On the effect of prior assumptions in Bayesian model averaging with applications to growth regression

We consider the problem of variable selection in linear regression models. Bayesian model averaging has become an important tool in empirical settings with large numbers of potential regressors and relatively limited numbers of observations. We examine the effect of a variety of prior assumptions on the inference concerning model size, posterior inclusion probabilities of regressors and on predictive performance. We illustrate these issues in the context of cross‐country growth regressions using three datasets with 41–67 potential drivers of growth and 72–93 observations. Finally, we recommend priors for use in this and related contexts. Copyright © 2009 John Wiley & Sons, Ltd.     

Whose call is it? Targeting universal service programs to low-income households’ telecommunications preferences

Do universal service programs give customers what they want? This paper uses new survey data to study low-income households’ telecommunications choices in the United States and to consider the degree to which such households’ preferences are addressed by existing universal service programs. The research shows that households that choose only one form of telecommunications increasingly are choosing a mobile phone, while those that choose to have both modes of communications are shifting their usage towards their mobile phones. These trends are less pronounced among higher-income households. One implication for universal service policy is that traditional subsidies for landline phones are increasingly ineffective in reaching low-income households such subsidies are designed to help; subsidies for acquiring and using mobile phone services might be more beneficial to low-income households than traditional subsidies for landline phones.     

Comments on ‘Jointness of growth determinants’

We consider the measures of jointness proposed by Doppelhofer and Weeks ( 2009 ) and Strachan ( 2009 ) in the context of variable selection. Using the general criteria suggested in Ley and Steel ( 2007 ), we identify some shortcomings of these measures, which are illustrated with empirically relevant example cases. We argue that careful examination of the properties of any jointness measure is critical before using it to inform decisions, and favour the use of the measures proposed in Ley and Steel ( 2007 ). Copyright © 2009 John Wiley & Sons, Ltd.     

Determinants of Government Aid to Katrina Survivors: Evidence from Survey Data

We analyze survey data from Mississippi coastal communities where Katrina made its final landfall. Logistic regressions indicate that government aid is helpful in dealing with one- to two-month economic disruption and long-term rebuilding but is less helpful with regard to short-term rebuilding and mitigating longer-term disruption. Our analysis (including a basic risk assessment) finds evidence that individuals receiving government aid and/or having a disability predisaster are likely to incur severe economic hardship postdisaster and that individuals with greater predisaster economic and/or social network capital seem to be less at risk. Our results underscore the importance of housing in the resumption of basic economic activity.     

Mending the family tree a reconciliation of the linearization and levels schools of AGE modelling

This paper offers a critical comparison between the North American levels school of applied general equilibrium modelling, and the Norwegian/Australian school of linearizers. The paper develops both the levels and linearized representations of a neoclassical multiregion trade model. This is used as a vehicle for focusing attention on similarities and differences between the two schools. The main conclusions are: (i) that both the levels and linearized versions of non-linear AGE model offer a valid starting point for obtaining accurate solutions to the non-linear equilibrium problem; (ii) when update formulae are not employed, linearized representations are particularly prone to erroneous welfare conclusions; (iii) the levels representation offers a more natural starting point for expressing accounting conditions, whereas behavioural relationships are relatively more easily expressed in a linearized representation. Recent software developments now make both approaches to the representation and solution of non-linear AGE models convenient to implement. Thus the choice of which approach to use will depend on the special needs and preferences of the economist implementing the model. The two traditions of AGE modelling have a great deal in common and both would benefit from greater cooperation.     

The road to cooperative supply‐chain environmental management: trust and uncertainty among pro‐active firms

The pressures on firms to improve their environmental performance have caused them to look outside their boundaries towards their supply chains. In such approaches, firms work with vendors to develop the environmental profile of supplied materials (for example) by reducing materials' toxicity or the amount of packaging used. While large firms can mandate that their suppliers comply with such initiatives, more cooperative approaches are generally likely to be more fruitful. This article presents the results from an exploratory, two phase study of the conditions under which firms engage in cooperative supply‐chain environmental management. First, the authors conducted interviews with 14 leading‐edge firms. In the second phase, the authors conducted a theoretical‐sample survey to examine a model of the antecedents of cooperative supply‐chain environmental management. The results suggest that inter‐firm trust, uncertainty and pro‐active environmental management most directly affect the extent to which firms engage in cooperative supply‐chain environmental management. Copyright © 2007 John Wiley & Sons, Ltd and ERP Environment.