July 20, 2023
Every month, our QIS Macro Team provides a complementary analysis of trends and predictions based on RavenPack's Inflation Nowcasting model.
Figure 1 presents the real-time density prediction of core Consumer Price Index (CPI) Month-Over-Month (MoM) inflation for May and June 2023 on an annual basis. The first column represents the prediction for May, while the second column represents the prediction for June. The chart also includes the actual inflation data released on June 13th and July 12th, 2023, corresponding to May and June, respectively.
In May, the median of the predictive distribution indicates an inflation rate of 3.51%. However, the mode of the distribution, which represents the most frequent value, is closer to the actual inflation rate of 5.22%. Notably, there is a lack of overlap between the median and mean of the predictive density, suggesting the presence of left or negative skewness. This implies a potential downward shift in MoM inflation. Indeed, the subsequent release indicated a decrease in core inflation for June (1.89%), approaching the target aimed by the U.S. Federal Reserve (Fed).
In June, the median of the predictive distribution indicates an inflation rate of 2.82%. It is worth noting that compared to May, there is a reduction in the negative skewness of inflation but an increase in its volatility. Interestingly, this decline in inflation was unexpected by the Fed backcast, which predicted an inflation rate around 5% for June in their July 11th update (see Inflation Nowcasting from clevelandfed.org ). By leveraging the predictive distribution of inflation, it is possible to estimate the probabilities associated with different inflation states using the Empirical Cumulative Distribution Function (ECDF). These states are categorized as Low (<0%), Medium Low (>0% and <2%), Medium High (>2% and <4%), and High (>4%). In Figure 2, displayed in the second row, the probabilities of observing inflation within these predefined states are depicted. The ECDF reveals a decrease in the probabilities of the Medium High and High states, along with an increase in the probabilities of the Low and Medium Low states of inflation for June compared to May.
The reported print for headline CPI in May was 1.48% on an annual basis. In May, the median of the predictive distribution was 2.42%, displaying left skewness (see Figure 2 , first row). For June, the median of the predictive distribution forecasted an inflation rate of 2.13%, closely aligned with the actual inflation rate of 2.16%, which is in proximity to the inflation target. Additionally, there was a reduction in skewness for June (from -0.58 to -0.2). However, it is important to note that the predictive density for June exhibited an increase in the standard deviation (from 2.11% to 2.48%) and the excess kurtosis (from 0.7 to 0.9). These changes indicate an increase in uncertainty surrounding the inflation forecast for June.
In June, the Empirical Cumulative Distribution Function (ECDF) demonstrated an increase in the probability of observing negative inflation or the Low state. Additionally, there is a slight increase in the probabilities associated with the Medium High and High states of inflation ( Figure 2 , second row).
The inflation prediction model used in this study relies on the incorporation of sentiment analytics and standard macroeconomic variables. By estimating the impact of each variable on the predictive distribution of inflation, it becomes possible to assess their influence during periods of high inflation. The estimation is conducted using a sample period spanning from December 2021 (when inflation surged and the Fed announced an interest rate hike) to July 2023.
Figure 3. illustrates the aggregated impact of both sentiment and standard macroeconomic variables for core CPI. Notably, both sentiment and standard macro data exhibit a more pronounced marginal contribution on the left tail of the distribution. Furthermore, during this specific time period, the sentiment data appears to have a more substantial impact on the predicted inflation compared to the standard macroeconomic variables.
During this specific time period, the sentiment analytics that have the greatest impact on the predicted inflation are as follows:
In terms of macroeconomic indicators, the following variables have the greatest impact on the predicted inflation during this time period:
Figure 4 displays the collective impact of both sentiment and standard macroeconomic variables on headline CPI. It is observed that both sentiment and standard macro data have a more significant marginal contribution on the right tail of the distribution, indicating their influence on higher inflation levels. Notably, during this specific period, sentiment data appears to have a greater impact compared to standard macroeconomic variables. This suggests that sentiment analytics play a prominent role in shaping the predicted dynamics of headline CPI during this time period.
In contrast to core CPI, the impact of variables on headline CPI is relatively sparse. Among the notable predictors are:
The model's effective prediction of the decrease in core CPI for June underscores its ability to timely capture the key drivers of inflation. By incorporating pertinent variables related to economic activity, labor market conditions, and financial indicators, the model can track significant changes that influence inflation. Notably, its timely reflection of the recent increase in US unemployment further strengthens its predictive capabilities for anticipating changes in core CPI.
Moreover, the model focusing on headline inflation demonstrates its ability to promptly capture the recent rise in unemployment and decrease in energy prices, which have resulted in decreased inflation rates. It also emphasizes the significance of variables like interest rates and CPI indicators guidance, which provide forward-looking sentiment. This indicates that the model adeptly incorporates influential factors that impact headline inflation, enabling it to offer valuable insights for forecasting and decision-making needs.
Please use your business email. If you don't have one, please email us at info@ravenpack.com.
By providing your personal information and submitting your details, you acknowledge that you have read, understood, and agreed to our Privacy Statement and you accept our Terms and Conditions. We will handle your personal information in compliance with our Privacy Statement. You can exercise your rights of access, rectification, erasure, restriction of processing, data portability, and objection by emailing us at privacy@ravenpack.com in accordance with the GDPRs. You also are agreeing to receive occasional updates and communications from RavenPack about resources, events, products, or services that may be of interest to you.
Your request has been recorded and a team member will be in touch soon.