[ABSTRAK
Penelitian ini dilakukan untuk menguji pengaruh variabel CAR, NPFN, ROE,NOM, QR, GWM dan kondisi makroekonomi dalam membedakan danmemprediksi probabilitas kinerja relatif baik bank berdasarkan kelompok sekawandan tanpa kelompok sekawan. Bank syariah yang menjadi sampel penelitian adalahBank Umum Syariah dengan periode penelitian 2005 triwulan 4 sampai dengan2011 Triwulan 3.Teknik analisis yang digunakan adalah regresi panel logit dengan modelrandom effect. Variabel terikat dalam penelitin ini merupakan data biner 1 (kinerjarelatif baik bank) dan 0 untuk (kinerja relatif kurang baik bank). Pendefinisian suatuobservasi memiliki kinerja relatif “baik” dan kinerja relatif “kurang baik” dilakukandengan cara memperingkat masing-masing rasio dari 15 rasio keuangan yangdipublikasikan pada laporan keuangan triwulan berdasarkan Surat Edaran BankIndonesia No 7/56/DBPS 2005. Jika suatu observasi memperoleh jumlah hasilpredikat kinerja “relatif baik” berada pada range 1/3 tertinggi, maka hasil akhir yangdiperoleh suatu observasi adalah 1 (relatif baik), selainnya 0 (relatif kurang baik).Hasil temuan menunjukkan bahwa pada analisis panel logit dengan kelompoksekawan, berdasarkan hasil estimasi variabel-variabel yang signifikan untukmembedakan dan memprediksi kinerja realatif bank adalah CAR, ROE dan GWMdengan arah hubungan yang positif, NPFN dengan arah hubungan yang negatif.Sementara variabel QR, Inflasi dan pertumbuhan PDB tidak signifikan dalammembedakan dan memprediksi kinerja relatif suatu bank. Pada analisis tanpakelompok sekawan, variabel yang negatif signifikan adalah NPFN, sementara ROEdan QR signifikan dengan arah hubungan yang positif membedakan danmemprediksi dan kinerja relatif suatu bank.Namun demikian, penelitian ini hanya terbatas dengan jumlah sampel 11Bank Umum Syariah dan periode pengamatan yang pendek secara kwartalan.Untuk penelitian selanjutnya disarankan agar menambah jumlah sampel dan rasiorasiokeuangan lainnya untuk mengetahui variabel-variabel yang dapatmembedakan dan memprediksi kinerja relatif suatu bank.

---

ABSTRACT
The study was conducted to examine the effect of variable CAR, NPFN,ROE, NOM, QR, GWM and macroeconomic conditions to distinguish and predictthe probability of the good relative performance of the Bank based on Peer Groupanalysis and without Peer Group Analysis. Sharia banks which became a sampleresearch are The Commercial Syariah Bank’s with time period 2005 Quarter 4th to2011 Quarter 3rd .The analysis technique used is logit panel random effect model. DependentVariable in this study is binary data 1 (good relative performance bank) and 0 for(less performance of bank). Defining an observation which has a good relativeperformance and the relative less performance by ranking each ratio of 15 financialratios which published quarterly financial report was based on Bank of IndonesiaAnnouncement Letter No. 7/56/DBPS 2005. If the observation which was obtainedthe sum of good relative performance is on the range of 1/3 the highest, the finalresult of the observation is 1 (relatively good), other is 0 (less performance).The findings result that the logit panel analysis with Peer Group, based onresult significantly variables to distinguish and predict the relatif performance ofthe Bank is CAR, ROE and GWM, with positively direction. NPFN by negativedirection of relation. QR, inflation and GDP growth does not significantlydifferentiate and predict the relative performance of the banks. In analysis withoutPeer Group Analysis, the significant variable is NPFN, meanwhile ROE and QRare significant by positive relation so those variables able to distinguish and predictrelative performance of bankHowever, the study was limited only with total 11 Sharia Banks and the shortperiod of observation which was quarterly. For further research, it is suggested thatresearchers increase the sample amount and other financial ratios to determinevariables to be able to distinguish and predict the relative performance of a bank., The study was conducted to examine the effect of variable CAR, NPFN,ROE, NOM, QR, GWM and macroeconomic conditions to distinguish and predictthe probability of the good relative performance of the Bank based on Peer Groupanalysis and without Peer Group Analysis. Sharia banks which became a sampleresearch are The Commercial Syariah Bank’s with time period 2005 Quarter 4th to2011 Quarter 3rd .The analysis technique used is logit panel random effect model. DependentVariable in this study is binary data 1 (good relative performance bank) and 0 for(less performance of bank). Defining an observation which has a good relativeperformance and the relative less performance by ranking each ratio of 15 financialratios which published quarterly financial report was based on Bank of IndonesiaAnnouncement Letter No. 7/56/DBPS 2005. If the observation which was obtainedthe sum of good relative performance is on the range of 1/3 the highest, the finalresult of the observation is 1 (relatively good), other is 0 (less performance).The findings result that the logit panel analysis with Peer Group, based onresult significantly variables to distinguish and predict the relatif performance ofthe Bank is CAR, ROE and GWM, with positively direction. NPFN by negativedirection of relation. QR, inflation and GDP growth does not significantlydifferentiate and predict the relative performance of the banks. In analysis withoutPeer Group Analysis, the significant variable is NPFN, meanwhile ROE and QRare significant by positive relation so those variables able to distinguish and predictrelative performance of bankHowever, the study was limited only with total 11 Sharia Banks and the shortperiod of observation which was quarterly. For further research, it is suggested thatresearchers increase the sample amount and other financial ratios to determinevariables to be able to distinguish and predict the relative performance of a bank.]