History Just Called - It's Time For Another Recession

|
Includes: BXUB, BXUC, CAPX, DHVW, DMRL, EPS, FTA, IVE, IVV, IVW, PPLC, RPG, RPV, RSP, RVRS, RYARX, SDS, SFLA, SH, SPDN, SPLX, SPUU, SPVU, SPXE, SPXL, SPXN, SPXS, SPXT, SPXU-OLD, SPXV, SPY, SPYG, SPYV, SSO, UPRO, USMC, VFINX, VOO, VOOG, VOOV
by: The Macro Investor
Summary

Below is a simple (and hopefully thought-provoking) analysis on anticipating the market impact and length of the next recession.

We are way overdue for a recession using historical analysis and near the all-time longest period between recessions.

The model indicates that the length of time between recessions has a unique correlation with the time of the recession that follows.

The model will later add additional parameters and data correlations to include: total U.S. market capitalization, Tobin's q, U.S. GDP, and U.S. government debt.

I highly suspect that updating the model will point toward a longer than expected recession, contrary to my conclusion drawn from the current model.

Editor's note: Seeking Alpha is proud to welcome The Macro Investor as a new contributor. It's easy to become a Seeking Alpha contributor and earn money for your best investment ideas. Active contributors also get free access to the SA PRO archive. Click here to find out more »

History is calling, and it says we are due for another recession. In fact, we are way overdue for a recession using historical data for our analysis. Starting in 1900, the average number of years between recessions is slightly under four. Post-WWII, the average is roughly five years. We are currently approaching nine years since the end of the Great Recession in June 2009. That is the second greatest number of years between recessions, with the longest being the dot-com bubble (10 years).

The goal of this post is to display a model for predicting how U.S. markets will react to the next recession, using the S&P as the example index. The model also conveniently provides us with information to predict on how harsh and long the next recession might be.

The model and charts shown are all driven off of historical data and simple correlations; I later discuss that the current model should be expanded upon so better analysis and conclusions can be drawn. This model explores how the S&P has reacted to U.S. recessions in the past, and what we can learn from this history.

Graph of CPI Adjusted S&P

Take a look at historical CPI adjusted prices of the S&P during each and every recession since 1901:

Historical Adjusted S&P Graphs like this are not rare or unique, although they are interesting to look at. Is it helpful in terms arriving at specific conclusions? I would say to a certain extent. To make the graph easier to draw conclusions from, I created a table displayed below.

I start with historical analysis using CPI adjusted highs and lows to see how the S&P reacts to different recessions.

S&P Performance
Looking at the table above, we still aren't able to derive fair amounts of valuable information from the holding period return. Recessions occur, and we see dips. This is where I decided we need to see correlations based on time between recessions, and the severity of the market corrections (or lack thereof).

Correlation Analysis

Let's now look at a table that includes the above information, but incorporates time and different correlations into the analysis:
Correlation This chart examines the following correlations:

  1. Length of time prior to recession versus length of recession
  2. Length of time prior to recession versus length of recession post-1960
  3. Length of time prior to recession versus percent change in the S&P
  4. Length of recession versus percent change in the S&P

Here are our conclusions from the correlation analysis:

We see a weak negative correlation in No. 1 above. This negative correlation in "length of time prior to recession versus the length of recession" is simple to understand from looking at the table. The "Time from last recession" column grows significantly with time, while the "Length of recession" column actually decreases slightly. Looking at only these two columns, this is an economically positive and impressive relationship.

I wanted to run the correlation in No. 2 to see how correlation No. 1 would change if we viewed a more recent economy. If you look at the number of recessions over the last 118 years, almost 75 percent occurred before 1960. However, even in No. 2 where we exclude a majority of the recessions, we get a similar value -- which discredits that the first half of the 20th century skewed our values about the notable ability for our economy to exhibit longer periods of growth and shorter periods of recessionary periods.

Simply put, longer periods between recessions don't necessarily lead to longer recessions. One explanation can be found in a paper published by Fed (although it uses U.S. state recessions for its analysis) titled "Countercyclical Policy and the Speed of Recovery After Recessions." Through proper government analysis and actions taken, "... the level of overall policy accommodation during ... recessionary periods do appear to help speed up recoveries." The more time the government has to prepare for a recession, the more it can do to help it recover.

In our third correlation, "Length of time prior to recession versus percent change in the S&P," we calculate what is effectively zero correlation. It is hard to draw conclusions here without looking deeper into how GDP fluctuated during times when a recession wasn't present. I plan to explore this more in depth in a later post.

In our fourth and final correlation, we see yet another negative and weak correlation between the length of a recession and percent change in the S&P. This correlation makes sense, in my opinion. In periods where it took longer for the U.S. economy to recover, you would assume stock prices suffered greater losses from their previous highs. Therefore, I accept the output value and deem it logical.

Final Thoughts

While this model accomplished deriving the high-level information I wanted to seek out, I do believe excluding GDP figures prevents the model from being anything more than a starting point for further research. I am going to be adding many more parameters to this model, if not changing it entirely, though the current model did allow some conclusions to be drawn about forecasting the impact and duration of a future recession. To some extent, this makes me more bullish on the overall future economic health of the U.S.

This is not to say I do not have some worries. How effective can the government be in speeding up the recovery of the inevitable upcoming recession when many of its effective market policies will be hindered by a multitude of factors? Just look at the current market for U.S. Treasuries and our federal deficit (not to mention the new 2018 budget deal).

Going forward with this model, I plan to explore many other historical correlations. Additional analysis using total U.S. market capitalization, Tobin's q, U.S. GDP, and U.S. debt levels will be my next step.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: I am long SPY and SPX Options