Calculated Risk: Predicting the Next Recession

In that 2013 post, I wrote:

The next recession is likely to be caused by one of the following factors (from least likely to most likely):

3) An exogenous event such as a pandemic, a significant military conflict, a power outage for any reason, a major natural disaster (meteorite fall, supervolcano, etc.) and a number of other unlikely causes. All of these events are possible, but they are unpredictable, and the likelihood that they will occur in the next few years or even decades is small.
added accent

Unfortunately, one of the unlikely events happened in 2020 (pandemic), and this led to a recession in 2020.

2) Significant policy error. Two examples: failing to reach a financial agreement and crossing the “fiscal cliff” would likely lead to a recession, and Congress’ refusal to “pay the bills” would be a political mistake that would send the economy into recession.

We have seen several policy mistakes, mostly related to immigration and trade during the previous administration, but none that would send the economy into recession.

1) Most post-war recessions were caused by Fed tightens monetary policy to slow inflation. I think this is the most likely cause of the next recession. Usually, when inflation starts to become a concern, the Fed tries to plan for a soft landing, and often the result is a recession.

As well as this most common cause of recession is a current problem. As inflation rose, mainly due to the pandemic (stimulus spending, supply restrictions) and the invasion of Ukraine, the Fed embarked on a tightening cycle to slow inflation down.

The Fed cannot ease the pandemic-related supply restrictions (other than curbing demand), and the Fed cannot stop the war. Thus, there is a possibility that the Fed will tighten policy too much, resulting in a “hard landing” (also known as a recession).

The key will be watching housing. Housing is the main transmission mechanism of the Fed’s policy. We have already seen how mortgage rates have risen so much that slow withdrawal of mortgage capital and slow down housing activity.

One of my favorite business cycle forecasting models uses new home sales (also housing construction and home investment). I also look at the yield curve, but I have found that selling a new home is generally more rewarding. (See my post in 2019: Don’t worry about the yield curve)

As for the economy, what I’m focused on is starting one family and selling a new house. For the lows and highs of key housing construction activity, a graph of single-family housing additions, new home sales, and housing investment (RI) as a percentage of GDP is provided.

Note. The pandemic has skewed the economic data and, as I have repeatedly pointed out, we cannot be slaves to any model.

Startups, sale of new homes, investment in housing Click on the graph to enlarge the image.

Arrows point to some of the earlier highs and lows for these three measures.

The purpose of this chart is to show that these three indicators usually peak and bottom together. Please note that residential real estate investments are quarterly and new home sales are monthly.

Sales of new homes and single-family homes have declined, but this has been partly due to the huge increase in sales during the pandemic. Housing investment may peak.

Annual Change New Home SalesThe second graph shows the annual change in new home sales, according to the Census Bureau. Currently, sales of new homes (average for 3 months) are down 10% compared to last year.

Note. New home sales data smoothed using three month average before calculating the annualized change. Census Bureau data begins in 1963.

Some observations:

1) When the annual change in new home sales falls by about 20%, a recession usually follows. The exception to this data series was the mid-1960s, when increased investment in Vietnam saved the economy from recession. Other the exception was the recent situation We saw a pandemic-related decline in new home sales year on year and a spike in new home sales in the second half of 2020. Ignore this pandemic distortion!

Also note that the sharp decline in 2010 was due to the 2009 Housing Tax Credit policy and was simply a continuation of the housing crisis.

2) It is also interesting to look at the periods of 86/87 and the mid-90s. Sales of new homes fell in both these periods, although not by 20%. As I have noted in previous posts, the mid-80s saw a surge in defense and MEW spending that more than offset the decline in new home sales. In the mid-1990s, investment in non-residential premises remained strong.

If the Fed’s tightening cycle leads to a recession, we should see housing go first (new home sales, single-family construction, housing investment). It can happen, but usually it’s ahead of the economy by a year or more. Thus, we can expect a recession in 2023.

There are other indicators such as the yield curve and sale of heavy trucks – but mostly I’ll be looking at housing. I don’t follow the recession at the moment.