Semiparametric Specification Research Articles

Welfare implications of a tax on electricity: A semi-parametric specification of the incomplete EASI demand system

We perform a welfare analysis due to a tax on electricity consumption based on the incomplete exact affine Stone index (EASI) model using a novel data set in the Colombian economy. We provide a novel inferential framework based on a non-parametric specification of the stochastic errors using Dirichlet processes mixtures that allows handling non-normal errors, gaining efficiency, and taking into account, microeconomic restrictions, censoring, simultaneous endogeneity and non-linearity. We find that there is a 95% probability that the equivalent variation of the representative household is between US¢34.1 and US¢34.3, given an approximately 0.8% tariff increase (US¢0.12 per kWh). In addition, we observe that the welfare loss of the representative household of the lowest socioeconomic characteristics is approximately twice the loss of the representative household of the highest socioeconomic characteristics.

Spatial Income Inequality, Convergence, and Regional Development in a Lower Middle‐Income Country: Satellite Evidence from the Philippines

This paper examines spatial income inequality and convergence in the Philippines—a lower middle‐income country with historically high inequality—and its course over the process of economic development. Combining higher quality nighttime lights (NTLs) with gridded population data, I construct subnational measures of spatial inequality for the 17 administrative regions in the country in the period 2000–2020. Using this unique dataset, I first document the tremendous improvement in income disparities over the last two decades. Income per capita across provinces has converged rapidly, and income dispersion within administrative regions has narrowed markedly. Then, I uncover a U‐shaped relationship between spatial inequality and economic development, which is robust; across alternative measures of inequality; to the outlier effects of highly urbanized cities; across parametric and semiparametric specifications; to business‐cycle effects; and to persistence of spatial inequality. Finally, I confirm that structural transformation acts as a transmission channel of this U‐shaped link.

More Is Better, Or Not? An Empirical Analysis of Buyer Preferences for Variety on the E-Market

This paper examines the effect of the number of sellers of a good on buyers’ demand, using an extensive dataset from an e-commerce platform (PriceMinister.com) in France. Accounting for seller and product characteristics constitutes variety amongst the same product. Although buyers may prefer a wide variety, it can introduce a search cost. Using a flexible semiparametric specification, I find that the demand of buyers does not monotonically increase in relation to the number of sellers (variety), contrary to assumptions in the literature. I illustrate the consequences of misspecification of these network effects through a counterfactual simulation.

Highly robust causal semiparametric U-statistic with applications in biomedical studies.

With our increased ability to capture large data, causal inference has received renewed attention and is playing an ever-important role in biomedicine and economics. However, one major methodological hurdle is that existing methods rely on many unverifiable model assumptions. Thus robust modeling is a critically important approach complementary to sensitivity analysis, where it compares results under various model assumptions. The more robust a method is with respect to model assumptions, the more worthy it is. The doubly robust estimator (DRE) is a significant advance in this direction. However, in practice, many outcome measures are functionals of multiple distributions, and so are the associated estimands, which can only be estimated via U-statistics. Thus most existing DREs do not apply. This article proposes a broad class of highly robust U-statistic estimators (HREs), which use semiparametric specifications for both the propensity score and outcome models in constructing the U-statistic. Thus, the HRE is more robust than the existing DREs. We derive comprehensive asymptotic properties of the proposed estimators and perform extensive simulation studies to evaluate their finite sample performance and compare them with the corresponding parametric U-statistics and the naive estimators, which show significant advantages. Then we apply the method to analyze a clinical trial from the AIDS Clinical Trials Group.

Interactive R&D spillovers: An estimation strategy based on forecasting-driven model selection

This paper proposes an estimation strategy that exploits recent non-parametric panel data methods that allow for a multifactor error structure and extends a recently proposed data-driven model-selection procedure, which has its roots in cross validation and aims to test whether two competing approximate models are equivalent in terms of their expected true error. We extend this procedure to a large panel data framework by using moving block bootstrap resampling techniques in order to preserve cross-sectional dependence in the bootstrapped samples. Such an estimation strategy is illustrated by revisiting an analysis of international technology diffusion. Model selection procedures clearly conclude in the superiority of a fully non-parametric (non-additive) specification over parametric and even semi-parametric (additive) specifications. This work also refines previous results by showing threshold effects, non-linearities, and interactions that are obscured in parametric specifications and which have relevant implications for policy.

Identification of semiparametric model coefficients, with an application to collective households

We prove identification of coefficients for a set of semiparametric specifications that are related to multiple index models. Potential applications of these results include models of observed heterogeneity in production functions and in consumer demand systems. We then generalize these results to identify a class of collective household consumption models. We extend the existing literature by proving point identification, rather than the weaker generic identification, of all the features of the collective household model, including price effects. We estimate the model using Japanese consumption data, and find substantial variation in resource shares and indifference scales across households of different sizes.

Semiparametric Bayes instrumental variable estimation with many weak instruments

We develop a new semiparametric Bayes instrumental variables estimation method. We employ the form of the regression function of the first‐stage equation, and the disturbances are modeled nonparametrically to achieve better predictive power of the endogenous variables, whereas we use parametric formulation in the second‐stage equation, which is of interest in inference. Our simulation studies show that under small sample sizes, the proposed method obtains more efficient estimates and very precise credible intervals compared with existing IV methods with smaller mean squared error. We applied our proposed method to a Mendelian randomization dataset where a large number of instruments are available and semiparametric specification is appropriate. We obtained statistically significant results that cannot be obtained by the existing methods, including standard Bayesian IV.

Characterizing Life-Cycle Dynamics of Annual Days of Work,Wages, and Cross-Covariances

This paper investigates the dispersions in days worked and wages by adapting a novel semi-parametric specification that minimizes assumptions about life-cycle labor income dynamics. Data for Italy shows a substantial increase in income inequality after age 50 for males over the time span from 1985 to 2012, which is remarkably driven by the variations in days worked rather than wages. Results show that this increase is determined by permanent changes in the number of days worked. I also introduce an empirical strategy to decompose the cross-covariances of wages and working days to quantify the permanent and transitory responses of days worked to wage shocks. A one-percent increase in permanent wages increases the permanent days worked by 0.8% at the age of 28, while this increase is about 0.3% at the age of 55. Despite the strong reaction of days of work to wage shocks early in careers, the correlation coefficients are small, indicating that only a small share of variation in permanent days worked – which shapes the permanent income inequality – is explained by the changes in wages.

Positivity properties of the ARFIMA[formula omitted] specifications and credibility analysis of frequency risks

Allowing for the seniority of claims and of risk exposure in the prediction of frequency risks necessitates dynamic random effects in Poisson mixtures. Non-life insurance data show evidence of long memory in stationary random effects. This paper proves that the ARFIMA(0,d,0) mixtures of Poisson distributions ensure nonnegative credibilities per period in the affine prediction of frequency risks. This is true regardless of the risk exposure. This property is maintained if the random effect is the product of a time-invariant component (which provides the highest level of memory in the data) and of a component that follows an ARFIMA(0,d,0) specification. The proof uses approximations of the ARFIMA(0,d,0) time series by AR(p) time series, which result from truncations of the filtering equations that define the former ones. Every given ARFIMA(0,d,0) specification inherits the positivity properties of the truncations because the supremum of the spectral densities of these truncations is integrable on the frequency domain. These semiparametric specifications are easily estimated from longitudinal count data, with the generalized method of moments.

Directional technology distance functions through duality

I provide an estimation procedure for directional technology distance functions (DTDF) through duality to avoid the unnecessarily often-used restrictive assumption that the DTDF is quadratic. I use a semi-parametric specification which avoids the imposition of restrictive functional forms on the profit function. Through the new procedure, one can obtain estimates of the distance function, its derivatives with respect to directions at observed points, as well as estimates of profit inefficiency. The dual formulation is estimated using the Bayesian version of Generalized Method of Moments of Gallant et al. (2017). An empirical application to U.S. banks reveals important aspects of technical as well as allocative inefficiency.

The impact of financial inclusion on consumers saving and borrowing behaviours: A retrospective cross-sectional evidence from the UAE and the USA

This paper highlights the impact of financial inclusion on borrowing and saving decisions in the United States (USA) and the United Arab Emirates (UAE). It does so, using data from the 2014 Global Findex database, and an empirical strategy that contrasts fully parametric and semi-parametric specifications of the saving and borrowing functions. The semi-parametric bivariate probit specification is found to better describe the two processes of saving and borrowing, with a 12.3% correlation coefficient. Although no significant difference is found in the likelihood of borrowing in the two economies, we found however, that consumers in the US are 31.4% more likely to save than their UAE counterparts. In addition, and in line with the Permanent Income Hypothesis, the results do not reveal any income based saving nor borrowing gradient in the two economies. Conversely, a gender based saving (12.4%) and borrowing (13.8%) inequality in favour of the male gender in found to prevail. Furthermore, access to a bank account and a debit card, companies' policy of direct salary deposits, and government transfer programs are all financial inclusion strategies that are found to significantly raise the likelihood of saving and borrowing in the two nations.

Arriving at a decision: A semi-parametric approach to institutional birth choice in India

The Multinomial Logit (MNL) model is popular, but a semi-parametric specification of its link/utility function has seldom been used in empirical applications. This is primarily because of the resource intensive nature of semi-parametric estimation. In this paper we propose and implement a parallel computation algorithm to estimate the semi-parametric kernel MNL model. This algorithm reduces model estimation time by a factor of 2–10, depending on the size of the dataset and the available resources for computation. These computational gains make the estimation of this model feasible for large datasets. Additionally, using a Monte Carlo study we show that the kernel MNL outperforms the traditional linear MNL model in terms of fit and predicted choice probabilities. We demonstrate how kernel-based specification can unearth important heterogeneities in the effect of covariates through an empirical exercise. We use data from a nationally representative household survey (N = 157,804) to analyze the factors associated with institutional births (as opposed to home births) in India. Our revealed-preference results indicate that maternal education, household assets, distance to formal health facility, and birth order play an essential role in determining birth location choice. Although the directions of impact are similar across both the linear and the kernel MNL specifications, there are significant differences in the marginal effects of different factors across the two models. These differences, which arise due to the flexibility afforded by the semi-parametric specification, potentially bring additional nuance to policy discussions.

Arriving at a Decision: A Semi-Parametric Approach to Institutional Birth Choice in India

We use nationally representative household survey data (N=157,804) to analyze the factors determining institutional births in India using the kernel multinomial logit model -- a state-of-the-art semi-parametric specification. In this study, we propose and implement a parallel computation algorithm to estimate the resource-intensive kernel MNL. Furthermore, we layout an empirical procedure to analyze the semi-parametric estimates and also to compare them with those of a conventional linear MNL model. Results from both models show that maternal education, asset level, distance to a formal health facility, and birth order play an essential role in determining birth location choice. However, we show significant differences in the marginal effect estimates derived from the two models. In fact, using a Monte Carlo study we show that the kernel MNL outperforms the linear MNL model. We argue that the flexibility offered by kernel MNL is insightful for policymakers.

A Bayesian Semiparametric Approach for Trend–Seasonal Interaction: an Application to Migration Forecasts

SummaryWe model complex trend–seasonal interactions within a Bayesian framework. The contribution divides into two parts. First, it proves, via a set of simulations, that a semiparametric specification of the interplay between the seasonal cycle and the global time trend outperforms parametric and non-parametric alternatives when the seasonal behaviour is represented by Fourier series of order bigger than 1. Second, the paper uses a Bayesian framework to forecast Swiss immigration, merging the simulations’ outcome with a set of priors derived from alternative hypotheses about the future number of incomers. The result is an effective symbiosis between Bayesian probability and semiparametric flexibility that can reconcile past observations with unprecedented expectations.

Multistep quantile forecasts for supply chain and logistics operations: bootstrapping, the GARCH model and quantile regression based approaches

In this paper, we discuss and compare empirically various ways of computing multistep quantile forecasts of demand, with a special emphasis on the use of the quantile regression methodology. Such forecasts constitute a basis for production planning and inventory management in logistics systems optimized according to the cycle service level approach. Different econometric methods and models are considered: direct and iterated computations, linear and nonlinear (GARCH) models, simulation and non-simulation based procedures and parametric as well as semiparametric specifications. These methods are applied to compute multiperiod quantile forecasts of the monthly microeconomic time series from the popular M3 competition database. According to various accuracy measures for quantile predictions, the best procedures are based on simulation techniques using predictive distributions generated by either the quantile regression methodology combined with random draws from the uniform distribution or parametric and nonparametric bootstrap techniques. These methods lead to large reductions in the total costs of logistics systems as compared with non-simulation based procedures. For example, in the case of forecasting 12 months ahead, relatively short time series and a high cycle service level, the quantile regression based simulation approach reduces the average supply chain cost per unit of output by about 70–85%. At the shortest horizons, the GARCH model should be seriously considered among the preferred forecasting solutions for production and inventory planning.

Private monetary transfers between households: Who is helped and by whom?

We propose using a system of three equations with binary responses to explore the determinants of a household receiving a private monetary transfer from three different types of informal lenders in Italy. First, we observe that a semi-parametric specification of the system of equations is preferred to a fully parametric modelling approach. Second, we detect the existence of an error correlation structure that characterises the dependence of informal lenders after accounting for a number of observable covariates. Third, we find that the network of close family relationships (parents or adult children) represents the main source of an informal transfer, especially for households in debt to financial intermediaries or who are in arrears with payments and whose household head is unemployed or in poor health. Finally, we propose estimating the entity of monetary transfer in terms of expected value using a simulation that combines the joint response probabilities obtained from the system of equations with the empirical distribution of monetary transfers.

Identification and estimation of semi‐parametric censored dynamic panel data models of short time periods

Summary In this paper, we present a semi-parametric identification and estimation method for censored dynamic panel data models of short time periods and their average partial effects with only two periods of data. The proposed method transforms the semi-parametric specification of censored dynamic panel data models into a parametric family of distribution functions of observables without specifying the distribution of the initial condition. Then the censored dynamic panel data models are globally identified under a standard maximum likelihood estimation framework. The identifying assumptions are related to the completeness of the families of known parametric distribution functions corresponding to censored dynamic panel data models. Dynamic tobit models and two-part dynamic regression models satisfy the key assumptions. We propose a sieve maximum likelihood estimator and we investigate the finite sample properties of these sieve-based estimators using Monte Carlo analysis. Our empirical application using the Medical Expenditure Panel Survey shows that individuals consume more health care when their incomes increase, after controlling for past health expenditures.

Is Productivity Growth in Electricity Distribution Negative? An Empirical Analysis Using Ontario Data

Electricity industries are experiencing upward cost pressures in many parts of the world. This paper focuses on productivity trends in electricity distribution. We apply two methodologies for estimating productivity growth—an index based approach, and an econometric cost based approach—to data on 73 Ontario distributors for the period 2002 to 2012. The resulting productivity growth estimates are approximately -1% per year, suggesting a reversal of the positive estimates that have generally been reported in previous periods. We implement flexible semi-parametric specifications to assess the robustness of these conclusions and discuss the use of such statistical analyses for calibrating productivity and relative efficiency within a price-cap framework.

Long-Term Quantile Forecasting for Supply Chain and Logistic Operations

In this paper, we discuss and compare empirically various ways of computing multistep quantile forecasts of demand, with a special emphasis on the use of the quantile regression methodology. Such forecasts constitute a basis for production planning and inventory management in logistic systems optimized according to the cycle service level approach. Different econometric methods and models are considered: direct and iterated computations, linear and nonlinear (GARCH) models, simulation and non-simulation based procedures and parametric as well as semiparametric specifications. These methods are applied to compute multiperiod quantile forecasts of the monthly microeconomic time series from the popular M3 competition database. According to various accuracy measures for quantile predictions, the best procedures are based on simulation techniques using predictive distributions generated by either the quantile regression methodology combined with random draws from the uniform distribution or parametric and nonparametric bootstrap techniques. These methods lead to large reductions in the total costs of logistic systems as compared with non-simulation based procedures. For example, in the case of forecasting 12 months ahead, relatively short time series and a high cycle service level, the quantile regression based simulation approach reduces the average supply chain cost per unit of output by about 70%–85%. At the shortest horizons, the GARCH model should be seriously considered among the preferred forecasting solutions for production and inventory planning.

Semiparametric versus parametric hedonic wine price models: an empirical investigation

ABSTRACTIn this article, we conduct formal statistical tests to compare a semiparametric hedonic wine price model with its parametric counterpart using a Canadian data set. The relevant test results turn out to be overwhelmingly in favour of the semiparametric specification. The estimated semiparametric model also provides clear evidence of nonlinearity between wine prices and quality when other wine attributes are controlled for.