Quantum Computing

QC is on the verge of exciting technological breakthroughs and will eventually present multiple sound investment opportunities. However, the industry is still in its early innings, and buying pure-play QC stocks is akin to buying lottery tickets. Big Tech may end up dominating the space. We recommend a cautious stance.

HighlightsThe current surge in US measured productivity looks very unlike what occurred in the mid-to-late 1990s. A detailed breakdown of labor productivity growth points to atypical labor market compositional effects – namely a significant decline in services employment – as being responsible for the apparent rise in productivity. In addition, technological disinflation, a major ingredient of the late 1990s “disinflationary boom”, is absent today.A cross-country comparison of the growth in output per worker during the pandemic can be mostly explained by differences in the fiscal response to the crisis. US output per worker surged compared to other countries, but the US fiscal response also generated a significant amount of excess income to support economic activity – unlike in the euro area, UK, and Japan.Micro-level arguments and some academic studies argue against the idea that work from home arrangements will ultimately be productivity-enhancing. Remote work makes it more difficult for firms to train the next generation of senior employees, which will raise the staffing risks for many businesses.While the long-term outlook for technologically-driven productivity growth is positive, projected commercialization timelines for several well-known technologies under development do not point to an imminent, inflation-offsetting boom in potential output.If inflation remains significantly above target after the pandemic is over, the Fed’s long-term interest rate projections may rise. US stocks would suffer potentially large losses in a scenario where 10-year US Treasury yields rise towards the potential growth rate of the economy. Investors should consider reducing their equity exposure if 5-year, 5-year forward US Treasury yields break above 2.5%. We do not expect that to occur this year, which for now justifies an overweight stance towards risky assets.Feature Chart II-1A Pandemic-Driven Productivity Surge? A Pandemic-Driven Productivity Surge? A Pandemic-Driven Productivity Surge?  The behavior of US labor productivity during the COVID-19 pandemic has raised several questions among investors. As defined by output per hour worked, US productivity accelerated significantly over the first six quarters of the COVID-19 pandemic, but then fell sharply in Q3 2021 (Chart II-1). While some market participants have questioned the cause of the recent decline, investors have generally been more interested in the question of whether the US is in the middle of a long-lasting productivity surge that will help alleviate inflationary pressure – akin to what occurred in the second half of the 1990s.In this report, we review the recent surge in US labor productivity in contrast to what occurred in the late-1990s, and then compare it with what has occurred globally. While we are not pessimistic about the pace of technological advancement and its potential to drive long-run productivity, we conclude that the US is not likely experiencing a sustained productivity boom driven by technological adoption during the pandemic. This underscores why investors should not expect a significant increase in potential output owing to the pandemic or its effects. It also highlights that, if elevated inflation in response to strongly positive output gaps were to occur over the coming few years, it would likely be met by significantly tighter fiscal or monetary policy.Today Versus The 1990s: Total Factor Productivity Versus Capital Intensity Chart II-2The Technologically-Driven US Productivity Surge In The 1990s Was A Major Macro Event The Technologically-Driven US Productivity Surge In The 1990s Was A Major Macro Event The Technologically-Driven US Productivity Surge In The 1990s Was A Major Macro Event  A technologically-driven surge in productivity growth in the second half of the 1990s was a highly significant macroeconomic event. Chart II-2 highlights that US labor productivity surged to over 3% from 1995 to 2000, alongside a significant deceleration in core PCE inflation and a sizeable acceleration in potential GDP growth.Given the acceleration in measured productivity during the pandemic, and the accompanying rapid adoption (or broader use) of technology, it is easy to see why some investors have questioned whether a 1990s-style productivity boom is underway. However, a detailed breakdown of the 2020 rise in labor productivity growth highlights substantial differences between the current environment and that of the late 1990s, which points instead to compositional effects as the main driver.Improvements in labor productivity can come from smarter workers, an increase in the amount of capital employed per worker, or from technological innovations and better working practices. The US Bureau of Labor Statistics provides a breakdown of the annual change in labor productivity that attempts to capture these three components:The contribution from shifts in labor composition: This measures the productivity impact of changes in the age, education, and gender structure of the labor force.The contribution from capital intensity: This measures the productivity impact of shifts in the amount of capital equipment available per worker.Total factor (or “multifactor”) productivity: This measures the changes in output per hour that cannot be accounted for by the above two factors. Thus, it includes the effects of technological changes, returns to scale, shifts in the allocation of resources, and other changes in operating procedures.Examining the 2020 rise in labor productivity growth along these three factors underscores key differences between the current environment and that of the late 1990s.The first point for investors to note is that the acceleration in labor productivity in 2020 occurred alongside a contraction in total factor productivity (TFP) growth, in contrast to the 1990s when TFP drove labor productivity (Chart II-3). The fact that TFP growth fell in 2020 means that the increase in labor productivity must have occurred either because of labor composition or capital intensity effects.In 2020, labor composition contributed somewhat to accelerating labor productivity, but that most of the increase was caused by a sharp increase in capital intensity. Some of the increase in overall capital intensity occurred because of an increase in the intensity of information processing equipment and intellectual property products (supporting the idea of an increase in pandemic-driven capital deployment), but this was outstripped by the contribution of “other” capital services (Chart II-4). Chart II-3Total Factor Productivity Collapsed In 2020, Unlike In The 1990s Total Factor Productivity Collapsed In 2020, Unlike In The 1990s Total Factor Productivity Collapsed In 2020, Unlike In The 1990s   Chart II-4The Surge In US Capital Intensity Reflects A Rapid Compositional Shift In The Labor Market February 2022 February 2022  The concept of capital intensity refers to the amount of capital available per worker, but in practice it is measured as the ratio of the amount of capital used relative to the amount of labor hours used to produce output. Thus, a surge in capital intensity that is not accounted for by an increase in the amount of tech-related capital available to workers points to a rapid compositional shift in the economy from relatively low capital-intensive industries to relatively high-intensive industries.Under less extreme economic circumstances we would be more inclined to search for other potential causes of a rapid increase in measured capital intensity, but a shift in employment from less to more capital-intensive industries is exactly what has occurred during the pandemic. Services jobs tend to be much more labor-intensive than goods-producing jobs; Chart II-5 highlights that the former fell far more than the latter during the pandemic, in sharp contrast to what normally occurs during a recession (Chart II-6). This phenomenon is also reflected in a highly unusual decline in services spending compared with very strong goods spending relative to their pre-pandemic trend. Chart II-5Employment In Low Capital Intensity Services-Producing Industries Fell Far More Than Goods-Producing Employment In Low Capital Intensity Services-Producing Industries Fell Far More Than Goods-Producing Employment In Low Capital Intensity Services-Producing Industries Fell Far More Than Goods-Producing   Chart II-6The Sharp Decline In Services Jobs During The Pandemic Was Unprecedented The Sharp Decline In Services Jobs During The Pandemic Was Unprecedented The Sharp Decline In Services Jobs During The Pandemic Was Unprecedented  The takeaway for investors is that the nature of the pandemic and its unique impact on the economy has created the appearance of an acceleration in productivity, when in reality true productivity has fallen and the standard measure of productivity is being flattered by enormous changes in the composition of the labor market.Today Versus The 1990s: IT Investment, And Technological DisinflationThe trends in IT investment and prices highlight another major difference between the current environment and that of the late 1990s. Charts II-7 and II-8 highlight recent trends in comparison to those of the 1990s, with the following notable points: Chart II-7There Are Major Differences Between IT Investment And Prices Today Versus The 1990s There Are Major Differences Between IT Investment And Prices Today Versus The 1990s There Are Major Differences Between IT Investment And Prices Today Versus The 1990s   Chart II-8A One-Off Move A One-Off Move A One-Off Move  The recent pace of real investment in total IT does not point to the pandemic as a sustained source of productivity growth. Real investment in IT has already slowed significantly, in contrast to the 1990s when it accelerated on a sustained basis for years.IT investment as a % of GDP and of total plant and equipment spending has already stopped rising (or is now falling), exhibiting clear signs of a one-off shift and thus undermining the view that IT investment has significantly raised potential output.In pronounced contrast to the mid-1990s when IT equipment prices were collapsing, computing equipment inflation has recently risen into positive territory – to the highest levels recorded since the data became available in 1959.Higher prices for IT equipment clearly reflect, at least in part, pandemic-driven pressure on global supply chains and the production of semiconductors. So we do not expect sustained increases in the price of computing equipment. But the key point for investors is that a major ingredient of the late 1990s “disinflationary boom” is missing today.The US Versus The WorldWe have presented Chart II-9 in previous reports to highlight that there is certainly no evidence of a global productivity surge, using output per worker as a proxy for the standard measure of labor productivity (output per hour worked). Some investors have countered that the US is a more dynamic economy, and that a sustained productivity boom would be more apparent in the US prior to its emergence in other countries. Or simply that the US alone is experiencing a productivity boom that will help reduce very elevated US inflation, with strong implications for Fed policy. Chart II-9During The Pandemic, Cross-Country Changes In Real Output Per Worker… February 2022 February 2022   Chart II-10…Are Mostly Explained By Different Fiscal Responses February 2022 February 2022   Chart II-11High US Real Output Per Worker Also Reflects A Lagging Jobs Recovery Relative To Pre-Pandemic Levels High US Real Output Per Worker Also Reflects A Lagging Jobs Recovery Relative To Pre-Pandemic Levels High US Real Output Per Worker Also Reflects A Lagging Jobs Recovery Relative To Pre-Pandemic Levels  Charts II-10 and II-11 present a different cross-country comparison that reinforces the view that the US is not likely experiencing a long-lasting productivity surge that will help reduce inflation. Chart II-10 highlights that in the face of a significant decline in employment, US output was supported by a substantial amount of “excess income” – the cumulative amount of household disposable income earned over the course of the pandemic in excess of what would have been predicted based on the pre-pandemic trend.Other major DM economies (such as the UK and euro area) either saw negative excess income or a modestly positive amount (Japan), underscoring that the fiscal response to the pandemic in most advanced economies was aimed at stabilizing income rather than raising it. In combination with Chart II-11 – which highlights that the US labor market recovery has significantly lagged behind the European and Canadian economies in terms of returning to the pre-pandemic employment trend – this would appear to explain why the US has experienced stronger real output per worker than other countries. Chart II-12Given A Similar Fiscal Response, Would The US Have Canada's Job Recovery If It Had Less COVID Cases? Given A Similar Fiscal Response, Would The US Have Canada's Job Recovery If It Had Less COVID Cases? Given A Similar Fiscal Response, Would The US Have Canada's Job Recovery If It Had Less COVID Cases?  Canada stands out as the outlier compared with the US, in the sense that it’s growth in real output per worker has been much lower but Canadian fiscal policy created a similar amount of excess income. However, it may be the case that the Canadian experience highlights that the US labor market recovery is the outlier, which could imply that the surge in US labor productivity may in fact have inflationary rather than disinflationary consequences at the margin.We discussed the factors that we believe are driving the slow recovery in the US working-age population in our 2022 annual outlook report, and how they are strongly linked to the pandemic. However, Canada has also clearly been affected by COVID-19, and yet it has experienced a more significant recovery in jobs.Chart II-12 highlights that there has been one major difference between the US and Canada during the pandemic: a substantial gap in the burden of disease from COVID-19. This raises the question of whether Canada has outperformed the US in terms of its labor market recovery, despite a similarly impactful fiscal response, because of a smaller labor shortage stemming from long-term COVID symptoms.Over the past two years, there have been many reports about people who have recovered from COVID but who continue to experience some symptoms of the disease. The medical community has labeled this condition as post-acute sequelae of SARS-CoV-2 infection (PASC), colloquially referred to as “long COVID.” Chart II-13Long-COVID Might Help Explain The US’ Lagged Return To Pre-Pandemic Employment February 2022 February 2022  The medical community’s understanding of long COVID is currently poor, and doctors do not know why some people get the condition or what treatment options are likely to be the most effective. Given this, it is possible that some reports of long COVID are, in fact, related to other conditions.But a recent research report from Brookings estimated that the US labor market may be missing 1.6 million workers because of long COVID’s effects (Chart II-13), which alone would account for 1 percentage point (or roughly 1/4th) of the growth in US real output per worker since the pandemic began. This circumstance would be inflationary rather than disinflationary on the margin, as it would imply that accelerating first and second quartile US wage growth may be sticky even as the pandemic recedes.Is Working From Home Positive For Productivity?We have noted above that the macro data argues against the idea of a sustained rise in US productivity stemming from the pandemic. A more micro-level perspective, one that examines the working-from-home (WFH) experience, also appears to support our case.It is true that surveys of employees highlight that their experience of WFH has been significantly better on average than workers expected and report their being more productive while working from home during the pandemic. Chart II-14 emphasizes that, based on the running surveys from Barrero, Bloom, and Davis (“BBD”), 60% of workers have conveyed better WFH outcomes relative to expectations, versus just 14% reporting worse outcomes. In addition, Chart II-15 clearly highlights that workers prefer at least some form of hybrid WFH arrangement, with just 22% of survey respondents reporting the desire to work from home either rarely or never. Chart II-14Remote Workers Have Reported Better Work-From-Home Outcomes Than What Was Expected February 2022 February 2022   Chart II-15Remote Workers Clearly Prefer A Hybrid Work Model February 2022 February 2022  However, worker preferences do not necessarily correlate with productivity gains, at least not to the same degree. Chart II-16 from the BBD surveys highlights that the share of workers reporting more efficiency while working from home is not as large as those reporting better outcomes relative to expectations, suggesting that employees are considering whether WFH arrangements are benefiting them personally when responding to their desired post-pandemic level of remote work. Chart II-17 also shows that employees working from home only spend a third of the time ordinarily allocated to commuting to working on their primary job; the rest is spent on childcare, leisure, home improvement, or working on a second job (which may or may not be a sustainable source of income). Chart II-16Less Than Half Of Workers Report Being More Efficient While Working Remotely February 2022 February 2022   Chart II-17Only 1/3rd Of Time Saved Commuting Is Spent On Primary Employment February 2022 February 2022  There is also some evidence from academic studies that indicates productivity fell during the pandemic for some remote workers. Michael Gibbs, Friederike Mengel, and Christoph Siemroth (2021) surveyed 10,000 professionals at a large Asian IT services company, and found that productivity declined because of a slight decline in average output and a rise in hours worked.1 Admittedly, elements of the study did point to some factors potentially impacting this decline in productivity that were more prominent in the earlier phase of the pandemic, specifically the issue of childcare (which would not likely be a drag on remote worker productivity in a post-pandemic environment).But it also noted that employees with a longer company tenure fared better, which in our view is an often overlooked element of remote work that points to less future productivity gains from WFH arrangements than may be recognized by investors. The outperformance of senior staff in a WFH environment is not particularly surprising: once employees have accrued significant experience, they spend less of their working time learning and more (or all) of their working time “doing.” It makes sense that employees who predominantly “platform” their existing experience may fare the same or better in a WFH arrangement, but it is highly questionable whether it is sustainable, because it makes it much more difficult for businesses to train the next generation of senior employees.The Gibbs, Mengel, and Siemroth study noted that higher communication and coordination costs featured prominently in their findings of reduced remote worker productivity. Importantly, they found that employees communicated with fewer individuals and business units, both inside and outside the firm, and received less coaching and one-to-one meetings with supervisors. While some firms may be able to mitigate these risks to the advancement and development of more junior staff while maintaining a hybrid on-site / WFH model, we suspect that many firms will fail to do so fully.Future Productivity: Pessimism Unwarranted, But No Inflation SalvationThe fact that the US is not likely in the middle of a pandemic-driven productivity boom does not mean that the outlook for productivity is poor. In fact, we would point to two factors that lead us to believe that productivity growth will be better in the future than it has been over the past decade:The pronounced consumer deleveraging phase that existed for several years following the global financial crisis is over, andThere are several identifiable technologies currently under development that are likely to have legitimate commercial applications and productivity-enhancing benefits in the futureOn the first point, we have contended in previous reports that the weak productivity growth observed during the first half of the last economic expansion was because of demand rather than supply-side factors. This notion is jarring for many investors, who are accustomed to think of productivity trends as being exclusively driven by supply-side phenomena. This is typically correct, in that the cyclical impact of fluctuating aggregate demand on measured productivity – particularly during and immediately after recessions – is usually temporary in nature.However, the 2008/2009 recession was highly atypical, in the sense that it was a household “balance sheet” recession rather than a normal “income” recession. This led to a prolonged period of US household deleveraging, below-average corporate sales growth, and poor growth in output per hour worked. In effect, the post-2008 deleveraging phase created a long-lasting, multi-year cyclical effect on measured productivity growth.In early-2009, pessimistic investors held to an understandable reason for why they doubted the sustainability of the economic recovery: there could be no meaningful labor market recovery if businesses expected several years of weak demand because of the likelihood of consumer deleveraging. In this respect, the post-2008 period served as an important natural experiment for macroeconomists and investors: we have learned that the response of firms to a durable but shallow economic recovery is, on the one hand, to hire additional workers, but, on the other hand, also to control wage and salary costs aggressively. Chart II-18Slow Productivity Growth Last Cycle Was A Demand Story, Not A Supply Story February 2022 February 2022  Chart II-18 encapsulates the point that weak productivity during the last economic cycle was closely tied to US household deleveraging. The chart highlights that the decline in total factor productivity due to goods-producing industries – heavily concentrated in manufacturing – was much larger than for private services from 2007 to 2019. Since there was no technological slowdown that disproportionally impacted the manufacturing industry during the period, this clearly points to demand-side rather than supply-side factors as the main driver of the post-GFC productivity slowdown.On the second point about future productivity growth, Table II-1 outlines five well-known technologies that are in various stages of development and are likely to lead to significant applications at some point in the future: artificial intelligence, automated driving (a specific application of AI), quantum computing, augmented/virtual reality and human-machine interface, and CRISPR/gene editing. The table outlines the nature of potential future applications, as well as projections from McKinsey Global Institute about the most likely commercialization timeline. Table II-1Technological Advancement Is Ongoing. It Won’t Likely Help Fight Inflation Over The Next Few Years February 2022 February 2022  A detailed analysis of each of these technologies is beyond the scope of this report, but Table II-1 underscores two key points for investors. The first is that further, technologically-driven productivity growth is not just possible, it is likely. It is clear what advancements will probably drive these productivity gains, and Table II-1 highlights only the most well-known technologies to which experts in the field would point to.The second point is that most major changes from these technologies are projected to occur beyond 2025, and, in many cases, beyond this decade. In the case of quantum computing, while it could potentially lead to an explosion of algorithmic power that would almost certainly have major commercial implications, it is even possible that this technology will initially subtract from total factor productivity growth before contributing positively. This is because of its potential to render much of the existing global internet security and privacy infrastructure useless, as highlighted by a NIST Cybersecurity White Paper last April:“Continued progress in the development of quantum computing foreshadows a particularly disruptive cryptographic transition. All widely used public-key cryptographic algorithms are theoretically vulnerable to attacks based on Shor’s algorithm, but the algorithm depends upon operations that can only be achieved by a large-scale quantum computer. Practical quantum computing, when available to cyber adversaries, will break the security of nearly all modern public-key cryptographic systems.”2Some experts believe that the preparation required to avoid this outcome may dwarf that of the millennium bug (“Y2K”) problem of the late-1990s,3 which cost roughly 1% of GDP to fix – and thus was clearly not productivity-enhancing.The bottom line for investors is that while the long-term outlook for technologically-driven productivity growth is bright, it is unlikely to save the US and/or global economies from elevated inflation over the next several years if output gaps in advanced economies rise to strongly positive levels in the wake of the pandemic.Investment ConclusionsOur analysis above has highlighted that the current surge in measured productivity looks very unlike what occurred in the mid-to-late 1990s, and that very atypical labor market compositional effects are likely responsible for the apparent rise in labor productivity. We have also highlighted that a cross-country comparison of the growth in output per worker during the pandemic can be mostly explained by differences in the fiscal response to the pandemic, and that there are micro-level arguments against the idea that work from home arrangements are productivity-enhancing. Finally, while the long-term outlook for technologically-driven productivity growth is positive, projected commercialization timelines for several well-known technologies under development do not point to an imminent, inflation-offsetting boom in potential output.While we believe that the COVID-19 pandemic will recede in importance this year, it is not yet over. As such, investors do not yet know how strong the output gap in the US and other advanced economies will be on average over the coming two to three years, or what the pace of consumer price inflation will look like in the face of strong aggregate demand but substantially lower (or no) pressure from the supply-side of the economy (as we expect). Chart II-19There Is A Lot Of Downside For Stocks If Bond Yields Rise To Potential Growth Rates There Is A Lot Of Downside For Stocks If Bond Yields Rise To Potential Growth Rates There Is A Lot Of Downside For Stocks If Bond Yields Rise To Potential Growth Rates  In a scenario in which aggregate demand remains strong next year and inflation remains above-target, even in the face of Fed tightening and a normalization in services/goods spending, we would expect to see significantly tighter fiscal or monetary policy. This is a scenario in which the secular stagnation narrative, which underpins the Fed’s low long-term interest rate projection, would likely be aggressively challenged by investors. Chart II-19 highlights that US equities would potentially suffer a 24% contraction in the forward P/E in a scenario in which the equity risk premium is in line with its historical average and 10-year US Treasury yields rise to the potential growth rate of the economy.We do not yet believe that a significant rise in long-term interest rate expectations will occur this year, meaning that investors should still be overweight stocks versus government bonds over the coming 6-12 months. But as we noted in last month’s report, we may recommend that investors reduce their equity exposure if 5-year, 5-year forward Treasury yields break above 2.5% (the FOMC’s long-run Fed funds rate projection), which we noted in Section 1 of our report is 50 basis points above current levels.Jonathan LaBerge, CFAVice PresidentThe Bank Credit AnalystFootnotes1 Michael Gibbs, Friederike Mengel, and Christoph Siemroth. “Work from Home & Productivity: Evidence from Personnel & Analytics Data.” Working Paper No. 2021-56. July 13, 2021. Pp. 1-30.2 William Barker, William Polk, and Murugiah Souppaya. “Getting Ready for Post-Quantum Cryptography: Exploring Challenges Associated with Adopting and Using Post-Quantum Cryptographic Algorithms.” National Institute of Standards and Technology, US Department of Commerce. April 28, 2021. Pp. 1-7.3 Jonathan Ruane, Andrew McAfee, and William Oliver. “Quantum Computing for Business Leaders.” Harvard Business Review, January-February 2022.