Real GDP still has mind of its own

 

We have discussed this before, but the LEI is still at extremely low levels. It may be bouncing off a low and did not reach the critical down 10% YoY, but it has been indicating a major slowdown. The real GDP on the other hand is moving higher. 

The hard landing story is likely over and there may not even be a soft landing. We have seen major divergences between the LEI and GDP in the past. There is no reason, given construction of an index and GDP, that these two should have a percentage connection, but large LEI down moves are associated with a recession. The 2001 period may be similar today but there was a NBER recession. 

Are real rates too high? Does not seem like it

 

10-year real rates have exploded from below zero to over 2%. Some are thinking that this level is too high and requires the Fed to lower rates. Relative to the post-GFC period, the current real rates are well above normal; however, we should place this real rate in context to the more normal period prior to the GFC. 

Of course, is there ever really a normal period? So, what is a normal real rate of interest? Is it close to zero or is it somewhere over 2% and closer to the long-term real rate of growth.  The post-GFC period was dominated by the QE which pushed rates lower, so with QE lowering every month, it is hard to say whether real rates are too high or just returning to normal. 

WEF risk assessment - Points for discussion

 

The WEF Global Risk Report is always a good read highlighting the perceptions by global leaders on what are the greatest risks in the short and long-run. It appears the risk assessments are ripped from the current headlines and are not based on deep thinking; nonetheless, their interconnection map and rankings provide a useful starting point for risk discussion. You may not agree with their risk assessments and presentation, but it does provide context of what is on the minds of major political, business, and thought leaders. 

The number one short-term issue is misinformation or disinformation although it is not exactly clear what that means. If you listen and act on misinformation, there will be clear risks when you find out that what you thought was true is untrue, but what will that lead to? How are economies and societies impacted by misinformation? Clearly, the misinformation of scientific facts are extremely dangerous but there are controls in place to limit risks. There is misinformation of what are verifiable truths, but in this case, the harm can be limited. Is the problem misinterpretation? I find this top risk odd and can be solved through a high degree of skepticism which is good. 

Extreme weather is also an interesting global risk since extremes are often localized. The real weather issue is systemic changes from the norm. Armed conflict only makes it into the top five although the impact can be large and immediate. Most of the biggest macro risks of the last 100 years have been armed conflict. This should always but a top pick. 

The long-term risks are all environmental which also seems odd. These risks are a variation on negative views embedded in Malthusian ideas. In ten years, is it possible that there can be technological changes that can mitigate these risks? Can there be adaptation to these risks? 

Surprising, economic issues of debt and inflation are not front and center in the list albeit in the top ten. Perhaps focused but economics are always a top risk. Solutions to environmental risks and misinformation are always driven by economic impacts and costs.

Purchasing Manager Index (PMI) is not what you think

 

The Purchasing Manager's Index (PMI) has often been considered one of the premier macroeconomic indices on the state of the economy, yet a more exhausting analysis suggests that there is not a strong link between the PMI and equity returns. This goes back to the key issue that the real economy is often not linked with equity returns. Recent work from Citibank reported in the FT discusses the problem. 

The general rule is that the economy is in contraction if the PMI is below 50 and in expansion if it is above 50 for this diffusion index. Further refinements can be made by looking at the 3-month change in the PMI, so there are four states: recovery when the PMI is below 50 and rising, expansion if the PMI is above 50 and rising, contraction if below 50 and falling, and slowdown when the PMI is above 50 but falling. 

The Citibank works suggests that sometimes the PMI indicator works, but it is also the case the equity markets lead PMI. This real indicator based on recent survey information with limited lags does not always provide any early sign on the return and risk in markets.

As usual, working with macro data is messy and there are no clear-cut rules that can be applied to the PMI data. It may tell you something about where you or where you may be going but it is not always a useful map. 

"Information is the resolution of uncertainty." Claude Shannon revised

I have always loved that quote from Claude Shannon. Shannon was the developer of entropy laws of communication which is critical to all phone and information networks. 

Better information should reduce uncertainty, but we can also think of the opposite. Poor information can lead to uncertainty. Misinformation can create uncertainty. Poor interpretation of information can lead to wrong action and results. Information that is revised crates uncertainty. Poorly defined information crates 

Uncertainty can come in many forms based on the type of information that is acquired and used. The problem with information is that more is not always better, and not all information is created equal or of equal value to users. Information is the wrong hands or handled poorly with add uncertainty. Start with the right information and you can solve most problems. Start with the wrong information and you only make matters worse. Deeply think about the inputs (information) before taking any action.

The choice of machine learning model - LSTM vs RNN

 

source: Analyticsvidhya.com

The choice of which ML model to use for financial time series is important and should not be taken lightly. The LTSM or Long Short-term Memory model has become a workhorse for many quant financial analysts. The LTSM model is a type of recurrent neural network that can account for long-term dependencies by gates to control inputs, outputs, and memory (a forget gate).  

RNNs (recurrent neural networks) are useful for processing sequential data such as time series because they can process sequential data both backwards and forwards. The RNN can use context and dependencies between time steps which is critical for financial data which may have some form of autocorrelation.  LSTM is form or type of RNN which uses a memory cell connected to gate which can be activated when long-term information is a useful input. The simplest case of a memory gate is a GRU, or gated recurrent unit which can help improve predictions. 

All these memory gates attempt to address what is referred to as the "vanishing gradient" problem, when the gradient of the weights in the model become small and there is limited learning or improvement. By pulling in or activating long-term memory, there is an attempt to boosted model performance through steepening the gradient for weight changes. Using longer-term data and not just current input can help improve forecasting or prediction context.

Macro trading and the jobs revision problem

 

This simple table for 2023 shows the original jobs report estimate and then the revised number. What is clear is that for eleven of twelve months, the original estimate was revised down. The job report was not as rosy as first expected. On an absolute number, the jobs data for 2023 look good. There has been a slight decline if you follow a six-month moving average which is expected given the increases in rates; however, if you want to trade the headline or the surprise you will be caught in a problem of adjustments for revision in the past. This month may look good but last month is revised down, the combination of the last two months may generate a different picture than what was expected. These effects have to be accounted for with any back-test using jobs data.

MInsky and Kindleberger - Kindred spirits concerning bubbles and credit instability

 

Charles Kindleberger set the tone for any discussion of bubbles in his book, Manias, Panics, and Crashes, through his deep narrative and useful framework which is similar to the pioneering work of Minsky. Normally, these two are separated as different thinker about bubbles, yet there are closer similarities than many may suspect, and many refer to their general explanation of bubbles as the Minksy-Kindleberger model although there is no formal mathematical approach.

This alignment of Minksy and Kindleberger is well described in a new article by Perry Mehrling called, "The Minsky-Kindleberger Connection and the Making of Manias, Panics, and Crashes". These two are kindred spirits and Kindleberger may have been influenced by the early work of Minsky on the instability of credit. He may not have been a believer of Minksy's views on the business cycle, but they are economic brothers with respect to their view that credit can be unstable those create extremes that can generate crashes. The inherent instability of credit requires superior monetary institutions that can control credit and serve as the lender of last resort if there is a crisis after a bubble.

Bubbles are all about credit extremes and Kindleberger used his strong knowledge of economics history and institutions to provide extensive analysis on how panics and bubbles may take root and create the vexing problems that continue into the 21st century.

Both economists come out of a pre-WWII institutionalist background which was forsaken in the post-WWII move to high theory in economics. Institutions matter and the credit frameworks that are created in the modern economy can lead to financial instability that is not directly modeled in most presentations of monetary economics.

The Cantillon Effect - What is driving survey differences?

Why are so many consumers unhappy in the current economic environment? We can think about the Cantillon Effect, names after the 18th century French economist. If there is new money in the system that can create inflation, it may first impact the rich who can increase their wealth. There may not be a general rise in prices as usually taught, albeit all prices may be increasing, but there are relative price changes which will affect different households differently. Because inflation can be localized and can be gradual, different groups will respond and be impacted by a shock to money that can lead to inflation. For example, increases food prices will have less impact on rich households because food is a smaller portion of their consumption basket.

Consumers who do not have wealth or do not have the knowledge or the capability to exploit increases in money may not be able to adjust or adapt to higher inflation pressures created from those who were able to exploit greater money earlier. These poorer households will be more impacted by the increase in goods without the ability to exploit the money increases. 

Hence, there is a distributional effect from inflation that is often not avoided in the inflation discussion. The inflation shock over the last two years has had a disproportional effect on lower income household who are not able to generate a wealth effect or are not able to take advantage of a monetary shock. 

The wealthy are feeling good about the economy. Poorer households who are unable to protect themselves from inflation may have a different view. 

The magnificent seven versus all the rest - what will drive 2024 returns?

 

2023 was all about the "magnificent seven", those highflyer stocks that dominate the SPX index (AAPL, AMZN, GOOG GOOGL, META, MSFT, NVDA, and TSLA). The ratio of the mag 7 to the Pax moved from about 2 at the beginning of the year to a high of 3.5 at the end of November to the current level of 3.37. Can this continue? Increasing valuation will be difficult but we must look at what is happening to the rest of the SPX. The equal-weighted index returned about half the market cap weighted SPX. If we have a market correction, it is not clear that the mag 7 will be the driver or the place to hide. 

The SPX benchmark will be driven by the mag 7, so any view on what will happen to equities should include a specific view on this large cap subindex. 

There is no certainty for 2024

 

“Certainty comes from believing we have learned all there is to know. Confidence comes from the effort to learn all we can."  - Madeline Albright Always good to think about the certainty issue as we make predictions for 2024. We are not good predictors and we do not have certainty. Those that feel they have certainty are foolish. There is still a lot to learn about what may happen in 2024, and we will be surprised perhaps even in January.  We can build our confidence, but research has shown that just because we have more information or facts does not mean that more confidence is warranted. Confidence is not always tied to more knowledge; however, less knowledge is a clear sign that we should be less confident.  So let's start 2024 with healthy skepticism.