Highlights The risk of a US recession has increased sharply over the past several months. We have not yet concluded that a recession over the coming year is inevitable, but substantial (further) supply-side and pandemic-related disinflation is likely needed for the US economy to avoid a contraction in output. The increased risk of a contraction has caused investors to ponder what the next recession might look like. One very important question concerns the likely behavior of short-term interest rates during the next recession, especially if it occurs sooner rather than later. The historical experience suggests that the Fed may cut interest rates to zero during the next recession, but that the re-establishment of a long-lasting zero interest rate policy and the associated resumption of large-scale asset purchases seem quite unlikely unless the recession is severe. In the post-WWII environment, severe US recessions have been accompanied by aggravating factors that appear to be missing in the current environment. In addition, there are several arguments pointing to the next US recession being a mild one. For fixed-income investors, the implication is that investors should not overstay their welcome in a long-duration position during the next US recession, and should be looking to reduce their duration exposure earlier rather than later. For equity investors, our findings underscore that meaningful downside risk exists for stocks even in a mild recession environment, because the decline in bond yields is not likely to offset a rise in the equity risk premium. Feature Over the past several months, investors have been faced with a sharp increase in the odds of a US recession. Gauging the risk of a recession has featured prominently in our recent reports, and we have concluded, for now, that a US recession over the coming year is not yet inevitable. Still, we acknowledge that the risks are quite elevated, and that substantial (further) supply-side and pandemic-related disinflation is likely needed for the US economy to avoid a contraction in output. Economic expansions do not last forever. This means that the US economy will eventually succumb to a recession at some point over the coming few years. One very important question for investors concerns the likely behavior of short-term interest rates during the next recession, especially if a contraction occurs sooner rather than later. A key aspect of this question is whether the Fed is likely to be forced back towards a zero or negative interest rate policy, and whether it will need to employ asset purchases as part of its stabilization efforts as it has during the last two recessions. If so, long-maturity bond yields are likely to fall significantly during the next recession; if not, investors may be surprised by how modestly long-maturity yields decline. In this report, we examine the historical record of short-term interest rates during recessions and discuss whether the next US recession is likely to be severe or mild. We conclude that the next US recession is more likely to be mild than severe, and that the 10-year Treasury yield is unlikely to fall below 2% during the recession (or fall below this level for very long). In the case of a more severe recession driven by unanchored inflation expectations, the implications would be clearly bearish for bonds. Within a fixed-income portfolio, one conclusion of our analysis is that investors should not overstay their welcome in a long-duration position during the next recession and should be looking to reduce their duration exposure earlier rather than later. For equity investors, our findings underscore that meaningful downside risk exists for stocks even in a mild recession environment, because the decline in bond yields is not likely to offset a rise in the equity risk premium. The Historical Recessionary Path Of Short-Term Interest Rates When projecting how the Fed funds rate is likely to evolve during the next US recession, most investors typically look to the average decline in short-term interest rates during previous recessions as a guide. Based on that approach, Table II-1 highlights that the Fed would likely have to cut rates into negative territory if a recession occurred over the coming 12-18 months, unless it is able to hike interest rates significantly more over the coming year than the market is currently expecting and the FOMC itself is projecting. But in our view, focusing on the historical recessionary decline in interest rates from their peak is not the right approach, because it ignores the fact that recessions typically occur when monetary policy is tight. If a recession occurs within the next 18 months, it will have happened in large part because of a collapse in real wage growth, not just because of the increase in interest rates that has occurred. Chart II-1 highlights that short-term interest rates remain well below potential GDP growth, highlighting that monetary policy would still be easy today – despite the quick pace of increase in short rates – if real wages were growing rather than contracting sharply. In our view, the right approach is to examine how much short-term interest rates have typically fallen during recessions relative to potential or average historical GDP growth. This method captures the degree to which monetary policy easing has typically been required relative to neutral levels to catalyze an economic recovery. Table II-1Based Only On The Historical Decline In Short-Term Interest Rates, The Fed Would Ostensibly Have To Cut Rates Into Negative Territory During The Next Recession
September 2022
September 2022
Chart II-1Monetary Policy Would Still Be Easy Today If Real Wage Growth Was Positive
Monetary Policy Would Still Be Easy Today If Real Wage Growth Was Positive
Monetary Policy Would Still Be Easy Today If Real Wage Growth Was Positive
Based on this approach, Chart II-2 highlights that the Fed might have to cut the target range for the Fed funds rate to 0-0.25% during the next recession, but there are some examples (like the 1990-1991 recession) that point to a cut to just 0.25-0.5%. The goal of this exercise is not to be specific about the exact level to which the Fed will have to cut the Fed funds rate, but rather whether the de facto re-establishment of a long-lasting zero interest rate policy and the associated resumption of large-scale asset purchases is likely. Chart II-2The Fed May Have To Cut To Zero During The Next Recession, But Probably Not Into Negative Territory
September 2022
September 2022
Structural bond bulls might note that there are five recessions in the post-war era that could potentially point to that outcome based on Chart II-2. However, these episodes involved circumstances that we doubt would be present during the next US recession, especially if one were to emerge over the coming 12-18 months. The 1950s Recessions The recessions of 1953-54 and 1957-58 were fairly sizeable based on the total rise in the unemployment rate, but the monetary policy stance at that time was wildly stimulative in a way that is very unlikely to repeat itself today. In the 1950s, the level of interest rates was still an artifact of WWII (with the Treasury-Fed accord having only been agreed upon in March 1951). Monetary policy was both overly responsive to a period of pent-up disinflation following the initial burst of government spending associated with the Korean war and insufficiently responsive to a strongly positive output gap (Chart II-3). This was meaningfully compounded by a poor understanding of the size of the output gap at that time; the deviation of the unemployment rate from its 10-year average was significantly smaller than its deviation from today’s estimate of NAIRU (Chart II-4). In sum, the economic and monetary policy conditions that existed in the 1950s and that contributed to an interest rate level that was well below the prevailing rate of economic growth do not exist today. As such, we strongly doubt that the Fed’s response to the next US recession would resemble what occurred during that decade. Chart II-3We Strongly Doubt The Fed's Response To The Next US Recession Would Resemble What Occurred In The 1950s
We Strongly Doubt The Fed's Response To The Next US Recession Would Resemble What Occurred In The 1950s
We Strongly Doubt The Fed's Response To The Next US Recession Would Resemble What Occurred In The 1950s
Chart II-4Low Interest Rates In The 1950s Were Partly Caused By Wrong Output Gap Estimates
Low Interest Rates In The 1950s Were Partly Caused By Wrong Output Gap Estimates
Low Interest Rates In The 1950s Were Partly Caused By Wrong Output Gap Estimates
1973-1975 The recession that began in 1973 occurred because of a huge energy shock that proved to be stagflationary in the true sense of the word. Excluding the 2020 recession, this was the third largest rise in the unemployment rate of any recession since WWII, following 2008/2009 and the 1981/1982 recessions. There are some parallels between this recession and the current economic environment, but the stability of inflation expectations so far does not point to a truly stagflationary outcome. As such, we do not see the 1973-74 recession as a reasonable parallel to today’s environment. In addition, manufacturing employment – which was heavily impacted by the permanent rise in oil prices due to the sector’s energy intensity – stood at 24% of total nonfarm employment in 1973, compared with 8% today. Finally, the weight of food and energy as a share of total consumer spending today is roughly half of what it was during the 1970s (Chart II-5). 2001 Of the five recessions potentially implying that the Fed may have to cut interest rates into negative territory during the next US recession, the 2001 recession is the most relevant parallel to today. It was a modern recession in which the Fed maintained very easy monetary policy for a significant amount of time, in response to concerns about a significant tightening in financial conditions and the impact of prior corporate sector excesses on aggregate demand. The total rise in the unemployment rate during this recession was not very large, but it took some time for the unemployment rate to return to NAIRU. Still, even though this justified a later liftoff, a Taylor rule approach makes it clear that the Fed overstimulated the economy in response to the recession – a view that is reinforced by the enormous rise in household debt that fueled the housing market bubble during that period (Chart II-6). The Fed was very concerned about the negative wealth effects of the bursting of the equity market bubble, which had been caused by a massive decline in the equity risk premium in the second half of the 1990s. These conditions are simply not present today. Chart II-5Today's US Economy Is Meaningfully Less Impacted By Energy And Food Prices
Today's US Economy Is Meaningfully Less Impacted By Energy And Food Prices
Today's US Economy Is Meaningfully Less Impacted By Energy And Food Prices
Chart II-6The Fed Clearly Overstimulated In Response To The 2001 Recession
The Fed Clearly Overstimulated In Response To The 2001 Recession
The Fed Clearly Overstimulated In Response To The 2001 Recession
2008/2009 Chart II-7A Repeat Of The 2008/2009 Recession In The US Is A Totally Implausible Scenario
A Repeat Of The 2008/2009 Recession In The US Is A Totally Implausible Scenario
A Repeat Of The 2008/2009 Recession In The US Is A Totally Implausible Scenario
Chart II-2 highlighted that the Fed would have to cut interest rates to -1% were the 2008/2009 recession to repeat itself, but we judge that to be a totally implausible scenario given the improvement in US household sector balance sheets and financial sector health since the global financial crisis (Chart II-7). As we discuss below, the next US recession is likely to be meaningfully less severe than the 2008/2009 and 2020 recessions, which we believe carries important significance for the path of interest rates and the response of long-maturity bond yields. The bottom line for investors is that, based on the historical experience of rate cuts during recessions, the Fed may end up cutting interest rates back to or close to the zero lower bound in response to the next recession. But the de facto re-establishment of a long-lasting zero interest rate policy and the associated resumption of large-scale asset purchases seems quite unlikely unless the recession is severe, which we do not expect. Will The Next US Recession Be Severe Or Mild? Chart II-8The Most Severe US Recessions Have Had Aggravating Factors That Do Not Appear To Be Present Today
September 2022
September 2022
How drastically the Fed will be forced to cut interest rates during the next recession will be driven by its severity. Chart II-8 presents the total rise in the unemployment rate during post-WWII recessions (excluding 2020), in order to gauge whether the factors that have led to severe recessions in the past are likely to be present during the next contraction in output. From our perspective, the most severe US recessions in the post-WWII era have been driven by factors that are very unlikely to repeat themselves in the current environment. We noted above that a repeat of the 2008/2009 recession is a totally implausible scenario, leaving the 1981-1982, 1973-1975, and 1950s recessions as potential severe recession analogues. In three of these four cases we see clear signs of an aggravating factor that we do not (yet) believe will be present during the next US recession. Chart II-9Inflation Expectations Have Not Yet Unanchored To The Upside, In Sharp Contrast To The 1970s
Inflation Expectations Have Not Yet Unanchored To The Upside, In Sharp Contrast To The 1970s
Inflation Expectations Have Not Yet Unanchored To The Upside, In Sharp Contrast To The 1970s
In the 1981-1982 recession, the unemployment rate rose significantly as the Federal Reserve confronted the fact that inflation expectations had become severely unanchored to the upside, causing a persistent wage/price spiral. While unanchored inflation expectations is a risk today, so far the evidence suggests that both households and market participants expect that currently elevated inflation will not persist over the long run (Chart II-9). If inflation expectations do become unanchored to the upside at some point over the coming 12-18 months (or beyond), we are very likely to change our view about the severity of the next recession. However, this would be a bond bearish outcome (at least initially), as it would imply sharply higher yields at both the short and long end of the yield curve in order to tame inflation and re-anchor inflation expectations. As noted above, in the 1973-74 recession, the unexpected and permanent rise in oil prices and outright energy shortages rendered a significant amount of capital and labor uneconomic, which is different than what has been occurring during the pandemic. Were the recent rise in natural gas prices to be permanent and no alternatives available, Europe’s current energy situation would be more reminiscent of the 1973-1974 recession than the pandemic-driven price pressures and supply shortages affecting the US and other developed economies. Chart II-10The US Is Currently Experiencing Fiscal Drag, But That Will Lessen Next Year
The US Is Currently Experiencing Fiscal Drag, But That Will Lessen Next Year
The US Is Currently Experiencing Fiscal Drag, But That Will Lessen Next Year
Finally, while the 1957-58 recession appears to be somewhat of an anomaly driven by a mix of factors, the 1953-54 recession was clearly exacerbated by a sharp slowdown in government spending following the end of the Korean war. It is true that the US is currently experiencing fiscal drag (Chart II-10), but this has occurred against the backdrop of a strong labor market, and IMF forecasts imply that the drag will be significantly smaller over the coming year than what the US is currently experiencing. There are several additional points suggesting that the next US recession will be comparatively mild: Chart II-11The Milder US Recessions Were All Seemingly Triggered By Tight Monetary Policy (As Would Be The Case Today)
The Milder US Recessions Were All Seemingly Triggered By Tight Monetary Policy (As Would Be The Case Today)
The Milder US Recessions Were All Seemingly Triggered By Tight Monetary Policy (As Would Be The Case Today)
Chart II-11 highlights that the milder recessions, those which have seen the unemployment rate rise by less than 3% from their previous low, have generally been the recessions that appear to have simply been triggered by monetary policy becoming tight or nearly tight. This would likely be the case during the next US recession. In the lead up to the 1970, 1990-91, and 2001 recessions, short-term interest rates approached or exceeded either potential growth or the rolling 10-year average growth rate of nominal GDP. The 1960-61 recession stands out slightly as an exception to this rule, in that interest rates were still moderately easy, which is based on our definition of the equilibrium short-term interest rate. But interest rates had risen close to 400 basis points from 1958 to 1960 (suggesting a change in addition to a level effect of interest rates on aggregate demand), and it is notable that the 60-61 recession was the mildest in post-war history, based on the total rise in the unemployment rate. Chart II-12Labor Scarcity May Mean That Firms Will Be Somewhat More Reluctant To Shed Labor During The Next Recession
Labor Scarcity May Mean That Firms Will Be Somewhat More Reluctant To Shed Labor During The Next Recession
Labor Scarcity May Mean That Firms Will Be Somewhat More Reluctant To Shed Labor During The Next Recession
We argued in Section 1 of our report that monetary policy is not currently restrictive on its own, and that the recessionary risk currently facing the US is the result of a combination of the speed of adjustment in interest rates, the fact that real wages have fallen sharply, and the fact that the Fed is determined to see inflation quickly return to target levels. However, what this also highlights is that a recession would likely cause a rise in real wages via a significant slowdown in inflation (at least for a time); this would likely help stabilize aggregate demand and cause a comparatively mild rise in the unemployment rate. While the odds and magnitude of this effect are difficult to quantify, the fact that the labor market has been so tight over the past year and that the participation rate has yet to recover to its pre-pandemic levels suggests that some firms may be reluctant to shed labor during a recession (Chart II-12), suggesting that the total rise in unemployment in the next recession could be relatively small. Finally, Chart II-13 shows that the excess savings that have accumulated over the course of the pandemic, now primarily the result of reduced spending on services, dwarf the magnitude of precautionary savings that were generated in the prior three recessions as a % of GDP. We agree that the savings rate would likely still rise during the next recession, but the existence of excess savings implies that the rise in the savings rate may be surprisingly small – which would, in turn, imply a comparatively mild rise in the unemployment rate. We noted above that the household sector has deleveraged significantly, which is strong evidence against an outsized or long-lasting decline in consumer spending as a possible driver of an above-average rise in the unemployment rate during the next recession. One question that we often receive from clients is whether excessive corporate sector leverage could cause a more severe decline in economic activity once a recession emerges. Chart II-14 illustrates that the answer is “probably not.” The chart presents one estimate of the US nonfinancial corporate sector debt service ratio, based on national accounts data. The chart highlights that the current debt burden for the nonfinancial corporate sector is very low, underscoring that elevated corporate sector debt would not likely act as an aggravating factor driving an outsized rise in the unemployment rate were a recession to occur today. The chart also shows that even if the 10-year Treasury yield were to rise to 4% and corporate bond spreads were to widen in the lead up to a recession, the nonfinancial corporate sector debt service burden would rise to a lower peak than seen in the last three recessions. One key risk to a mild recession view is a scenario in which inflation does not return to or below the Fed’s target during the recession. In that kind of environment, the Fed would not likely cut interest rates to as low a level as they have in the past relative to potential growth. But the historical record is clear that recessions cause a deceleration in inflation, and if a recession emerges over the coming 12-18 months it will likely happen after supply-side and pandemic-related disinflation has already occurred. That means that inflation is likely to move back to or below the Fed’s target in a recessionary environment. We should note that this assessment differs somewhat from the scenario described by my former colleague Martin Barnes, who wrote a guest report on inflation published in our July Bank Credit Analyst.1 Chart II-13Today’s Pandemic-Related Excess Savings Dwarf Precautionary Savings During The Prior Three Recessions
September 2022
September 2022
Chart II-14US Corporate Sector Debt Unlikely To Lead To A More Severe Recession, Even In A Higher Yield Environment
US Corporate Sector Debt Unlikely To Lead To A More Severe Recession, Even In A Higher Yield Environment
US Corporate Sector Debt Unlikely To Lead To A More Severe Recession, Even In A Higher Yield Environment
Long-Maturity Bond Yields And The Next US Recession What does our analysis imply for long-maturity bond yields and the duration call over the coming few years? In order to judge what is likely to happen to long-maturity bond yields in a recession scenario over the coming 12-18 months, we first project the fair value of the 5-year Treasury yield based on the following hypothetical circumstances: The onset of recession in March 2023 and a peak in the Fed funds rate at a target range of 3.75-4%. A recession duration of eight months, over which time the Fed steadily cuts the policy rate to 0-0.25%. An initial Fed rate hike in September 2024, nine months following the end of the recession, consistent with a relatively short return of the unemployment rate to NAIRU as an expansion takes hold. A rate hike pace of eight quarter-point hikes per year, with the Fed again raising rates to a peak of 4% A longer-term average Fed funds rate of 3%, which we regard as a low estimate. Chart II-15The 5-Year Treasury Yield Would Not Fall Enormously In A Mild Recessionary Scenario
The 5-Year Treasury Yield Would Not Fall Enormously In A Mild Recessionary Scenario
The 5-Year Treasury Yield Would Not Fall Enormously In A Mild Recessionary Scenario
Chart II-15 highlights the fair value path for the 5-year Treasury yield in this scenario. Not surprisingly, the fair value today is lower than the current level of the 5-year yield, highlighting that a shift to a long duration stance will be warranted at some point over the coming year if the US economy enters a non-technical, typical income-statement recession. However, the chart also highlights that a long duration position is not likely to be warranted for very long, given that the lowest level of the 5-year fair value path is substantially higher than it was in 2020 and 2021 and is also higher than its 10-year average. Chart II-16 reveals the importance of forecasting the near-term path of interest rates when predicting the likely behavior of long-maturity bond yields. Even though near- and long-term interest rate expectations should be at least somewhat differentiated, the chart highlights that the real 5-year/5-year forward Treasury yield is very closely explained by the real 5-year Treasury yield and a 3-year lag of our adaptive inflation expectations model (which is highly consistent with BCA’s Golden Rule of bond investing framework). Chart II-16 shows that long-maturity bond yields should be higher than they are based on the current level of real 5-year yields and lagged inflation expectations, underscoring the point that we made in Section 1 of our report that significant upside risk exists for long-maturity bond yields in a non-recessionary outcome over the coming year. In a recessionary outcome, it is clear that bond yields will fall as the Fed cuts interest rates, as Chart II-15 demonstrated. But, Chart II-17 highlights that during recessions, there is little precedent for a negative 5-10 yield curve slope outside of the context of the persistently high inflation environment of the late 1960s and 1970s. Applying that template to the fair value path that we showed in Chart II-15 suggests that the 10-year Treasury yield will not fall below 2% during the next recession. As we noted in our August report,2 a 10-year Treasury yield decline to 2% would result in significant performance for long-maturity bonds, but it would not end the structural bear market in bonds that began two years ago – a fact that we suspect would be very surprising to bond-bullish investors. Chart II-165-Year Bond Yields Strongly Explain Yields 5-Years/5-Years Forward
5-Year Bond Yields Strongly Explain Yields 5-Years/5-Years Forward
5-Year Bond Yields Strongly Explain Yields 5-Years/5-Years Forward
Chart II-17There Is Not Much Precedent For A Negative 5/10 Yield Curve During Modern Recessions, Suggesting 10-Year Yields Will Not Fall Below 2% During The Next Recession
There Is Not Much Precedent For A Negative 5/10 Yield Curve During Modern Recessions, Suggesting 10-Year Yields Will Not Fall Below 2% During The Next Recession
There Is Not Much Precedent For A Negative 5/10 Yield Curve During Modern Recessions, Suggesting 10-Year Yields Will Not Fall Below 2% During The Next Recession
It is true that bond yields may deviate from the fair value levels shown in Chart II-15 if investors expect a different outcome for the path of the Fed funds rate than we described. However, it is worth noting that changes in our assumed post-recession peak Fed funds rate and the long-term average do not substantially change the outcome shown in Chart II-15. If investors instead assume that the Fed funds rate will peak at 3% during the next expansion, that lowers the fair value path for the 5-year yield by approximately 5 basis points. Changing the long-term average Fed funds rate to 2.4%, the Fed’s current neutral rate expectation, would reduce it by about 25 basis points. These levels would still be significantly above the lows reached in 2011-2013 and in 2020, underscoring that the length of the recession and the speed at which the Fed begins to raise interest rates will be far more important determinants of the path of US Treasury yields. We strongly suspect that investors will recognize that a comparatively mild recession will not result in the same hyper-accomodative monetary policy stance that occurred during the past two recessions, implying that long-maturity bond yields will have less downside during the next recession than may be currently recognized. Investment Conclusions As we have presented, the historical experience suggests that the Fed may cut interest rates to zero during the next recession, but that the re-establishment of a long-lasting zero interest rate policy and the associated resumption of large-scale asset purchases seem quite unlikely unless the recession is severe. In the post-WWII environment, severe US recessions have been accompanied by aggravating factors that appear to be missing in the current environment. In addition to this, there are several arguments pointing to the next US recession being a mild one. In a mild recession scenario, we doubt that the 10-year Treasury yield would fall below 2%, or fall below this level for very long. For fixed-income investors, while bond yields will fall for a time if a recession emerges, the implication is that investors should not overstay their welcome in a long-duration position during the recession and should be looking to reduce their duration exposure earlier rather than later. For equity investors, our findings underscore that meaningful downside risk exists for stocks even in a mild recession environment, because the decline in bond yields is not likely to offset a rise in the equity risk premium. We noted in our July report that if a recession occurred within the coming 6-12 months, that the S&P 500 would likely fall to 3100, even if the recession were average. A mild recession may see the S&P 500 decline less severely than this, but stocks are still likely to incur significant losses during the next recession unless investors price in a much shallower path for short-term interest rates than we believe will be warranted. As noted in Section 1 of our report, we have not yet concluded that a US recession is inevitable over the coming 6-12 months. Still, we acknowledge that the risks are quite elevated, and that substantial (further) supply-side and pandemic-related disinflation is likely needed for the US economy to avoid a contraction in output. Additional changes to our recommended cyclical allocation may thus occur over the coming few months, in response to incoming data, our assessment of the likely implications for monetary policy, and the response of long-maturity government bond yields. Jonathan LaBerge, CFA Vice President The Bank Credit Analyst Footnotes 1 Please see The Bank Credit Analyst "Inflation Whipsaw Ahead," dated June 30, 2022, available at bca.bcaresearch.com 2 Please see The Bank Credit Analyst "August 2022," dated July 28, 2022, available at bca.bcaresearch.com