New Economics Papers
on Computational Economics
Issue of 2008‒04‒15
seven papers chosen by

  1. Credit scoring with boosted decision trees By Bastos, Joao
  2. Estimation of substitution elasticities for CGE models By Azusa OKAGAWA; Kanemi BAN
  3. Modelling the Economic Effects of Population Ageing By James Giesecke; G.A. Meagher
  4. Growth and structural change in the Vietnamese economy 1996-2003: A CGE analysis By Tran Hoang Nhi; James A. Giesecke
  5. Optimization in non-standard problems. An application to the provision of public inputs By A. Jesus Sanchez Fuentes; Diego Martinez Lopez
  6. Will drought erode the competitiveness of Australia's wine industry? By Glyn Wittwer
  7. Educational Attainment, Growth and Poverty Reduction within the MDG Framework: Simulations and Costing for the Peruvian Case By Gustavo Yamada; Juan F. Castro; Arlette Beltran; Maria A. Cardenas

  1. By: Bastos, Joao
    Abstract: The enormous growth experienced by the credit industry has led researchers to develop sophisticated credit scoring models that help lenders decide whether to grant or reject credit to applicants. This paper proposes a credit scoring model based on boosted decision trees, a powerful learning technique that aggregates several decision trees to form a classifier given by a weighted majority vote of classifications predicted by individual decision trees. The performance of boosted decision trees is evaluated using two publicly available credit card application datasets. The prediction accuracy of boosted decision trees is benchmarked against two alternative data mining techniques: the multilayer perceptron and support vector machines. The results show that boosted decision trees are a competitive technique for implementing credit scoring models.
    Keywords: Credit scoring; Boosting; Decision tree; neural network; support vector machine
    JEL: C44 G32
    Date: 2008–04–01
  2. By: Azusa OKAGAWA (Graduate School of Economics, Osaka University); Kanemi BAN (Graduate School of Economics, Osaka University)
    Abstract: Many studies of climate policy are based on computable general equilibrium (CGE) modeling. The simulation results and conclusions reached by these models depend on the size of the parameters specified. In particular, the substitution elasticities between production factors have a major influence. Therefore, in order to obtain reliable simulation results we should employ empirical evidence gathered on the substitution elasticities. Unfortunately, in many instances, the lack of econometric analysis means we must specify these key parameters based on existing work or borrow them from prominent modeling exercises. In this study, we estimate nested constant elasticity of substitution (CES) production functions using panel data for OECD countries to help improve the reliability of CGE models for climate policy. Our results show higher values for substitution elasticities closely related to energy inputs for energy-intensive industries and lower values for other industries compared to the conventional values often used in existing models. With the new parameters estimated, we find that conventional parameters could overestimate the necessary carbon price by 44%, and obtain evidence of different distributions of CO2 emission reduction costs across industries.
    Keywords: Substitution elasticities, CGE modeling, Climate policy
    JEL: D58 Q43 Q54
    Date: 2008–04
  3. By: James Giesecke; G.A. Meagher
    Abstract: In March 2005, the Productivity Commission released a report on the Economic Implications of an Ageing Australia. The report describes projections for a number of economic variables including population, labour force participation rates, labour supply, employment and hours worked per week. The present paper describes a number of simulations with the MONASH model designed to extend the range of the Commission's earlier analysis. The first is a base case forecast for the Australian economy for the twenty-year period 2004-05 to 2024-25. As far as possible, it is specified so as to maintain consistency with the Commission's projections. The others are alternative forecasts for the same period in which various effects of population ageing have been removed. The alternative forecasts separately identify: (a) a taste effect due to the removal of age-related shifts in the commodity composition of household consumption; (b) a public effect due to the removal of age-related shifts in public consumption; (c) a skill effect due to the removal of age-related shifts in hours of employment distinguished by skill (with total hours of employment unchanged); (d) a scale effect due to the removal of age-related shifts in total hours of employment (with the skill composition of employment unchanged); and (e) a total effect due to the simultaneous removal of all the above age-related shifts. To accommodate the simulations, the MONASH model itself is reconfigured such that labour by qualification group can be converted into labour by occupation according to Constant Elasticity Transformation (CET) functions. Labour by occupation in its turn can be converted into effective units of industry-specific labour according to Constant Elasticity Substitution (CES) functions. Labour of a partcular skill is then distributed between occupations and industries according to relative wage rates. The scheme incorporates 67 qualification groups, 81 occupations (the ASCO minor groups) and 107 industries (the input-output classification).
    Keywords: computable general equilibrium modelling, population ageing, labour market forecasting
    JEL: C53 C68 J11
    Date: 2008–02
  4. By: Tran Hoang Nhi; James A. Giesecke
    Abstract: We use MVN - a dynamic CGE model of the Vietnamese economy - to investigate the Vietnamese economy's rapid growth and structural change over the period 1996 to 2003. We do this in two steps. First, we estimate changes in variables representing production technologies, consumer preferences, government policy and other structural features of the economy. Movements in these structural and policy variables are then used to explain the recent history of Vietnam's rapid growth and structural change. We find the most important sources of growth and change to be technical improvements, favourable shifts in foreign preferences for Vietnamese goods and employment growth. Other important factors include movement in household preferences away from primary products and towards manufactures and services, expansion in agricultural land supply, and tax reform.
    Keywords: CGE model, Vietnam
    JEL: C68 D58 F14 O12
    Date: 2008–02
  5. By: A. Jesus Sanchez Fuentes (Department of Economics, Universidad Pablo de Olavide); Diego Martinez Lopez (Department of Economics, Universidad Pablo de Olavide)
    Abstract: This paper describes a new method for solving non-standard constrained optimization problems for which standard methodologies do not work properly. Our method (the Rational Iterative Multisection -RIM- algorithm) consists of different stages that can be interpreted as different requirements of precision by obtaining the optimal solution. We have performed an application of RIM method to the case of public inputs provision. We prove that the RIM approach and comparable standard methodologies achieve the same results with regular optimization problems while the RIM algorithm takes advantage over them when facing non-standard optimization problems.
    Keywords: direct search, constrained optimization, multisection, optimal taxation, public input.
    JEL: C6 H21 H3 H41 H43
    Date: 2008–03
  6. By: Glyn Wittwer
    Abstract: Australia's 2007 vintage was smaller than the previous few vintages as a consequence of drought and frost in 2006. Then the worst happened in the following year with an unprecedented second year of drought that resulted in severe cuts to water allocations for irrigators in the southern Murray-Darling Basin. This paper examines the medium-term prospects for the Australian wine industry using global projections to 2016 with the World Wine Model. The international growth prospects of commercial-premium wine matter to growers in irrigated regions of Australia. Overall, demand for commercial-premium wine is set to grow. This is despite a slowing of growth in Australia's number one market, the United Kingdom.
    Keywords: CGE modelling, wine consumption
    JEL: C68 Q13
    Date: 2008–03
  7. By: Gustavo Yamada; Juan F. Castro; Arlette Beltran; Maria A. Cardenas
    Abstract: We propose a model that accounts for the potential feedback between schooling performance, human capital accumulation and long run GDP growth, and links these results with poverty incidence. Our simulation exercise takes into account targets for education indicators and GDP growth itself (as arguments in our planner's loss function) and provides two conclusions: (i) with additional funds which amount to 1 percent of GDP each year, public intervention could, by year 2015, add an extra 0.89 and 1.80 percentage points in terms of long-run GDP growth and permanent reduction in poverty incidence, respectively; and (ii) in order to engineer an intervention in the educational sector so as to transfer households the necessary assets to attain a larger income generation potential in the long run, we need to extend the original set of MDG indicators to account for access to higher educational levels besides primary education.
    Keywords: Millennium development goals, education, human capital, GDP growth, poverty, Peru
    JEL: O41 I32 C25 C41 C61
    Date: 2008

General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.