Macro Research
SUMMARY
Building on Edward Leamer’s 2007 NBER Working Paper “Housing Is The Business Cycle,” we examine the relationship of past housing cycles with their respective business and employment cycles to inform our own views on the subject.
A robust debate around the timing of the business cycle slowdown in response to Fed policy tightening has at its center a conversation about the significance of the housing cycle. Building on Edward Leamer’s 2007 NBER Working Paper “Housing Is The Business Cycle,” we examine the relationship of past housing cycles with their respective business and employment cycles to inform our own views on the subject.
Stategy Capital’s research process focuses on analyzing both trends and cycles across the global macroeconomy. We rely on time-tested, empirically-reviewed statistical methods for all our analysis. The objective of this study is to better understand how the housing sector relates to the overall U.S. business cycle. In this analysis we will rely on the NBER defined business-cycle that is thought to last between two and eight years.
Our methodology for measuring cycles uses band-pass filters to analyze time series data. Band-pass filters allow one to decompose any time series into a long-run trend and a cyclical component. This is especially useful when a time series is trending over time. Figure 1 shows U.S. real GDP from 1960-2022 (recessions designated with shading). If we define a full business cycle as lasting on average 2-8 years, it is difficult to identify how the business-cycle evolves using unfiltered data.
However, if we apply the Christiano-Fitzgerald asymmetric band-pass filter to U.S. real GDP, we can decompose the time series into both a long-run trend and cyclical component of frequency 2-8 years. The business-cycle component of real GDP is shown below (Figure 2). The zero line represents the trend component. Any movement above or below represents standard deviations from trend. This represents the business-cycle component.
Band-pass filtering is an intuitive approach to analyzing cyclical movements of a time series with several key advantages. This approach provides insight into the timing of when a variable is above or below the long-run trend and provides insight into turning points in the business cycle. In addition, band-pass filtering can be applied to any financial or macroeconomic variable allowing one to compare both frequency and amplitude across variables in a consistent and interpretable way.
Applying the same Christiano-Fitzgerald band-pass filter methodology, we evaluated the business-cycle components of U.S. real GDP, housing starts, and non-farm payroll dating from 1960-2022 (Figure 3). Using band-pass filters, we have decomposed all three series into their long-run trend and business-cycle component. The frequency of the business cycle was defined as 2-8 years and follows the NBER convention.
Using the cycle decomposition in Figure 3, we observe the “U.S. Great Moderation” between 1985-2007, characterized by a dampening of the amplitude of U.S. business cycles. This era was followed by a unique period from 2009-2019, which lacked typical business-cycle dynamics. This is presumably due to sustained implementation of Quantitative Easing policies. The 2020 Covid-19 shutdown represented a return to high amplitude cycles akin to those observed in the 1970’s.
In addition, we observe consistent evidence of the U.S. housing cycle as a leading indicator of growth and employment across business cycles. This is timely as it relates to the current discussion of if or when the U.S. will move into a recession in the coming quarters. Resilient employment conditions in 2022 has made the timing of recession estimates difficult, but this analysis is a reminder that we are following a similar pattern seen in past business cycles. Figure 4 provides a closer look at the current business cycle.
Stategy research provides evidence that housing is once again leading the business-cycle and indicating an oncoming period of below-trend growth. The cyclical component peaked in Q421 and is now moving towards below-trend growth. Non-farm payroll employment as a function of the business cycle is only now peaking and is expected to begin to move towards below-trend growth in the coming quarters.
It is reasonable to assume that the weakening in employment and growth will be seen in early-to-mid 2023. In our minds, the more interesting question is whether the Fed will allow the economy to go through a full business-cycle contraction or if instead we will see similar dynamics to the period from 2009-2019. As always, time will tell, and investors must evaluate and consider both possibilities while constructing portfolios.
Stategy Capital’s approach to measuring cycles is unique from traditional modeling approaches. Our use of statistical methods ensures we adhere to our beliefs of first letting data inform the analysis rather than using data to support an already decided outcome. This ensures our decision-making process is internally consistent and let’s data drive the investment process. We believe in continual learning and, as always, strive to apply responsible data science to tough problems.
Lawrence J. Christiano, Terry J. Fitzgerald, The Band Pass Filter, International Economic Review, Vol. 44, No. 2 (May, 2003), pp. 435-465
Leamer, E. E. (2007, September 1). Housing IS the Business Cycle. National Bureau of Economic Research.
