Michael Cembalest Chairman of Market and Investment Strategy for J.P. Morgan Asset & Wealth Management Oct 19, 2022
Reruns: how equity declines precede the fall in earnings, growth and employment during recessions; new US semiconductor export policies on China and the clash of empires; and another press article extolling the renewable energy virtues of a country with little relevance for anyone else
Reruns. When bear markets occur and the investment mistakes of the prior cycle are revealed, bearish investment commentary tends to intensify. There is a confessional, self-flagellating quality to some of this research, as if its authors are trying to atone for having missed the signals and risks during the prior boom. I read around 1,500 pages of research each week and the most consistent message now is a litany of gloom on earnings, valuations, wage and price inflation, Central Bank policy normalization, housing, trade, energy, the surge in the US$, China COVID policy, etc. I am not saying that these things are not important, since of course they are (see Appendix charts). But for investors, there is a remarkable consistency to the patterns shown below: equities tend to bottom several months (at least) before the rest of the victims of a recession.
Let’s start with the Eisenhower recession, which is notable for the lack of monetary and fiscal stimulus deployed in what at the time was a pretty severe recession. Equities bottomed in December 1957, while earnings did not bottom until a year later. GDP and payrolls also didn’t start to improve until the middle of 1958. You will see the same pattern during the 1970’s stagflation, the 1980’s double dip recession, the S&L crisis of the 1990’s, the Global Financial Crisis and the COVID pandemic.
The indexed line chart compares the S&P 500, GDP, Earnings, and Nonfarm Payrolls throughout the Eisenhower recession from June 1957 to June 1959. The S&P 500 bottomed in December 1957, followed by GDP in March 1958, Payrolls in May 1958 and Earnings in September 1958.
The indexed line chart compares the S&P 500, GDP, Earnings, and Nonfarm Payrolls throughout the Stagflation era recession from June 1973 to December 1976. The S&P 500 bottomed in December 1974, followed by GDP in March 1975, Payrolls in April 1975 and Earnings in September 1975.
The indexed line chart compares the S&P 500, GDP, Earnings, and Nonfarm Payrolls throughout the double-dip recession from December 1980 to December 1985. The S&P 500 bottomed in July 1982, followed by GDP in September 1982, Payrolls in December 1983 and Earnings in March 1983
The indexed line chart compares the S&P 500, GDP, Earnings, and Nonfarm Payrolls throughout the S&L crisis from December 1989 to June 1993. The S&P 500 bottomed in October 1990, followed by GDP in March 1991, Payrolls in September 1991, and Earnings in December 1991.
The indexed line chart compares the S&P 500, GDP, Earnings, and Nonfarm Payrolls throughout the Global financial crisis from March 2007 to June 2013. The S&P 500 bottomed in February 2009, followed by GDP in June 2009, Earnings in September 2009 and Payrolls in February 2010.
The indexed line chart compares the S&P 500, GDP, Earnings, and Nonfarm Payrolls throughout the Global COVID pandemic from December 2019 to December 2021. The S&P 500 bottomed in March 2020, followed by Payroll in April 2020, GDP in June 2020 and Earnings in December 2020.
The Dotcom collapse is the outlier: the earnings decline preceded the equity market decline, there was barely a recession at all, and the mini-recession of 0.3% preceded the equity market bottom by more than a year.
As for the latest bear market, it appears on the right. I see no reason why this cycle will not end up looking like most of the other ones. If so, the bottom in equities will occur even as news on profits, GDP and payrolls continues to get worse. When will that be? We will be watching the ISM manufacturing survey very closely. It has a good track record of roughly coinciding with equity market bottoms, as shown in the table. I would consider 3200-3300 on the S&P 500 index good value for long term investors, if such levels were reached sometime this fall/winter.
The indexed line chart compares the S&P 500, GDP, Earnings, and Nonfarm Payrolls throughout the dot-com bubble recession from December 1999 to March 2004. Earnings bottomed in June 2001, followed by the S&P 500 in September 2002. However, GDP bottomed in March 2001, while Payrolls began declining in April 2001.
The indexed line chart compares the S&P 500, GDP, Earnings, and Nonfarm Payrolls in our current time period. The S&P 500 is leading the decline, beginning in January 2022, followed by GDP in March 2022 and Earnings in June 2022. Meanwhile, Payrolls are still rising.
Thucydides Trap update: US semiconductor policy on China deepens the rift
In May 2018, I published the chart below on economic ties between the US and China as a counterpoint to Graham Allison’s “Thucydides Trap” book. Allison’s work analyzed rising empires competing for power with incumbents. By Allison’s account, in 12 of 16 historical examples, competing empires ended up in military conflict, and Allison sees the US-China relationship as a rerun of these precedents. My chart was meant to highlight the unique economic ties between the US and China when compared to historical adversaries. As things stand now, certain fragments of the US-China relationship support my thesis: US imports from China in 2021 reached $540 billion (close to the pre-trade war peak in 2018), and China still holds ~$1 trillion of US Treasuries and Agencies. However, the latest developments on US semiconductor policy do not support it, and represent the most comprehensive change in US-China trade policy in decades.
Chart shows economic linkages between actual and potential adversaries of the last 100 years by plotting each pairing’s combined GDP in a specified year (as estimated by the sum of their bilateral central bank holdings, bilateral foreign direct investment, and bilateral annual trade). US-China economic linkages far surpass any other pairings, exhibiting a combined GDP of over 7% in 2014. The next runner-ups, France-Germany, exhibit a combined GDP of just over 1% in 1930.
The Trump administration took steps against ZTE and Huawei but left open the possibility of future engagement, perhaps in exchange for Chinese cooperation in other geopolitical areas. The latest moves by the Biden administration appear to close those doors, and for a very long time. A summary of the latest US actions1:
- The new semiconductor policies focus on four areas in which the US has a current strategic advantage over China: AI chip designs, electronic design automation software, semiconductor manufacturing equipment and related equipment components
- There are four interlocking elements of the new policy: (1) impede China’s AI industry by restricting access to high-end AI chips; (2) block China from designing AI chips domestically by cutting off China’s access to US-made chip design software; (3) block China from manufacturing advanced chips by cutting off access to US-built semiconductor manufacturing equipment; and (4) block China from domestically producing semiconductor manufacturing equipment by cutting off access to US-built components
- China’s fusion of military and civil activities make it difficult to target the former without affecting the latter (prior bans on the sale of high-end Intel Xeon chips to China’s military didn’t work, since shell companies reportedly acted as an intermediary; Chinas military is reportedly still actively using US chip technology2)
- The new policy: high-end AI chips can no longer be sold to any entity operating in China, whether that is the Chinese military, a Chinese tech company or a US company operating a data center in China. The new rules set a performance standard of what kind of chips can be sold; anything above that level requires an export license from the Dep’t of Commerce which may be subject to “presumption of denial”
- The US is invoking something called the “foreign direct product rule”. The best way to explain it is to give an example of how it works. Any chip manufacturing operation anywhere in the world that seeks to build high end Chinese chip designs will risk losing its access to US semiconductor manufacturing equipment. As a result, Chinese chip design companies will not be able to outsource manufacturing abroad for advanced AI and supercomputing chips, and for the 28 Chinese organizations on the BIS Entity List, they will be blocked from outsourcing the manufacturing of any types of chips at all. [Note: Almost every advanced semiconductor fabrication facility in the world is critically dependent on US technology companies for initial purchase and ongoing onsite advice, troubleshooting and repair]
- New licensing requirements for logic chip equipment sales for chips at 16 nanometers (nm) or less ends up blocking the sale of equipment to China that is several years old and already in use. For DRAM (short term memory) the limit will be 18 nm, and for NAND (long term memory) the limits are 128 layers and higher. Translating the semiconductor jargon: the new policies threaten the viability of some of China’s largest memory chip companies (some of whom have been hoarding as much US made equipment as they can)
- These policies also require all “US persons” (citizens, residents and green card holders) to obtain a license to continue working on development, production or use of integrated circuits at certain Chinese semiconductor facilities
The scale and scope of these restrictions do appear to be unprecedented. While they only apply to a subset of high performance AI-related chips right now (such as those sold by NVIDIA which accounts for 95% of AI chip sales in China), the new chip performance benchmarks are being held constant. In other words, over time, more and more of the semiconductor market will be subject to these restrictions. These new policies complement the recently passed CHIPS Act and its $52 billion for US semiconductor research, manufacturing and workforce development.
What might happen now? US National Security Advisor Jake Sullivan gave a speech last month that suggests that this may not be a one-time thing3:
“Earlier this year, the United States and our allies and partners levied on Russia the most stringent technology restrictions ever imposed on a major economy. These measures have inflicted tremendous costs, forcing Russia to use chips from dishwashers in its military equipment. This has demonstrated that technology export controls can be more than just a preventative tool. If implemented in a way that is robust, durable, and comprehensive, they can be a new strategic asset in the US and allied toolkit to impose costs on adversaries, and even over time degrade their battlefield capabilities”.
Next up: possible restrictions on US entities investing in Chinese technology companies as the focus shifts from the transfer of technology to the transfer of capital4.
Line chart plots the price change since September 2021 of four semiconductor equities and indices: NVIDIA, Advanced Micro Devices, PHL (Philadelphia) Semiconductor Index, FS (FactSet) China Semiconductor Index. The basket of securities performed well until the end of 2021 (with some securities reaching almost +50% returns in November 2021) before declining throughout 2022. Currently, all four securities have returned between -35% to -50% since September 2021.
The Pitcairn Problem: the limited relevance of most small countries with high shares of renewable energy
More reruns: yet another article on tiny countries with high shares of renewable power generation as paragons of the future that are “leading the way” on sustainability without any mention of the factors that limit their relevance for larger developed and developing economies. The latest piece is on Uruguay, from the New York Times5. What’s missing is the broader context of how to understand Uruguay, Iceland, Norway, Costa Rica and other “Pitcairn” examples6.
- Such countries generally have small shares of global GDP, small populations and lower population density. Low density countries face fewer challenges in terms of siting low density wind and solar power
- They tend to be much smaller in terms of land area, an important consideration when thinking about the cost of transmission investment to load centers required for onshore and offshore wind, and solar power
- They often have lower economic complexity (a measure of a country’s ability to produce a wide range of complex products across industries), which requires less developed energy systems
- Most have abundant hydroelectric resources which contribute the lion’s share of electricity generation. Developed countries have already built out most of their suitable hydroelectric resources
- Some have unique geothermal resources (Iceland, Costa Rica, New Zealand), or sugarcane-based biomass whose energy return on investment (“EROI”) is 7x higher than corn-based ethanol (Guatemala and Uruguay)
- Uruguay is interesting in its use of biomass for backup power when wind conditions are low, and in this regard it shares a lot in common with Denmark, another small country with excellent coastal wind resources (40% capacity factors, similar to Uruguay) that uses biomass (manure, wood waste, energy crops and industrial feedstock) as a replacement for thermal power. Both small countries also benefit from proximity to much larger ones for grid stabilization services (Uruguay and Brazil, Denmark and Germany)
The combination of these attributes generally makes the Pitcairn group of countries much less relevant for larger, industrialized, urbanized countries. Most of the latter are pursuing a renewable future based more on wind and solar power, which requires raw materials, project siting, transmission investment and plenty of backup thermal power. More on all of this in next year’s energy paper.
Appendix charts: market indicators are generally downbeat
Line chart shows year-over-year changes in the US dollar from 1990 to present. The dollar surges during each market crisis. Currently, the dollar is up about 20% year-over-year.
Line chart plots the year-over-year change in percent of three consumer price inflation (CPI) measures from 1985 to present: Cleveland Fed median CPI, Cleveland Fed trimmed mean CPI, and Atlanta Fed sticky CPI. All three CPI indicators are up about 7% year-over-year, their highest rate in the time period shown.
Line chart compares actual trailing earnings growth year-over-year to two earnings leading indicator models from 2001 to now. The models closely follow actual trailing earnings growth and point to more earnings declines ahead.
Line chart compares the change in the Fed funds target rate since the start of a Fed tightening cycle for each cycle from 1958-2022. The chart demonstrates that we are currently in the fastest tightening cycle in over 60 years.
Line chart plots the US housing affordability index from 1970 to present. The index hovered around 160-180 for most of the 2010s, indicating greater housing affordability. Currently, the index has dipped to around 100, which is its lowest affordability since 2007.
Line chart plots job openings plus employment as a percent of labor force from 1951 to July 2022, with any point above 0% signaling more jobs than workers. The chart shows that the US is experiencing the largest worker shortage in the post war era.
1 “Choking Off China’s Access to the Future of AI”, Gregory Allen, Director, AI Governance Project and Senior Fellow, Strategic Technologies Program, Center for Strategic and International Studies, October 11, 2022
2 “Managing the Chinese Military’s Access to AI Chips”, June 2022, Center for Security and Emerging Technology
3 “Remarks by National Security Advisor Jake Sullivan at the Special Competitive Studies Project Global Emerging Technologies Summit”, September 16, 2022
4 “White House Weighs Order to Screen US Investment in Tech in China”, WSJ, September 8, 2022
5 “What Does Sustainable Living Look Like? Maybe Like Uruguay”, NYT, October 7, 2022
6 Pitcairn Island, current population 47, where HMS Bounty mutineers settled in 1790