本文以實地實證研究方法，探討「非財務性績效衡量指標」與「工作設計」此二項誘因工具對員工績效的影響。實地研究對象為一家商業銀行，研究期間為1993-1997；該行於1995年7月開始採用非財務性績效衡量指標，並自1996年1月選擇五家分行實施徵、授信分工。 本文首先利用個案銀行17位授信員連續60個月所構成的縱查資料（panel data），以固定影響模型（fixed-effects model）分析非財務性績效衡量指標誘因工具對員工績效的影響，發現銀行授信人員之績效考核制度中，加入授信不良率與脫落率非財務性衡量指標後，雖然授信員努力的重分配導致其授信金額與利息收入財務績效變差，但授信員的授信不良率與脫落率非財務績效變佳，整體而言，授信利潤有顯著的改善。 本文另以實施徵、授信分工前後，51名授信員的績效資料，採中斷的時間序列設計（interrupted time-series design），檢定非財務性績效衡量指標與工作設計配合採用後，員工的績效是否較佳。分析結果顯示，存在非財務績效衡量指標之獎酬制度下，配合實施徵、授信分工的授信員，其授信金額及授信不良率績效顯著較徵、授信合一之授信員為佳。 / This thesis reports the results of a longitudinal field study examining the impact of nonfinancial performance measures and job design on employee performance. A commercial bank was the research site. The bank adopted nonfinancial performance measures in July 1995 and redesigned jobs in retail banking credit underwriting at five branches in January 1996. Prior to the job redesign, sales representatives were jointly responsible for all credit underwriting functions including credit extension and credit servicing. After the change, sales representatives were only responsible for credit extension while credit servicing was handled by credit decidion-makers. Panel data for 17 sales representatives over 60 months (1993 through 1997) were obtained to assess the effects of adopting nonfinancial measures on employee performance. Using fixed-effects regression analysis, the results are consistent with the predictions of agency theory: although loan volume and interest revenue decreased following adoption of nonfinancial measures to evaluate and compensate the performance of sales representatives, loan profit increased significantly. This thesis also hypothesizes that employee performance increases with the fit between incentive systems based on nonfinancial measures and job design. The empirical tests of this prediction rely on an interrupted time-series design, using performance data of 51 sales representatives in the 1995-1997 period. After the change in job design-separation of the credit underwriting jobs-alogn with introduction of incentive systems based on nonfinancial measures, the treatment group demonstrated greater improvements in productivity and quality performance. This result supports the hypothesis developed.