ABSTRAK
Skripsi ini bertujuan untuk menguji teori efisiensi pasar modal di Indonesia padatingkat semistrong-form, dengan menganalisis pengaruh karakteristik perusahaanatau internal perusahaan yaitu rasio book-to-market, rasio dividend-price, rasiodividend-payout, rasio earning-price, rasio cash flow-price, variabelmakroekonomi yaitu inflasi, serta faktor dari pasar modal yaitu market riskpremium, terhadap excess return dari 45 perusahaan di Indonesia yang tergabungdalam indeks LQ45, dengan membagi dua periode penelitian yaitu periode 2006-2011 sebagai in-sample period serta periode 2012-2014 sebagai out-of-sampleperiod. Dengan menggunakan analisis regresi FQGLS, secara umum, ditemukanbahwa prediktor-prediktor secara statistik signifikan dapat mempengaruhi excessreturn. Uji in-sample menunjukkan bahwa kekuatan prediktor heterogen, bersifatspesifik terhadap karakteristik perusahaan. Selanjutnya, uji out-samplemenunjukkan bahwa akurasi prediksi juga bersifat heterogen, dimana akurasimodel prediksi untuk beberapa saham kuat sampai akhir periode out-of-sample, danbeberapa saham lagi tidak bisa diprediksi, dimana nilai excess return regresi denganaktual memiliki perbedaan yang besar. Secara khusus, terdapat indikasi bahwapasar modal di Indonesia belum efisien pada tingkat semistrong, namun padajangka panjang, akurasi prediktabilitas semakin lemah sehingga pasar bergerak kearah efisien pada long-run.

---

ABSTRACT
This study tests the efficient market hypothesis in Indonesia at semistrong-form, byexamining the impact of company specific or internal factors such as book-tomarketratio, dividend-price ratio, dividend-payout ratio, earning-price ratio, andcash flow-price ratio, macroeconomic variable which is inflation, and marketfactors which is the market risk premium, on the excess return from 45 companieslisted in LQ45 of Indonesian Stock Exchange Market, by dividing the research intotwo periods which is year 2006-2011 as the in-sample period, and year 2012-2014as the out-of-sample period. Using FQGLS regression, the results show that thesepredictors are statistically significant affecting excess return. In-sample test showsthat the power of these predictors to explain excess return is heterogeneous.Furthermore, out-of-sample test shows that the prediction accuracy is alsoheterogeneous, that there are strong accuracies in some firms until the end of theout-sample period, and there are weak accuracies that the value between actualexcess return with predictive excess return are clearly different. In particular, thereis an indication that Indonesia stock market is not fully efficient at semistrong level,but for the long run, the prediction accuracies become weaker and the market moveto be efficient.