Differentiating Between Dimensionality and Duration in Multidimensional Measures of Poverty: Methodology with an Application to China

We develop a multidimensional poverty measure that is sensitive to the within‐individual distribution of deprivations across dimensions and time. Our measure combines features from a static multidimensional measure (Alkire and Foster, 2011a ) and a time‐dependent unidimensional measure (Foster, 2009 ). The proposed measure separately identifies—and can therefore be decomposed according to—the proportion of the poverty score attributable to: (i) the concentration of deprivations within periods; (ii) the concentration of deprivations within dimensions. In doing so it allows for a poverty ranking that is robust to assumptions about the trade‐off between the two components. Previous measures have not allowed for the features proposed here due to the inability to calculate the exact contribution of each dimension to overall poverty. We overcome this by adapting to our measure the Shapley decomposition proposed in Shorrocks ( 2013 ) (based on Shapley, 1953 ). The measure is applied to data from China, 2000‐2011.     

Dimensionality reduction in forecasting with temporal hierarchies

Combining forecasts from multiple temporal aggregation levels exploits information differences and mitigates model uncertainty, while reconciliation ensures a unified prediction that supports aligned decisions at different horizons. It can be challenging to estimate the full cross-covariance matrix for a temporal hierarchy, which can easily be of very large dimension, yet it is difficult to know a priori which part of the error structure is most important. To address these issues, we propose to use eigendecomposition for dimensionality reduction when reconciling forecasts to extract as much information as possible from the error structure given the data available. We evaluate the proposed estimator in a simulation study and demonstrate its usefulness through applications to short-term electricity load and financial volatility forecasting. We find that accuracy can be improved uniformly across all aggregation levels, as the estimator achieves state-of-the-art accuracy while being applicable to hierarchies of all sizes.     

A decomposition clustering ensemble learning approach for forecasting foreign exchange rates

A decomposition clustering ensemble (DCE) learning approach is proposed for forecasting foreign exchange rates by integrating the variational mode decomposition (VMD), the self-organizing map (SOM) network, and the kernel extreme learning machine (KELM). First, the exchange rate time series is decomposed into N subcomponents by the VMD method. Second, each subcomponent series is modeled by the KELM. Third, the SOM neural network is introduced to cluster the subcomponent forecasting results of the in-sample dataset to obtain cluster centers. Finally, each cluster's ensemble weight is estimated by another KELM, and the final forecasting results are obtained by the corresponding clusters' ensemble weights. The empirical results illustrate that our proposed DCE learning approach can significantly improve forecasting performance, and statistically outperform some other benchmark models in directional and level forecasting accuracy.     

A general method for comparative dynamics in continuous-time models

This paper develops a general and applicable method for the computation of comparative dynamics in continuous-time perfect foresight models. The key technique in our method is the Jordan decomposition of the Jacobian matrix. This enables us to derive analytical solutions when dealing with high-dimensional systems with repeated eigenvalues. In an application, we compute comparative dynamics of an unanticipated expansionary monetary policy in scenarios with and without repeated eigenvalues. We find that the short-run effects on the social welfare are opposite in the two scenarios, while the long-run effects are similar.     

Has agricultural labor restructuring improved agricultural labor productivity in China? A decomposition approach

This paper presents a framework for analyzing the changes in agricultural labor productivity with regards to the structural, land intensity, and land productivity effects. This approach allows for the residual-free decomposition of data from different levels of aggregation. The logarithmic mean Divisia index was applied for the analysis and a data envelopment analysis model was constructed to identify potential gains in agricultural labor productivity due to the optimization of input use and output production. The proposed approach was applied to the case of China over the period of 1997–2017. Province-level data were used to identify the major driving factors behind agricultural labor productivity change. Land productivity change appeared to be the major source of agricultural labor productivity gains in China. The structural change was rather negligible, suggesting that the reallocation of the agricultural labor force did not add to the agricultural labor productivity growth in China. A frontier analysis indicated that agricultural labor productivity could increase by some 45% on average in case full technical efficiency is achieved.     

Continuous‐time mean–variance portfolio selection: A reinforcement learning framework

We approach the continuous‐time mean–variance portfolio selection with reinforcement learning (RL). The problem is to achieve the best trade‐off between exploration and exploitation, and is formulated as an entropy‐regularized, relaxed stochastic control problem. We prove that the optimal feedback policy for this problem must be Gaussian, with time‐decaying variance. We then prove a policy improvement theorem, based on which we devise an implementable RL algorithm. We find that our algorithm and its variant outperform both traditional and deep neural network based algorithms in our simulation and empirical studies.     

A martingale representation theorem and valuation of defaultable securities

We consider a financial framework with two levels of information: the public information generated by the financial assets, and a larger flow of information that contains additional knowledge about a random time. This random time can represent many economic and financial settings, such as the default time of a firm for credit risk, and the death time of an insured for life insurance. As the random time cannot be seen before its occurrence, the progressive enlargement of filtration seems tailor‐fit to model the larger flow of information that incorporates both the public flow and the information about the random time. In this context, our interest focuses on the following challenges: (a) How to single out the various risks coming from the financial assets, the random time, and their correlations? (b) How these risks interplay and lead to the formation of any risk in the larger flow of information? It is clear that understanding how risks build‐up and interact, when one enlarges the flow of information, is vital for an efficient risk management and derivatives' evaluation in those informational markets. Our answers to these challenges are full and complete no matter what the model for the random time is and no matter how the random time is related to the public flow. In fact, we introduce “pure default” risks, and quantify and classify these risks afterward. Then we elaborate our martingale representation results, which state that any martingale in the large filtration stopped at the random time can be decomposed into orthogonal local martingales (i.e., local martingales whose product remains a local martingale). This constitutes our first principal contribution, while our second contribution consists in evaluating various defaultable securities according to the recovery policy, within our financial setting that encompasses any default model, using a martingale “basis.” Our pricing formulas explain the impact of various recovery policies on securities and determine the types of pure default risk they entail.     

The role of credit supply shocks in pacific alliance countries: A TVP-VAR-SV approach

This paper analyzes the effect of loan supply shocks on the real economic activity of Pacific Alliance countries using a Time-Varying Parameter VAR with Stochastic Volatility (TVP-VAR-SV) model which is identified by sign restrictions. Two main results arise from the analysis. First, loan supply shocks have an important impact on real economic activity in all Pacific Alliance countries: about 1% in Colombia, Mexico, and Peru, and about 0.5% in Chile. Thus, its contribution to business cycle fluctuations is similar to that of aggregate supply shocks and aggregate demand shocks in both stability and slowdown periods. Second, the power of loan supply shocks to affect economic activity do not remain constant over the time and its evolution across periods is heterogeneous among all Pacific Alliance countries. The sensitivity analysis indicates that the results of the model are robust to different priors specifications and to multiple sets of sign restrictions.     

Measuring the gender wage gap—a methodological note

ABSTRACT

We propose to estimate the Blinder-Oaxaca decomposition by a single-equation model augmented with interactions between the group membership and other predictors. The relative importance of predictors on the discriminatory wage gap is examined by the interaction coefficients, which may lead to very different conclusions than the usual percentage calculations using the detailed decomposition method. Comparisons are made between the traditional interpretations and those suggested here using wage data from Finland. The decomposition analysis suggests that the discriminatory male-female wage gap is largely related to work experience, while our preferred model points to the importance of family gap and working industry.     

Performance and informed trading. Comparing foreigners,institutions and individuals in an emerging stock market

Do local institutions, local individuals or foreigners perform better in a small emerging market? We seek answers to this question by examining the investment performance of each group of investors using two unique databases of Colombian stocks. Our analysis serves as an out-of-sample test of previous findings. We find that local institutions have better outcomes than foreigners and local individuals on all performance measures. Furthermore, local individuals perform better than foreigners on trading execution, while foreigners generate better long-term returns. Using a bid-ask spread decomposition, we provide additional evidence that the better performance of institutions is consistent with informational advantages. Our findings reconcile contradictory statements from previous studies.