Design and cost optimization for hierarchical data

In this paper, we consider balanced hierarchical data designs for both one‐sample and two‐sample (two‐treatment) location problems. The variances of the relevant estimates and the powers of the tests strongly depend on the data structure through the variance components at each hierarchical level. Also, the costs of a design may depend on the number of units at different hierarchy levels, and these costs may be different for the two treatments. Finally, the number of units at different levels may be restricted by several constraints. Knowledge of the variance components, the costs at each level, and the constraints allow us to find the optimal design. Solving such problems often requires advanced optimization tools and techniques, which we briefly explain in the paper. We develop new analytical tools for sample size calculations and cost optimization and apply our method to a data set on Baltic herring.     

集成化供应链绩效评价的数据包络分析应用

利用数据包络分析（DEA）处理多输入多输出、无需考虑数据量纲影响的非参数统计估计的＂天然＂优势,并利用主成分分析法对参考绩效指标进行筛选,对多条供应链绩效进行横向评价,不但可以区分被评价者绩效水平的优劣,而且还可以得出处于尚未达到最优水平的被考评者应该优化的方向;同时对有效决策单元进行排序,获得所有被测评者绩效水平的总排序,从而体现了该方法在供应链绩效评价当中应用的优越性.     

Pooling data for the analysis of dynamic marketing systems

Vector autoregressive (VAR) models have become popular in marketing literature for analyzing the behavior of competitive marketing systems. One drawback of these models is that the number of parameters can become very large, potentially leading to estimation problems. Pooling data for multiple cross-sectional units (stores) can partly alleviate these problems. An important issue in such models is how heterogeneity among cross-sectional units is accounted for. We investigate the performance of several pooling approaches that accommodate different levels of cross-sectional heterogeneity in a simulation study and in an empirical application. Our results show that the random coefficients modeling approach is an overall good choice when the estimated VAR model is used for out-of-sample forecasting only. When the estimated model is used to compute Impulse Response Functions, we conclude that one should select a modeling approach that matches the level of heterogeneity in the data.     

Do social media enhance constructive employee voice all of the time or just some of the time?

Social media are becoming widely adopted by organisations to encourage collaboration and communication. We seek to understand how social media can enhance employee voice and employees' willingness to engage in constructive dialogue with both colleagues and managers. By drawing on literature on employee voice, signalling theory and personal control to analyse qualitative data from research into three strategic business units in a major global telecommunications corporation, we find that (a) employee perceptions of personal control and autonomy influence whether and how employees' exercise voice through social media, and (b) these perceptions vary according to different organisational/field‐level contexts evident in the corporation.     

Robustness issues in multilevel regression analysis

A multilevel problem concerns a population with a hierarchical structure. A sample from such a population can be described as a multistage sample. First, a sample of higher level units is drawn (e.g. schools or organizations), and next a sample of the sub‐units from the available units (e.g. pupils in schools or employees in organizations). In such samples, the individual observations are in general not completely independent. Multilevel analysis software accounts for this dependence and in recent years these programs have been widely accepted. Two problems that occur in the practice of multilevel modeling will be discussed. The first problem is the choice of the sample sizes at the different levels. What are sufficient sample sizes for accurate estimation? The second problem is the normality assumption of the level‐2 error distribution. When one wants to conduct tests of significance, the errors need to be normally distributed. What happens when this is not the case? In this paper, simulation studies are used to answer both questions. With respect to the first question, the results show that a small sample size at level two (meaning a sample of 50 or less) leads to biased estimates of the second‐level standard errors. The answer to the second question is that only the standard errors for the random effects at the second level are highly inaccurate if the distributional assumptions concerning the level‐2 errors are not fulfilled. Robust standard errors turn out to be more reliable than the asymptotic standard errors based on maximum likelihood.     

The market price of Low-Income Housing Tax Credits

The Low-Income Housing Tax Credit (LIHTC) program awards a subsidy to private developers who construct and operate housing units with income-targeted rent controls for at least 15 years. The program allocated $6.6 billion to developers in 2006, and over 1.6 million units have been subsidized under the program since its inception in 1987. A historical literature suggests place-based housing subsidies, such as the LIHTC program, will be more expensive in providing the same level of housing support to the poor than tenant-based strategies (i.e., housing vouchers). This paper uses an administrative data series of LIHTC subsidized properties in California to show the program encourages developers to construct housing units that are an estimated 20% more expensive per square foot than average industry estimates. It is additionally shown that due to liquidity constraints faced by LIHTC primary developers in how the subsidy is allocated, virtually all developers sell the tax credits at a substantial discount below their statutory value immediately after construction. This price is estimated to be $0.73 per $1 of tax credit, or $1.8 billion annually, as compared to alternatively allocating a lump sum grant to developers.     

The Problem of Time Dependent Explanatory Variables at the Context-Level in Discrete Time Multilevel Event History Analysis: A Comparison of Models Considering Mobility Between Local Labour Markets as an Example

The paper deals with the question of how to include time dependent explanatory variables at the context-level in multilevel event history models. In general, context-level explanatory variables in multilevel models are assumed to be time constant. Only time constant context-level explanatory variables perform the task of reducing context-level error variance. Thus, it will be suggested that the analysis should be extended to a three-level model. In this model, time periods of persons constitute level 1 units, time periods of contexts constitute level 2 units and the contexts themselves constitute level 3 units – in which in turn level 2 units are clustered. Considering mobility between local labour markets as an example, four different ways of modelling time varying context-level variables are compared. The result is that the proposed three-level model leads to the most conservative results.     

Pay level satisfaction and employee outcomes: the moderating effect of autonomy and support climates

The present study examined autonomy climate (AC) and support climate (SC) as moderators of the relationship between pay level satisfaction (PLS) and employee outcomes (i.e. job satisfaction, affective commitment and intention to stay). Survey data were collected from 5801 Belgian employees, representing 148 units. The hypotheses derived from distributive justice theory and from research on the meaning of money received partial support. Multilevel analyses revealed that AC buffered the negative effects of low PLS on all three outcomes, and that SC exacerbated the negative effects of low PLS on intention to stay. Theoretical and practical implications of this differential moderating effect are discussed.     

Sources of downward bias in estimating the demand income elasticity for urban housing

Housing demand studies, whether relying upon individual or grouped data, have limited their observations to similar housing units and/or similar housing consumers to help control for product heterogeneity. Yet similar housing units tend to locate in clusters; tenants tend to segregate by race and income. The unintended results may be: (1) for grouped data, selection of a product subgroup with a supply price elasticity small enough to matter; (2) for individual data, selection of a consumer subgroup possessing a lower income elasticity than all housing consumers. Evidence is given that the supply price elasticity is sometimes small enough to matter when grouped data are used.     

科学划分我国城市的职能类型 建立分类指导的扩大内需政策

基于我国第五次人口普查和第二次基本单位普查数据(企业级),综合采用聚类分析和判别分析等方法,科学划分了我国649个县级以上城市的职能类型,将其划分为3个大类、15个亚类和37个职能组。建议下一步以分类结果为参考,以不同职能组城市为单元建立扩大内需投资效果评价和信息反馈机制,重点以大区级和省区级综合性城市为扩大内需着力点,积极发展城市群,促进我国宏观经济平稳启动。     

Best fit,best practice,or stuck in the middle? The impact of unit ownership on unit HR performance in franchise systems

Franchising is an important form of entrepreneurial wealth creation in many retailing and service industries. Since Human Resource Management (HRM) is a critical factor in such industries, it is important to understand how franchisees—as semi-autonomous entrepreneurs—deal with HRM in their units and how this ultimately affects performance at the unit level. However, the very few studies linking franchisee HRM behaviors to performance have not included multi-unit franchising (MUF) as a type of unit ownership. Given the ever-increasing popularity of MUF and the unique characteristics of MUFs, this represents an important knowledge gap. We aim to fill this gap by building a theoretical framework on how the type of unit ownership affects unit HR performance within franchise systems. Building on agency, resource and entrepreneurship perspectives, we propose that units owned by single-unit franchisees (SUFs) and small MUFs (i.e., franchisees with a very small number of units) adopt a ‘best fit’ system regarding HRM, whereas company-owned units (COs) and units owned by larger MUFs (i.e., franchisees with a large number of units) typically adopt a ‘best practice’ system. Each system has its own advantages and disadvantages, which results in two contrasting propositions regarding their effects on unit performance. Moreover, we expect the units owned by medium-size MUFs to have the lowest performance since they are ‘stuck in the middle’ regarding their HRM system.     

Uncertainty Estimation for Pseudo‐Bayesian Inference Under Complex Sampling

Social and economic studies are often implemented as complex survey designs. For example, multistage, unequal probability sampling designs utilised by federal statistical agencies are typically constructed to maximise the efficiency of the target domain level estimator (e.g. indexed by geographic area) within cost constraints for survey administration. Such designs may induce dependence between the sampled units; for example, with employment of a sampling step that selects geographically indexed clusters of units. A sampling‐weighted pseudo‐posterior distribution may be used to estimate the population model on the observed sample. The dependence induced between coclustered units inflates the scale of the resulting pseudo‐posterior covariance matrix that has been shown to induce under coverage of the credibility sets. By bridging results across Bayesian model misspecification and survey sampling, we demonstrate that the scale and shape of the asymptotic distributions are different between each of the pseudo‐maximum likelihood estimate (MLE), the pseudo‐posterior and the MLE under simple random sampling. Through insights from survey‐sampling variance estimation and recent advances in computational methods, we devise a correction applied as a simple and fast postprocessing step to Markov chain Monte Carlo draws of the pseudo‐posterior distribution. This adjustment projects the pseudo‐posterior covariance matrix such that the nominal coverage is approximately achieved. We make an application to the National Survey on Drug Use and Health as a motivating example and we demonstrate the efficacy of our scale and shape projection procedure on synthetic data on several common archetypes of survey designs.     

Estimating economic relations from incomplete cross-section/time-series data

Authors dealing with combined cross-section/time-series data usually assume that complete time-series exist for all units under observation. In the context of micro data, however, this may be a very restrictive assumption. The paper is concerned with problems of model specification and estimation when the data at hand are incomplete time-series from a sample of micro units. Particular attention is paid to a situation where the sample of micro units ‘rotates’ over time. The main results are compared with those derived by Nerlove and others for the standard specification with complete cross-section/time-series data. Some illustrative examples based on data from Norwegian household budget surveys are also given.     

Shadow Price Approach to Total Factor Productivity Measurement: With an Application to Finnish Grass-Silage Production

This paper explores an intermediate route between the Fisher and the Malmquist productivity indexes so as to minimize data requirements and assumptions about economic behavior of production units and their production technology. Assuming quantity data of inputs and outputs and the behavioral hypothesis of allocative efficiency, we calculate the exact value of the Fisher ideal productivity index using implicit shadow prices revealed by the choice of input–output mix. The approach is operationalized by means of a nonparametric data envelopment analysis (DEA) model. Empirical application to Finnish grass silage farms suggests that the Malmquist and the Fisher productivity indices yield similar results when averaged over firms, but there can be major differences in the results of the two approaches at the level of individual firms.     

Detailed population projections for small areas: the Massachusetts experience.

There is an urgent and growing need for population projections that are not only detailed by geographic unit, but are also detailed by demographic characteristics. A high level of disaggregation allows the planner considerable flexibility to aggregate the projected data in a variety of geographic and demographic configurations. This study reports on the methodology and results of the Massachusetts experience in developing detailed population projections. Combining conventional with some not-so-conventional techniques, population projections are developed for 351 geographic units, as well as 108 demographic categories. The projections are currently in use in a large variety of public and private planning activities.     

Sensing endogenous seasonality in the case of a coffee supply chain

Rogue seasonality, or endogenously generated cyclicality (in variables), is common in supply chains and known to adversely affect performance. This paper explores a technique for sensing rogue seasonality at a supply chain echelon level. A signature and index based on cluster profiles of variables, which are meant to sense echelon-level generation and intensity of rogue seasonality, respectively, are proposed. Their validity is then established on echelons of a downstream coffee supply chain for five stock keeping units (SKUs) with contrasting rogue seasonality generation behaviour. The appropriateness of spectra as the domain for representing variables, data for which is daily sampled, is highlighted. Time-batching cycles which could corrupt the sensing are observed in variables, and the need to therefore filter them out in advance is also highlighted. The knowledge gained about the echelon location, intensity and time of generation of rogue seasonality could enable timely deployment of specific mitigation actions.     

Time-series–cross-section Data

This article treats the analysis of 'time-series–cross-section' (TSCS) data. Such data consists of repeated observations on a series of fixed units. Examples of such data are annual observations on the political economy of OECD nations in the post-war era. TSCS data is distinguished from 'panel' data, in that asymptotics are in the number of repeated observations, not the number of units.
The article begins by treating the complications of TSCS data in an 'old-fashioned' manner, that is, as a nuisance which causes estimation difficulties. It claims that TSCS data should be analyzed via ordinary least squares with 'panel correct standard errors' rather than generalized least squares methods. Dynamics should be modeled via a lagged dependent variable or, if appropriate, a single equation error correction model.
The article then treats more modern issues, in particular, the modeling of spatial effects and heterogeneity. It also claims that heterogeneity should be assessed with 'panel cross-validation' as well as more standard tests. The article concludes with a discussion of estimation in the presence of a binary dependent variable.     

Unit nonresponse errors in income surveys: a case study

A survey on the economic and social conditions of households in the city of Modena was carried out in 2002 and in 2006 (two waves) by the CAPP (Centre for Analyses of Public Policies). In the first wave of 2002, each designated sampling unit (i.e., the family) had three units as reserves. If the first refused to be interviewed, the interviewer contacted the three reserves, one after the other, until obtaining either one respondent or four non-participant units. At the end of the survey four categories of units were distinguished: interviewees, refusals, noncontacts, and unused reserves. All units were matched with their corresponding record in the databases of the Ministry of Finance of 2002 and the Census of 2001. The resulting data set permitted the analysis of unit or total nonresponses. The distribution of fiscal income showed different shapes for the four categories, implying a selective participation of the families. The interviewees yielded a positive bias of about 600?, holding constant other factors. The selection of the significant factors affecting nonresponse was performed via backward elimination in a logit model and with the lasso method. Participation increased as fiscal income and age increased and by education level (secondary school and university degree), while it decreased among entrepreneurs, independent workers, managers, and medium-to-low skilled workers.     

Hierarchies and Groups in DEA

Conventional applications of data envelopment analysis (DEA) presume the existence of a set of similar decision making units, wherein each unit is evaluated relative to other members of the set. Often, however, the DMUs fall naturally into groupings, giving rise first to the problem of how to view the groups themselves as DMUs, and second to the issue of how to deal with several different ratings for any given DMU when groupings can be formed in different ways. In the present paper we introduce the concept of hierarchical DEA, where efficiency can be viewed at various levels. We provide a means for adjusting the ratings of DMUs at one level to account for the ratings received by the groups (into which these DMUs fall) at a higher level. We also develop models for aggregating different ratings for a DMU arising from different possible groupings. An application of these models to a set of power plants is given.     

Data aggregation in stochastic frontier models: the closed skew normal distribution

The effect of aggregation on estimates of stochastic frontier functions is considered. Inefficiency is assumed associated with the individual units being aggregated. In this case, the aggregated data have a closed skew normal distribution. Estimating the parameters of a closed skew normal distribution is difficult and so we focus mostly on the biases created by ignoring the fact that the data are aggregated. The conclusions are based on both analytical and Monte Carlo results. When data for firms are aggregates over smaller units and the inefficiency is associated with the units and not the firm, empirical work that does not consider the effect of aggregation will attribute the inefficiency of large firms to diseconomies of scale.