Playing the the inflation regime game

 

Inflation is not going away as measured by the PCE core inflation rate. It is sticky which means that we will not be transitioning into a new inflation regime just yet. While we may not be in a regime of fire or ice but a simmering burn that does not lead to a risk-on environment. It is not clear that bonds are a safe asset because inflation expectations have not fallen. Inflation is still impacting real wages and savings are being eroded. Equities are supposed to be immune from inflation but at depends on whether firms have control of pricing.  

The Man Institute in their work, "Varying by Degrees: Fire & Ice 2.0" describes the different regimes for inflation and shows what will be the average returns for several asset classes. This is all good work; however, we still need to know what is the regime and what are the points of transition. These transitions are often not easy to measure. A further Fed increase is likely, but the link between Fed policy and inflation does not seem too well-defined in the current environment. 

Is now the time for international diversification?

 

Has international diversification been one of the big loser trades for investors? If you look at returns for US equities and compares to the rest of the world, the answer is yes. You cannot eat diversification, but the returns of the past are not the basis for making future decisions. A recent paper by AQR states the case for international diversification and their arguments are compelling. See "International Diversification - Still Not Crazy After All These Years".

The argument for international diversification is based on five points. One, basic theory has not changed. There is value from diversifying across countries and business cycles. Two, while markets become more correlated when there is a crash, there is still diversification benefit albeit less than expected. Three, changes in valuation can drive returns higher and cause investors to make wrong decisions. The gains in the US equity markets were driven by changes in valuation not the underlying quality of the markets. Four, given the high relative valuation today, there is a bias against the US. Look forward not backward when making the diversification decision. Five, there is a value for active investors across international markets because factor variation is more disperse. 

Again, this may be the time to think about international diversification.

Time series variation of factors -The zoo is dynamic

There is a growing zoo of factors that have significant explanatory power through time, but a closer look of this factor zoo suggests that these factors are dynamic and go in and out of favor or significance. Markets are more complex and difficult to understand because the agent who trade are dynamics with changing objective functions. The significance of any factor changes with the firms that constitute the market, the investors and their objectives, and the market environment. It is no easy task to say which factors will be important in explaining cross-sectional returns and out of sample Sharpe ratios. Nevertheless, the recent paper "Time Series Variation in the Factor Zoo" provides a wealth of analysis on the behavior of factor risks.

This is a. complex paper given the wealth of information provided and the subtle conclusions that are generated. It is worth a close read. Perhaps it will be needed to read more than once.  For example, the importance of the classic 3 to 6 factor core Fama-French model varies significantly through time and leads other factors having an important role both in and out of sample. These factors of varying importance will be associated with economic conditions, especially recessions, and with the diversity in firm characteristics. The environment changes and factor importance changes. The diversity of firms requires more factor to explain variation and allows for more factors being important at any time. 

The cross-sectional pricing of stocks is not easy to solve and requires a wide set of factors that will varying in importance. Employing only six factors will not solve the problem as the research suggests that 30-40 factors may be relevant. This research is critical if you want to use factors as a driver for stock-picking and portfolio construction. 

Multi-factor rotation now seen in an ETF

 

The rotation of sector and factors based on the economic environment has been take to a new level with an active ETF that will allocate among risk factors based on the business cycle.  These rotation strategies state the risk for different sectors and factors change with economic growth. Defensive sectors or quality firms will do better in a contractionary period while cyclicals and value will do better during recoveries. Risk premia for sectors and factors are time varying and can be exploited. Of course you have to know the relationship between these risk premia and the business cycle and you need to make a good forecast on the business cycle. 

The core model for most of this factor work is based on the idea that the business cycle can be divided into four regimes: recovery, expansion, slowdown, and contraction. The risk factor set includes: size, value, momentum, volatility and quality. The economic regime is defined by a set of key macro indicators to measure economic growth and market sentiment. (See IMFL from Invesco.)

It has been found that low volatility has done poorly before a recession but is a good defensive strategy when the contraction happens. Momentum will do well during an expansion, and value will do well during a recovery. Unfortunately, there are few events to test these rotation strategies. It is hard to make an empirical argument when there have only been three major business cycle events in the last 25 years: the 2000 tech bubble, the GFC, and the pandemic. 

Tinbergen Rule and Trying To Do Too Much

 

The Tinbergen Rule, named after one of the first winners of the Nobel Prize in economics, states that "n" independent policy targets need "n" policy instruments. There are actually 3 main variables in the Tinbergen set-up, data which is independent, the target variable which could be inflation, GDP, unemployment, and/or financial stability, and the policy instrument which is in a broad sense monetary and fiscal policy. More specifically, the monetary policy instrument could be the short rate controlled by the Fed. Fiscal policy may be tax and spending policies. Governments can also use regulation to reach their targets. In this case, regulation can enhance financial stability.

Think about the Tinbergen Rule in the context of the current market environment.  A monetary policy target is 2% inflation, but if you use monetary policy to raise rates, you will have a problem with growth or financial stability.  You can have a policy instrument that may be at odds with a given target. Raising rates may increase the likelihood of a recession or increase financial instability. 

A quick review of the Tinbergen Rule suggests that a given policy mix can be ineffective at reaching specific targets.  This is old school thinking, but it shines a strong light in the limitations of monetary and fiscal policy.

Factor Investing Risks - They Are Variable

 

Factor investing is not as easy as some may state. Factor returns are variable, do not show consistent Sharpe ratios, and are subject to drawdowns. They are driven by different rationales and different trading groups. Factors will also behave differently across the business cycle and different economic environments. 

While there have been hundreds of studies of factors, it still seems to be the case that there are a set of core factors that show persistence; nevertheless, these factors can still fall in and out of favor based in the changing market environment. Investing in factors is risky as seen by the wide variation in Sharpe ratios across asset classes and through time. Drawdowns for even factors that show strong persistence can be very deep (beyond 20%). Factor investing provides diversification benefits, yet factor returns will be lower when there is an overall market decline.

See "Fact, Fiction, and Factor Investing".

Our core view is that factor investing should be integrated with price trend and global macro investing. Follow trends in factors given the return dispersion through time. Follow the key macro factors given the sensitivity of factors to macro trends. Exploit the time series, exploit the behavior across macro regimes, and use the rich/cheapness of factors to gain an edge. There is no guarantee of excess returns, but the strong cyclical behavior suggests that avoiding periods of factor underperformance will be helpful for any factor investor.

Trends in Macro Variables Support Trends in Prices

 

There are trends in price because there is a slow response to new information about the drivers of valuation. A response to new information is often immediate, but it may not always be complete. When there is more uncertainty about new macro information, there is a slower response to these macro trends.

There are also trends in macroeconomic information because the collected information on the business cycle and policy responses often shows a slow adjustment processes. Central banks are often cautious with their actions, and they make forecast mistakes. Policy adjustment will not be immediate to the macro environment. In tracking macro variables, the adjustments of macro sectors take time. For example, the housing sector of the economy does not just adjust to rates immediately. It is a process that takes time, so the impact on financial prices will also take time. Unemployment may be lagging indicator and show a slow adjustment to slower economic activity.

Being more specific, increases in GDP will have a positive impact on equities, currencies, and commodities and a negative impact on fixed income. Inflation will have a positive impact on commodities but negative impact on fixed income. Monetary policy as measured by rising short rates (2-year) will have a negative impact on traditional assets. Trends in macro variables have an impact on the tilt in trends for financial assets or put differently, they will have additive uncorrelated signals that can reinforce price trends or provide unique signals. See "Economic Trend".

The global macro perspective for trend-following states that following trends in macro can enhance the trends found in prices. Global macro trends can be followed by asset class, macro themes, or as combination. In all cases there is value in using macro signals. This is especially the case when economic downturns are identified. Use macro signals to avoid drawdowns in equities and fixed income.

While the premise of macro trends is that they provide better insight on price trends, the two signals are not always correlated and their drawdowns may not occur at the same time. Macro change points will not occur when prices peak or bottom. Hence, there is diversification benefit from holding a diversified portfolio of price and economic trends.

If you are worried about equity and fixed income drawdowns and you are worried about the risks from holding a price-based system, a holding a combination portfolio of price and economic trend signals will be supportive and rational as a diversification strategy. 

 

Robert Lucas and Reality

 

“I guess everyone’s a Keynesian in the foxhole,” ruminated Robert Lucas.

Robert Lucas, the Nobel Prize winning economist, who drove the rational expectations revolution in macroeconomics and was famous for his Lucas Critique died this week. He was likely the most influential macro economist for the last 50 years based on his key work on rational expectations and policy. Whether you liked his work, or whether you believed it was an abstraction from reality, you must read and understand his thinking to study macroeconomics. 

Nevertheless, his comment on Keynesians in foxholes is a telling indictment on the transformation of economic theory to policy reality. Policymakers like to do something. They are activists, and crisis requires action. Lucas may not argue for no action, but his foundational thinking is that if action is expected the impact will be muted or at least unclear. 

More on risk and uncertainty in asset prices

 

Market risk and uncertainty are connected with the uncertainty we see in the macro world. They are connected. Spikes in the VIX index and realized volatility will be tied to uncertainty with macro variables and their link with future values. This is evident with the JLN uncertainty indices and financial risk measures. See Uncertainty Data from Sydney Ludvigson for the details on the macro uncertainty index.

Focusing on this link is critical because it makes investors more macro aware of risks. Macro uncertainty spikes during recessions. Financial risk may spike at other times, but the key driver is financial risk is macro risk.

In 2023, we have seen banking risks, debt ceiling risks, monetary policy risks, and recession risk issues. All have created an environment that changes stock and bond risks and have impacted the return profits for investors. These periods of uncertainty can define a decade. See the return patterns from the AQR paper, "Certainly Uncertain".  While long-term Sharpe ratios seem stable, the returns over a given uncertain environment may be significantly different. 

  

Volatility Across Asset Classes Scored

 

Many investors look at the VIX index as a proxy for market fear, yet there are several issues that limit its usefulness. Most important, it does not represent the risks in all market. It is an index of equity volatility. The VIX also has tendency to stay rangebound for long periods. 

To solve these problems, we look at the volatility of bonds through the MOVE index, and the JP Morgan currency volatility index. For each of these indices, we z-score the values to produce a rolling 60-day index. A positive number represent an increase in risk while a negative number represents a decrease in risk. To get a broader index of risk, we form a weighted average of stock, bond, and currency risk components using a 50%, 30%, and 20% weighting scheme. 

Of course this is not perfect. It does not incorporate global equity volatility and it may underrepresent the risk associated with bonds, nevertheless, we believe that this is an improvement over using a single asset classes without scoring. 

We find that the overall index shows significant changes through time and corresponds to periods of high market risk. It captures the Fed policy changes and uncertainty and the banking crisis over the last 18 months. We have looked at a longer history and find that it does a good job of tracking overall risk  events.

How should you rank hedge funds

 

Most investors would say that you should rank hedge funds based on their Sharpe ratio, but there are fundamental problems with Sharpe that may not adequately measure risk. The Sharpe can be gamed with option strategies and does not account for characteristics beyond volatility. Hence, there are many who argue that using drawdown-based measures may be a better way. Research on the surface suggests that there is little difference in rank measures based on the high rank correlation across different performance measures, but a closer look says that these metrics are not all the same and picking the right performance measure matters.

Investors can classify performance measures based distributional assumption (Sharpe or Sortino) or investors can focus on drawdown analysis such as the Calmar or Sterling. Finally, there also are measures that account for the whole distribution of performance returns. Research suggest that ranking firms based on different performance is not always going to give you the same answer. See "Assessing Hedge Fund Performance: Does The Choice of Measures Matter?". This an older study that provides useful insights for investors. 

The rankings of hedge funds are highly correlated. it is found that all ranking across metrics seem to be the same with correlations well above .95. However, when there is an assessment of the equality of ranks, there is a different story.  There can be large divergences in the ranking based on different metrics. The performance measures are not all the same. Investors may mis-rank managers based on their choice of performance measure. About 1/5 of firms will move to a different rank decile based on a change in the performance measure. Nevertheless, in some cases, the rankings from some performance metrics show persistence when looked at in quartiles; however, there is no persistent when focused directly on ranks.

The traditional measures like Sharpe, Sortino, and Calmar all show persistence and seem to lead to stable rankings over time. Investors may use other measures, but an ensemble of performance metrics may be a helpful tool for rankings managers. 

Commodity prices and the dollar - The changing relationship

 

The dollar and commodity prices usually move in opposite directions. When the dollar gets stronger, there is a fall in commodity prices given most commodities are priced in dollar. The dollar rise makes commodities more expensive for importing countries and thus there is a fall in demand. However, over the last two years there has been a change in this relationship. The dollar strengthening has been matched by an increase in commodity prices.  See "The changing nexus between commodity prices and the dollar: causes and implications".

The rolling correlation has moved to a positive number over the last two years; however, the relationship is trending back to negative and is currently close to zero.  While some of this change is a function of recent shocks, there is also the key difference in US oil flows. The United States is now a net energy exporter.  Higher oil prices have increased the US terms of trade. 

Commodity exporters have done better in the current environment which has turned to commodity (oil) dollar relationship in favor of the positive correlation. This upends some traditional views between commodity importers and exporter currencies. 

Dominant Currency - Depends on a how you count dominance

There has been significant talk about the dollar losing its dominant position in global trade and capital flows. This talk happens on a regular basis when the dollar declines or there is some global event that creates questions on dollar hegemony.  It is always good to go to data to see which currencies are dominant. 

Many investors look first at the foreign exchange reserves held in dollars. That reserve number has fallen over the long-run but seems to be stable at around 60%. Capital flows are a critical number given these flows dwarf trade. International debt issued in dollars is above 60%. More countries have increased their local currency international debt given their stronger sovereign balance sheets, but the dollar is still supreme. International loans and deposits, capital flows not in the form of issued bonds, is still slightly below 60%. The dollar is still dominant for trading; however, increased stable bilateral trade allows for trading outside of the dollar. Exports using the dollar as an invoicing currency is also above 50% and the global payment currency through SWIFT is also dollar dominant. 

The second dominant currency is the Euro, yet its second place position is based on the flows across countries in the Eurozone and EMS. French trade with Spain in Euros will be booked as international currency transaction. As trade within Europe increases, the Euro grows in importance; however this may not translate to trade with other parts of the world. The yen and  Renminbi hold the third and fourth positions.

The Renminbi is likely to grow as an invoicing and payment currency based on the growing trade between China and many EM countries. This is also a place where the RMB can be used for trade financing. Given the export flow of commodities to China and the import flow of goods from China, there is a natural play for this trade to use Renminbi, yet even here, the invoicing and payment choice will be based on the particular company doing the business. A commodity firm may not want payment in RMB if they cannot convert to make their payments in local or another currency like dollar. 

The currency system is complex with network inertia causing friction against change. Will dollar hegemony decline? Yes, but this process is going to take some time. 

Sherman Kent and the problem between warner and warnee

A key part of the job of an analyst is to warn his clients about potential risks and danger. Warning also can come in the form of alerting investors about upside events.  Realize that there are two parties for any warning. There is the warner, the sender of the message, and the warnee, the receiver of the message. The form of any warning as to account for how it will be received. This problem was aptly described by Sherman Kent, the father of intelligence analysis for the CIA in his final comments. He states the problem simply, "Warning is like love - it takes two to make it."

The single central issue of warning is that it is a multi-step process which involves two parties: the Warner and the Warnee. Warning is not complete until: (1) The Warner warns (2) The Warnee hears, believes, and acts.

[There is] no warning if [the analysts’ assessment] (1) is not read; (2) is read but not believed; (3) believed but not really taken aboard.

The Warner tries to watch everything in the world and issues a warning when in his opinion the thing he sees coming up is: (1) Of considerable importance to the national security. (2) Highly likely (or likely) to take place. (3) The right time interval away - Not this afternoon. The analyst goofed—too late; Not next year—the analyst is too early. 

In examining a something [i.e., a prospective event] to tell how it meets these criteria ,you [the analysts] realize that you are judging, weighing, estimating. 

When making the next investment decision, think about the comments from Sherman Kent. There are two parties for every piece of analysis, the analyst and the investor.  What does the investor truly need to make a good decision? What is the form that the analysis should take? How can a narrative be developed that effectively communicates information on the likelihood, severity, and risk of an event. 

Perhaps being a quant saves us from this problem. Unfortunately, the forecast from analysis must be converted into a decision and that is not always easy to do.  The model provides a warning, but the PM as the warnee must use this output appropriately.

See other posts on Sherman Kent :

Sherman Kent - The analyst's analyst - Be precise with your forecasts

Sherman Kent - My guy for precision in forecasting language

Sherman Kent - the godfather of precision in forecasting language

"Likely" or "Probable" - there is a difference in language when talking about recommendations in the new year

Language, perception, and numbers - The translation problem

What does failed intelligence tell us about investing

Equity risk premium change with "good" and "bad" times

 

The macro market regime can be classified as risk tolerance, macro outlook, macro stability, and risk-on conditions. Each of these regimes can be either in a good or bad state. A set of macro state variables like the short rate, term spread, credit spread, dividend yield, effective spread, price impact and systematic volatility can be used to define these regime states.

The size of the equity risk premium will differ based on the market regime state. Conditioning on the market regime can create higher performing risk premium portfolios relative to any unconditional portfolio. See "Macroeconomic Risks in Equity Factor Investing: Part 2/2"

Factor risk premium can have low unconditional correlation but in reality have high sensitivity in different regimes. Momentum and high profitability are highly sensitive to the risk tolerance regime. These two risk premium have correlated regime sensitivity. What is surprising is that the size premium seems to be independent of macro regimes and hence will be a good diversifier in all states. Also surprising is the fact that the low volatility risk premium has high sensitivities to many market regimes. Low volatility is not a strategy for protecting a portfolio given this high regime sensitivities.

Understand your market regime and realize that equity risk premiums are regime sensitivity.  An unconditional correlation may not tell us about the true risk between equity risk premiums. 

Macroeconomic surprises impact equity risk premium

Macro risks impact equity factor investing. Equity factors (value, momentum, size, low risk, high profitability, and low investment) show cyclicality associated with some macro variables. Bundling risk factors together unconditionally will add value when forming a portfolio; however, accounting for macro risk surprises will provide further insights on the spread or sensitivity of these premium. Looking at short rates, the term premium, the credit premium, dividend yields, and market illiquidity can all provide better insights on the conditional movement in equity risk factors and are easily employed in any model of risk premium. See "Macroeconomic Risks in Equity Factor Investing". 

All the primary risk factors show a systematic relationship to one or more macro variables.  Short rates and the term premium seem to be the most important macro variables. Clearly, they represent changes in the economy and monetary policy.  Increases in short rates will negatively affect size, volatility, and the investment premium. Increases in the term premium will have a strong negative impact on momentum, profitability, and investments, but will have a positive impact on value. Default or credit spread will have a positive impact on profitability. 

The macro regime matters on equity factor risk premium. This has been known for some time. This papers quantifies the impact of macro surprises on various risk factors. If you don't account for where you are in the economic cycle, you factor exposures will harm your portfolio return. Of course, the problem is now determining what economic regime you are in or where you are headed. 

Skewness across asset classes - It can be exploited

 

There is a unique risk premium associated with skew and it exists to varying degrees across all asset classes, equities, bonds, commodities, and currencies. Creating rank weighted asset class portfolios that are long negative skew and short positive skew and bundled equally across all four asset classes can generate a portfolio that has a Sharpe of .72 over the period from 1990-2017.  The value of skew is shown in the paper, "Cross-Asset Skew" which creates a global skewness factor. 

The value of skew can be seen using several different statistical measures and is robust across different data sub-samples. The value of skew cannot be explained by other factors like momentum, carry or value. It is unique. Additionally, the skew risk premium seen in one asset class is not correlated with the skew in other asset classes.  Investors need to be compensated for the risk of negative skew and investors overpay for the lottery ticket embedded in a positive skewed asset. The combination of going long (short) skew and negative (positive) skew is pervasive Except for currencies across all asset classes. Even though asset classes may have most markets negatively skewed, the rank ordering shows the pervasive benefit from buying the lowest ranked skewed markets and selling the highest rank. 

Forming mean variance efficient multi-factor portfolios, the researchers find skewness is given a positive weight. Holding the skew factor is relevant for improving the efficient frontier especially at lower volatility. 

Different measures of skew which may have better intuition

Skew can be thought of as one side of the distribution being stretched to reflect the greater likelihood of events in either the left or right tail of the distribution. It is a distribution asymmetry. It is the third movement of the distribution and usually is measured as: 

However, the measure of skew as a difference cubed does not have immediate intuitive sense. You can measure it as either being positive or negative, and give it a level of intensity, but it does not have a good feel for most users. Investors will like positive skew, but it must be thought of in the context of price. How much do I have to pay for the skew I may like. 

Other alternatives to the traditional measure of skew that are easy to calculate for return include: 

Low moment skew which is the scaled difference between the sample mean and median.

Spread between the up and down semi-variances which is easy to calculate and has good intuition.

The high minus low measure which is the spread of absolute value of  the max and min scaled by the standard deviation.

All can be calculated on a rolling basis to provide insight on the changing skew of any asset.

Hegel and financial history - Don't Look for some linear process

We are facing another banking crisis. The same financial and banking mistakes are being repeated as institutions assume that the past will continue well into the future. Hence, banks bought long-dated assets with short-dated deposits and took the asset liability mismatch. Regulators fight the last battle and again miss the crisis that is facing them, blind to the obvious because coping is painful. Policymakers have the hubris to think they can forecast the future and have the tools to solve any problem. Again, they are being proved to be mistaken.  

Technology may be progressive, but the decisions of market participants do not follow a process that learns from the past. For Hegel the great 19th century philosopher, history is an intelligible process that moves toward a specific condition, design with purpose. The purpose is to have the freedom to pursue principles and moral law that look beyond personal desires. Nevertheless, he was pessimistic about whether anyone can learn from history. 

Hegel was not a financial historian. When we contrast his view with current financial behavior, there is little progress from learning and little linearity. Lessons are either forgotten or never learned, so there is constant cyclicality and repetition with financial markets.

This cyclicality may not be exploitable in the sense of being forecastable; however, it can be useful for seeing the financial world.  Assume that the behavior mistakes of the past will be repeated on both a micro and grand scale. 

On the micro level, assets prices are driven by behaviors biases. On a broader levels financial crises are driven by market excesses. Economies will go through business and credit cycles. World orders will change with dominance lost through a well-defined cyclical process. Accept cycles and end linear thinking and your investment decisions will be better.

Dedollarization - Rising talk but is it an issue?

 

Dedollarization is the current buzz with many investors given recent developments with the BRICS and their potential use of the Chinese renminbi as the currency for trade payments. Russia is adopting RMB invoicing for oil since the dollar is off limits. Brazil has used the RMB for some trade invoices given the large trade relations between the two countries. Saudi Arabia has started to invoice some oil exports in RMB.  Clearly, there has been some movement away from the dollar, but the reasons are mainly political and not based on the declining value of the dollar; nevertheless, regardless for the reason, some countries are choosing an alternative to the dollar. While a trickle, there is a growing view by non-US aligned countries that it may be in their interests to have new choices for invoicing, payments, and pricing.

We have heard about the dollar demise before. It is a recurring theme that has not played out in actual behavior.  Yet, if the US government and the Fed ignores the problem, the trickle can become  a deluge. 

Dollar debt still dominates the world's financial markets. Dollar reserves with central banks also are still dominate, and the desires for dollar swap lines are still strong when liquidity is needed. Figures are from Dedollarization is Not a Thing

Still, changes in trade invoicing and settlement are a canary in a coal mine signal that the trade and finance world may become multi-polar.  Trade invoicing and payment flows serve as a foundation for a switch in financial flows. 

Currency regimes through history - 100 years of dominance is about right

With all the talk about dedollarization, it is important to look at the history of past dominance of global reserves currencies. Generally, the dominance has lasted about 100 years and coincide with the dominance of an empire or country hegemony.  The end of a currency as a global reserve matches closely the loss of a war that bankrupts the reserve currency country. It certainly is not surprising to hear the talk about a dollar decline given the decline in US dominance as a world leader. The US is still dominant but there is a growing degree of polarization and country grouping. Of course, there has always been polarization. The Cold War formed a bipolar world between East and West with the Third World somewhere in the middle, but there was no competition to US and Western economic dominance. 

Of course, a country decline is not always abrupt. There is a period of transition, and it is during the transition period that long-term trends will emerge; nevertheless, these long-term trends are unlikely to lead to short-term profits. A long dollar decline will still see periods of dollar gains. 

The biggest impediment to a dollar decline is not from anything the US can or will do. The dollar is dominant by default. China runs large trade surpluses and is net saver, and it is unwilling to be convertible will all other countries. The RMB may be used for trade financing and payments, but that is not the same as serving as a store of value.

 

From Uncommon Cents