阿拉伯商业与管理评论杂志

???????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????

????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????

The study has potential limitations. A random effects model has been estimated considering log-transformed data seeking to obtain reliable measurements that allow a statistical analysis. The estimate model has conducted a panel data analysis considering the computer, communications and other services as a percentage of commercial service exports taken as a dependent variable (Y) and three explanatory variables related to the share of high-technology exports in manufactured exports (X₁), researchers (per million) (X₂) in research and development (R&D), share of computer, communications and other services in commercial service imports (X₃). A panel of eight developed countries has been chosen that are United States, Japan, China, France, Germany, United Kingdom, Singapore and Korea Republic covering the period of 1996-2017 (22 years). With three panel unit root tests, such that λ_B, IPS and CADF, all variables require a first difference to become stationary. The null hypothesis of no co-integration can not be rejected according to the seven tests proposed by Pedroni. The findings reveal a positive impact for both variables (X₂) and (X₃) but a negative impact for (X₁) on the variable (Y). The findings reveal that if (X₂) (respectively (X₃)) increases 1% then Y increases 0.115% (respectively 0.271%). While if the variable (X₁) increases 1% then Y decreases 0.119%.