Economic indicators play a vital role in measuring the health of an economy. Some important macroeconomic indicators are the unemployment rate, consumer price index or CPI which measures inflation, gross domestic product or GDP, disposable income, short-term and long-term interest rates (such as Fed Funds rate and Treasury rates), home price index, and stock market performance metrics.
The relationship between the indicators can be complex and dynamic due to changing monetary policy, geopolitical situation, population growth, etc. Let’s look at the interrelation between macroeconomic indicators in the US.
Historical Data for Macroeconomic Indicators
Our analysis is based on historical data covering 1990 to 2023. During that time the US economy experienced many ups and downs including four recessions (the Gulf War recession in 1990-1991, the dot-com bubble in 2001, the Great Recession in 2007-2009, and the pandemic in 2020), a near zero interest rate environment in 2009-2016 and 2020-2022, and record low unemployment rates in 2018-2019 and 2022-2023.
The unemployment rate and Treasury 3-month rate often move in opposite directions as we observe from the chart above. We can quantify their relationship using the correlation coefficient. Let’s look into the correlation between these and other economic indicators.
Correlation Between the Economic Indicators
We calculated the pairwise correlation between eleven macroeconomic indicators shown in the heatmap below.
As a quick reminder, for a pair of variables a correlation close to 1 implies a strong comovement in the same direction, a correlation close to -1 is an indication of a strong comovement in the opposite direction, and a correlation close to zero indicates a weak or absence of any (linear) relationship between two variables.
From the heatmap above, we can immediately observe that all interest rates have a strong positive correlation. This is expected as they generally move in tandem when reacting to changes in monetary policy.
Another observation is that the Treasury 3-month rate negatively correlates with the inflation rate. This is because the Treasury 3-month rate closely tracks the Fed Funds rate which the Federal Reserve Board (FRB) uses to control inflation. When inflation is high, the FRB increases the Fed Funds rate to make borrowing more expensive. As a result, it reduces the money supply for businesses and consumers and decreases their buying power eventually reducing the inflation rate.
We can also notice a negative correlation between the Dow Jones US Total Stock Market Index (DWCF) and VIX (stock market volatility). One may expect a stronger correlation between these two indicators since VIX spikes when the stock market sharply drops. The reason for the weaker correlation is the granularity of the data. We could observe a much stronger negative correlation between DWCF and VIX on daily or weekly data. However, with monthly or quarterly data, the market’s partial recovery during the period makes the relationship with the VIX less dramatic.
Rank-Based Correlation Between the Economic Indicators
In the heatmap above, the correlation between real GDP growth and real disposable income growth is negative which is counterintuitive. When the real disposable income growth is high, consumers have more money to spend on discretionary goods and services, which should positively impact GDP.
The reason for the negative correlation is rather technical. In the above heatmap, we used the most popular Pearson correlation method. However, the Pearson correlation is very sensitive to large outliers, which means a rare large movement in the historical data may distort the correlation between the two indicators.
We can see from the chart below that shortly after the pandemic lockdown GDP went sharply down. On the contrary, disposable income jumped due to the government stimulus. The magnitude of these movements was so strong that it distorted the (Pearson) correlation between real GDP and real disposable income.
To eliminate the impact of the outliers, we used the Spearman correlation method also known as rank-order correlation. We can observe that in the new heatmap, the correlation between real GDP growth and real disposable income growth is now positive which makes more sense.
Notice that in the rank-order heatmap, the correlation between the inflation rate and the real disposable income growth turned slightly negative. It is more intuitive since during high inflation periods wages usually increase at a slower pace and further adjustment by the inflation rate may result in a decreased real disposable income.
The correlation between the inflation rate and HPI growth also changed in the rank-order heatmap. Unlike disposable income, the relationship between the inflation rate and HPI is much more complex and depends on other factors too. For example, while increased material and labor costs during high inflation may increase property prices, higher interest rates may reduce the affordability for mortgage borrowers. Besides, supply and demand, population growth, and overall economic health are important factors impacting HPI too.
Conclusion
Macroeconomic indicators are highly interrelated; even from this simple correlation analysis, it is evident that their relationship is quite complex. For example, long-term relationships can be disrupted by a sudden shock or a short-term reversal, and the relationship between two variables may be impacted by a third variable. Besides, some indicators may lag others which introduces another layer of complexity in the relationship of economic indicators.
At Scenarios by AI, we developed proprietary generative AI models that capture the complex dynamics of microeconomic indicators and generate a wide range of coherent macroeconomic scenarios. Reach out to explore how these economic scenarios can help you enhance your quantitative risk management.