本文利用民國83到89年台灣電子業的新加入廠商,共6466筆資料為樣本,以可數資料(Count Data)模型研究三分位電子產業新廠加入數的決定因素。並進一步利用存活分析(Survival Analysis)的模型,探討電子業新廠商在83-89年間的加入趨勢及其影響因素,以了解電子廠商加入行為的特性。 三分位電子業新廠商加入數決定因子的模型,以負二項分配模型來估計較為適當。此外,更進一步考慮縱橫資料(Panel Data)的特性,以負二項分配中的隨機效果模型(NB Random effects Model)進行實證研究,結果發現影響新廠加入數的正向因素為預期利潤、產業成長率、經濟成長率、市場規模大小等因素。而負面的因素為資本的需求及最小效率的規模等加入障礙因素,這些結果大致與產經理論一致。 新廠加入的動態分析中,利用不同的存活分析模型研究83-89年時間過程中,新廠商加入的時點及趨勢,實證結果發現Log-logistic模型最合適,並發現電子產業的廠商在83-87年間加入率上升,到87-88年期間加入率達到最高,而88年之後則呈緩慢的降低,該結果與國內經濟景氣循環有關。另外,亦可發現電力機械器材製造業、照明設備製造業、家用電器製造業相對於電子零組件製造業、通信機械器材製造業、資料儲存與處理設備製造業、電池製造業加入速率較為快速,其原因可能是加入障礙不同所導致。 本文無論從三分位電子產業各年加入數的靜態分析,或廠商加入的動態分析均可驗證出台灣電子產業加入頻繁的正面因素為:產業的利潤率、市場的規模大小、台灣GDP的成長率、產業成長率。而加入障礙為最小效率規模、資本需求等因素有對於新廠加入有負向效果。而研究中發現:當產業內舊廠關廠越多,會導致新廠的加入減少,可能原因在於台灣電子業競爭激烈,產能供給過多,市場接近飽和,使潛在廠商抱持觀望態度。另外值得注意的是,中小企業比例、研發經費投入、外銷比例等變數,在產業別實證結果與個別廠商加入模型中間呈現不同的結果,需要更進一步研究,才能清楚了解這些變數對於新廠的加入效果為何。
In this thesis, I use the data of newly established firms in Taiwanese electronics industry during 1994 to 2000, summing up to 6466 samples. The Count Data model was specified to empirically study and find out determinants of the new comers. Additionally, survival analysis model is used to analyze trends and factors which are taken into consideration by newly established electronic firms. The empirical results from negative binomial model with random effects showed that growth rate of specific industries, growth rate of economy, market scale are key positive factors related to the numbers of newly established firms. The other factors such as required capital, minimum efficient scale and entry barriers are negative relations with firm entry. In the dynamic analysis, using different survival model and studying for 1994-2000 years, we found that Log-logistic model is most suitable to explain the behavior of firm entry. Entry rate of new firms in the electronic industry rises during the 1994-1998 period, and reaches highest during 1988-1989. These results are related to business cycle of Taiwan. In addition, there exist different entry rates in the different electronic industries. The fastest one is the electric machinery apparatus manufacturing industry, and the slowest one is the battery manufacturing industry. The different entry barriers in specific industry are probably reason to explain it. No matter the static behavior analysis about the new firm entry that is counted every year from the electronic industry of three digits, or the dynamic analysis, verifies and issues positive factors about the new firm entry in electronic industry: profit rate, market scale, Taiwan GDP growth rate, growth rate of specific industries, while the entry barriers such as the minimum efficient scale, required capital have negative effect on firm entry. We also found the counter results which showed that the more exit in previous year, the less the new firm entry in the current year. The reason is that the competition of electron industry of Taiwan is fierce, and the market is close to the saturation. Thus, it makes the potential entrants clasp the wait-and-see attitude. A noteworthy one is in addition, some parameters, such as proportion of small and medium-sized enterprises, researching and developing expenditure and export rate, show different effects on the specific industries model (NB model) and individual firms model (survival model). It will need further study and may understand clearly the behavior of the new firm entry.