curve models so that they can be compared against each other. Table 2Butterfly Strategy Valuation: Residuals Table 3Butterfly Strategy Valuation: Standardized Residuals Notice in Tables 2 and 3 that almost all of the numbers are negative. This means that bullet trades are currently expensive relative to barbell trades. Using our criteria of wanting to be short the 5-year or 7-year part of the curve, we can use the tables to see that a position short the 7-year bullet and long the duration-matched 2-year/30-year barbell has an attractive standardized residual of -1.00. This appears to be the most attractive curve trade for the current environment. As such, today we close our current yield curve recommendation to favor the 2-year bullet over the 1-year/5-year barbell for a gain of 2 bps. This recommendation had been in place since November 5. In its place, we initiate a recommendation to go long a duration-matched barbell consisting of the 2-year and 30-year maturities and short the 7-year note. Bottom Line: Correlations that have been in place since the financial crisis show that the 5-year and 7-year maturities are most sensitive to changes in near-term rate hike expectations. With the discounter likely to move higher in the coming months, investors should favor yield curve trades that are short that portion of the curve. With that in mind, we close our 2-year over 1-year/5-year trade and initiate a position short the 7-year bullet and long a duration-matched 2-year/30-year barbell. Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com Footnotes 1 We don’t expect to see sustained yield curve inversion until late this year. For further details please see U.S. Bond Strategy Special Report, “2019 Key Views: Implications For U.S. Fixed Income”, dated December 11, 2018, available at usbs.bcaresearch.com 2 Please see U.S. Bond Strategy Weekly Report, “Buy Corporate Credit”, dated January 15, 2019, available at usbs.bcaresearch.com 3 For further details on the models please see U.S. Bond Strategy Special Report, “More Bullets, Barbells And Butterflies”, dated May 15, 2018, available at usbs.bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification

Special Report

Highlights MLPs’ one-of-a-kind legal structure offers investors gaudy distribution yields and tax-saving advantages. They boomed alongside fracking, enjoying spectacular growth between 2009 and 2014. MLPs used to exhibit a high correlation with utilities, but since the 2014 oil bust, they have performed in step with the rest of the energy sector. Improved valuations have recently put MLPs back on investors’ radar. However, structural impediments and heterogeneous balance-sheet quality argue against broad index exposure. Investors would be better served by concentrating their efforts on picking individual stocks. Opportunities reside within smaller-cap MLPs and MLPs exposed to the Permian basin. Feature Dear Client, In place of a Weekly Report written from South Africa, where I have been meeting with clients, we are sending you this Special Report on Master Limited Partnerships (MLPs), written by my colleague Jennifer Lacombe.* Like mortgage REITs, which U.S. Investment Strategy followed from 2011 to 2013, MLPs are a yield play that investors might find to be an appealing bond alternative. We trust that you will find this report interesting and informative. Best regards, Doug Peta, Senior Vice President U.S. Investment Strategy * This report was initially published by our Global ETF Strategy service on November 15, 2018. It has been lightly revised to update charts and reflect subsequent market developments.   Q: What are MLPs and their tax benefits? Master Limited Partnerships (MLPs) are publicly listed partnerships involving two classes of partners. A General Partner (GP) controls the assets and manages the daily operations of the business. Limited Partners (LPs) - and public investors - provide the capital and collect cash flow distributions. Unlike corporations, which pay corporate taxes on their income, MLPs have the ability to pass through all of their income to their owners, along with deductible items like amortization and depreciation expenses. MLP investors, in turn pay income tax at their own individual marginal tax rates. MLP owners are thereby shielded from the double taxation that would otherwise apply when the corporation paid taxes on its income, and the shareholder paid taxes on the dividend distributed from the corporation’s income. Q: Why are they predominantly found in the energy sector? Concerns about the potential loss of federal income led Congress to limit MLP eligibility to companies in the energy and real estate sectors when it overhauled the tax code in 1986. Since the 1986 Act took effect, MLPs have had to generate at least 90% of “qualifying income” from their energy or real estate operations. Section 7704 of the Internal Revenue Code defines “qualifying income” as income derived from exploration, development, mining or production, processing, refining, transportation or marketing of any mineral or natural resource, as well as certain passive-type income including interest, dividends and real property rents. Over the years, the shale revolution and the rise of new technologies, such as horizontal drilling and fracturing, created elevated demand for energy infrastructure. Today, MLPs almost exclusively operate in the natural resources space (Chart 1). Q: Why did MLPs outperform assets of all stripes following the Great Financial Crisis? A combination of several factors led MLPs to record stunning returns between 2009 and 2014. The Alerian MLP Total Return Index grew by a whopping annualized rate of 38% during that time. Decreasing interest-rate environments are typically supportive of yield plays’ outperformance. Powered by high single-digit to double-digit distribution yields, MLPs led Treasuries, utilities stocks, high-yield bonds and even the S&P 500 over that six-year stretch (Chart 2). With the shale revolution in full swing, sustaining strong demand for pipelines and other energy infrastructure, investors’ funds flowed abundantly into the energy MLP space (Chart 3). Prices - a mathematical function of multiples and earnings - soared as money kept pouring in and P/E tripled in the first 7 years following the Great Financial Crisis (Chart 4). Chart 2Decreasing Interest Rates Are A Boon To Yield Plays   Chart 3Horizontal Drilling Attracted A Lot Of Money...   Chart 4...Sending Multiples Soaring Q: Why has such outperformance not attracted more institutional and foreign investors? Because of U.S. tax rules, MLPs are relatively unattractive to tax-exempt investors and non-U.S. investors. The tax rule for U.S. tax-exempt investors – institutional investors such as pension funds, university endowments, charities and IRAs – treats MLP earnings as unrelated business taxable income (UBTI), making them subject to income tax. Moreover, to retain their own pass-through status and tax shield, open-ended funds – like many mutual funds and ETFs – can allocate no more than 25% of their total holdings to MLPs, and no more than 10% to a single MLP. U.S. tax rules consider foreign owners of MLPs to be engaged in a business in the U.S., and require them to file and pay U.S. federal income tax. Therefore, only U.S. individuals can truly reap the full benefits of the MLP structure. Though they easily access these securities on public exchanges, the tax shield comes at the price of convoluted accounting treatments. Unitholders receive Schedule K-1 tax forms that can be complicated enough to result in significant accounting costs. They are most suited for high net worth investors’ portfolios, although smaller investors who are not daunted by accounting burdens have also embraced the vehicle. Q: Why are MLP yields so high? The typical MLP partnership agreement incentivizes a GP to distribute all available cash to unitholders, after retaining reserves for business operations and liabilities. Not only does the corporate tax exemption increase the amount of available cash, but the General Partner also has wide discretion over the amount of retained reserves. Because distributions are the main determinant of any yield play’s performance, GPs have historically emphasized distribution yields – sometimes at the expense of retained earnings. The more assurance investors have that they will receive reliable cash flows, the better the MLP will perform in the market. Q: Do MLPs trade like other bond proxies? The distribution model worked beautifully during the shale-oil boom. Low retained reserves never became an issue because MLPs collected steady revenues – a function of prices and volumes of oil or gas processed - and could fund distributions in excess of operating cash flow by issuing new debt or equity. Investors were so eager to invest that GPs found themselves at the controls of a positive feedback loop in which the more cash they distributed to investors, the more capital flowed in to fund even higher distributions. The infrastructure-heavy business model and high payout ratios echoed companies in the utilities sector and, indeed, MLP returns correlated strongly with utilities stocks. However, the discretion embedded in the MLP model reached a breaking point soon after the oil bust arrived in mid-2014. The price-led decline in revenues necessitated distribution cuts and severed the correlation with utilities (Chart 5). Chart 5A Utilities Proxy No More... Q: Were MLPs immune to energy price swings before the 2014 bust? Conventional investor wisdom maintains that MLPs are immune to commodity price swings in the aggregate because of their utility-like characteristics and because long-term contracts lock in selling prices. Actually, however, MLP revenue structures differ greatly from one line of activity to the other. Natural gas pipeline transportation accounts for a quarter of aggregate MLP activity. Prices per unit of volume transited are contractually locked in 5-to-20-year contracts, providing immunity to spot price moves during the entire duration of the contract. Storage (natural gas not immediately needed, or crude oil waiting to be refined) accounts for another quarter of aggregate activity and is subject to a similar pricing model as natural gas pipelines. Only the contract lengths are much shorter, ranging from 1 to 5 years. Petroleum pipeline transportation accounts for 44% of MLP activity. Contracts locking prices over the long run are not typical in this line of business. The Federal Energy Regulatory Commission (FERC) also imposes a yearly price increase amounting to the Producer Price Index for Finished Goods, plus a 1.23% adjustment. MLP revenue structures are therefore varied, and only natural gas pipeline transportation’s revenue streams - a quarter of the sector – are truly immune to fluctuations in spot prices, thanks to their long-term contracts. It follows that MLPs in aggregate are indeed correlated with energy price swings and trade closely in line with energy stocks (Chart 6). Chart 6...An Energy Proxy Instead Up until recently, their correlation to spot oil prices in particular was even more striking. However, they failed to match the 2017-18 recovery in oil markets (Chart 8). Because cash flow reliability is a key driver of the investment decision for any yield play, distribution cuts are bound to make any MLP investors skittish, and oil prices may have to enter an extended bull market before they overcome their fears (Chart 7).   Chart 8...Kept MLPs Depressed In Spite Of Oil Price Recovery Q: So, how cheap are they now? Since its peak in the summer of 2014, the Alerian MLP Total Return index has declined by 38% and is now flirting with the two-standard-deviation-cheap zone (Chart 9). Their profit margins have also strongly recovered (Chart 10). Chart 9Cheap Valuations...   Chart 10...Amid Recovering Profit Margins Because of the infrastructure-heavy nature of MLPs, traditional valuation metrics such as price-to-earnings can be misleading. High depreciation charges have significant impacts on earnings. Cash flows are an appropriate measure as they best inform a firm’s ability to maintain its distributions. Q: Great! So which ETF should I buy? The Alerian MLP index’s low multiples and recovering profit margins are not sufficient endorsements in themselves. An index is not an investible vehicle and even the best of index-tracking instruments can only imperfectly replicate an exposure. In the MLP space in particular, structural impediments reduce the attractiveness of exchange-traded products. Because ETFs are subject to the previously mentioned 25% cap on MLP holdings, many supplement their portfolios with regular pipeline or infrastructure stocks. Although the overall fund provides a decent exposure to the energy infrastructure sector, the diluted MLP exposure does not offer distribution yields anywhere comparable to the yields direct MLP owners receive. An alternative is to opt for a C-corporation structure. The flagship Alerian MLP ETF (ticker: AMLP) falls into this category. This structure allows for an undiluted exposure to MLPs, all the while relieving an ETF shareholder from having to deal with the complicated and costly accounting treatment that direct MLP ownership involves. However, C-corporations are subject to corporate income taxes, which cancels out the tax benefits of investing in MLPs in the first place. The resulting cumulative tax drag on returns can become substantial over time (Chart 11). Investors seek MLP exposure for the high distribution yields made possible by tax advantages. A fund will indeed provide diversification and accounting relief, but at the cost of surrendering either some yield or some of the tax advantages. This is not to mention that the bulk of the exchange-traded vehicles are Exchange Traded Notes (ETNs). Unlike ETFs, they do not own any underlying shares or units of securities. Instead, they are instruments issued and backed by financial institutions. Even in the case of well-established lending institutions, we shy away from these types of products, as we are not keen on taking unnecessary counterparty risk. Many MLP exchange-traded products are also illiquid, or have not gathered a significant mass of assets under management. The expense ratios are also high in the MLP exchange-traded product space, a result of the complicated accounting treatment of K-1 forms that are borne by the ETF or ETN sponsor (Table 1). Table 1ETNs Constitute Two Thirds Of A Relatively Illiquid Universe Q: What about the flagship Alerian MLP ETF? It’s clearly well-established. The flagship Alerian MLP ETF (ticker: AMLP) tracks the Alerian MLP Infrastructure Index and has gathered close to USD 10bn of AUM under its belt since its inception in 2010. Amid all the above limitations, it is the only viable option. However, it comes with its own set of yellow flags. Because it tracks a market-capitalization weighted index, half of the fund’s assets under management are concentrated in its five largest holdings. As we go to press, these are Magellan Midstream Partners LP, Enterprise Products Partners LP, Energy Transfer LP, Plains All American Pipeline LP and MPLX LP. These companies’ distribution yields have recovered since the 2014 oil crash, but the question of the sustainability of these cash flows is of utmost importance. Although retained earnings are at all-time highs, so is the level of debt (Chart 12). The fact that 50% of the fund is concentrated in these top 5 constituents dilutes the diversification benefits of index investing. Chart 12Distributions Are Financed By Cash Flows...And A Lot Of Debt Q: So, what are my options? The MLP universe is heterogeneous. Wide disparity in valuation (Chart 13), debt levels (Chart 14) and performance (Chart 15) indicate that opportunities reside further down the capitalization scale.     Because an index is a weighted average, a heterogeneous market does not warrant broad-index exposure, especially when the smallest constituents offer the best opportunities. Amalgamation is always a process of blending wheat and chaff together, but in this case it disproportionately favors the chaff. Stock picking thrives against this backdrop. Our expertise does not extend to evaluating individual energy MLPs. We leave the honor of recommending the best-in-class opportunities to the professional bottom-up analysts, backed by thorough and diligent review of company fundamentals and management capabilities. Where we can add value is in the analysis of economic cycles and secular macroeconomic forces. Despite the sharp fall in prices over the past two months, brought about by the surprise eleventh-hour waivers granted to Iranian oil importers, BCA’s Commodity & Energy Strategy service believes the global oil market remains tight. Our strategists expect that oil prices will recover in 2019 as OPEC producers, Russia, and Canada reduce output by an aggregate 1.4 million barrels a day, and the Iran-driven supply glut is worked off. While a 2019 oil spike would be a tailwind to petroleum pipeline MLPs, surging production in U.S. shales – led by the Permian Basin in West Texas – means the new pipeline capacity being built to accommodate higher output will find a ready market. Regardless of what happens with prices, our energy strategists foresee a localized surge in demand for transportation and other midstream services in the U.S. shales. In line with IEA projections, they expect U.S. crude oil production to grow by approximately 1.3 million barrels a day in 2019 once the constraints imposed by a lack of pipeline capacity in the fecund Permian basin ease. MLPs positioned to resolve the transportation bottleneck should be able to count on a bright near-term future. “Location, location, location” applies to pipelines as well as real estate, and reinforces a bottom-up focus when selecting MLPs.   Jennifer Lacombe, Senior Analyst jenniferl@bcaresearch.com