论文部分内容阅读
本文检验了收入分配对发展中国家增长的影响。我们基于世界银行和联合国开发计划署提供的数据,以28个发展中国家为样本研究发现,与艾莱斯纳和罗德里克(Alesina and Rodrik,1994)以往的研究结果不同,在这些国家中,收入分配并未对增长造成影响。我们也没有发现一个国家的民主程度对增长能产生统计意义上的影响。我们观察到,由于具有统计学意义的解释变量之间存在严重的多重共线性,单个独立变量的系数符号与预期的并不一致。回归结果表明,总生育率、人均GDP的原始水平,以及女性与男性识字率之比,这些合在一起,对发展中国家的增长会产生线性的影响。这种实证检验的结果会有助于那些国家的政府甄别哪些地方需要改进,以此来刺激经济发展。所有变量的数据均来自于1978年的世界发展报告,《世界2007年鉴》,以及1999年、2000年和2007/2008年人类发展报告。我们在多元线性回归中运用了最小二乘法。同时,我们注意到解释变量之间多重共线性的严重程度可能会导致其系数符号错误。
This article examines the impact of income distribution on growth in developing countries. Based on the data provided by the World Bank and the United Nations Development Program, we found in 28 developing countries that unlike the previous studies by Alesina and Rodrik (1994), in these countries , Income distribution did not affect growth. Nor have we found that the degree of democracy in a country can have a statistically significant effect on growth. We observe that, due to the severe multicollinearity between statistically significant explanatory variables, the sign of the coefficients of individual independent variables is not as expected. The regression results show that total fertility, the original level of per capita GDP and the ratio of female to male literacy rates, all together, have a linear effect on the growth of developing countries. The result of this empirical test will help governments in those countries identify what needs to be improved in order to stimulate economic development. Data for all variables are derived from the World Development Report 1978, the World Yearbook 2007, and the human development reports for 1999, 2000 and 2007/2008. We use the least-squares method in multivariate linear regression. At the same time, we note that the severity of the multicollinearity between explanatory variables may lead to the wrong sign of their coefficients.