August 28, 2023
A guide to inflation nowcasting and its integration with sentiment analysis.
Nowcasting (a combination of “now” and “forecasting”) aims to predict the current and very recent economic conditions, as well as the immediate future, by generating accurate estimates to fill in the missing data gaps. Traditionally, economic indicators are released with a delay, making it challenging for investors to make timely decisions based on the latest economic developments. Nowcasting tackles this issue by leveraging real time data to approximate lagging data and high-frequency data for more frequent revaluations of the proxy. By analyzing these contemporaneous data sources, nowcasting models make informed predictions about the upcoming official releases of key economic metrics.
An inflation nowcasting model is a predictive analytics tool designed to provide real-time or near-real-time estimates of inflation rates. It leverages various economic indicators, high-frequency data, and advanced statistical techniques to forecast inflation in the short term. Unlike traditional inflation forecasting models, which often rely on historical data and are used to predict inflation over longer time horizons, inflation nowcasting models focus on providing more immediate estimates of inflation, typically for the current or next few months.
Inflation is a complex economic phenomenon that has a significant impact on businesses, consumers, and financial markets. Accurately forecasting inflation is difficult, but inflation nowcasting models can provide timely insights into current and near-term inflation levels. These models are valuable for investors because they can help to calibrate investment strategies, manage risk, and analyze sectors. As inflation becomes more volatile, inflation nowcasting models are becoming increasingly important.
Our predictions target monthly core-CPI and headline-CPI for the US. The Month-Over-Month (MoM) measure of the core Consumer Price Index (CPI) is a more precise way to identify and track inflationary trends in financial markets. It offers a shorter time frame, which means it can capture immediate changes in inflation. This allows market participants to respond quickly to temporary inflation shocks, allowing more informed investment decisions.
Our sentiment-based inflation nowcasting models add a layer of sentiment analytics over a series of indicators and expert surveys on unemployment, labor costs and economic performance to predict or nowcast inflation trends. Our focus is on sectors closely tied to inflation changes, and we apply a meticulous set of filters to extract high-quality, accurate, and timely events. As our focus is the US inflation, we only consider content in English and related to six crucial policy institutions in the US, namely the United States, United States Federal Reserve, Government of the United States, Federal Open Market Committee, The White House, and the Department of the Treasury of the United States.
Sentiment analytics refer to quantifying subjective information resulting from financial textual data, such as news articles, earnings call transcripts or job postings. The goal is to have a quantified view of the perceptions about a particular stock, sector, market or an event (a merger, an earnings release) and then systematically use this information to make more informed investment decisions, manage risk, predict market trends or macro indicators.
Compared to traditional economic indicators, sentiment tends to react faster to changes in the markets. It is also a signal with a much higher frequency. When combined with other data sources, sentiment data can enhance the accuracy of inflation nowcasting models and provide investors with timely insights to make informed decisions.
There are multiple approaches to sentiment analysis, which often overlap. RavenPack models combine Natural Language Processing and Machine Learning capabilities with sentiment analysis based on expert consensus:
Sentiment analytics have proven to be a strong asset in inflation nowcasting. For instance, RavenPack uses a Bayesian neural network (BNN)* model that combines RavenPack Analytics and macroeconomic variables to predict monthly US inflation rates (Core and Headline CPI) on a daily basis. The model typically outperforms alternative methods. Specifically, it achieves a significant improvement of 20% in mean squared error (MSE) for predicting headline Consumer Price Index (CPI), and 4% for predicting core CPI. Moreover, the accuracy of the RavenPack BNN model increases as the release date of inflation data approaches.
A Bayesian neural network (BNN) is a computational model that gives predictions along with a measure of uncertainty
Neural networks are machine-learning approaches which base classification decisions on the highest probability. However, when data is limited, it is more relevant to consider not only the likely classification, but also the degree of confidence in that classification, so the algorithm can describe its decisions with a probability qualifier. That's what Bayesian networks achieve. The main benefit of these networks is to not rush to conclusion (a move described as 'overfitting' the data) but to accompany each classification decision with a distribution of likelihood. As a result, Bayesian neural networks provide more than just classification information: they can also convey uncertainty in a useful way.
Sentiment based Inflation nowcasting models aim to enhance accuracy and timeliness of short-term predictions of inflation rates, which can be valuable in various investment strategies. Here's how different types of investors can use these models:
Trading Strategies: Quantitative investors can use inflation nowcasting models to develop trading strategies that capitalize on short-term fluctuations in inflation rates. These models can help identify trading opportunities in inflation-sensitive assets, such as commodities, inflation-linked bonds, or certain equities.
Risk Management: Inflation nowcasting can aid quantitative investors in managing risk by incorporating short-term inflation predictions into their portfolio optimization process. By adjusting portfolio allocations based on expected inflation, they can create more robust and adaptive investment strategies.
Tactical Asset Allocation: Discretionary investors can use inflation nowcasting models as one of the inputs for tactical asset allocation decisions. By considering short-term inflation forecasts, they can adjust their portfolio exposures to various asset classes or industries that may benefit or suffer from inflationary pressures.
Sector and Industry Analysis: With inflation nowcasting, discretionary investors can analyze how different sectors and industries are likely to perform in the short term. For example, certain sectors, like energy and materials, tend to outperform during inflationary periods, while others, like technology, might face challenges.
Please use your business email. If you don't have one, please email us at firstname.lastname@example.org.
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 email@example.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.