Location decisions of Japanese new manufacturing plants in China: a discrete-choice analysis

Abstract

F 23, L 20, R 30,

FDI Location Decision: Evidence from Firms Investing in China

“…A discrete-choice model, namely, binary logit model was used with disaggregate observations to study the determinants of FDI location choice in China. Discrete choice models empower researchers to reveal each individual choice maker's preferences, some of which may be lost in the aggregate methodologies (such as OLS) (Cheng & Stough, 2006).…”

“…In the aggregate approach, ordinary least square (OLS) method is generally used and assumes that FDI stocks are normally distributed across cities and provinces and that the city or the province is able to accumulate any specific volume of FDI in any year and over years (Cheng & Stough, 2006). In the disaggregate approach, each individual firm or observation is examined against observable location characteristics.…”

“…Zhang, 1994;Schroath et al, 1993;Cheng & Stough, 2006), only few of them took into account the impact of the country of origin on foreign investors' location choice. While Schroath et al (1993) argued that cultural and geographical factors play an important role in the spatial concentration of FDI, Qu and Green (1997) suggested that the country of origin determines the location choice and demonstrated that FDI from USA, Japan, and European countries is interested in city sizes, consumption levels and infrastructure in its location decisions (Cheng & Stough, 2006). He (2003) concluded, in the same vein, that Japanese investors have special location preferences compared to other investors while FDI from Hong Kong is generally sensitive to geographical and cultural proximity.…”

Probabilistic forecasting with discrete choice models: Evaluating predictions with pseudo-coefficients of determination

“…Discrete choice models, such as conditional logit (CL) and multinomial probit, are ideal for performing this function and are widely applied in probabilistic forecasting. Their primary use has been 2 in predicting individuals' choices from a range of alternatives, so they have been employed in consumer choice and marketing (Lin & Sibdari, 2009) and econometrics (Maddala, 1983), but have also been adopted in such diverse fields as epidemiology (Breslow & Day, 1994), operations research (Cheng & Stough, 2006), and the forecasting of competitive events (Smith & Vaughan Williams, 2010). …”

“…To begin with, there are significant differences between R 2 s and pseudo-R 2 s. So, while pseudo-R 2 s are commonly reported (Cheng & Stough, 2006;Schnytzer et al, 2010), their usage is seldom justified (Veall & Zimmerman, 1996), and there are still many unresolved issues associated with them. First, unlike R 2 , there is no single definition of pseudo-R 2 that is universally employed.…”