ECB historical policy review - An institution in two regimes

Happy 20th birthday the EMU and the ECB. Not clear whether we should treat this birthday as a celebration. It has gone through significant growing pains and a large personality change. An institutional review is presented in the ECB working paper, "A tale of two decades: the ECB's monetary policy at 20". This is a long piece, over 300 pages, and may be a little biased given it is an ECB working paper, but it lays-out the significant changes in the thinking and policy responses of the central bank.

It will pay for all investors to think about the ECB as being two different institutions or having shifted personalities. In the first decade it acted like a traditional central bank with a 2% inflation ceiling, (not a target). A second shorter period is when it moved to be an activist disinflation fighter. This is not the central bank first envisioned, but it is the central that the EMU has now. 

The question going forward is whether the second decade tale will continue or whether ECB President Lagarde will find a new third decade path. The activist ECB is the current bias, but it is not a long-term monetary strategy but an increasingly complex set of policies to find a solution to a problem that does not seem to go away. 

Investors will be hard pressed to make strong decisions without a clear policy direction and a mandate for what the ECB wants to achieve. This institution needs a clarifying strategy.   

Keep decisions simple when there is uncertainty - The case of coronavirus

"So, in general, if you are in an uncertain world, make it simple. If you are in a world that’s highly predictable, make it complex."

- Gerd Gigerenzer "Instinct Can Beat Analytical Thinking"

The challenge for decision-making and model building is determining whether you are in an uncertain world or a predictable work. Or, as Robin Hogarth has referred to as "kind" or "wicked" learning environments. (See "Kind" versus "Wicked" learning environment - Financial markets are not kind) In a wicked model, the link between between the past and future is unclear or uncertain. It is hard to learn and model in this situation. Your modeling and decision response should be based on the environment you face.

I have investing years trying to think through the problem of using heuristics versus analytics and avoiding behavioral biases and focusing on quantitative rationality. Someone who has not grappled with this issue has avoided the crux of making good decisions under uncertainty. There is a form of comfort with being a model builder. You can avoid the challenges of uncertainty by attempting to increase complexity. There is also comfort with having heuristics because they are fast and simple to implement. Moving between these two extremes may separate the average from the very good investor.

I bring this issue to the forefront because the current focus on the coronavirus is a real world case of dealing with high uncertainty. No one could spell or even cared about this issue a few weeks ago, but now it is driving the financial markets. Someone can derive models for infection rates and shocks to different economies from a potential pandemic. This is rational and makes sense, but this may be a better time for using good heuristics like "one good reason decision-making". 

The one good reason heuristic is simple. There is an unknown risk that can have a large financial impact and is causing prices to fall now; therefore, sell risk exposure and get to the sidelines. There is no big complex model, the trend is down, get out. The risks are not easily measurable and growing, get out. This is a good reason and don't wait for a deeper model answer. As new information is added, the decision can change, but the quick answer may be the best when the inputs for modeling are unclear.   

AQR expected returns - Don't expect much return

Long-term asset return forecasts help temper changes in asset allocation and portfolio return expectations. After a strong return year, there is often return extrapolation. This extrapolation is not usually one for one, but a good year will give investors the view that it will continue into the next year; however, a close look at long-term expectations suggests that 2019 returns are unlikely to be repeated.

The AQR 5-10-year expected returns show a decline over all asset classes from the prior year. For equities, the higher valuations point to lower future returns. For bonds, lower current yields, the best indicator of future returns, point to lower fixed income returns. Inflation expectations may be lower, but real yields are not expected to receive a boost. Credit spreads are tight and will not provide significant return gains for investors. Long-term average returns for commodities will stay low without the threat of a supply shock.

Traditional assets in any simple asset allocation will not generate real returns even near 4%. The only way to gain higher returns is through alternative investment which will subject investors to liquidity issues or different risk premia. The challenge is to find the alternatives that can produce higher real portfolio returns while adding diversification. It can be done, but it requires active decisions and greater monitoring.  

Sahm recession indicator - Nothing to see here now

Quantitative indices should provide a lot of information in a simple form that is easy to read. The Sahm Indicator or rule seems to present a nice picture on an economic slowdown that is easy to create and has gotten a lot of recent interest. It is just the 3-month moving average of unemployment minus the low of the last 12 months. If the number moves above .5, there is the expectation that there will a recession. 

It was created by a former Fed economist to look for times when fiscal stimulus may be necessary; however, the question is whether this is a better than simple picture and whether this indicator has predictive power. It is pre-calculated in the St Louis FRED database. Of course, the standard of whether it can predict a recession based NBER dating is a fool's errand since the NBER dating comes well after a recession begins. 

The Sahm Indicator can be compared with the 3-month moving average to show how translation of data can have a significant impact on visualization.

Looking at a difference can rough up the data at key points. A spike is very clear that labor markets are hitting a rough patch. Hat tip to Klement on Investing as to whether this has predictive power. It is not predictive as measured by classic NBER dating, but investors have to think in terms of a nowcast. 

Labor markets are hard to employ as a leading indicator. Nevertheless, recasting data can be a form of private information that can be exploited, so data transformation should be a regular tool used by any global macro quant analyst.