The Index Investor
September 2019

Note: The language we use to describe our estimated likelihood of asset class over or undervaluation is based on US Intelligence Community Directive 203 on Analytic Standards, which includes the following table:

Market Stress Indicators (@31Aug19)

At the end of August, four out of our five indicators showed a higher level of underlying market stress. The lack of movement in the very low BB spread while gold prices significantly rose was particularly notable.

Macro Regime Forecast Probabilities (@31Aug19)

The Current State of Quantitative Regime Predictors

Our quantitative forecast methodology focuses on the level and change in three-month returns, over the most recent and previous three-month periods, for those asset classes, which should perform best under different regimes (in this sense, our regimes can be regarded as macro factors). We assume that higher returns are associated with a higher underlying probability for the relevant macro regime.

This month’s Evidence File (see below) summarizes the high value indicators and surprises we observed this month in the areas of technology, the economy, national security, society, politics, and investor and financial market behavior. Here are the most important ones that affected how we updated our regime probabilities this month.

As has often (but not always) been the case in my life, market action in August was relatively muted, which gave many observers time to reflect on bigger picture issues, and the overall state of the global system. Of the reflections that were published this month, few if any were uplifting or optimistic.

Instead, they have likely reinforced readers’ already heavy sense of foreboding, and fears that the complexity of today’s challenges outstrips the capacity of current political governance systems to successfully address them. As we wrote in last month’s feature article, a regime change is coming that will very likely be one the world has not seen for many years.

That said, there was not sufficient new evidence this month to justify any changes to our current probability forecasts for different macro regimes.

Tactical Asset Allocation Implications of Our Forecasts

Based on subscriber requests, this month we are re-introducing a feature from the previous version of The Index Investor: Tactical Asset Allocation Implications from our analyses.

These take two forms. The first takes a systematic approach, and is based on relative asset class valuations. Our starting point is our “neutral” model portfolio, which is equally weighted across nine broad asset classes, and also includes a 10% allocation to alpha strategies (equity market neutral and global macro) that are designed to have a low correlation to returns on broad asset classes. Based on asset class valuations, we systematically vary the asset class weights (but not the active strategy weight), increasing from 10% to 15% when an asset class is likely undervalued, and 15% when it is very likely undervalued. In the case of overvaluations, we go to 5% and then into cash, if there are no undervalued asset classes with room for an increase. In effect, this replicates the systematic rebalancing strategy we used for 15 years in our previous model portfolios.

The second tactical approach is based on our subjective view not only of current asset class valuations, but also of the implications of the broader macro trends and uncertainties that we analyze each month. Importantly, this subjective view reflects our primary goal of avoiding large downside losses, rather than seeking large upside gains.

Two final notes. First, with respect to US fixed income, we include credit products (investment grade and high yield) in the same asset class as government debt, and will shift into the former when their valuations become attractive. Second, we regard gold not as a separate asset class to be held long-term, but rather as a complement to cash, into which we shift in periods of substantial overvaluation across multiple asset classes.

Here are our tactical asset allocation views at the start of September:

Pre-Mortem Analysis

One of the most important forecasting disciplines is to ask yourself why your forecast could be wrong. Dr. Gary Klein’s research has shown that a very powerful and insightful way to do this is via a “pre-mortem analysis.” This method asks you to assume that it is a point in the future, and your forecast has been proven wrong (or your strategy or company has failed). You are then asked to look backward from this imagined point in the future, to explain why you failed, what you missed, and what you could have done differently to avoid your fate.

The pre-mortem method takes advantage of the fact that humans reason much more concretely and in more detail when explaining the past than they do when trying to forecast the future.

So let us assume that it is one year from now, and our current forecast has turned out to be wrong.

How did this happen? What developments did we fail to anticipate? Here are three possibilities:

• The leaders of the world’s three major powers – Xi Jinping, Donald Trump, and Vladimir Putin are all facing weakening economies and declining political popularity. History teaches us that this can lead to increased “foreign adventurism” to distract the public from worsening domestic conditions, as a nation rallies around its leader in a period of heightened external conflict. Should a “kinetic” conflict develop between China and the United States, or between Russia and one or more European countries, it would generate a sharp increase in uncertainty that would likely cause an equally sharp economic slowdown and, given high debt levels, speed the arrival of the Persistent Deflation Regime.

• As we have previously noted, while the probability is remote, a supply side shock of some type could produce a sudden increase in inflation – the most likely scenario being a reduction in oil supplies due to a kinetic conflict in the Middle East (e.g., escalation of the current US-Iran conflict) that produced a prolonged disruption in global oil supplies, or, less likely, an infectious disease pandemic or major crop failures (e.g., due to climate change and/or disease).

• While we believe it is very unlikely, we can envision a scenario in which for a range of possible reasons, both Xi Jinping and Donald Trump leave their current roles, and are replaced by leaders who are more committed to lessening conflicts both between China and the United States and in the international system as a whole. This would likely provide a strong boost to confidence (and thus lead to an equally strong reduction in uncertainty). Whether this would also create an opening for a reduction in domestic political conflict in the United States, and thus progress on policy reforms to address weak growth and rising inequality isn’t clear.

System Tipping Points/Critical Threshold Analysis

Like Professors Andrew Lo, Doyne Farmer and others, we regard financial markets as a complex adaptive system (CAS), that exist as part of a larger macro system comprised of other CAS between which there are multiple feedback loops. These other systems include those that produce technology innovations, and economic, environmental, national security (including cyber), social, demographic, and political outcomes.

We also find that these systems tend to operate and generate effects in a rough chronological sequence, albeit with many feedback loops between them. The following chart highlights that the changes we observe in different areas at any point in time are actually part of a much more complex evolutionary process.

While most media coverage of these systems focused on flows (e.g., the size of the government deficit), rapid non-linear change in complex adaptive systems is often caused by a key stock (e.g., the amount of outstanding government debt) exceeding a critical threshold.

The next table highlights the key macro system stocks that we monitor.

In the next section, we will discuss information received over the past month that is related to these stocks, and which we believe is significant to our assessment of the probabilities that a critical threshold will be reached and a regime change will occur. We will conclude with our estimate, at the end of this month, of how close the macro system is to these critical thresholds, and the implications for financial market regime change probabilities.

How Close is the Macro System to One or More Critical Thresholds?

As we have noted, the macro drivers of financial market regime changes typically follow a rough chronological sequence, from technology to economic, security, social, and political causes and effects. Yet there are many feedbacks loops between them, creating complex root causes for many of the critical thresholds we have identified.

Understanding the time dynamics in this complex system is critical to avoiding substantial downside investment risk.

We use the UK Met Office Warning Model to communicate our assessment of these time dynamics. We estimate the time remaining before a critical macro system threshold is reached that could trigger a regime change, which is usually accompanied by substantial changes in asset class valuations.

The model uses three increasingly serious levels of warning, from “Be Aware” (condition yellow), to “Be Prepared” (condition orange), to “Take Action” (condition red).

For our purposes, we denote as “Be Aware” (yellow) critical thresholds that we assess to be three or more years away. We estimate that “Be Prepared” (orange) thresholds could be reached within 1 to 3 years. “Take Action” thresholds are very likely to be reached within one year.

Given their nature, we also note that in our three “wildcard” areas (Environment and Energy related; Disease and Human Caused Bioevents; and Cyber and Electromagnetic Events), our forecasts have higher levels of uncertainty.

The following charts summarize our current estimate of the time remaining before different critical thresholds will be reached.

At the highest level, we believe the complex adaptive global macro system can be in one of four states, based on its degree of order versus disorder, and degree of social cooperation versus conflict. A very coarse-grained reading of history suggests that these states evolve in a predictable cycle, from ordered/cooperative, to disordered/cooperative, to disordered/conflicted, to ordered/conflicted.

We believe that the system is currently in its most uncertain state, characterized by high degrees of underlying disorder and social conflict, both domestically and internationally. Beyond some point, intensifying conflict eventually increases the degree of order in the system. That appears to be happening now, via the increasing conflict between China, Russia, and Iran and the United States and other Western nations.

Note: Combining Our Forecasts with Others From Other Sources and Extremizing the Result Should Increase Your Predictive Accuracy

Research has found that three steps can improve forecast accuracy. The first is seeking forecasts based on different forecasting methodologies, or prepared by forecasters with significantly different backgrounds (as a proxy for different mental models and information). The second is combining those forecasts (using a simple average if few are included, or the median if many are). The final step, which significantly improved the performance of the Good Judgment Project team in the IARPA forecasting tournament, is to “extremize” the average (mean) or median forecast by moving it closer to 0% or 100%.

Forecasts for binary events (e.g., the probability an event will or will not happen within a given time frame) are most useful to decision makers when they are closer to 0% or 100% than the uninformative “coin toss” 50%. As described by Baron et al in “Two Reasons to Make Aggregated Probability Forecasts More Extreme”, forecasters will often shrink their probability estimates towards 50% to take into account their subjective belief about the extent of potentially useful information that they are missing.

When you average multiple forecasters’ estimates, you are including more information, which should increase forecast confidence and push the mean estimate closer to 0% or 100%. However, this doesn’t happen when you use simple averaging. For this reason, forecast accuracy is increased when you employ a structured “extremizing” technique to move the mean estimate closer to 0% or 100%.

You can download an extremizing model from our website to use when combining the forecasts you use in your decision process.

The extremizing factors in our model are those that the Good Judgment Project found maximized the accuracy of combined forecasts. Note that the extremizing factor is lower when average forecaster expertise is higher. This is based on the assumption that a group of expert forecasters will incorporate more of the full amount of potentially useful information than will novice forecasters.

High Value Information Observed In August 2019

In our methodology, we classify new information as significant and highly valuable if either it (1) is an “indicator”, which reduces or increases our uncertainty about the value of a parameter in our mental model for making sense of the dynamic macro system, or (2) it is a “surprise” which increases our uncertainty about, and causes us to revaluate the structure of our mental model.

Feature Article: Future Labor Productivity Growth: The Supply Side Secular Stagnation Story

“Productivity isn’t everything, but in the long run it is almost everything.” – Paul Krugman, 1994

Consider what has happened over the twenty-five years since Krugman said this. In the United States, the average increase in the size of the potential labor force was 0.91% per year. Over the same period, the average increase output per hour worked (a measure of labor productivity) in private non-farm businesses was 2.1% per year. Combining these two, the potential output of private non-farm businesses grew by about 3% per year. The actual growth of real GDP averaged 2.7% per year.

In other words, actual growth was relatively close to the “speed limit” imposed by the growth of the labor force and labor productivity. Growing faster than this would risk triggering higher inflation. In point of fact, we were probably closer to the speed limit than these data indicate, because labor productivity growth in sectors of the economy outside of private non-farm businesses has actually been much lower – some have estimated that it may actually be negative in healthcare and education, which together account for almost 25% of America’s GDP.

Contrast this with the period 1960 – 1979, when US GDP growth averaged 3.9% per year, and ask yourself how much easier it would be to either avoid or solve many of the problems we face today if GDP had kept growing at that rate.

Today, labor force growth is projected to grow much more slowly than in the past, which makes increasing labor productivity more important than ever before as the driver of increased potential output and GDP growth.

Before delving into the challenges and uncertainties this entails, it is important to first clear up some of the definitional issues that surround the general concept of labor productivity.

Historical labor productivity has been calculated in terms of growth in output per labor hour, by dividing GDP by the total hours that were worked to produce it. However, measuring productivity this way introduces a number of uncertainties.

The first is the extent to which GDP accurately captures the total value of an economy’s output (e.g., the value of many free services provided by technology companies in exchange for their access to our browsing and social media data, or the value of outputs produced in the home by unpaid labor – e.g., loads of laundry done).

The second is the accuracy of the labor hours worked measure (e.g., it doesn’t capture the time people spend on work emails when they are doing that laundry). Studies that have examined these measurement issues have concluded that while they introduce a measure of uncertainty into labor productivity estimates, the size of potential measurement error is small relative to the slowdown in productivity growth that has occurred in recent years (e.g., see, “Does the United States have a Productivity Slowdown or a Measurement Problem?” by Byrne et al).

A potentially more serious problem is caused by differences in the rates at which GDP and hours worked change over time. For example, if GDP is falling, but businesses at first hesitate to lay off their employees, this will create the appearance of a fall in labor productivity in the data, even though the productivity of those worker in terms of their ability to potentially produce a maximum quantity of output (assuming demand for it exists) has not changed at all. Conversely, if GDP increases by management delays the rehiring of previously laid off workers, the data may give the appearance of an increase in labor productivity when, in potential output terms, none has occurred.

The key point is that we have to separate historical labor productivity growth from growth in potential output, or aggregate supply. In this analysis, we are focused on the latter (we covered aggregate demand related issues in last month’s feature on the coming economic downturn).

Let us turn now to the key uncertainties we confront when estimating the rate at which labor force productivity (and thus potential output) will likely grow in the future.

We’ll start with the health of what will be an aging labor force. Poor health reduces labor productivity.

Today, Medicare (for people 65 and older) and Medicaid (for people below a certain level of income) pay lower prices to health care suppliers that private sector insurers pay, and far less than what people without insurance pay. In healthcare, the law of one price doesn’t apply, and prices and quality measures are far from transparent. In effect, the current system facilitates the cross-subsidization of Medicare and Medicaid clients by people who purchase private health insurance, either directly or through their employer.

As the population ages, and if the gig economy grows and income inequality continues to worsen, this business model will come under even more strain that it is today. As Herbert Stein famously observed, “if something cannot go on forever, it will stop.” Time will tell what will come next, and whether the next US healthcare business model will result in a healthier workforce.

Beyond their health, the knowledge and skills of the labor force are critical to productivity growth. Recent research into productivity differences across companies has highlighted why this is the case. Companies with highly skilled and knowledgeable employees have a two-edged advantage. Not only can they absorb and apply new productivity enhancing technologies faster than other companies, but they are also better at creating innovative new technologies, products, and business models. This translates into a productivity advantage, whose benefits can be divided between customers (via more attractive offerings and/or lower prices), employees (via higher compensation), and investors (via higher returns to equity and less credit risk for lenders). Observers have also noted that, in so far as these benefits go to employees, they are also a driver of worsening income inequality.

This is why, in survey after survey, CEOs are placing their ability to attract and retain talent near the top of the list of the risks facing their companies. It is also why, across more and more industries, we are seeing the emergence of a top tier of “winner take most” firms, that are generating substantial economic value, and leaving most other firms in their industry (which lack the deep talent base, advanced technology, and high productivity) with economic returns that barely cover their cost-of-capital and leave little or no room for employee compensation increases (e.g., see, “The Global Productivity Slowdown, Technology Divergence And Public Policy: A Firm Level Perspective”, by Andrews et al from the OECD, “Management as a Technology” by Bloom et al, and “A Bad Time To Be Average”, by Burkner et al from BCG).

Unfortunately, the poor current state of student achievement in America’s K-12 education system (and its seemingly intractable resistance to substantial change), the questionable results being produced by significant parts of its university system, and its fragmented and ineffective workforce development and lifetime learning programs, collectively lead to a sense of pessimism that the average knowledge and skill level of the US labor force will see a dramatic improvement anytime soon.

Labor productivity can also be boosted by capital investment. Some observers have noted the decline in business capital investment in the wake of the 2008 Global Financial Crisis and continuing uncertainty about future demand as important contributors to weak labor productivity growth.

Others have claimed that this decline may be overstated, due to the lower capital needs of many businesses today (e.g., being able to use cloud services rather than purchase their own servers), and increasing spending on R&D (which is often treated as an expense rather than an investment, even though its benefits are expected to extend over many years). However, other researchers have pointed to the decline in government funding for basic scientific research, and a fall in the productivity of private sector R&D spending (e.g., see “Are Ideas Getting Harder to Find?” by Bloom et al).

A more optimistic argument is that, like general purpose technologies that substantially increased labor productivity in the past (e.g., electricity), it will take many years before the benefits of some of today’s technologies (e.g., automation and artificial intelligence) have a similar impact, due to the long time companies need to adjust processes, systems, structure, staff, and often business models to take full advantage of them (but again, that process is constrained by the availability of human talent).

A final uncertainty with respect to overall labor productivity growth has its roots in the sectoral shifts in employment that have taken place in the U.S. economy. As agriculture became mechanized and labor productivity increases, workers moved from the farm into the factories, where mechanization increased their productivity there too. More recently, however, increasing manufacturing productivity and offshoring have forced more and more workers into service sectors where labor productivity growth has been much slower. Today, agriculture accounts for only 1.4% of American employment, and manufacturing just 7.9%.

In contrast, healthcare and social assistance are at 12.4%, state and local government at 12.2% (to which educational services adds another 2.3%), leisure and hospitality at 10.2%, and retail trade at 9.8%. Collectively, these areas account for 47% of U.S. employment – and all are areas where labor productivity growth has lagged far behind agriculture and manufacturing.

For example, between 1987 and 2017, it averaged only 1.4% per year in hotels, 0.4% per year in restaurants, and was (some have claimed) negative in both healthcare and education.

These sectoral shifts in employment have significantly contributed to the lower overall level of labor productivity growth in the U.S. economy. Unfortunately, absent substantial changes in these sectors (particularly healthcare and education business models), it seems unlikely that we will see substantial aggregate labor productivity improvements in the future.

As this analysis has shown, the “secular stagnation” narrative is about supply as much as demand. After weighing the evidence that is available today, it seems very likely that U.S. labor productivity growth (and thus growth in potential GDP) will remain weak over the next 5 – 7 years. However, there is an even chance that it will accelerate thereafter, as the full impact of new technologies emerges across more sectors of the economy. Whether this will translate into an equally substantial increase in actual GDP growth will depend on whether the demand side challenges we analyzed last month are successfully met.

For more reading:

“The Case for An American Productivity Revival” by Branstetter and Sichel of the Peterson Institute for International Economics

“Is Slow Still the New Normal for GDP Growth?”, by Fernald and Li of the Federal Reserve Bank of San Francisco

“Productivity Puzzles”, by Andrew Haldane, Chief Economist of the Bank of England

“The Rise and Fall of American Growth”, by Robert J. Gordon