Retail Sales: A Closer Look at January’s Economic Signals

As we stepped into 2024, the U.S. economy presented a mixed bag of results in January, revealing a slowdown that has caught the eyes of consumers and economists alike. Following a robust end to 2023, marked by strong holiday shopping, retail sales experienced a surprising dip in January, falling 0.8% from December, according to the Commerce Department.This decline exceeded the modest expectations of a 0.3% decrease, signaling a cautious start to the year for American shoppers. 

January’s chill extended beyond the weather, with cold conditions potentially influencing spending habits. However, it’s clear that the weather alone doesn’t account for the observed retail slowdown. Notably, even e-commerce, typically resilient to meteorological changes, saw a 0.8% decrease in sales, underscoring broader economic factors at play. This period of economic “breathing” doesn’t spell doom; rather, it reflects a complex interplay of seasonal adjustments, consumer caution, and ongoing shifts in spending patterns. 

Despite this setback, there are silver linings that suggest resilience within the U.S. economy. Food services and drinking places, for instance, bucked the trend by registering a 0.7% increase from December, a robust sign amidst general reticence. This uptick, alongside notable gains in high-tech manufacturing output such as semiconductors, indicates pockets of strength and adaptation within the broader economic landscape. 

While acknowledging the slowdown, I anticipate a continued growth, albeit at a moderated pace. This perspective aligns with the broader expectation that the Federal Reserve’s past rate hikes will gradually exert more pronounced effects on economic activity. 

The juxtaposition of January’s retail sales decline with the ongoing resilience in specific economic sectors and consumer spending areas paints a nuanced picture. It suggests an economy that is navigating the complexities of post-pandemic adjustment, interest rate impacts, and shifting consumer preferences. As we move forward, the focus will likely remain on discerning the balance between caution and confidence among U.S. consumers and businesses alike.

One aspect I always consider prudent is to verify the importance of indicators as predictors of key economic drivers. For this reason, I conducted a Granger causality test to determine if changes in retail sales (month-over-month) can predict changes in other economic indicators. Essentially, I’m investigating whether knowledge of past retail sales data can forecast future movements in economic metrics, such as the Consumer Price Index (CPI) or Composite Leading Indicators (CLI).

Here’s how I approached it:

1. I prepared a dataset combining retail sales with various economic indicators.
2. I split this dataset into a training set (80% of the data) to fit my model and a test set (20%) to evaluate it.
3. I then used a statistical model called Vector Autoregression (VAR) on the training data to capture the relationships between retail sales and the other indicators.
4. After fitting the model, I performed the Granger causality tests for a number of lags (delays) to see if past values of retail sales are useful in predicting the other indicators.
5. Finally, I visualized the results by plotting the p-values (the probability that the results are due to chance) of the tests for different lags. If the p-value is below a certain threshold (0.05 is common), then I can state with confidence that retail sales do indeed have predictive power over the economic indicator in question.