Failures With Information Usage - Competing Models and a Solution

Investors do not always use all the information that is available to them; however this is a not a unique problem to finance but an issue that runs the gamut for all consumer decisions. The explanations for the problem of information usage or non-usage has fallen into two major camps or models of behavior and described nicely in a recent Journal of Economic Perspective article, "Frictions or Mental Gaps: What's Behind the Information We (Don't) Use and When Do We Care?" by Benjamin Handel and Joshua Schwartzstein. We present their framework with our view on how the problem can be solved.

One view is that the lack of information usage is associated with market frictions, transaction costs. There is a cost of gathering and processing information so it is not done. This seems unlikely for sophisticated investors, but there is evidence that even simple differences in the costs of index funds are not fully explored by investors. Of course, the cost of processing information may be much higher than the cost of gathering information. If there is a friction, it is with effectively weighting all the information available.

The alternative view is the mental gaps school of thinking which focuses on behavior biases. Investors gather all of the information but they do not use it effectively. It is the poor processing  because of mental heuristics which could be the problem with information usage.

A little of both models can explain decision-making problems in finance. The question is whether there is an effective way to avoid frictions and mental gaps. The problem could be solved through the issue of disciplined systematic decision-making. 

For the friction explanation, disciplined investment decisions can use all of the information available and can increase processing through the use of quantitative models. There can be errors with the models, but their effectiveness can be measured and any errors can be adjusted. 

For the mental gap explanation, disciplined decision-making can eliminate the problems identified in behavioral finance. Whether anchoring or recency biases, a model or rules-based system can eliminate some obvious behavioral biases. 

Markets will always react to surprises in new information, but investors should not be disadvantaged through not using information that is already in the marketplace. Costs can be minimized through effective information gathering and processing which can be done through quantitative tools. Mental gaps can be closed by using rules to hardwire good behavior. There is no reason why information inefficiencies cannot be closed. 

Portfolio trust in mean reversion not momentum - The contradiction of investing in trend-followers

Most trend-follower will say that they are "non-predictive".  While I think this is true in the sense they do not form forecasts or expectations, trend-following is also based on the prediction that the price direction through some set of price weighting from yesterdays and today will continue into tomorrow. Trend-followers do not try and forecast expected returns rather they extract signals from past data under the assumption that price moves will have memory of at least direction. 

These managers find trends across a large diverse set of markets and then invest long or short based on these trends. If the markets are moving higher, they are a buyers, and if price move lower they are sellers. Buy high under the assumption that prices will move higher and sell low under the view that prices will move lower. 

The relative value of any trend manager comes from their ability to better extract a signal from the noise versus their peers. Nevertheless, the question for many investors is how to predict when this strategy as a whole will make money which may form a contradiction within the trend strategy. 

One investment approach is to not make any prediction. The trend-following properties of diversification should lead to a constant allocation. Some may take a more active approach and invest in a performance trend. Others will take the opposite approach. The best time to buy is when returns are low and the best time to sell is when performance is high. Be a trend-following contrarian. There is evidence that this approach works.

The contrarian view suggests that trends in any one direction do not last forever; consequently, there will come time when strong performance will be reversed or at least the return to risk will turn down. Similarly, there will periods when poor performance will be reversed because risk-taking is reduced or the period of no trends is finished. Investing in trend-followers through thinking like a mean-reverter while paradoxical does work.

Stock-Bond Correlation - An inflation regime change will push it higher

A recurring global macro theme has been how investors should think about stock bond correlation. The negative correlation between stocks and bonds has been the single best diversification provider for any portfolio. There are very few alternative investments that have offered the same amount of diversification and provided a significant amount of alpha. This simple diversification is why variations on the 60/40 stock/bond portfolio mix have been such winners since the Financial Crisis. But times change, or more specifically, regimes change. 

A UBS Asset Management piece "Investment Troubles" suggests that the stock/bond correlation is loosely tied to inflation regimes. The negative correlation of today has not been a given through history and an often overlooked fact is that the negative stock/bond correlation seen in the US is not a given in other countries. For example, the US stock/bond correlations in the 70's, 80's and 90's were positive.  Canada has seen positive correlation between stocks and bonds during the entire period when the US correlation has been negative. This asset class correlation is dynamic and situational.

A long-term look at the US stock/bond correlation shows that there have been long period of both positive and negative correlation. These dynamics exist even though a good working hypothesis is that this correlation should be positive. A change in short-term interest rates should impact the valuations for stocks and bonds in the same direction through present valuing of cash flows. In reality, the key issue affecting correlation is the level and volatility of inflation.

When inflation is high, the short-term discounting factor which includes expected inflation seems to dominate the stock/bond correlation. When inflation is low and thus short rates are low, other factors such as economic growth seem to dominate and growth will have an negatively impact on the stock/bond correlation. A further review suggests that these inflation regimes will be associated with monetary policy.

The monetary regimes will respond to inflation and create a   different correlation environment as measured by the Taylor Rule.

Extrapolating this historical information suggests that the combination of restrictive monetary policy in a higher volatility environment will tilt diversification risks to higher correlation. It is unlikely that there will be a significant inflation risk shift in the near-term, but our priors suggests that investors will not continue to receive the diversification tailwind that was the great portfolio risk reducer in the post Financial Crisis period.


Some past posts on the issue:

The stock-bond correlation curve - risks from the Fed?

Yes, the stock-bond correlation is rising, but let's have some perspective

What do you really get from bond diversification in the 60/40 blend?

Normalcy for stock bond correlation says you need other forms of diversification

The problem with a 60/40 asset allocation - correlation

Inflation Protection Trade - TIPS Are Not Enough

Inflation is rising and is likely to be centered around 2% in the current environment. This market view may be a good null hypothesis for what many investors believe they will face in the coming year. Consequently, investors have shown renewed interest in TIPS (Treasury inflation protected securities), but we have concerns that TIPS may not provide enough return to help many pensions. Alternative investments that have  return advantages should be carefully reviewed. There are some simple disadvantages with TIPS.

First, many pensions have discount rates that are still around 7 percent, so a CPI-linked investment is not going to help gain ground versus the expected discount rate. Pension funding gaps only grow if portfolio returns are less than the discount rate. Pensions may be concerned about inflation, but on a relative basis, the matching or exceeding the discount rate is a more critical problem.

Second, the inflation that is faced by pensioners is higher than what is being stated in the CPI. The CPI-E (experimental CPI weighted for seniors) is increasing at a slightly higher rate given the mix in their consumption basket. TIPS may offer inflation protection but not for the inflation that is being faced by pensioners. 

Third, many state pension funds have COLA provisions that are often set at a rate that is higher than the CPI. In the case of many states, it could be at 3%. Hence, if there is not a TIPS premium, a TIPS return will underperform the COLA provision if COLA is higher than CPI. Additionally, any TIPS premium is present to deal with the risk with the TIPS and not as a means of supporting higher COLAs.

Unfortunately, the use of long-only commodity indices has not helped pensions because these indices have been in an extended drawdown since the financial crisis. Using commodities has not helped pensions but actually caused a performance drag.

The solution to inflation protection is to think outside the immediate inflation securities box. Three alternatives come to mind, real estate, systematic trading, and commodity risk premium portfolios. Each offers a slightly different approach to providing inflation protection. 

Real estate is a classic inflation hedge, but there are issues of liquidity and current valuations at this point in the business cycle. Leases should increase with inflation but investors will lock-in funds with limited flexibility. 

Systematic investment strategies that could include risk parity that explicitly has commodity exposure or managed futures that actively trade commodities is another alternative. Investors will have significant liquidity. There also are implicit adjustments across asset classes through active management. Dynamic asset class adjustments could provide inflation protection as allocations are changed during an inflation shock.

A relatively new strategy would be to invest in a commodity portfolio that is based on commodity risk premiums. Instead of investing in a long-only basket with fixed commodity weights, investors would build a commodity portfolio based on well-defined risk premiums such as carry (backwardation/contango), momentum (trend), value, and volatility. This portfolio will be uncorrelated with core traditional assets and should be positioned to take advantage of inflation increases.

Thinking outside the TIPS box may allow pensions to have inflation protection but also receive a return that will not be a drag relative to the expected discount factor. This is a win in almost any environment.

Old school "natural rate" relationship is a problem for macro managers

"Oh my god, the unemployment rate is below the natural rates, sell your bonds!"

This has been the usual bond market reaction for the last few decades, but the world has changed. There may be good reasons to sell bonds, but a low unemployment rate may not be the main driver. Investors need to kill off the idea of the natural rate of unemployment or at least modify their views. That does not mean that bonds will not react to low unemployment rates, but the current sensitivity is significantly different than what we have seen in the past. 

As stated by Keynes, "Practical men, who believe themselves to be quite exempt from any intellectual influences, are usually slaves of some defunct economist." 

All of macro investing has been embedded with the teaching of the natural rate hypothesis. See "Should we reject the natural rate hypothesis?" Olivier Blanchard Journal of Economic Perspectives Winter 2018. Any manager who is thinking about the inflation and interest rate problem is likely using the framework taught them in their introductory macroeconomics class. Simply put, there is a natural rate of employment which clears the market and that monetary policy cannot sustain employment below this level without incurring higher inflation. It is assumed that the natural rate can be identified and the pushing on the equilibrium level will lead to inflation acceleration.

Most of the hand wringing by bond traders is based on the idea that as we push unemployment lower we are causing more inflation pressure to build. The empirical evidence suggests that old theories are just not as robust in the current environment. The sensitivity between inflation and unemployment is lower and the expectation sensitivity of consumers and forecasters is different than in the past. See the statistics from Blanchard.

The natural rate can be a good null hypothesis, but investors have to keep an open mind and put some weight on alternatives factors or theories such as labor participation and output gaps to gauge current lack of inflation sensitivity to the natural rate. The current lack of sensitivity has harmed traders who have been grounded in past thinking. The big bond trade will come if the sensitivity of inflationary expectations changes back to the 90's; however, this type of change is unlikely to occur quickly. 

Libor-OIS, credit spreads and the change in the credit cycle

If you want to understand overall credit spreads you have to have both a macro and a micro view. The macro view looks at the business cycle and the chance of default for risky assets based on economic growth. The micro view looks at the credit supply coming to market, the demand for loanable funds at any time, and the structure of deals. The macro focuses on credit risk expectations and the micro will be more centered on the flow of funds. A macro-micro framework helps focus our interest in actual and perceived credit dislocations.

There is growing interest is the strong widening in the LIBOR-OIS spreads as measure of short-term financial credit risk. Let be clear that the dynamics of LIBOR are complex; a concentrated dealer community, a disperse set of global users, a multitude of strategies and uses, and a history of manipulation. There is no simple explanation for spread changes and a spread dislocation may just as easily represent a structural change as a macro expectation change. That said, LIBOR-OIS spreads have widened along with short-term corporate spreads and the TED spread. Credit risks are being repriced. 

We believe that macro credit risk should be repriced higher given higher equity and bond volatility and a potential slowdown in earnings later this year, but the current spike is more related to structural issues. However, the spread widening may be unrelated to bank risk. Financial conditions have tightened, but the absolute level is not suggestive of high spreads. 

What is an issue is the global flow of funds. The repatriation of dollars held offshore, an increase in Treasury bill supply and a tightening of dollar swap lines are all impacting short rates on the margin. Global credit structures matter.

Markets are adjusting under what we call the Fed 's new period of quantitative tightening (QT). What is the most powerful effect may be the shortage of dollar funding for global financial transaction. For investors who have following the research and commentary of the Bank of International Settlements (BIS), the demand for dollar funding is an ongoing structural problem given the amount of dollar denominated debt. This is not going away and a shortage of dollar will force spreads higher.

Forecasts are often biased - So do it yourself and be held accountable

Forecasts may tell you a great deal about the forecaster; they tell you nothing about the future. - Warren Buffett

There are forecast biases from any opinions sold on the street. Investors should expect biases and you are getting the opinions of an individual or group of individuals which are not fully accountable for the view they sell.

Some of these biases are from herding and trying to be close to the market consensus. No forecaster wants to be an outlier and have a career ender. Some biases are from the strategic behavior of the forecaster who is trying to either preserve or grow a reputation. It could also be caused by optimism or pessimism by the forecaster or irrationality, the lack of using all information available.  These biases can lead to asymmetric loss behavior which is hard to detect but has real effects on the users of the forecast.

The critical issue is whether the forecasters have "skin in the game", the current theme of Nassim Taleb. The best way for a manager or investor to have skin in the game is to develop their own forecasts and be held accountable for what the objective of the forecast. 

There are limited excuses for failure if the decision-maker is also the forecaster. The question is clear - did my forecast make money? An advantage of systematic (quant) fund managers is that they place their forecasting skill directly in the their program. The inputs of any forecast are perfectly known. There is a direct link between forecast quality and dollars made. Even trend-followers, which try to exploit price direction, are held accountable for their "forecasts" through the decisions and profits made. Forecasting bias may not be eliminated but the forecasts is held accountable. 

Is cash no longer trash?

There is a new meme going around that cash is no longer trash - the rising short rates make cash a real alternative to a over-valued equities and bonds. Cash is now a viable alternative with rate moving higher. The 3-month bill rate has moved up to 1.75 percent from .16 bps at the end of 2015. The impact on money funds is significant even if inflation is close to 2%. There are good alternative but they will have to compete against cash for investor attention.

• An alternative to bonds - The flatter curve makes cash not trash. The differential between the 10-year and 1-year Treasury is inside 100 bps. There is less term premium with holding longer duration.
• An alternative to dividend yield - The higher cash yields make cash not trash. The cash yield at one year is similar to the dividend yield (1.80%) on the stock market (SPY). The current earnings yield is 3.95, higher than one-year Treasuries but at a lower spread than over the last few years. 
• An alternative to international bonds - Positive yields make cash not trash. Euro area yields for one year are still negative at -.68%.
• Overvalued bonds and equities make cash not trash. There is still downside risk as measured by tail exposures in options for both stocks and bonds. CAPE P/E ratios are high relative to history and bond risk premiums are still close to zero.

Inflation makes cash still trash. Given the US inflation rate is close to 2%, the real rate is still at best zero to slightly negative. These are better than recent numbers but nothing that should get investors happy.

Hedge funds can make cash trash. In the current environment of falling correlation within stock indices, higher volatility, and more market dislocations, there is the opportunity for greater hedge fund returns. For CTAs, which hold most non-margin money in cash, rising rates are a nice tailwind, albeit rising rates has not always been a CTA friend.

Unfortunately, rising cash rates also means that credit is tightening. Corporate spreads are widening. This does not immediately translate to higher risk, but the credit environment is not changing for the better. 

Nevertheless, cash is now a viable investment option to marginal investment strategies.

Quants and uncertainty - Post-February thinking

As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality.  -Albert Einstein

Many quant firms took a return hit with February performance. It was not just for trend-following but also for many risk premium programs. Short-term traders who were able to actively trade the volatility shock were an exception. 

For most firms, these poor returns were unexpected. First, the programs may not have anticipated the price patterns associated with this volatility shock because they just do not occur frequently. This is a sampling and testing problem. Second, models often make risk management assumptions that are not grounded in reality. Whether liquidity shocks, fund failures, odd behavior by market participants or price moves after the close, models are often not built for all of the structural events that may occur. Models are based on simplifying assumptions and not formed to account for all of the realities of trading. 

Given the slippage between modeling and reality, firms have three options for program protection and it is important to find out how managers approach the model reality error problem. 

One option is the scenario approach to reality. Managers should stress a portfolio based on events that are possible but not actually found in the empirical data. Be ready for what has not yet occurred. 

Second, managers should use a precautionary principle to portfolio construction. Build a portfolio that can survive any extreme event. A precautionary principal can take the form of a barbell between low risk and high risk portfolios. The high risk portfolio can potentially take a devastating hit, but still leave the overall portfolio able to survive to some predetermined loss. 

Finally, a core principle can be to never lever a portfolio beyond a set level especially in a low volatility world. A key problem was with the target volatility crowd is that they continued to add positions to maintain volatility targets in a low volatility environment. A max leverage principle would have capped the leverage regard of target volatility levels. Volatility may stay below the target level but leverage caps are more important.

Working to account for extremes that may never be seem may seem cautious and clearly leave money on the table if the risk event never occurs, but it is the only way to protect against the slippage between models and reality that may occur in a February shock.

Expressing trend bets through futures or option - The problem of more moving parts

The choice between using futures versus employing options for a trend-following program is worth reviewing after the recent market events. Would trend-followers who used options have done better than those who expressed their directional bets with futures? The key to this answer is whether the trend-follower had a view on volatility.

Most trend-following is done through futures and not options, so there must be clear reasons for this preference. The simple reason for this focus on futures is that there are less moving parts. 

Most trend managers have not been able to include in their models all of the choices represented in options. For futures, the choice is simple. Find the direction. You don't have to forecast the size of the move or when it exactly  will occur. All you have to do is get the sign right. 

If options are used to express the trend view, the forecasting decision is much more complicated. You have to get the sign right just as with the futures trade. Additionally, you have to forecast the size of the move and express this view through picking a strike. You have to get the timing of the price move right because there is time decay (theta), and you need to have a view on volatility (vega).

A general rule of thumb is that most trend-followers will have winners only about a third of the time, another third are scratch trades, and the final third are losers. The size of the winners will offset the other 2/3rds of the trades that either are flat or have a loss. There is a lot of room for error, yet still an opportunity for profit if the winners are associated with big moves. 

There is less room for forecasting error with options given the choices that have to be made. This is especially true when futures positions are structured to act like an option through combine the trend decision with a stop. The stop-loss rule creates an option pay-off without the payment of a premium. These created synthetic options do not have time decay. They will still be affected by volatility but the creator does not have to pay for the volatility through a premium. 

The February volatility spike would have been costly for any wrong bet because the reversal of option positions would be more expensive given the embedded cost of volatility in the option price. Consequently options as a means of expressing trends hold only be undertaken if the trader has deeper view on the strength of trends and the market environment.

Buying commodities a play on China growth as well as inflation

We have previously posted the relationship between global growth and commodity returns as measured by the leading index (BCOM). Global growth above 3% is a good tailwind for overall commodity demand that will push prices higher. 

If we drive down one level deeper, we can see the strong link between China growth and commodities. A slowdown in China growth will provide a headwind that will limit any broad increase in commodity prices. 

The importance of China to commodity markets cannot be understated since its percentage of global demand for many commodities is so high. This strong demand exists not only for metals but also foodstuffs and is out-sized relative to their global GDP and population. This demand has been strengthened by the huge debt increases that have fueled   China growth; consequently, commodity returns are more complex than just a play on inflation. 

There are a number of factors that should be considered for any commodity exposure, global including China and EM growth and global inflation. Commodities can be used as an alternative way to gain growth factor exposures without taking the valuation risks currently found in stocks and bonds around the globe.

Commodity driven by global growth - Above 3% growth and prices push higher

Commodities are an effective way to hedge against inflation, but it can also be viewed as a simple way to play the global growth story. The average annual total return for the Bloomberg BCOM index (formerly the DJUBS commodity index) since 1992 has been 1.37% but if we look at returns when global growth is above 3%, the average annual return was 10.91%. 

The poor performance during the post Financial Crisis was mostly during an extended period of sub-3% growth between 2012 and 2016. The poor commodity returns were also closely tied to the slowdown in China growth. The strong declines in commodity prices match the transition from    9% growth in pre-2012 to the 7% growth from 2012 to 2016.  Growth matters especially for the oil and metal cyclical commodities.

A sustained increase in global growth above 3% will provide a strong tailwind for any broad-based commodity investing and provide an alternative to over-valued equities or bonds.

Looking back over commodities for the last ten to fifteen years - Currently in normalization phase

Commodities have not behaved like other asset classes since the Financial Crisis. In spite of all the monetary easing and the lowering of interest rates, commodities have marched to their own drummer, or, put differently, the commodity cycle has been independent of the business, financial, or monetary cycles. It is because of this difference that commodity exposure may look more attractive than traditional assets.

To place in context, let's look at the major trends in the two leading commodity indices the SPGSCI and the BCOM (Bloomberg Commodity index formerly the Dow-Jones UBS index or DJP). The major difference between these indices is the higher energy exposure in the SPGSCI. 

If we had to breakdown the behavior of the indices since the Financial Crisis we see four or five major trends. First, there is the pre-Crisis super-cycle which peaked in mid-2008 before the worst of the Financial Crisis. The steep decline was consistent with the economic slowdown that was already occurring before September 2008. The second leg of the major commodity trends for the last 15 years was the financial crisis. All risky assets correlated to one as there was a single factor that was driving the market declines. The third part was the initial recovery after the market bottom in March of 2009 The pre-crisis excess, Financial Crisis fall, and partial recovery are consistent with the behavior of financial assets. 

The next phase of the price moves separates commodity indices from financial assets. The combination of the European debt crisis, the decline of China growth from 12% in 2010 to 7% in 2015, and the ongoing below trend growth in the rest of the world curtailed demand in a commodity world of excess capacity. Commodities were more linked to economic activity than financial assets which responded to excess central bank liquidity. 

Our current stage in commodity trends is recovery and stabilization. Energy prices have moved off lows and metals prices have moved with increased economic growth. The global economy has now both stabilized and moved to a new footing of growth which has allowed for a new increase in commodity prices. Call this "normalization" but commodity prices have moved beyond the super-cycle excesses. 

Correlation up - a beta not alpha world for now

We follow the dispersion, volatility and correlation indices generated by Standard & Poor's which show intra-index stock behavior over time. When the correlation across stocks within in an index is high, there is a clear sign that the market is facing a macro shock. Performance should be biased toward beta risks. Fund returns will be driven by their beta exposure and timing skill not by their stock-picking skill. When correlation is low across stocks within an index, we can say it is a stock picker's environment because skill-based traders will be rewarded for exploiting differentiation across firms. This lower correlation will generate more dispersion in returns if there is additional volatility.

Given the volatility shock, February was a beta-timing environment. Those managers with higher beta exposure suffered. Those with lower beta exposure had more muted returns. Those managers who adjusted quickly to the environment or did not get whipsawed showed better performance. 

The last year showed a longer period of low correlation which should have been better for stock-pickers, but the low volatility kept return dispersions relatively stable so opportunities were limited or not significantly different from earlier periods of higher volatility tied with higher correlation. The Goldilocks conditions for stock-pickers would be a combination of low correlation with higher volatility. Investors await those conditions.

Commodities - A good time to hold or increase at this point in the business cycle

Where are we in the business cycle? This is never an easy question to answer but we might get agreement to the fact that we are in a new expansion in later stage business cycle. We are in a new expansion since we have seen surprise upside in growth around the globe in the last year, but we also are in the later stages of the business cycle given the time since the last recession and the fact that central banks are leaning toward tightening. Time by itself is not a good indicator of a business cycle reversal, but as the business cycle lengthens there is greater likelihood for price and inventory tightening. 

This point in the economic cycle is usually a good time for holding commodities. Coupled with the large dispersion between stocks and commodity price levels, commodities are also attractive relative to financial assets.

There is additional value with holding commodities through futures given the positive roll characteristics with the largest market component of most commodity indices - crude oil. The reason for holding commodities is only enhanced if more markets are in backwardation. In fact, the research, which has focused on the advantages of holding a commodity index, shows that the gain from backwardation or commodity carry is one of the key drivers for futures index returns.

Looking back in history since the Financial Crisis would suggest that holding diversified commodity would be a bad bet, but the critical thinking is to look beyond any poor performance over a three or five year period and focus on the opportunity going forward. A forward-looking analysis coupled with expectations that inflation may continue to rise makes holding commodities attractive as an opportunity for  diversification and return enhancement.