State of the Business Cycle — Part I
Learn how to identify the length and magnitude of the U.S. Business Cycle using different methods in Julia.
The beginning of the coronavirus pandemic in February ended a 128 months expansion. After its peak in February 2020, U.S. GDP declined dramatically by a historical 9.5% in the second quarter of 2020. Moreover, the pandemic has ended the longest business cycle in U.S. history. When and how much does it take for the U.S. economy to recover? We will take one step back and ask:
What is a business cycle, and how can we identify its stages?
This piece is the first part of a short series on how to define, measure, and understand the cyclical variations in the U.S. business cycle using Julia. Whereas this part is more of an introduction, later parts will explain how to use statistical and econometric methods on U.S. GDP.
Spoiler Alert: The multitude of opinions you encounter on that issue over a dinner talk (with economists) is as numerous as the methods used to determine a business cycle. It is science in itself.
We take the most common time-series used in economic analysis: Real Gross Domestic Product or GDP. It reflects the value of all goods and services produced by an economy within a year, corrected for annual changes of the aggregate price level. Let’s take a look:
Looking at this chart, it’s not hard to make two striking observations:
First, U.S. Real GDP is more than seven times higher than it was after WWII. It indicates that GDP follows a long-term positive trend or the economy’s long-term potential output. Second, during some years, the expansion of GDP either accelerates or falls — sometimes even dramatically, as we have seen throughout the second quarter of 2020. That is, the economy is either above or below its most efficient production capacity.
Hence, in addition to long-run growth-related drivers of GDP, short-run fluctuations move GDP either up or downwards. Having established that idea, let’s think of GDP as the product of a trend, a cyclical, and other unobserved components. We will use that later.
So how do we get to an identification of the U.S. Business Cycle? We could partition the business cycle into stages of boom and bust, inevitably following each other. That means, when the economy is expanding in a boom phase, annual GDP growth is positive. Otherwise, when the economy is in a recession or a bust phase has begun, GDP growth is negative.
Is this correct? Almost — or better, not quite.
To understand why this simple characterization of a business cycle is still incomplete, let’s look at the chart below. It depicts annualized GDP growth and episodes of recessions, as classified by the National Bureau of Economic Research (NBER), in shades:
The grey-shaded areas mark the beginning and duration of a recession. Moreover, peak growth rates (in absolute numbers) occur either within boom phases (white areas) or recession phases (grey shades). We observe that the U.S. economy entered into downturns while either still growing or into expansions while still declining in absolute terms.
Let’s take a more positive approach towards the definition of a recession:
We generally define an economic downturn as two consecutive quarters of negative real GDP growth (quarter-on-quarter). Besides, the NBER speaks of a recession as a
[…] significant decline in economic activity spread across the economy, normally visible in production, employment, and other indicators.
Where it gets more difficult are magnitudes. Let’s dive right into it:
How severe is the recent downturn due to the CoVid19 pandemic likely to be?
The answer to this question is the economist’s favorite answer: It all depends. As unsatisfying as this might sound, more detailed reasoning will provide the intuition as to why we need better tools to tackle this question.
On the one hand, economic production or corporate profits can fall off a cliff because of a lack of aggregate demand: People stop consuming certain goods. Either because they are no longer available (mainly due to temporary health restrictions, like concerts, festivals, or clubs). Or, people adjust their behavior. They could quit going to restaurants for some time in response to a worsening of the health crisis in their district or in anticipation of it. Furthermore, people might bump up their savings out of a precautionary motive— because they either lost their job recently or fear to do so soon. Finally, some goods might temporarily be unavailable due to a breaking down of local supply chains. These are all examples (far from a complete description of all effects) of short-run, business-cycle driven fluctuations that affect economic output.
On the other hand, the economy can take long-term hits from the pandemic as well: International migration flows have fallen to all-time lows. Travel bans and restrictions on mobility have severely impeded the tourist and airline industries. Countries openly discuss renationalizing the production of certain types of goods, thus destroying established and cost-efficient supply-chains. Closing borders and policies of self-interest instead of cooperation post a threat to a globalized world and economic well-being worldwide. These examples are long-run factors that will hamper long-run growth prospects and potential output.
Both factors come together when you want to adress the -9.5% drop in Q2 2020.
How can we entangle these two effects and quantify them?
In the subsequent articles of this series, we will learn to apply several statistical detrending and filtering tools that separate the GDP time series into a trend and a cyclical component. The latter determines how far above or below the economy is from its efficient production capacity or potential output.
The article finishes with the code used to reproduces the figures discussed above. We will use Julia and the Fred API® to download a long-run time-series of U.S. real GDP. It builds on a previous article of mine.
Thanks for bearing with me. I hope you enjoyed this introduction.
