|
on Computational Economics |
|
Issue of 2022‒07‒25
thirteen papers chosen by |
| By: | Buckmann, Marcus (Bank of England); Joseph, Andreas (Bank of England) |
| Abstract: | We propose a generic workflow for the use of machine learning models to inform decision making and to communicate modelling results with stakeholders. It involves three steps: (1) a comparative model evaluation, (2) a feature importance analysis and (3) statistical inference based on Shapley value decompositions. We discuss the different steps of the workflow in detail and demonstrate each by forecasting changes in US unemployment one year ahead using the well-established FRED-MD dataset. We find that universal function approximators from the machine learning literature, including gradient boosting and artificial neural networks, outperform more conventional linear models. This better performance is associated with greater flexibility, allowing the machine learning models to account for time-varying and nonlinear relationships in the data generating process. The Shapley value decomposition identifies economically meaningful nonlinearities learned by the models. Shapley regressions for statistical inference on machine learning models enable us to assess and communicate variable importance akin to conventional econometric approaches. While we also explore high-dimensional models, our findings suggest that the best trade-off between interpretability and performance of the models is achieved when a small set of variables is selected by domain experts. |
| Keywords: | machine learning; model interpretability; forecasting; unemployment; Shapley values |
| JEL: | C14 C38 C45 C52 C53 C71 E24 |
| Date: | 2022–06–01 |
| URL: | http://d.repec.org/n?u=RePEc:boe:boeewp:0984&r= |
| By: | Callum Tilbury |
| Abstract: | Agent-based computational macroeconomics is a field with a rich academic history, yet one which has struggled to enter mainstream policy design toolboxes, plagued by the challenges associated with representing a complex and dynamic reality. The field of Reinforcement Learning (RL), too, has a rich history, and has recently been at the centre of several exponential developments. Modern RL implementations have been able to achieve unprecedented levels of sophistication, handling previously-unthinkable degrees of complexity. This review surveys the historical barriers of classical agent-based techniques in macroeconomic modelling, and contemplates whether recent developments in RL can overcome any of them. |
| Date: | 2022–06 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2206.08781&r= |
| By: | Farmer, J. Doyne; Axtell, Robert L. |
| Abstract: | Agent-based modeling (ABM) is a novel computational methodology for representing the behavior of individuals in order to study social phenomena. Its use is rapidly growing in many fields. We review ABM in economics and finance and highlight how it can be used to relax conventional assumptions in standard economic models. In economics, ABM has enriched our understanding of markets, industrial organization, labor, macro, development, environmental and resource economics, as well as policy. In financial markets, substantial accomplishments include understanding clustered volatility, market impact, systemic risk and housing markets. We present a vision for how ABMs might be used in the future to build more realistic models of the economy and review some of hurdles that must be overcome to achieve this. |
| Keywords: | agent-based computational economics, multi-agent systems, agent-based modeling and simulation, distributed systems |
| JEL: | C00 C63 C69 D00 E00 G00 |
| Date: | 2022–06 |
| URL: | http://d.repec.org/n?u=RePEc:amz:wpaper:2022-10&r= |
| By: | Kevin Kamm; Michelle Muniz |
| Abstract: | In this paper, we introduce a novel methodology to model rating transitions with a stochastic process. To introduce stochastic processes, whose values are valid rating matrices, we noticed the geometric properties of stochastic matrices and its link to matrix Lie groups. We give a gentle introduction to this topic and demonstrate how It\^o-SDEs in R will generate the desired model for rating transitions. To calibrate the rating model to historical data, we use a Deep-Neural-Network (DNN) called TimeGAN to learn the features of a time series of historical rating matrices. Then, we use this DNN to generate synthetic rating transition matrices. Afterwards, we fit the moments of the generated rating matrices and the rating process at specific time points, which results in a good fit. After calibration, we discuss the quality of the calibrated rating transition process by examining some properties that a time series of rating matrices should satisfy, and we will see that this geometric approach works very well. |
| Date: | 2022–05 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2205.15699&r= |
| By: | Karim Barhoumi; Jiaxiong Yao; Tara Iyer; Seung Mo Choi; Jiakun Li; Franck Ouattara; Mr. Andrew J Tiffin |
| Abstract: | The COVID-19 crisis has had a tremendous economic impact for all countries. Yet, assessing the full impact of the crisis has been frequently hampered by the delayed publication of official GDP statistics in several emerging market and developing economies. This paper outlines a machine-learning framework that helps track economic activity in real time for these economies. As illustrative examples, the framework is applied to selected sub-Saharan African economies. The framework is able to provide timely information on economic activity more swiftly than official statistics. |
| Keywords: | Sub-Saharan Africa; Economic Activity; GDP; Machine Learning; Nowcasting; COVID-19; machine learning approach; data sparsity; GDP statistics; crisis in Sub-Saharan Africa; learning framework; Oil prices; Real effective exchange rates; Africa; Global |
| Date: | 2022–05–06 |
| URL: | http://d.repec.org/n?u=RePEc:imf:imfwpa:2022/088&r= |
| By: | Kaiser, Caspar; Oparina, Ekaterina; Gentile, Niccolò; Tkatchenko, Alexandre; Clark, Andrew E.; De Neve, Jan-Emmanuel; D’Ambrosio, Conchita |
| Abstract: | There is a vast literature on the determinants of subjective wellbeing. Yet, standard regression models explain little variation in wellbeing. We here use data from Germany, the UK, and the US to assess the potential of Machine Learning (ML) to help us better understand wellbeing. Compared to traditional models, ML approaches provide moderate improvements in predictive performance. Drastically expanding the set of explanatory variables doubles our predictive ability across approaches on unseen data. The variables identified as important by ML – material conditions, health, social relations – are similar to those previously identified. Our data-driven ML results therefore validate previous conventional approaches. |
| Date: | 2022–06 |
| URL: | http://d.repec.org/n?u=RePEc:amz:wpaper:2022-11&r= |
| By: | Sara B. Heller; Benjamin Jakubowski; Zubin Jelveh; Max Kapustin |
| Abstract: | This paper shows that shootings are predictable enough to be preventable. Using arrest and victimization records for almost 644,000 people from the Chicago Police Department, we train a machine learning model to predict the risk of being shot in the next 18 months. We address central concerns about police data and algorithmic bias by predicting shooting victimization rather than arrest, which we show accurately captures risk differences across demographic groups despite bias in the predictors. Out-of-sample accuracy is strikingly high: of the 500 people with the highest predicted risk, 13 percent are shot within 18 months, a rate 130 times higher than the average Chicagoan. Although Black male victims more often have enough police contact to generate predictions, those predictions are not, on average, inflated; the demographic composition of predicted and actual shooting victims is almost identical. There are legal, ethical, and practical barriers to using these predictions to target law enforcement. But using them to target social services could have enormous preventive benefits: predictive accuracy among the top 500 people justifies spending up to $123,500 per person for an intervention that could cut their risk of being shot in half. |
| JEL: | C53 H75 I14 K42 |
| Date: | 2022–06 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:30170&r= |
| By: | Diana Gabrielyan; Lenno Uusküla |
| Abstract: | We extract measures of inflation expectations from online news to build real interest rates that capture true consumer expectations. The new measure is infused to various Euler consumption models. While benchmark models based on traditional risk-free returns rates fail, models built with novel news-driven inflation expectations indices improve upon benchmark models and result in strong instruments. Our positive findings highlight the role played by the media for consumer expectation formation and allow for the use of such novel data sources for other key macroeconomic relationships. |
| Keywords: | Euler equation, expectations, media, machine learning |
| Date: | 2022 |
| URL: | http://d.repec.org/n?u=RePEc:mtk:febawb:142&r= |
| By: | Naphtal Hakizimana; John Karangwa; Jesse Lastunen; Aimable Mshabimana; Innocente Murasi; Lucie Niyigena; Michael Noble; Gemma Wright |
| Abstract: | This paper assesses the feasibility of developing a tax and benefit microsimulation model in Rwanda. Tax-benefit microsimulation can be used to explore ways in which national development goals can be achieved in a cost-effective manner, and to assess the distributional effects of more comprehensive social security arrangements. |
| Keywords: | Tax-benefit microsimulation, Revenue, Microsimulation, Tax-benefit policy, Tax benefits |
| Date: | 2022 |
| URL: | http://d.repec.org/n?u=RePEc:unu:wpaper:wp-2022-72&r= |
| By: | Daniel Bj\"orkegren; Joshua E. Blumenstock; Samsun Knight |
| Abstract: | When a policy prioritizes one person over another, is it because they benefit more, or because they are preferred? This paper develops a method to uncover the values consistent with observed allocation decisions. We use machine learning methods to estimate how much each individual benefits from an intervention, and then reconcile its allocation with (i) the welfare weights assigned to different people; (ii) heterogeneous treatment effects of the intervention; and (iii) weights on different outcomes. We demonstrate this approach by analyzing Mexico's PROGRESA anti-poverty program. The analysis reveals that while the program prioritized certain subgroups -- such as indigenous households -- the fact that those groups benefited more implies that they were in fact assigned a lower welfare weight. The PROGRESA case illustrates how the method makes it possible to audit existing policies, and to design future policies that better align with values. |
| Date: | 2022–06 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2206.00727&r= |
| By: | Valerio Astuti (Bank of Italy); Marta Crispino (Bank of Italy); Marco Langiulli (Bank of Italy); Juri Marcucci (Bank of Italy) |
| Abstract: | Text data gathered from social media are extremely up-to-date and have a great potential value for economic research. At the same time, they pose some challenges, as they require different statistical methods from the ones used for traditional data. The aim of this paper is to give a critical overview of three of the most common techniques used to extract information from text data: topic modelling, word embedding and sentiment analysis. We apply these methodologies to data collected from Twitter during the COVID-19 pandemic to investigate the influence the pandemic had on the Italian Twitter community and to discover the topics most actively discussed on the platform. Using these techniques of automated textual analysis, we are able to make inferences about the most important subjects covered over time and build real-time daily indicators of the sentiment expressed on this platform. |
| Keywords: | text as data, Twitter, big data, sentiment, Covid-19, topic analysis, word embedding |
| JEL: | C55 C14 C81 L82 |
| Date: | 2022–06 |
| URL: | http://d.repec.org/n?u=RePEc:bdi:opques:qef_692_22&r= |
| By: | Majid Ahmadi; Nathan Durst; Jeff Lachman; John List; Mason List; Noah List; Atom Vayalinkal |
| Abstract: | Recent models and empirical work on network formation emphasize the importance of propinquity in producing strong interpersonal connections. Yet, one might wonder how deep such insights run, as thus far empirical results rely on survey and lab-based evidence. In this study, we examine propinquity in a high-stakes setting of talent allocation: the Major League Baseball (MLB) Draft. We examine draft picks from 2000-2019 across every MLB club of the nearly 30,000 players drafted (from a player pool of more than a million potential draftees). Our findings can be summarized in three parts. First, propinquity is alive and well in our setting, and spans even the latter years of our sample, when higher-level statistical exercises have become the norm rather than the exception. Second, the measured effect size is important, as MLB clubs pay a real cost in terms of inferior talent acquired due to propinquity bias: for example, their draft picks appear in 25 fewer games relative to teams that do not exhibit propinquity bias. Finally, the effect is found to be the most pronounced in later rounds of the draft (after round 15), where the Scouting Director has the greatest latitude. |
| Date: | 2022 |
| URL: | http://d.repec.org/n?u=RePEc:feb:artefa:00758&r= |
| By: | Veli Andirin; Yusuf Neggers; Mehdi Shadmehr; Jesse M. Shapiro |
| Abstract: | We develop a measure of a regime's tolerance for an action by its citizens. We ground our measure in an economic model and apply it to the setting of political protest. In the model, a regime anticipating a protest can take a costly action to repress it. We define the regime's tolerance as the ratio of its cost of repression to its cost of protest. Because an intolerant regime will engage in repression whenever protest is sufficiently likely, a regime's tolerance determines the maximum equilibrium probability of protest. Tolerance can therefore be identified from the distribution of protest probabilities. We construct a novel cross-national database of protest occurrence and protest predictors, and apply machine-learning methods to estimate protest probabilities. We use the estimated protest probabilities to form a measure of tolerance at the country, country-year, and country-month levels. We apply the measure to questions of interest. |
| JEL: | C55 D74 |
| Date: | 2022–06 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:30167&r= |