Throughout history, the economy of the United States has suffered down-turns of different magnitudes, from the stock-market crash of 1929 (followed by the Great Depression) to the deep recession of 2008 and many others in between.

In this text and the accompanying Jupyter notebook, I explore visually the recoveries that took place after these crises were over. In particular, I compare several indicators across the different post-recession periods: GDP, Unemployment, Gini (wealth inequality) and Debt-to-GDP ratio.

We start by importing a dataset containing changes in Gross Domestic Product (GDP). These are seasonally-adjusted annual rates for each quarter.

This is a good place to start because we're looking at the overall size of the economy. A common standard for defining recessions is two consecutive quarters of negative GDP growth as originally proposed by Commissioner of the Bureau of Labor Statistics Julius Shiskin in 1974.

The dataset comes from the Bureau of Economic Analysis at the US Department of Commerce.

The distribution above makes sense, since the the economy of the United States has, for the most part, grown at a steady pace between 1-5%. In fact, half of all GDP change values fall between 1.2% and 5.1%. The minimum (-31.4%) and maximum (16.7%) values are surprisingly extreme. Let's take a quick look at them.

The shrinkage of -31.4% corresponds to the economic shutdown during the 2020 COVID-19 pandemic. At the time of this writing, we're not out of the woods, yet, so we'll leave this one for future analyses.

Let's visualize the GDP changes for the whole dataset:

We have 8 post-recession recoveries that we will look at. Here are the last quarters for each of those recessions: 1949Q2, 1954Q1, 1958Q1, 1970Q1, 1975Q1, 1982Q1, 1991Q1 and 2009Q2.

Now we plot the GDP change rate per quarter for each post-recession recovery. Note that the shaded area to the left of each plot represents the last 2 quarters of each recession.

We can observe that in general, the GDP growth rate is high for 3-4 quarters after the end of a recession, with a dip at quarters 5-6.

Now we look at the rate of unemployment after each crisis. The data comes from the Bureau of Labor Statistics.

It's time to take a look at the rate of unemployment in the period after each recession:

As intuition would tell us, the rate of unemployment decreases during post-recession recoveries. However, this effect is delayed. The peak of unemployment for these 8 recessions occurred as follows:

1949 - unemployment peaked 2 quarters after the recession ended

1954 - 1 quarter

1958 - 2 quarters

1970 - 6 quarters

1975 - 1 quarter

1982 - 3-4 quarters

1991 - 6 quarters

2009 - 2 quarters

The Gini coefficient (or Gini index) describes the income or wealth inequality within a country. It was developed by the Italian statistician and sociologist Corrado Gini. A Gini coefficient of 0 expresses full equality and 1 would be maximal inequality (1 person holding all of the wealth or income). The dataset that we will use comes from the United States Census Bureau. Here are the first 5 entries:

We can now visualize all of our Gini values. This amounts to a visualization of wealth inequality in the United States over the last 50 years or so.

Clearly, wealth inequality increased after each recession. However, there's a big caveat: wealth inequality may have increased anyway, whether there was a recession or not. For this we look at the previous plot ("Wealth inequality after recessions in the US") and we see that the Gini index has been increasing steadily in the US during periods of growth and contraction.

We end by looking at the Debt to GDP ratio after recessions in the US. This is a measurment of a country's government debt in relation to the total size of the economy.

It's not uncommon for this ratio to increase after disruptive negative events like wars, natural disasters, recessions and others, as the economic activity might slow down or decrease while government accrues more debt to cover budget deficits or stimulate the economy.

We start by looking at data from the Treasury of the United States. This dataset contains the total outstanding debt amount for each year since 1790!

It's difficult in that graph to see what happened before the middle of the 20th century, as the values are too small relative to recent ones and the expansion of the debt is clearly not linear. It may be better to convert the y-scale to logarithmic.

Here is the period that interests us for this analysis:

As interesting as it is to look at the absolute values for government debt, what we're really after is the debt-to-gdp ratio. There is another dataset that contains this data from 1900 onwards and is provided by the Congressional Budget Office.

Let's visualize the change in Debt-to-GDP ratio over time. The dashed line is the present time and the shaded area represents projected values.

One can see that there is no definite trend here. The first 4 recoveries saw a decrease in Debt-to-GDP ratio, while the latest 4 saw an increase. This may be explained by the fact that GDP growth after the first 4 recessions was very strong, with peaks over 10%. The most recents post-recessions recoveries where tamer, with peaks under 10% (and under 5% for the last two recessions). Also, as we visualized with the first Debt dataset, the government debt in the US grew slowly before 1970 and grew exponentially after.

Finally, let's visualize all indicators for each post-recession recovery. Because the indicators are on such different scales, we will normalize the values. The graphs won't be informative of particular numbers, but we can better visualize tendencies in different columns all at once.