|
on Econometrics |
| By: | Bruno Bosco; Paolo Maranzano |
| Abstract: | Difference-in-Differences (DiD) is a useful statistical technique employed to estimate the effects that exogenous events can have on the outcome of some response variables. The latter are obtained from a random sample of treated units (i.e. units exposed to the event) ideally drawn from an infinite population. The comparable random sample/s of untreated units serve as control comparison group/s. The event is termed “treatment†, but it could be equally termed “causal factor†to emphasise that with DiD we are not estimating a mere statistical association among phenomena. With DiD we try to evaluate whether a precise causal link between causes and effects –defined according to a model based on a proper identification of the relationship among variables– is actually consistent with the data, and to estimate how intensive and statistically robust the causal-effect link actually is. This Guide will present the DiD techniques starting from the very basic methods used to estimate the Average Treatment Effect upon Treated (ATET) originally developed for the 2–period and 2–group case and covers many of the issues that have recently emerged in the multi units and multi period context. Particular attention will be devoted to the correct definition of the identification process of the causal-effect relationship in the multi period case, namely to the parallel trend and to the no anticipation assumption. Some space will be devoted to the developments associated to the techniques employed with either treatment homogeneity or treatment heterogeneity. Also, extensions of the DiD estimators accounting for complex data structures are discussed. The Guide includes a brief presentation of some policy-oriented applications of DiD. Areas covered are income taxation, migration, regulation and environment management. |
| Keywords: | Difference-in-Differences (DID); Guide for causal inference; Applied and empirical economics; Treatment and control; Extensions of the DID estimator. |
| JEL: | C23 C50 C54 D04 E6 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:mib:wpaper:549 |
| By: | NAKAJIMA, Jouchi |
| Abstract: | This study discusses a general approach to dynamic modeling using the local projection (LP) method. Previous studies have proposed time-varying (TV) parameters in LPs; however, they did not address possible variations in error variances. Overlooking this could introduce significant bias in the estimate of the TV parameter, and consequently, the estimated impulse response. We develop an estimation strategy for LPs with stochastic volatility (SV) and illustrate the importance of SV inclusion using simulated data. Application to a topical macroeconomic time-series analysis illustrates the benefits of the proposed approach in terms of improved predictions. |
| Keywords: | Local projections, Time-varying parameters, Stochastic volatility |
| JEL: | C15 C22 C53 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:hit:hituec:761 |
| By: | Dimitris Korobilis (Adam Smith Business School, University of Glasgow, UK; BI Norwegian Business School, Norway; Rimini Centre for Economic Analysis); Emmanuel C. Mamatzakis (Birkbeck Business School, UK); Vasileios Pappas (Surrey Business School, UK) |
| Abstract: | This paper introduces a novel econometric framework for identifying and modeling bank business models (BBMs), which dynamically evolve in response to changing financial and economic conditions. Building on the stochastic frontier literature, we extend the traditional cost-efficiency models by decomposing inefficiency into persistent and transient components. We propose a Bayesian nonparametric approach that adapts to the data through an infinite mixture model with predictor-dependent clustering, enabling a flexible classification of banks into distinct business models. Our method, based on the Logit Stick-Breaking Process (LSBP), provides a data-driven way to capture the heterogeneity in bank strategies, allowing for dynamic transitions between business models over time. This model offers a significant advancement over existing parametric and kernel-based approaches by combining the scalability of nonparametric methods with efficient computational routines. We apply the model to a dataset of European banks and identify four distinct business model clusters, providing novel insights into the evolution of bank performance and efficiency. Our findings contribute to the growing literature on the identification and measurement of bank business models, offering valuable implications for policy and regulatory frameworks. |
| Keywords: | Stochastic Frontier Analysis, Bayesian Nonparametrics, Logit Stick-Breaking Process, Dynamic Clustering, Cost Efficiency |
| JEL: | C11 C14 C23 D24 G21 G28 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:rim:rimwps:25-03 |
| By: | Yufei Li (King's Business School, King's College London,); Liudas Giraitis (School of Economics and Finance, Queen Mary University of London); Genaro Sucarrat (BI Norwegian Business School) |
| Abstract: | The presence of autocorrelated nancial returns has major implications for investment decisions.Unsurprisingly, therefore, numerous studies have sought to shed light on whether returns areautocorrelated or not, to what extent, and when. Standard tests for autocorrelation rely onthe assumption of strict stationarity of returns, possibly after a suitable transformation. Recentstudies, however, reveal that intraday nancial returns are often characterised by a subtle formof non-stationarity that cannot be transformed away, namely non-stationary periodicity in thezero-process. Here, we propose tests for autocorrelation that are valid under this (and otherforms) of non-stationarity. The tests are simple to implement, and well-sized and powerful asdocumented in our Monte Carlo simulations. Next, in a study of the intraday returns of stocksand exchange rates, our robust tests document that returns are rarely autocorrelated. This is insharp contrast to the standard benchmark test, which spuriously detects a substantial numberof autocorrelations. Moreover, stability analyses with our robust tests suggest the signi canceof the autocorrelations is short-lived and very erratic. So it is unclear whether the short-livedautocorrelations can be used to inform decision-making. |
| Keywords: | robust correlation testing, zero-process, non-stationary periodicity |
| JEL: | C01 C12 C22 |
| Date: | 2024–02–26 |
| URL: | https://d.repec.org/n?u=RePEc:qmw:qmwecw:987 |
| By: | Santiago Burone;; Lukas Leitner; |
| Abstract: | Willingness to pay (WTP) has become an important tool in economic analysis, despite the difficulty to obtain reliable estimates. This paper investigates the occurrence of starting point bias when eliciting WTP for health, a domain where this phenomenon has received limited attention, and illustrates its effect on equivalent consumption, a preference-based well-being measure. In an online experiment, three experimental groups responded to two dichotomous choice questions, with varying initial bids. The treatment groups then provided exact estimates for their WTP in an open-ended question. We find strong evidence for the existence of the bias using both non-parametric and parametric tests, and estimate a sizeable overall effect. Different parametric specifications yield point estimates between 29 and 43 percent for the first bid, whereas the effect of the second bid, which we estimate using an instrumental variable approach, is not statistically different from zero. We propose two ex post approaches to address this effect when using WTP data for interpersonal well-being comparisons. Although the percentage of rankings reversals is relatively small across all feasible comparisons, it becomes notable when examining comparisons for individuals within the same consumption deciles. |
| Date: | 2025–02 |
| URL: | https://d.repec.org/n?u=RePEc:hdl:wpaper:2501 |