Learning Social Meta-knowledge for Nowcasting Human Mobility in Disaster

Human mobility nowcasting is a fundamental research problem for intelligent transportation planning, disaster responses and management, etc. In particular, human mobility under big disasters such as hurricanes and pandemics deviates from its daily routine to a large extent, which makes the task more challenging. Existing works mainly focus on traffic or crowd flow prediction in normal situations. To tackle this problem, in this study, disaster-related Twitter data is incorporated as a covariate to understand the public awareness and attention about the disaster events and thus perceive their impacts on the human mobility. Accordingly, we propose a Meta-knowledge-Memorizable Spatio-Temporal Network (MemeSTN), which leverages memory network and meta-learning to fuse social media and human mobility data. Extensive experiments over three real-world disasters including Japan 2019 typhoon season, Japan 2020 COVID-19 pandemic, and US 2019 hurricane season were conducted to illustrate the effectiveness of our proposed solution. Compared to the state-of-the-art spatio-temporal deep models and multivariate-time-series deep models, our model can achieve superior performance for nowcasting human mobility in disaster situations at both country level and state level. © 2023 ACM.

The Asymmetric Impact of COVID-19: A Novel Approach to Quantifying Financial Distress across Industries.

This paper assesses corporate financial distress in terms of liquidity and risk of insolvency due to the COVID-19 pandemic. We develop a novel multivariate approach to obtain monthly data on industry turnover, exploiting real time data to capture the atypical character of industry-specific disturbances. By combining the estimated set of industry revenue shocks with pre-pandemic financial statements, we quantify the impact of the pandemic on the risk of insolvency in the EU non-financial corporate sector. Our definition of risk of insolvency takes into account not only the equity position of firms, but also risks relating to overindebtedness. The analysis controls for firms that were financially vulnerable already before the pandemic, thus being prone to become at risk of insolvency also in absence of the COVID-19 turmoil. We find that, for the EU as a whole, 25% of firms exhausted their liquidity buffers by the end of 2021 (a practical cut-off date of the analysis, not an assumed end of the pandemic). Furthermore, 10% of firms which were viable before the pandemic, appear to have shifted into risk of insolvency as a result of the COVID-19 crisis. The magnification of financial vulnerability in the hardest-hit industries mainly occurs among firms with no legacy issues, i.e. firms with positive profitability pre-pandemic. A similar finding is reported for some of the hardest-hit countries, such as Italy and Spain. In other countries, such as Germany or Greece, the magnification of financial vulnerability mainly occurs among firms with negative profitability pre-pandemic.

US weekly economic index: Replication and extension

We revisit the US weekly economic index (WEI) put forth by Lewis, Mertens, Stock and Trivedi (2021). In a narrow sense, we replicate their main results with data gathered from its original sources. In a wide sense, we apply the methodology established in Wegmüller, Glocker and Guggia (2023) to adjust the weekly input series for seasonal patterns, calendar day effects, and excess volatility. In a long sense, we show that our proposed data adjustment significantly improves the nowcasting performance of the WEI. © 2023 John Wiley & Sons, Ltd.

Nowcasting food inflation with a massive amount of online prices

The consensus in the literature on providing accurate inflation forecasts underlines the importance of precise nowcasts. In this paper, we focus on this issue by employing a unique, extensive dataset of online food and non-alcoholic beverages prices gathered automatically from the webpages of major online retailers in Poland since 2009. We perform a real-time nowcasting experiment by using a highly disaggregated framework among popular, simple univariate approaches. We demonstrate that pure estimates of online price changes are already effective in nowcasting food inflation, but accounting for online food prices in a simple, recursively optimized model delivers further gains in the nowcast accuracy. Our framework outperforms various other approaches, includ-ing judgmental methods, traditional benchmarks, and model combinations. After the outbreak of the COVID-19 pandemic, its nowcasting quality has improved compared to other approaches and remained comparable with judgmental nowcasts. We also show that nowcast accuracy increases with the volume of online data, but their quality and relevance are essential for providing accurate in-sample fit and out-of-sample nowcasts. We conclude that online prices can markedly aid the decision-making process at central banks.(c) 2022 International Institute of Forecasters. Published by Elsevier B.V. All rights reserved.

Nowcasting and Forecasting COVID-19 Cases and Deaths Using Twitter Sentiment

Real-time access to information during a pandemic is crucial for mobilizing a response. A sentiment analysis of Twitter posts from the first 90 days of the COVID-19 pandemic was conducted. In particular, 2 million English tweets were collected from users in the United States that contained the word 'covid' between January 1, 2020 and March 31, 2020. Sentiments were used to model the new case and death counts using data from this time. The results of linear regression and k-nearest neighbors indicate that sentiments expressed on social media accurately predict both same-day and near future counts of both COVID-19 cases and deaths. Public health officials can use this knowledge to assist in responding to adverse public health events. Additionally, implications for future research and theorizing of social media's impact on health behaviors are discussed. © 2022 IEEE Computer Society. All rights reserved.

Nowcasting Japan's GDP.

This paper backtests a nowcast of Japan's real GDP growth. It has three contributions: (i) use of genuine real-time data, (ii) implementation of a new method for the revision analysis that relates the revision of the nowcast to not only new observations but also data revisions, and (iii) a benchmarking of the nowcast to a market consensus forecast at monthly forecasting horizons. Our nowcast's forecast accuracy is comparable to that of the consensus at most, but not all, monthly horizons. Our revision analysis of the March 2011 earthquake finds the nowcast reacting to a steep post-quake decline in car production. In contrast, the consensus hardly budged, most likely because the decline was correctly viewed as temporary. The onset of COVID-19 triggers the consensus to take a precipitous descent. The nowcast, despite timely red flags from "soft" (i.e., survey-based) indicators, does not respond immediately in full, because it took a month or more for "hard" (i.e., non-survey-based) indicators to register sharply reduced economic activities.

Nowcasting the state of the Italian economy: The role of financial markets

This paper compares several methods for constructing weekly nowcasts of recession probabilities in Italy, with a focus on the most recent period of the Covid-19 pandemic. The common thread of these methods is that they use, in different ways, the information content provided by financial market data. In particular, a battery of probit models are estimated after extracting information from a large dataset of more than 130 financial market variables observed at a weekly frequency. The accuracy of these models is explored in a pseudo out-of-sample nowcasting exercise. The results demonstrate that nowcasts derived from probit models estimated on a large set of financial variables are, on average, more accurate than those delivered by standard probit models estimated on a single financial covariate, such as the slope of the yield curve. The proposed approach performs well even compared with probit models estimated on single time series of real economic activity variables, such as industrial production, business tendency survey data or composite PMI indicators. Overall, the financial indicators used in this paper can be easily updated as soon as new data become available on a weekly basis, thus providing reliable early estimates of the Italian business cycle. © 2023 John Wiley & Sons Ltd.

Aggregate Learning for Mixed Frequency Data

Large and acute economic shocks such as the 2007-2009 financial crisis and the current COVID-19 infections rapidly change the economic environment. In such a situation, real-time analysis of regional heterogeneity of economic conditions using alternative data is essential. We take advantage of spatio-temporal granularity of alternative data and propose a Mixed-Frequency Aggregate Learning (MF-AGL) model that predicts economic indicators for the smaller areas in real-time. We apply the model for the real-world problem;prediction of the number of job applicants which is closely related to the unemployment rates. We find that the proposed model predicts (i) the regional heterogeneity of the labor market condition and (ii) the rapidly changing economic status. The model can be applied to various tasks, especially economic analysis. © 2022 IEEE.

Forecasting recovery from COVID-19 using financial data: An application to Vietnam.

We develop a new methodology to nowcast the effects of the COVID-19 crisis on GDP and forecast its evolution in small, export-oriented countries. To this aim, we exploit variation in financial indexes at the industry level in the early stages of the crisis and relate them to the expected duration of the crisis for each industry, under the assumption that the main shocks to financial prices in 2020 came from COVID-19. Starting from the latest official information available at different stages of the crisis on industry-level trend deviations of GDP, often a few months old, we predict the ensuing recovery trajectories using the most recent financial data available at the time of the prediction. The financial data reflect, among other things, how subsequent waves of infections and information about new vaccines have impacted expectations about the future. We apply our method to Vietnam, one of the most open economies in the world, and obtain predictions that are more optimistic than projections by the International Monetary Fund and other international forecasters, and closer to the realised figures. Our claim is that this better-than-expected performance was visible in stock market data early on but was largely missed by conventional forecasting methods.

Weekly economic activity: Measurement and informational content

We construct a composite index to measure the real activity of the Swiss economy on a weekly frequency. The index is based on a novel high-frequency data set capturing economic activity across distinct dimensions over a long time horizon. We propose a six-step procedure for extracting precise business cycle signals from the raw data. By means of a real-time evaluation, we highlight the importance of our proposed adjustment procedure: (i) our weekly index significantly outperforms a comparable index without adjusted input variables;and (ii) the weekly index outperforms established monthly indicators in nowcasting GDP growth. These insights should help improve other recently developed high-frequency indicators. © 2021 International Institute of Forecasters

Nowcasting in a pandemic using non-parametric mixed frequency VARs

This paper develops Bayesian econometric methods for posterior inference in non-parametric mixed frequency VARs using additive regression trees. We argue that regression tree models are ideally suited for macroeconomic nowcasting in the face of extreme observations, for instance those produced by the COVID-19 pandemic of 2020. This is due to their flexibility and ability to model outliers. In an application involving four major euro area countries, we find substantial improvements in nowcasting performance relative to a linear mixed frequency VAR. © 2020 The Author(s)

Assessing the informational content of card transactions for nowcasting retail trade: Evidence for Latvia

We use a novel card transaction data maintained at the Central Bank of Latvia to assess their informational content for nowcasting retail trade in Latvia. During the COVID-19 pandemic in Latvia, the retail trade turnover dynamics underwent drastic changes reflecting the various virus containment measures introduced during three separate waves of the pandemic. We show that the nowcasting model augmented with card transaction data successfully captures the turbulence in retail trade turnover induced by the COVID-19 pandemic. The model with card transaction data outperforms all benchmark models in the out-of-sample nowcasting exercise and yields a notable improvement in forecasting metrics. We conduct our nowcasting exercise in forecast-as-you-go manner or in real-time squared;that is, we use real-time data vintages, and we make our nowcasts in real time as soon as card transaction data become available for the target month. © 2023 The Authors. Journal of Forecasting published by John Wiley & Sons Ltd.

Nowcasting the output gap

We propose a way to directly nowcast the output gap using the Beveridge–Nelson decomposition based on a mixed-frequency Bayesian VAR. The mixed-frequency approach produces similar but more timely estimates of the U.S. output gap compared to those based on a quarterly model, the CBO measure of potential, or the HP filter. We find that within-quarter nowcasts for the output gap are more reliable than for output growth, with monthly indicators for a credit risk spread, consumer sentiment, and the unemployment rate providing particularly useful new information about the final estimate of the output gap. An out-of-sample analysis of the COVID-19 crisis anticipates the exceptionally large negative output gap of −8.3% in 2020Q2 before the release of real GDP data for the quarter, with both conditional and scenario nowcasts tracking a dramatic decline in the output gap given the April data. © 2022 The Authors

NOWCASTING REGIONAL ECONOMIC GROWTH IN INDONESIA

This study aims to nowcast gross regional domestic product at the provincial level for Indonesia. The dynamic factor model and mixed data sampling were applied to three sets of variables;namely, macroeconomic, financial, and Google Trends. We find that both methods captured several economic expansions and contractions, including the recent downturn during the COVID-19 pandemic. By including the pandemic period, accuracy across the same set of variables and provinces was slightly reduced. © 2022 The authors.

SOS incomes: simulated effects of COVID-19 and emergency benefits on individual and household income distribution in Italy

Using a static microsimulation model based on a link between survey and administrative data, this article investigates the effects of the pandemic on income distribution in Italy in 2020. The analysis focuses on both individuals and households by simulating through nowcasting techniques changes in labour income and in equivalized income, respectively. For both units of observations, we compare changes before and after social policy interventions, that is, automatic stabilizers and benefits introduced by the government to address the effects of the COVID-19 emergency. We find that the pandemic has led to a relatively greater drop in labour income for those lying in the poorest quantiles, which, however, benefited more from the income support benefits. As a result, compared with the ‘No-COVID scenario', income poverty and inequality indices grow considerably when these benefits are not considered, whereas the poverty increase greatly narrows and inequality slightly decreases once social policy interventions are taken into account. This evidence signals the crucial role played by cash social transfers to contrast with the most serious economic consequences of the pandemic.

Nowcasting COVID-19 Statistics Reported with Delay: A Case-Study of Sweden and the UK.

The COVID-19 pandemic has demonstrated the importance of unbiased, real-time statistics of trends in disease events in order to achieve an effective response. Because of reporting delays, real-time statistics frequently underestimate the total number of infections, hospitalizations and deaths. When studied by event date, such delays also risk creating an illusion of a downward trend. Here, we describe a statistical methodology for predicting true daily quantities and their uncertainty, estimated using historical reporting delays. The methodology takes into account the observed distribution pattern of the lag. It is derived from the "removal method"-a well-established estimation framework in the field of ecology.

Nowcasting Chinese GDP in a data-rich environment: Lessons from machine learning algorithms

The ability to estimate current GDP growth before official data are released, known as "nowcasting”, is crucial for the Chinese government to effectively implement economic policy and manage economic uncertainties;however, there is limited research on nowcasting China's GDP in a data-rich environment. We evaluate the performance of various machine learning algorithms, dynamic factor models, static factor models, and MIDAS regressions in nowcasting the Chinese annualised real GDP growth rate in pseudo out-of-sample exercise, using 89 macroeconomic variables from years 1995 to 2020. We find that some machine learning methods outperform the benchmark dynamic factor model. The machine learning method that deserves more attention is ridge regression, which dominates all other models not only in terms of nowcast error but also in effective recognition of the impacts of the Global Financial Crisis and Covid-19 shocks. Policy-wise, our study guides practitioners in selecting appropriate nowcasting models for China's macroeconomy.

Bayesian Spatio-Temporal Prediction and Counterfactual Generation: An Application in Non-Pharmaceutical Interventions in COVID-19.

The spatio-temporal course of an epidemic (such as COVID-19) can be significantly affected by non-pharmaceutical interventions (NPIs) such as full or partial lockdowns. Bayesian Susceptible-Infected-Removed (SIR) models can be applied to the spatio-temporal spread of infectious diseases (STIFs) (such as COVID-19). In causal inference, it is classically of interest to investigate the counterfactuals. In the context of STIF, it is possible to use nowcasting to assess the possible counterfactual realization of disease in an incidence that would have been evidenced with no NPI. Classic lagged dependency spatio-temporal IF models are discussed, and the importance of the ST component in nowcasting is assessed. Real examples of lockdowns for COVID-19 in two US states during 2020 and 2021 are provided. The degeneracy in prediction over longer time periods is highlighted, and the wide confidence intervals characterize the forecasts. For SC, the early and short lockdown contrasted with the longer NJ intervention. The approach here demonstrated marked differences in spatio-temporal disparities across counties with respect to an adherence to counterfactual predictions.

Nowcasting India's Quarterly GDP Growth: A Factor-Augmented Time-Varying Coefficient Regression Model (FA-TVCRM).

Governments, central banks, private firms and others need high frequency information on the state of the economy for their decision making. However, a key indicator like GDP is only available quarterly and that too with a lag. Hence decision makers use high frequency daily, weekly or monthly information to project GDP growth in a given quarter. This method, known as nowcasting, started out in advanced country central banks using bridge models. Nowcasting is now based on more advanced techniques, mostly dynamic factor models. In this paper we use a novel approach, a Factor Augmented Time Varying Coefficient Regression (FA-TVCR) model, which allows us to extract information from a large number of high frequency indicators and at the same time inherently addresses the issue of frequent structural breaks encountered in Indian GDP growth. One specification of the FA-TVCR model is estimated using 19 variables available for a long period starting in 2007-08:Q1. Another specification estimates the model using a larger set of 28 indicators available for a shorter period starting in 2015-16:Q1. Comparing our model with two alternative models, we find that the FA-TVCR model outperforms a Dynamic Factor Model (DFM) model and a univariate Autoregressive Integrated Moving Average (ARIMA) model in terms of both in-sample and out-of-sample Root Mean Square Error (RMSE). Further, comparing the predictive power of the three models using the Diebold-Mariano test, we find that FA-TVCR model outperforms DFM consistently. In terms of out-of-sample forecast accuracy both the FA-TVCR model and the ARIMA model have the same predictive accuracy under normal conditions. However, the FA-TVCR model outperforms the ARIMA model when applied for nowcasting in periods of major shocks like the Covid-19 shock of 2020-21.