|
on Econometrics |
| By: | Niko Hauzenberger; Florian Huber; Karin Klieber; Massimiliano Marcellino |
| Abstract: | We develop Bayesian neural networks (BNNs) that permit to model generic nonlinearities and time variation for (possibly large sets of) macroeconomic and financial variables. From a methodological point of view, we allow for a general specification of networks that can be applied to either dense or sparse datasets, and combines various activation functions, a possibly very large number of neurons, and stochastic volatility (SV) for the error term. From a computational point of view, we develop fast and efficient estimation algorithms for the general BNNs we introduce. From an empirical point of view, we show both with simulated data and with a set of common macro and financial applications that our BNNs can be of practical use, particularly so for observations in the tails of the cross-sectional or time series distributions of the target variables. |
| Date: | 2022–11 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2211.04752&r=ecm |
| By: | Eoghan O'Neill |
| Abstract: | This paper introduces Type I Tobit Bayesian Additive Regression Trees (TOBART-1). Simulation results and applications to real data sets demonstrate that TOBART-1 produces more accurate predictions than competing methods. TOBART-1 provides accurate posterior intervals for the conditional expectation and other quantities of interest. A Dirichlet Process mixture of normal distributions can flexibly model the error term for improved uncertainty quantification. |
| Date: | 2022–11 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2211.07506&r=ecm |
| By: | Clint Harris |
| Abstract: | Many treatment variables used in empirical applications nest multiple unobserved versions of a treatment. I show that instrumental variable (IV) estimands for the effect of a composite treatment are IV-specific weighted averages of effects of unobserved component treatments. Differences between IVs in unobserved component compliance produce differences in IV estimands even without treatment effect heterogeneity. I describe a monotonicity condition under which IV estimands are positively-weighted averages of unobserved component treatment effects. Next, I develop a method that allows instruments that violate this condition to contribute to estimation of treatment effects by allowing them to place nonconvex, outcome-invariant weights on unobserved component treatments across multiple outcomes. Finally, I apply the method to estimate returns to college, finding wage returns that range from 7\% to 30\% over the life cycle. My findings emphasize the importance of leveraging instrumental variables that do not shift individuals between versions of treatment, as well as the importance of policies that encourage students to attend "high-return college" in addition to those that encourage "high-return students" to attend college. |
| Date: | 2022–11 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2211.13132&r=ecm |
| By: | Michael P. Leung; Pantelis Loupos |
| Abstract: | This paper studies nonparametric estimation of treatment and spillover effects using observational data from a single large network. We consider a model of network interference that allows for peer influence in selection into treatment or outcomes but requires influence to decay with network distance. In this setting, the network and covariates of all units can be potential sources of confounding, in contrast to existing work that assumes confounding is limited to a known, low-dimensional function of these objects. To estimate the first-stage nuisance functions of the doubly robust estimator, we propose to use graph neural networks, which are designed to approximate functions of graph-structured inputs. Under our model of interference, we derive primitive conditions for a network analog of approximate sparsity, which provides justification for the use of shallow architectures. |
| Date: | 2022–11 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2211.07823&r=ecm |
| By: | Giovanni Mellace (University of Southern Denmark); Alessandra Pasquini (Bank of Italy) |
| Abstract: | The synthetic control method (SCM) allows estimating the causal effect of an intervention in settings where panel data on a small number of treated and control units are available. We show that the existing SCM, as well as its extensions, can be easily modified to estimate how much of the “total†effect goes through observed causal channels. Our new mediation analysis synthetic control (MASC) method requires additional assumptions that are arguably mild in many settings. We illustrate the implementation of MASC in an empirical application estimating the direct and indirect effects of an anti-smoking intervention (California's Proposition 99). |
| Keywords: | Synthetic Control Method, mediation analysis, causal mechanisms, direct and indirect effects |
| JEL: | C21 C23 C31 C33 |
| Date: | 2022–11 |
| URL: | http://d.repec.org/n?u=RePEc:bdi:wptemi:td_1389_22&r=ecm |
| By: | Peter Arcidiacono; Attila Gyetvai; Arnaud Maurel; Ekaterina S. Jardim |
| Abstract: | This paper applies some of the key insights of dynamic discrete choice models to continuous-time job search models. We propose a novel framework that incorporates preference shocks into search models, resulting in a tight connection between value functions and conditional choice probabilities. Including preference shocks allows us to establish constructive identification of all the model parameters. Our method also makes it possible to estimate rich nonstationary job search models in a simple and tractable way, without having to solve any differential equations. We apply our framework to rich longitudinal data from Hungarian administrative records, allowing for nonstationarities in offer arrival rates, wage offers, and in the flow payoff of unemployment. Longer unemployment durations are associated with substantially worse wage offers and lower offer arrival rates, which results in accepted wages falling over time. |
| JEL: | C59 J62 J64 |
| Date: | 2022–11 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:30655&r=ecm |
| By: | Nguyen, Loc PhD, PostDoc |
| Abstract: | Regression analysis is an important tool in statistical analysis, in which there is a demand of discovering essential independent variables among many other ones, especially in case that there is a huge number of random variables. Extreme bound analysis is a powerful approach to extract such important variables called robust regressors. In this research, I propose a so-called Regressive Expectation Maximization with RObust regressors (REMRO) algorithm as an alternative method beside other probabilistic methods for analyzing robust variables. By the different ideology from other probabilistic methods, REMRO searches for robust regressors forming optimal regression model and sorts them according to descending ordering given their fitness values determined by two proposed concepts of local correlation and global correlation. Local correlation represents sufficient explanatories to possible regressive models and global correlation reflects independence level and stand-alone capacity of regressors. Moreover, REMRO can resist incomplete data because it applies Regressive Expectation Maximization (REM) algorithm into filling missing values by estimated values based on ideology of expectation maximization (EM) algorithm. From experimental results, REMRO is more accurate for modeling numeric regressors than traditional probabilistic methods like Sala-I-Martin method but REMRO cannot be applied in case of nonnumeric regression model yet in this research. |
| Date: | 2022–11–18 |
| URL: | http://d.repec.org/n?u=RePEc:osf:osfxxx:wjnz3&r=ecm |
| By: | Khushboo Surana |
| Abstract: | I present a nonparametric method to empirically identify the degree of heterogeneity in individual preferences. Using revealed preference conditions for rational consumption behaviour, the method estimates interpersonal preference heterogeneity as the distance between individual preference rankings over a finite set of choice alternatives. An application to US consumption data drawn from the Panel Study of Income Dynamics shows that the method yields informative empirical results on the distance-based heterogeneity measure. I further show the usefulness of the method for applied demand analysis through three empirical applications. Specifically, I take the recovered estimates to form groups of individuals with similar preferences. I demonstrate that employing these preference types as separate units of analysis obtains more informative demand predictions, welfare evaluations and detection of functional misspecification in the case of parametric estimation. |
| Keywords: | Preference heterogeneity, Revealed preference analysis, Kemeny distance |
| JEL: | C14 C60 D11 D12 |
| Date: | 2022–11 |
| URL: | http://d.repec.org/n?u=RePEc:yor:yorken:22/09&r=ecm |
| By: | Todd E. Clark; Gergely Ganics; Elmar Mertens |
| Abstract: | We develop a model that permits the estimation of a term structure of both expectations and forecast uncertainty for application to professional forecasts such as the Survey of Professional Forecasters (SPF). Our approach exactly replicates a given data set of predictions from the SPF (or a similar forecast source) without measurement error. Our model captures fixed horizon and fixed-event forecasts, and can accommodate changes in the maximal forecast horizon available from the SPF. The model casts a decomposition of multi-period forecast errors into a sequence of forecast updates that may be partially unobserved, resulting in a multivariate unobserved components model. In our empirical analysis, we provide quarterly term structures of expectations and uncertainty bands. Our preferred specification features stochastic volatility in forecast updates, which improves forecast performance and yields model estimates of forecast uncertainty that vary over time. We conclude by constructing SPF-based fan charts for calendar-year forecasts like those published by the Federal Reserve. |
| Keywords: | Term Structure of Expectations; Uncertainty; Survey Forecasts; Fan Charts |
| JEL: | C53 E37 |
| Date: | 2022–11–23 |
| URL: | http://d.repec.org/n?u=RePEc:fip:fedcwq:95170&r=ecm |
| By: | Steven Campbell; Ting-Kam Leonard Wong |
| Abstract: | Convex PCA, which was introduced by Bigot et al., is a dimension reduction methodology for data with values in a convex subset of a Hilbert space. This setting arises naturally in many applications, including distributional data in the Wasserstein space of an interval, and ranked compositional data under the Aitchison geometry. Our contribution in this paper is threefold. First, we present several new theoretical results including consistency as well as continuity and differentiability of the objective function in the finite dimensional case. Second, we develop a numerical implementation of finite dimensional convex PCA when the convex set is polyhedral, and show that this provides a natural approximation of Wasserstein geodesic PCA. Third, we illustrate our results with two financial applications, namely distributions of stock returns ranked by size and the capital distribution curve, both of which are of independent interest in stochastic portfolio theory. |
| Date: | 2022–11 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2211.02990&r=ecm |
| By: | Katarzyna Kopczewska (Faculty of Economic Sciences, University of Warsaw) |
| Abstract: | This paper is a methodological guide on using machine learning in the spatial context. It provides an overview of the existing spatial toolbox proposed in the literature: unsupervised learning, which deals with clustering of spatial data and supervised learning, which displaces classical spatial econometrics. It shows the potential and traps of using this developing methodology. It catalogues and comments on the usage of spatial clustering methods (for locations and values, separately and jointly) for mapping, bootstrapping, cross-validation, GWR modelling, and density indicators. It shows details of spatial machine learning models, combined with spatial data integration, modelling, model fine-tuning and predictions, to deal with spatial autocorrelation and big data. The paper delineates "already available" and "forthcoming" methods and gives inspirations to transplant modern quantitative methods from other thematic areas to research in regional science. |
| Keywords: | spatial machine learning, clustering, spatial covariates, spatial cross-validation, spatial autocorrelation |
| JEL: | C31 R10 C49 |
| Date: | 2021 |
| URL: | http://d.repec.org/n?u=RePEc:war:wpaper:2021-16&r=ecm |
| By: | Duo Qin (Department of Economics, SOAS University of London) |
| Abstract: | Infiltration of machine learning (ML) methods into econometrics has remained relatively slow, compared with their extensive applications in many other disciplines. The bottleneck is traced to two key factors – a communal nescience of the theoretical foundation of ML and an outdated probability foundation. The present study ventures on an overhaul of the probability approach by Haavelmo (1944) in light of ML theories of learnibility, centred upon the notion of probably approximately correct (PAC) learning. The study argues for a reorientation of the probability approach towards assisting decision making for model learning and selection purposes. The first part of the study is presented here. |
| Keywords: | probability; uncertainty; machine learning; hypothesis testing; knowledge; representation |
| JEL: | C10 C18 B40 |
| Date: | 2022–03 |
| URL: | http://d.repec.org/n?u=RePEc:soa:wpaper:249&r=ecm |