Meta analysis of Taiwanese family firms' performances

家族企業在全世界皆是商業活動中的重要組織型態。因此在過去就有許多國內外研究針對家族企業的特色以及其對績效的影響作探討，但結果相當分歧。因此，本研究希望藉由後設分析(meta analysis)研究方法將過去個別研究的結果，透過計量方法以獲得一特定結論。本研究蒐集23篇過去研究台灣家族企業績效之文獻共29個樣本，以後設分析效應量觀察台灣家族企業與非家族企業之績效是否有顯著差異；以及哪些干擾變數會影響家族企業與非家族企業的績效。在實證分析中，以敘述統計觀察，文獻結論多為家族企業績效優於非家族企業；而以後設分析的效應量觀察，發現總體效應量與個別以ROA、ROE、Tobin’s Q作為績效的效應量皆顯示家族企業的績效優於非家族企業，其中總體效應量和Tobin’s Q的效應量為統計上顯著，說明此研究範圍中家族企業績效較佳的發現。而觀察干擾變數的效應量，以ROA、ROE作為績效衡量時，文獻樣本數為顯著的干擾變數，當文獻樣本數在700以下時，家族企業績效優於非家族企業，樣本數700以上時，反之；在以ROE為績效指標時，文獻類別為具影響的干擾變數，當文獻為學術期刊時，家族企業績效遜於非家族企業，當文獻為學位論文時，反之；以Tobin’s Q作為績效指標的實證發現，以家族成員持股率為家族企業定義時，家族企業績效顯著優於非家族企業。以上實證結果證實了利益收斂假說與家族企業富不過三代之說法。Family firm is a significant type of business organizations in the whole world. Therefore, many researches have been done for discussing the characteristic of family firms and how family firms influence organization’s performance. However, the results are varied. Hence, the purpose of this thesis is to get a generalized conclusion about family firm’ performance by Meta analysis’s quantitative method which collect the results from past related researches. The samples of this thesis are from 23 past researches based on Taiwanese family firms. This thesis calculates Meta analysis’s effect sizes to examine the difference of performances between family firms and non-family firms and examine which are the moderator variables that impact the performances. Consequently, description statistics shows that most researches result the performances of family firms are better than non-family firms. In addition, Meta analysis also shows the same finding in total mean effect and Tobin’s Q which are statistical significances on effect size. With regard to moderator variables’ effect sizes, sample size is a moderator variable with statistical significance when ROA and ROE as performance index. While sample sizes are below 700, the performances of family firms are better than non-family firms, vice versa. Additionally, type of sample is a moderator variable when ROE as performance index. While simples are Journal articles, the performances of family firms are worse than non-family firms, vice versa. Furthermore, definition of family firm is a moderator variable as well when Tobin’s Q as performance index. While simples define family firms by family members holding rate, family firms’ performance are better. In conclusion, above findings prove that convergence of interest hypothesis and great men's sons seldom do well