Oxford Bulletin of Economics and Statistics最新文献

We reconcile dense and sparse modelling by exploiting the positive aspects of both. We employ a high-dimensional, approximate static factor model and assume the idiosyncratic term follows a sparse vector autoregressive model (VAR). The estimation is articulated in two steps: (i) factors and loadings are estimated via principal component analysis (PCA); (ii) a sparse VAR is estimated via the lasso on the estimated idiosyncratic components from (i). Step (ii) allows to model cross-sectional and time dependence left after the factors estimation. We prove the consistency of this approach as the time and cross-sectional dimensions diverge. In (ii), sparsity is allowed to be very general: approximate, row-wise, and growing with the sample size. However, the estimation error of (i) needs to be accounted for. Instead of simply plugging-in the standard rates derived for the PCA estimation of the factors in (i), we derive a refined expression of the error, which enables us to derive tighter rates for the lasso in (ii). We discuss applications on forecasting & factor-augmented regression and present an empirical application on macroeconomic forecasting using the Federal Reserve Economic Data - Monthly Database (FRED-MD).

In this paper, we propose a two-step procedure based on the group LASSO estimator in combination with a backward elimination algorithm to detect multiple structural breaks in linear regressions with multivariate responses. Applying the two-step estimator, we jointly detect the number and location of structural breaks and provide consistent estimates of the coefficients. Our framework is flexible enough to allow for a mix of integrated and stationary regressors, as well as deterministic terms. Using simulation experiments, we show that the proposed two-step estimator performs competitively against the likelihood-based approach in finite samples. However, the two-step estimator is computationally much more efficient. An economic application to the identification of structural breaks in the term structure of interest rates illustrates this methodology.

We investigate the health outcome effects of a reduction in years of schooling in Egypt in 1988, a policy change that moves in the opposite direction in relation to the extant literature. We exploit this policy change as a natural experiment and employ a fuzzy regression discontinuity design to investigate a range of objectively measured health outcomes and behaviours. Despite the policy's adverse effect on students' years of schooling and ability to complete educational milestones, there is no statistical effect on any health outcomes. Results also suggest that the reduction in years of schooling had no effect on labour market outcomes, thus providing a rationale for the lack of effect on health outcomes. While increasing education often leads to improved health outcomes, our findings suggest that decreasing education does not necessarily result in poorer health outcomes.

We propose and apply a framework that analyses the distributional impact of episodes of overall health growth or decline. We use local smoothing to model the relationship between health outcomes and relative income rank and we introduce Health Growth Incidence Curves (HGIC) to graph the rate of change of health outcomes across the income distribution. We illustrate our framework with an analysis of the distributional dynamics of health and well-being among those aged 60–85 in the years 2004 and 2015 using pooled SHARE data for selected European countries. Overall we observe a pattern that favours individuals with higher socioeconomic status, with a pro-rich distribution of improvements in outcomes. Only cognitive capacity growth has been uniform across income groups.

A rubella infection in early pregnancy poses a significant risk of damage to the foetus. In this paper, we examine the later-life impact of a rubella outbreak that occurred in Ireland in 1956. Matching the outcomes of individuals born in 1954–1957 in the 2016 Irish Census of Population to the county-level rubella incidence rate that was prevailing when respondents were in utero in early pregnancy, we find that one extra rubella case per 10,000 population is associated with between 0.4% and 1.2% point increases in the probability of having lower levels of educational attainment, being in poor health and having a disability in later life.

Using a novel subject pool from Botswana, we experimentally examine the impact of task difficulty on willingness to compete, an economic indicator that may partially explain gender disparities in the labour market. Our experiment closely follows the paradigm of Niederle and Vesterlund (2007, QJE) and is concerned with treatments where a task is either relatively easy or more difficult. We report evidence of a larger gender gap in willingness to compete when the math task is easy in comparison to when the math task is more difficult. Our results have implications for future experiments examining gender differences.

We study differences in teacher evaluations of student performance relative to those measured by test scores. While much literature is concerned with estimating various types of teacher biases, we show conceptually that there is no single ‘teacher bias’ effect. Even if teachers have no group bias, teacher evaluation differences by group masystematically deviate from test score differences if the distribution of test scores differs across groups. Commonly used approaches are not equivalent and can lead to different conclusions as they target different estimands. We demonstrate our findings using Monte Carlo simulations and, using two recent UK cohort surveys, we show that these conceptual issues matter in practice when we evaluate whether teachers are likely to over-estimate female performance in English. Finally, we use the methods to examine an issue of substantive importance, gender differences in teacher perceptions in comparative advantage in English relative to mathematics. Our findings suggest that it is unlikely that teacher misperceptions of comparative advantage by gender are an important cause of the gender gap in STEM.