Developed Countries

Highlights Fed: With financial conditions easing and core inflation more likely to rise than fall, the majority of Fed officials will feel justified lifting rates again in the second half of this year. The best way to position for the resumption of rate hikes is to sell the 5-year or 7-year part of the Treasury curve and buy a duration-matched barbell consisting of the short and long ends of the curve. These sorts of positions currently offer positive carry, meaning you get paid as you wait for the market to price rate hikes back in. Corporate Spreads: Maintain an overweight allocation to corporate bonds (both investment grade and high-yield) with the exception of the Aaa credit tier. But be prepared to reduce exposure when spreads reach our target levels. Economy: Tracking estimates for 2018 Q4 and 2019 Q1 real GDP have fallen significantly during the past two weeks. The decline in tracking estimates is heavily influenced by an abnormal December retail sales report. That impact will reverse in 2019. Feature The Federal Reserve’s “on hold” strategy is now well known and has been completely discounted in the market. In fact, the overnight index swap curve is priced for 9 bps of rate cuts during the next 12 months and 21 bps of cuts during the next 24 months (Chart 1). Chart 1Primary Dealers Still Looking For Hikes At this point, the only thing that’s unclear is how the Fed will respond to the economic data going forward. Will it be eager to re-start rate hikes at the first sign of calm? Or perhaps the Fed is leaning toward a strategy where the next move will be a rate cut in the face of flagging economic growth? Survey Says Unfortunately, last month’s FOMC meeting was not accompanied by an updated Summary of Economic Projections. We therefore don’t know how policymakers have revised their rate hike expectations since December. However, the New York Fed’s Survey of Primary Dealers was updated in January, and it shows that the median primary dealer still expects two rate hikes this year. The only change between the December and January surveys is that the median primary dealer now expects one of the 2019 rate hikes in June and the other in December. In the December survey, both 2019 rate hikes were anticipated before the end of June (Chart 1). Typically, the median primary dealer and the median FOMC participant have very similar views on the future interest rate trajectory. Counting The Minutes The next stop on our search for clarity is the minutes from the January FOMC meeting, which were released last week. The January minutes provide a lot of insight into the thought processes of different FOMC participants. Unfortunately, they also reveal a serious lack of cohesion amongst the group. All in all, the document might confuse more than it clarifies. A few key excerpts from the document drive this point home. Referring to “global economic and financial developments”: Many participants observed that if uncertainty abated, the Committee would need to reassess the characterization of monetary policy as “patient” and might then use different language. This suggests that many Fed participants view the pause in rate hikes as a result of slower non-U.S. growth and tighter financial conditions. They also suggest that if global growth improves and financial conditions ease it would be appropriate to abandon a “patient” stance. … several […] participants argued that rate increases might prove necessary only if inflation outcomes were higher than in their baseline outlook. This second statement is much more dovish than the first. It suggests that several participants think that even improving global growth and an easing of financial conditions would not be sufficient to re-start rate hikes. They would also need to see inflation come in stronger than expected. Several other participants indicated that, if the economy evolved as they expected, they would view it as appropriate to raise the target range for the federal funds rate later this year. Finally, this last statement reveals that several other participants disagree with the view that an unexpected rise in inflation is a pre-condition for further rate hikes. What can we make of all this mess? The first thing that seems clear is that all Fed members view easier financial conditions as a pre-condition for further rate hikes. In this regard, we are already well on our way. Financial conditions have eased considerably since the start of the year, with the stock-to-bond total return ratio up sharply and credit spreads, the VIX and the dollar all off their highs (Chart 2). Chart 2Financial Conditions Are Easing Second, all FOMC participants need more confidence that inflation will return to target before re-starting rate hikes, but this bar seems higher for some than for others. Year-over-year core and trimmed mean CPI are currently running at 2.15% and 2.19%, respectively. This is slightly below the 2.4% level that is consistent with the Fed’s inflation target (Chart 3).1 The minutes suggest that some FOMC participants would be comfortable re-starting rate hikes as long as core inflation moves higher in the next few months and approaches the Fed’s target from below. Some others, however, may need to see an overshoot of the Fed’s inflation target before recommending rate hikes. Chart 3Core Inflation Needs To Move Higher Depressed inflation expectations, as seen in the TIPS market or the Michigan Consumer Sentiment survey, are a related issue (Chart 3, bottom 2 panels). The Fed will probably want to see upward movement in both of these measures before resuming rate hikes. In fact, New York Fed President John Williams warned last week that the “persistent undershoot of the Fed’s [inflation] target risks undermining the 2 percent inflation anchor.” He added that “the risk of the inflation expectations anchor slipping toward shore calls for a reassessment of the dominant inflation targeting framework.”2 Williams has long been an advocate for a monetary policy framework where the Fed targets an overshoot of its inflation target in the future to “make up” for undershooting its target in the past, i.e. some form of price level targeting. The Fed is currently conducting a year-long investigation into whether it should switch to this sort of regime and we learned last week that the Fed will announce the results of its investigation in the first half of 2020. Our own sense is that the Fed will eventually adopt some sort of “history dependent” inflation target as a way to avoid continuously bumping up against the zero-lower bound on interest rates. But this change will not occur this year and maybe not even next year. Of course, the more immediate concern for bond investors is whether inflation pressures will be meaningful enough in the next few months for the Fed to resume rate hikes in 2019. We expect they will be. We have previously shown that base effects alone will pressure year-over-year core CPI higher as we head toward mid-year.3 Meanwhile, other signs also point toward rising core inflation (Chart 4): Chart 4Inflation Pressures Building The New York Fed’s Underlying Inflation Gauge is running close to 3% (Chart 4, top panel). The ISM Manufacturing PMI is off its highs, but is still consistent with rising year-over-year core CPI (Chart 4, panel 2). Our CPI Diffusion Index is deep in positive territory, pointing to further near-term upside in the core measure (Chart 4, bottom panel). Bottom Line: With financial conditions easing and core inflation more likely to rise than fall, the majority of Fed officials will feel justified lifting rates again this year. January’s FOMC minutes imply that several Fed members want to see an overshoot of the inflation target before advocating for the resumption of rate hikes, but until the Fed changes its inflation targeting regime they will likely be out-voted. The Best Way To Trade The Fed We continue to recommend a below-benchmark duration bias in U.S. bond portfolios, on the view that rate hikes will exceed depressed market expectations on a 12-month horizon. However, this is not the most attractive way to position for the resumption of Fed rate hikes. The best way to trade the Fed in the current environment is by initiating a duration-neutral yield curve trade where you buy a barbell consisting of the long and short ends of the curve, and sell the 5-year or 7-year maturity. In a prior report we demonstrated that the 5-year and 7-year Treasury yields are most sensitive to changes in our 12-month fed funds discounter.4 That is, when the market starts to price-in more Fed rate hikes, the 5-year and 7-year Treasury yields increase more than other maturities. Similarly, the 5-year and 7-year yields fall the most when our discounter declines. Clearly, this means that if you are short the 5-year/7-year part of the curve versus the wings, you will make money as rate hikes are priced back into the market. Usually the problem with implementing such a trade is that it has negative carry. That is, the 5-year or 7-year bullet typically offers a greater yield than what you would earn on a duration-matched 2/10 or 2/30 barbell. If you don’t time the trade properly, you end up losing money waiting for Fed rate hike expectations to move. However, this is not a problem at the moment. In fact, duration-matched barbells are now positive carry propositions relative to 5-year and 7-year bullets (Chart 5). Chart 5 Barbell Yields Greater Than Bullet Yields In other words, if you think rate hikes will resume at some point, you are currently getting paid to wait for the market to catch on. The only way to lose money in this sort of trade is if our 12-month fed funds discounter falls further from its current -9 bps level. We view that as an unlikely scenario. Bottom Line: The best way to position for the resumption of Fed rate hikes is to sell the 5-year or 7-year part of the Treasury curve, and buy a barbell consisting of the long and short ends of the curve. We currently recommend being short the 7-year and long the 2/30 barbell. This trade has positive carry, meaning that you will earn money as you wait for rate hikes to get priced back in.  Corporate Spread Targets As we have discussed in prior reports, we think the Fed’s pause opens up a window where corporate bond spreads have room to tighten during the next few months.5 However, we also acknowledge that the window for outperformance is limited. Once financial conditions ease and the Fed resumes rate hikes, the environment will quickly become more difficult for corporate bonds. For this reason, in last week’s report we presented Chart 6. The diamonds in Chart 6 show where corporate 12-month breakeven spreads are today relative to past “Phase 2” periods, which are environments similar to today when the yield curve is quite flat but still positively sloped.6 We argued that we would be quick to reduce corporate bond exposure when the breakeven spreads reach the historical median for Phase 2 periods, i.e. when the diamonds fall to the 50% line in Chart 6. However, we acknowledge that this is not a helpful guide for investors who don’t have timely access to our valuation metrics. So this week we present Charts 7A and 7B. These charts estimate the option-adjusted spread (OAS) levels for each credit tier of the Bloomberg Barclays corporate bond indexes that would be consistent with the 50% line in Chart 6. To make these estimates we need to assume that the average duration of each index remains constant. The results show the following spread targets: For Aa we target 55 bps. The current OAS is 61 bps. For A we target 84 bps. The current OAS is 94 bps. For Baa we target 128 bps. The current OAS is 161 bps. For Ba we target 186 bps. The current OAS is 236 bps. For B we target 298 bps. The current OAS is 391 bps. For Caa we target 571 bps. The current OAS is 813 bps. We do not recommend an overweight allocation to Aaa-rated corporate bonds, where spreads are already expensive relative to past Phase 2 periods (Chart 7A, top panel). Chart 7aInvestment Grade Spread Targets Chart 7BHigh-Yield Spread Targets   Bottom Line: Maintain an overweight allocation to corporate bonds (both investment grade and high-yield) with the exception of the Aaa credit tier. But be prepared to reduce exposure when spreads reach our target levels. Economic Update We will finally receive GDP data for the fourth quarter of 2018 on Thursday, and investors should ready themselves for a weak number. In fact, the most recent tracking estimates from the New York Fed have real GDP coming in at 2.35% in Q4 and a mere 1.20% in 2019 Q1 (Chart 8). Chart 8Poor GDP Tracking Estimates ... It will come as no surprise that the trend in GDP growth is vital to our interest rate call. In fact, we showed in a recent report that when year-over-year nominal GDP growth falls below the 10-year Treasury yield it is often a good signal that monetary policy has turned restrictive and that interest rates have peaked for the cycle.7 With that in mind, if we add 1.2% expected real growth in Q1 to the 1.7% average growth rate of the GDP deflator (Chart 8, bottom panel), we can roughly estimate nominal GDP growth of 2.9% in Q1. This remains above the current 10-year Treasury yield, suggesting that monetary conditions would still be accommodative, but just barely. However, we expect the Q1 tracking forecast to improve as new data come in. According to the New York Fed’s model, the weak December retail sales report trimmed 0.41% from its Q1 growth forecast and this report increasingly looks like an aberration. In contrast to the retail sales number, the Johnson Redbook index of same-store sales is growing at a rate close to 5%, and indexes of consumer confidence remain elevated (Chart 9). Chart 9...Driven By Abnormal Retail Sales Even the Fed staff’s economic report, as presented in the January FOMC minutes, suggests that December should have been a good month for consumer spending: The release of the retail sales report for December was delayed, but available indicators – such as credit card and debit card transaction data and light motor vehicle sales – suggested that household spending growth remained strong in December. Bottom Line: However, we expect the Q1 tracking forecast to improve as new data come in. According to the New York it seems likely that the partial government shutdown influenced the collection of the December retail sales data and led to an abnormal print. Since the retail sales data feed directly into GDP, the impact will be felt in the next GDP report. But the impact will prove fleeting.   Ryan Swift, Vice President U.S. Bond Strategy rswift@bcaresearch.com Footnotes 1 The Fed’s target is for 2% PCE inflation. CPI tends to run about 0.4% above PCE. 12-month core PCE is currently 1.88%, but data only go to November. This is why we refer to CPI in this report, which has data through January. 2 https://www.newyorkfed.org/newsevents/speeches/2019/wil190222 3 Please see U.S. Bond Strategy Weekly Report, “Caught Offside”, dated February 12, 2019, available at usbs.bcaresearch.com 4 Please see U.S. Bond Strategy Weekly Report, “Don’t Position For Curve Inversion”, dated January 22, 2019, available at usbs.bcaresearch.com 5 Please see U.S. Bond Strategy Weekly Report, “Buy Corporate Credit”, dated January 15, 2019, available at usbs.bcaresearch.com 6 For more detail on the different phases of the economic cycle 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 7 Please see U.S. Bond Strategy Weekly Report, “Running Room”, dated January 29, 2019, available at usbs.bcaresearch.com Fixed Income Sector Performance Recommended Portfolio Specification

Highlights Low Bond Volatility: Weakening non-U.S. growth and a more dovish Fed have crushed global government bond volatility, especially in Europe and Japan where yields are struggling to stay above 0%. Treasury-Bund and Treasury-JGB spreads, which now largely reflect long-run real growth differentials between the U.S and Europe/Japan, are likely to stay range bound. USTs vs Bunds/JGBs: Stay overweight Bunds & JGBs versus Treasuries, on a hedged basis in U.S. dollars, given the boost to returns from hedging into higher-yielding dollars. Feature Bond Yields Are In Winter Hibernation Developed market (DM) government bonds, never the most exciting of asset classes to begin with, have become boring of late. While benchmark 10-year yields since the end of January have moved in line with our recommended country allocations - lower in Germany (-7bps), Japan (-3bps), the U.K. (-5bps) and Australia (-11bps) where we are overweight, higher in the U.S. (+5bps), Canada (+2bps) and Italy (+19bps) where we are underweight – government bonds have settled into trading ranges and lack direction. The proximate trigger for the muted yield volatility was the Federal Reserve shifting to a neutral stance on U.S. monetary policy in January. Investors have priced out any possibility of a Fed rate hike over the next year, and now even discount a modest rate cut, according to the U.S. Overnight Index Swap (OIS) curve. Yet while most of the attention for bond investors have been focused on the U.S., there are developments in other major economies that are also depressing yields – namely, weakening economic momentum and sluggish inflation. In particular, the downturn has shown no signs of stabilizing in the eurozone and Japan, with the latest readings on manufacturing PMIs now below the 50 line, signaling a contraction (Chart of the Week). The latest data in both regions still shows that core inflation is nowhere near the inflation targets of the European Central Bank (ECB) and Bank of Japan (BoJ). The story is much different in the U.S, with the manufacturing PMI still well above 50 and core inflation hovering close to the Fed’s 2% inflation target. Yet Treasury yield volatility has collapsed, with the MOVE index of Treasury options prices now back to the lows of this cycle. Chart Of The WeekAre Treasuries Leading Or Following? For the time being, non-U.S. factors are driving the direction of global bond yields. We think that will change later this year, as steady U.S. growth and surprisingly firm U.S. inflation readings will prompt the Fed to begin hiking rates again. Yet until there are signs that non-U.S. growth is stabilizing, the low yields in Europe and Japan will act as an anchor on U.S. Treasury yields, particularly given how wide U.S./non-U.S. yield differentials already reflect faster growth and inflation in the U.S. Decomposing Treasury-Bund & Treasury-JGB Spreads When looking at the pricing of the “Big 3” DM government bond markets – the U.S., Germany and Japan – there are some major differences but also some similarities as well. Even with the benchmark 10-year U.S. Treasury sitting at 2.68% compared to a mere 0.11% and -0.03% on the 10-year German Bund and 10-year Japanese government bond (JGB), respectively. Simply looking at the breakdown of those nominal 10-year yields into the real and inflation expectations components, there is not much of a comparison (Chart 2). The real 10-year Treasury yield is in positive territory at 0.6%, compared to -1.4% and +0.2% for JGBs and German bunds, respectively. Inflation expectations, measured by 10-year CPI swap rates, are 2.1% in the U.S., 1.5% in Germany and 0.2% in Japan. Thus, the current wide 10-year Treasury-Bund spread (just under +260ps) can be broken down into a real yield spread of +200bps and an inflation expectations gap of +60bps. In the case of the 10-year Treasury-JGB spread (just under +270bps), that breaks down into a real yield differential of +80bps and an inflation gap of +190bps. Chart 2Big Differentials Here... So while the Treasury-Bund and Treasury-JGB spreads are of similar magnitude, the valuation components driving the spread are much different. The former is more of a real yield gap, while the latter is more of an inflation expectations gap. That is no surprise given the BoJ’s Yield Curve Control policy that maintains a ceiling on the 10-year JGB yield of between 0.1% and 0.2%, limiting how much real yields can move (there are no BoJ restrictions on the level of CPI swap rates). Yet the U.S.-Japan inflation expectations gap is not too far off the spread between realized headline and core inflation measures in both countries - both are 1.4 percentage points higher in the U.S. as of January. Looking at other valuation metrics, the cross-county differentials are less pronounced (Chart 3). Chart 3...But Less So For Other Yield Measures Yield curves are quite flat, with the 2-year/10-year slope a mere +16bps in the U.S., +14bps in Japan and only +66bps in Germany. Our estimates of the term premia on 10-year government debt are negative for all three markets, most notably in the countries that have seen quantitative easing in recent years (-10bps in the U.S., -90bps in Germany and -60bps in Japan). Perhaps most importantly, our preferred measure of the market pricing of the real terminal policy rate – the 5-year OIS rate, 5-years forward minus the 5-year CPI swap rate, 5-years forward – is +0.2% in the U.S., -0.5% in Germany and 0.0% in Japan. That means the market is pricing in only a +70bp differential, in real terms, between the neutral policy rates of the Fed and ECB. That gap is only +20bps between market pricing of the neutral real rates for the Fed and BoJ. That narrower gap between the market-implied pricing of the real neutral rate is consistent with the theoretical macroeconomic drivers of real rate differentials, like growth rates of potential GDP and labor productivity. According to OECD estimates, potential GDP growth is 1.8% in the U.S., 1.5% in the overall euro area and 1.2% in Japan (Chart 4). This implies a long-run real yield gap between the U.S. and Germany of +60bps and the U.S. and Japan of +30bps – very close to the market pricing for the real terminal rate differentials.1 When looking at the 5-year annualized growth rates of labor productivity data from the OECD, there is no difference between the three regions with all growing at a mere 0.5% (suggesting that either a faster growth rate of the labor input, or greater productivity of capital, accounts for the higher potential growth rate in the U.S.). Chart 4No Major Differences In Long-Run Real Growth With the cross-country yield spreads now effectively priced for the long-run real growth differentials between the U.S. and Europe/Japan, this will limit the ability for nominal Treasury-Bund and Treasury-JGB spreads to widen much further. Right now, U.S. inflation expectations are rising faster than those of Europe and Japan, in response to the Fed’s more dovish stance. Yet if those expectations continue to rise, likely in the context of stickier realized U.S. inflation alongside solid U.S. growth, then the Fed will return to a hawkish bias. That ultimately means higher U.S. real yields and, most likely, some pullback in U.S. inflation expectations since the markets would begin to price in the implications of the Fed moving to a restrictive policy stance (including a stronger U.S. dollar that will help dampen U.S. inflation, at the margin). So that means inflation differentials between the U.S. and Germany/Japan can move wider now but will narrow later; and vice versa for real yield differentials (narrower now and wider later). The main investment implication: nominal UST-Bund and UST-JGB spreads are unlikely to move much wider, likely for the remainder of this business cycle/Fed tightening cycle. The main takeaway is that bond yields in core Europe and Japan are effectively anchoring global yields, in general, and U.S. yields, in particular. Treasury yields will not be able to break out of the current narrow trading ranges until there are signs that growth has stabilized in Europe and Japan. Reduced global trade tensions and faster Chinese growth (and import demand) are necessary conditions to reflate the export-heavy economies of Europe and Japan. Yet even if that scenario does unfold in the months ahead (which is BCA’s base case scenario), there is still a case to prefer Bunds and JGBs over U.S. Treasuries on a currency-hedged basis in U.S. dollars. Given the wide short-term interest rate differentials between the U.S. and Europe/Japan, those near-zero 10-year Bund and JGB yields, after hedging into U.S. dollars, are actually higher than 10-year Treasury yields, which benefits the relative hedged performance of the low-yielders versus the U.S. (Chart 5) Chart 5Stay Overweight Bunds & JGBs Vs. USTs (Hedged Into USD) Thus, we continue to recommend an overweight stance on core Europe and Japan, versus an underweight tilt on the U.S., in global U.S. dollar-hedged government bond portfolios. Bottom Line: Weakening non-U.S. growth and a more dovish Fed have crushed global government bond volatility, especially in Europe and Japan where yields are struggling to stay above 0%. Treasury-Bund and Treasury-JGB spreads, which now largely reflect long-run real growth differentials between the U.S and Europe/Japan are likely to stay range bound. Stay overweight Bunds & JGBs versus Treasuries, on a hedged basis in U.S. dollars, given the boost to returns from hedging into higher-yielding dollars.   Robert Robis, CFA, Senior Vice President Global Fixed Income Strategy rrobis@bcaresearch.com    Footnotes 1      We are using the full euro area data for these economic comparisons, even though we are discussing U.S.-German yield differentials in this report. We think this is reasonable given the status of German government bonds as the benchmark for the euro area, and with the ECB setting its monetary policy for the overall euro area. The differences between the data for Germany and the overall euro area are modest, with German potential GDP and 5-year productivity growth both only 0.3 percentage points higher. 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 Phillips curve, which encouraged economic policymakers of the sixties and early seventies to believe in a mechanical tradeoff between inflation and unemployment, fell into disrepute once stagflation strangled the U.S. economy. We do not view the idea that there is an inverse relationship between the unemployment rate and wage gains as controversial. This weak form of the Phillips curve simply formalizes the interplay between supply and demand in the labor market. We have found, however, that any reference to the Phillips curve has the potential to provoke strong reactions from investors. The criticism that the link between compensation gains and consumer prices is questionable has merit. Over the last 30 years, changes in compensation have exhibited a sporadic correlation with changes in consumer prices. Even if the empirical evidence between labor market tightness and inflation is somewhat wobbly, the Fed remains squarely in the Phillips curve camp, and its take on the relationship is the only one that matters for monetary policy. The investment implication is that labor market strength will prove self-limiting. An unemployment rate bound for 3.5% or lower will pull the Fed back off the sidelines, ultimately bringing down the curtain on the expansion and the equity bull market. Feature The stagflation of the seventies was a near-death experience for the Phillips curve and its proposition that unemployment and inflation are inversely related. As both Milton Friedman and Edmund Phelps had predicted, the trade-off could not survive beyond the short term because workers would adjust their expectations as they caught on to the pattern, demanding wages that kept pace with inflation even when unemployment was high. Duly modified, the Phillips curve’s appeal was rekindled, and the Phelps-Phillips expectations-augmented version has gone mostly unchallenged within the economics profession ever since. The Fed and other policymakers may have given up on the notion that they could manage their economies via a mechanical tradeoff between inflation and unemployment, but the inverse relationship remains a pillar of their macroeconomic models. We don’t find the idea that the unemployment rate and wage inflation are inversely related the least bit controversial, as it fully accords with the laws of supply and demand. Unemployment’s link to consumer price inflation is uncertain, however, and even the narrow unemployment/wages form of the Phillips curve relationship we favor often invites controversy. Discussing upward wage pressures within the context of consumer price inflation and the Fed’s reaction function can elicit spirited resistance. As one client put it in a January meeting, “it is unbecoming for BCA to subscribe to these sorts of cost-plus notions of inflation.” This Special Report examines the record in an effort to determine the influence the Phillips curve thesis will have on policy and markets going forward. It asks the following questions along the way: What is the Phillips curve? Where does inflation come from? Is there a relationship between wage inflation and price inflation? Where does the Fed stand? What impact will a falling unemployment rate have on the economy and financial markets? A Brief History Of The Phillips Curve The Phillips curve arose from a study of the unemployment rate and wages in the U.K. from the mid-nineteenth to the mid-twentieth centuries. William Phillips discovered a consistent inverse relationship between the unemployment rate and changes in wages: high unemployment was associated with muted wage gains, and low unemployment was associated with robust wage gains. He posited that the unemployment rate revealed the level of tightness in the labor market, and the extent to which employers had to compete to attract workers. Other researchers extended the relationship from wage inflation to price-level inflation and suggested that policy makers could use the tradeoff between unemployment and inflation to fine-tune the course of the economy. The stagflation of the seventies blew up the notion of a mechanical tradeoff, but a modified form of the inverse relationship between unemployment and wage gains resides at the heart of mainstream macroeconomic forecasting models. Those models have become more sophisticated, and now include the concept of a natural rate of unemployment, but the inverse relationship between unemployment and inflation remains at their core. Investor skepticism aside, the Phillips curve is deeply embedded in orthodox economic narratives relating inflation and unemployment. As New York Fed President Williams put it last Friday in the first line of a speech discussing the issues raised in a new Phillips curve paper, “The Phillips curve is the connective tissue between the Federal Reserve’s dual mandate goals of maximum employment and price stability.1” Where Does Inflation Come From? Thousands of dissertations have grappled with this subject without providing a definitive solution, but there are two broad explanations we find most compelling. The first is that inflation responds to the level of slack in the economy. That’s to say that inflation is a by-product of the relative balance between aggregate supply and aggregate demand. When the output gap is wide (demand falls well short of the economy’s capacity), inflation is unlikely to find a footing. When the output gap is closed (demand and capacity are in balance) or negative (demand exceeds capacity), inflation will gain traction unless imported capacity bridges the gap. For the second, we combine the idea that inflation expectations play a central role with Milton Friedman’s always-and-everywhere admonition. The stable inflation of the last couple of decades has coincided with stable inflation expectations. The causation mostly appears to run from (trailing) inflation to expectations (Chart 1), but expectations surely influence economic actors’ price negotiations and open the door to a monetary influence. Inflation expectations are likely to be well anchored under a central bank that convinces households and businesses of its commitment to price stability. When the monetary authority lacks inflation credibility, inflation expectations may become unmoored and impel economic actors to insist upon higher wages and selling prices to keep pace with a rising price level. Chart 1Seeing The Future In The Recent Past The expectations-augmented Phillips curve makes it clear that inflation is a function of inflation expectations just as surely as it is a function of the unemployment rate. The more firmly expectations are anchored, the more unemployment has to drift from its natural rate (NAIRU, or u-star (u*)) to move the inflation needle. In other words, when expectations are as well-anchored as they have been since the crisis, wages will be so unresponsive to changes in the unemployment rate that the Phillips curve will appear to be broken. Believing that inflation will permanently remain at 2% or lower, workers feel no urgency to press for larger wage/salary increases. The Empirical Record – Unemployment And Wages The seventies played havoc with the Phillips curve, but over the last twenty-five years, the inverse relationship between changes in the unemployment rate and wage gains has held up very well once the unemployment rate has reached threshold levels at or near u-star. When there is ample slack in the labor market, wages are nearly insensitive to changes in the unemployment rate. When the unemployment rate moves from 10% to 9%, 9% to 8%, or 8% to 7%, there are multiple qualified candidates for every job opening and employers have no reason to bid wages higher (Chart 2, top panel). Below 5%, roughly around u*, employers have to compete for workers and wage gains are very sensitive to moves in the unemployment rate (Chart 2, bottom panel). Chart 3 illustrates the threshold concept, segmenting the last 30 years of observations by their relationship to the unemployment gap. Observations for which the unemployment gap is greater than or equal to 2% are shown in gray; their best-fit line with wage gains is nearly flat. Positive, but small, unemployment-gap observations are shown in orange; their best-fit line is steeper and indicates a more robust correlation with moves in wages. Negative unemployment-gap observations are colored blue; they have the steepest best-fit line and exhibit the tightest correlation with changes in wages. A skeptic might seek more convincing evidence, but period-to-period noise in the data limits the amount of variation in wages explained by the unemployment rate (just under 40% over the last 30 years). Noting that the unemployment gap tends to persist in negative and positive territory for extended periods, we measured the annualized rate of wage gains for negative-gap and positive-gap phases. The results were robust, with wage gains in negative-gap phases consistently topping gains in positive-gap phases (Chart 4). Both groups exhibited remarkably consistent growth rates – the three complete negative-gap phases featured wage gains of 3.8%, 3.8% and 3.9%, while the three positive-gap phases had wage growth of 2.7%, 2.5% and 2.4%. At 3%, the current negative-gap phase has already separated itself from the last three decades’ positive-gap phases, though the 3.8% level is still a ways away. Chart 4Mind The Gap The Empirical Record – Wage Inflation And Price Inflation If businesses were omniscient, omnipotent and able to adjust selling prices in real time – something like Amazon, in another words – they might seek to preserve their profit margins by instantaneously raising prices to offset wage gains. Wage inflation and price inflation would then move together in lockstep without any lags. Businesses do not have unlimited power or unlimited knowledge, however, and neither do workers. There are information and expectation lags, and price-making/price-taking status is fluid. The empirical record over the 50-plus years covered by the average hourly earnings series shows that the wage-price relationship is constantly shifting. Under a cost-push inflation framework, tightness in the labor market shows up in consumer prices after employees negotiate raises, and employers subsequently raise prices to recoup lost profits. In a demand-pull model, businesses perceiving signs of excess demand take the opportunity to raise prices, spurring employees to demand raises to preserve their purchasing power. There is room for both models, as BCA’s analysis of wage/price dynamics over the years has shown that leadership between prices and wages regularly shifts. For the purposes of this report, it is sufficient to note that the wage/price skeptics have a point. A decade-by-decade review of year-on-year gains in average hourly earnings (“AHE”) and core CPI shows that correlations between AHE and consumer prices regularly make big swings. The ‘60s, ‘80s and ‘00s were pretty good to Phillips curve adherents (Chart 5), but the ‘70s, ‘90s and the current decade mocked them, featuring repeated instances of outright decoupling (Chart 6). The bottom line is that the direction of causation between wages and consumer price inflation, as well as the sensitivity of the relationship, is fluid. The empirical record does not support the idea that wage inflation translates to overall inflation in a consistent and timely fashion. Chart 5Moving In Lockstep One Decade... Chart 6... Decoupling The Next The Fed’s Reaction Function Wage gains exhibit little sensitivity to changes in the unemployment rate when there is a lot of slack in the labor market. Even at lower levels of unemployment, inflation expectations can temper wages’ sensitivity to the unemployment rate. There is assuredly an inverse relationship between wages and unemployment, nonetheless, and wage gains are especially sensitive when the unemployment gap is negative. The jury is out on the relationship between unemployment and inflation, however. The direction of causation is not constant and the response lags between the series can be quite long. Inflation expectations play a sizable role, and are capable of smothering wage gains in times of low unemployment if they’re well-anchored, or goosing them even in times of high unemployment if they’re spiraling upward. Believing in the Phillips curve relationship requires a lot of assumptions, and if the theory were brand-new today, it might have a hard time surviving peer review. Markets don’t take their cues from peer-reviewed journals, however. When it comes to interest rates and the entire gamut of financial assets impacted by monetary policy, the Fed has the last word. What it believes about the Phillips curve is much more important than whether or not its conclusions have iron-clad empirical support. It has long been BCA’s view, informed by our contacts within the Fed, the former central bankers who sat on our Research Advisory Board, the Bank of Canada veterans who have worked at BCA, and careful observation of the Fed’s own comments and research, that the Fed maintains a Phillips curve view of the world. The Fed has plenty of company in this regard. Nearly all central banks are Phillips curve believers; in the absence of a mainstream alternative model of inflation, they all have to fall back on the expectations-augmented hypothesis. Investors and economics enthusiasts can rail against the Phillips curve’s empirical shortcomings, and posit that globalization, robotics/AI, Amazon and the gig economy have rendered it null and void. Those theories have not been confirmed by the data,2 however, and until the profession unites behind an alternative narrative, the Phillips curve will continue to heavily influence monetary policy. New York Fed President Williams clearly subscribes to the tell-‘em-what-you’re-gonna-tell-‘em/tell-‘em/tell-‘em-what-you-just-told-‘em method of constructing speeches. One need look no further than his remarks last Friday, when discussing a paper co-authored by former Fed governor Frederic Mishkin, for his view. “[T]he Phillips curve is very much alive in very tight labor markets,” he said near the beginning of his remarks. “[T]he Phillips curve is alive and kicking,” he said more than halfway through. “In summary, the Phillips curve is alive and well,” he said in conclusion, in case anyone in the audience had been napping. The bottom line for an investor today is that the Fed’s reaction function ensures that labor market strength will ultimately prove to be self-limiting. Assuming that Baby Boomer retirements will stifle further gains in the labor force participation rate, the unemployment rate is likely to ratchet lower across 2019.3 As it dips further and further below NAIRU, the Fed can be counted upon to remove accommodation, ultimately triggering a recession (Chart 7). Chart 7Expansions End When Unemployment Rises Investment Implications As the Fed’s pause allows the economy to regather momentum, hiring and wage growth should be well supported. The accompanying decline in the unemployment rate will drive the Fed to revive its tightening campaign. The irony is the longer the Fed grants the economy, and investors, a respite by holding its fire, the more accommodation it will have to remove to stamp out inflation pressures. It will take until 2020 for the Fed to complete its tightening campaign, but we expect the terminal fed funds rate in this cycle will be at least 3.25 to 3.5%, far above the OIS curves’ projection that fed funds will end 2020 at 2.25%. Such a wide disparity between our expectations and market expectations leaves considerable room for the Treasury curve to shift out along all maturities. We expect the curve will ultimately invert, but the process will follow a bear-flattening course, and long maturities will suffer the worst capital losses. We therefore advocate underweighting Treasuries in all fixed-income portfolios, while maintaining below-benchmark duration in all bond sleeves. We expect that Fed tightening will bring the curtain down on the equity bull market before the recession officially begins (Chart 8). Until it does, however, we expect the Fed’s forbearance to help the economy generate evident momentum, pushing risk-asset values higher. We continue to recommend that investors overweight equities and spread product for now, but the clock is ticking. Watch the unemployment gap for the cue to position portfolios more defensively. Chart 8Inducing A Recession Is Tantamount To Inducing A Bear Market Doug Peta, CFA, Senior Vice President U.S. Investment Strategy dougp@bcaresearch.com     Footnotes 1      Williams, John C., “Discussion of ‘Prospects for Inflation in a High Pressure Economy: Is the Phillips Curve Dead or Is It Just Hibernating?’” Remarks at the U.S. Monetary Policy Forum, New York City, February 22, 2019. https://www.newyorkfed.org/newsevents/speeches/2019/wil190222 2      Please see the September 2017 Bank Credit Analyst Special Report, “Did Amazon Kill the Phillips Curve?” available at bcaresearch.com. 3      Holding the participation rate constant, the U.S. economy has to create 110,000 jobs a month to keep the unemployment rate at a steady state. Please see the Atlanta Fed’s online jobs calculator at https://www.frbatlanta.org/chcs/calculator.aspx.

Overweight General Electric took the S&P industrial conglomerates index higher yesterday on news it had agreed to sell its BioPharma business to Danaher, the former home of GE’s relatively new CEO. The deal, which will raise more than $21 billion for the firm, was celebrated by investors who delivered one of the best share price moves for the company in a decade. Importantly, GE promised to use the proceeds of the transaction to pay down the debt load that has weighed on the stock since the GFC. Further, the deal will see Danaher assuming the pension obligations of the group, another source of shareholder angst. We share investor sentiment with respect to deleveraging; more is better for this highly indebted sector and yesterday’s news is another step in the right direction (third panel). Tack on the catalyst that relief in the trade war with China adds (please see our previous Insight Report) and the recent rally in the still reasonably valued S&P industrial conglomerates index (bottom panel) looks very sustainable; stay overweight. The ticker symbols for the stocks in this index are: BLBG: S5INDCX - GE, MMM, HON, ROP.

Overweight Year-to-date, industrials stocks are the best performing GICS1 sector, outperforming the SPX by a massive 650bps. While such a breakneck pace is unsustainable and a short term breather is likely, from a cyclical perspective more gains are in store in this still underowned sector. In Monday’s Weekly Report, we highlighted the top five reasons it still pays to be overweight this deep cyclical sector. With Sunday’s news that the Trump administration is planning to delay the slate of additional Chinese tariffs that were scheduled to begin on March 1 and optimistic tweets on the progress of negotiations, one reason in particular seems most relevant: an easy credit, fiscal & monetary policy trifecta in China. Beyond the positive resolution in the U.S./China trade dispute, China has opened up its central bank liquidity tap to complement ongoing easy monetary policy. Tack on the recent monster loan origination and reaccelerating infrastructure spending and factors are falling into place for a pick up in end demand, which is a boon for U.S. capitals goods producers. Bottom Line: Stay overweight the S&P industrials sector and see Monday’s Weekly Report for more details and the other four reasons we still like industrials.