Tujuan dari penelitian ini meliputi lima hal. Pertama, menguji apakah terdapat hubungan simultan (bidirectional) antara Indeks Harga Saham Gabungan (IHSG) dan nilai tukar rupiah terhadap dolar Amerika Serikat (ldrus). Kedua, menguji apakah terdapat hubungan relasional jangka panjang (kointegrasi) antara Idrus dan lHSG. Ketiga, mempelajari pola hubungan kointegrasi antara Idrus dan IHSG, jika terdapat hubungan kointegrasi antara keduanya. Keempat, mempelajari pengaruh variabel-variabel lag IHSG dan variabel-variabel lag ldrus terhadap IHSG dan ldrus. Kelima, mempelajari pola pengaruh Dow Jones Industrial Averages (DJIA) ienhadap IHSG dan Idrus.Penelitian ini memasukkan IHSG dan ldrus sebagai variabel-variabel endogenus, serta DJIA (indeks saham Dow Jones) sebagai variabel eksogenus. Semua data ditransformasikan ke dalam logaritma natural (Ln). Sampel Penelitian dibagi menjadi dua periode yaitu periode sebelum krisis ekonomi (19 Mei 1996-29 Juni 1997) dan periode selama krisis ekonomi (4 Januari 1998-9 September 2001). Analisis data meliputi tujuh langkah. Pertama, uji simultanitas. Kedua, pra-uji umuk menentukan order integrasi. Ketiga, penentuan panjang lag dalam VARL. Keempai, uji kointegrasi dengan metode Johansen. Kelima, analisis hasil estimasi. Keenam, uji kausalitas Granger. Ketujuh, innovation accounting.Untuk periode sebelum krisis ekonomi, data generating process menggunakan Vector Autoregressive in Differenced (VARD) lag 3 atau VARD(3). Sedangkan, untuk periode selama krisis ekonomi menggunakan Vector Error Correction Model (VECM) lag 7 atau VECM(7).Hasil uji simultanitas series dalam level menunjukkan adanya pengaruh timbal balik antara nilai tukar rupiah terhadap dolar Amerika Serikat dan IHSG pada kedua periode. Pada periode sebelum krisis ekonomi tidak ada hubungan kointegrasi antara kedua variabel endogenus, karena nilai tukar selama periode tersebut berdasarkan managed floating. Hasil VARD(3) menemukan tidak ada hubungan timbal balik antara perubahan IHSG atau D(LnIHSG) dan perubahan nilai tukar rupiah terhadap dolar AS atau D(Lnldrus) dalam jangka pendek. Variabel eksogenus atau D(LnDJIA) tidak berpengaruh terhadap D(Ln1}-ISG) maupun D(Ln1drus) karcna perbedaan metode perhitungan indeks maupun karena nilai tukar berdasarkan managed floating.Pada periode selama krisis ekonomi terdapat hubungan kointegrasi antara kedua variabel endogenus. Hasil VECM(7) menunjukkan pengaruh negatif LnIdrus terhadap LnIHSG dalam jangka panjang. Hal ini menunjukkan bahwa hubungan antara nilai tukar rupiah terhadap dolar AS dalam jangka panjang lebih rentan terhadap infaltion disturbance. Hasil uji kausalitas Granger menemukan hubungan searah dimana IHSG mempengaruhi Lnldrus. Hal ini berarti IHSG merupakan leading indicator dalam jangka pendek. Perubahan indeks saham Dow Jones atau D(LnDJIA) berpengaruh negatif terhadap perubahan nilai tukar rupiah terhadap dolar Amerika Serikat atau D(Lnldrus) dimana perhitungan nilai tukar menggunakan free floating. Namun, D(LnDJIA) tidak berpengaruh terhadap perubahan IHSG atau D(LnlHSG), karena perbedaan metode perhitungan. Hasil ini konsisten dengan periode sebelum krisis ekonomi.

---

This research has five objectives. First, investigate whether simultaneous is in Jakarta Composite Stock Price Index (JKSE) and rupiah - U.S. dollar exchange rate (ldrus) relationship. Second, investigate whether co-integrating relationship between Idrus and JKSE. Third, study the relationship between ldrus and JKSE in the long run. Fourth, study the eiiects of lagged values of JKSE and lagged values of ldrus on JKSE and Idrus, respectively. Fifth, study the effect of Dow Jones Industrial Averages (DJIA) on JKSE and Idrus, respectively.JKSE and ldrus are treated as endogenous variables. DHA is treated as exogenous variable. All of series are transformed to natural logarithmic (Ln). Period of research is divided to two periods, before crisis period (19 May 1996-29 June 1997) and during crisis period (4 January 1998-9 September 2001). The analysis of data will be done in seven steps. First, test of simultaneous. Second, pre-test for determining the order of integration for each series. Third, determine lags of Vector autoregressive in Level (VARL). Fourth, Johansen co-integration test. Fifth, analyze the estimate parameters. Sixth, test of Granger causality. Seventh, do innovation accounting.For before economic crisis period, data generating process use Vector autoregressive in Differenced (VARD) lag 3, while, for during economic crisis period, it use Vector Error Correction Model (VECM) lag 7.Test of simultaneous with series in level found bidirectional relationship between Jakarta Composite Stock Price Index (JKSE) and rupiah - US dollar exchange rate (Idrus) in both period. There is no co-integrating relationship between Lnldrus and LnJKSE in before economic crisis period because of managed floating exchange rate. Analysis of parameters of VARD lag 3 found no bidirectional short-run relationship between rupiah - US dollar differential or D(LnIdrus) and IKSE differential or D(LnJKSE). DJIA differential or D(LnDJIA) doesn?t affect D(LnJKSE) because of difference in determination method. D(LnDJIA) doesn?t affect D(LnIdrus) also because rupiah - US dollar exchange rate was managed.There is co-integrating relationship between Lnldrus and LnJKSE in during economic crisis period. Analysis of parameters of VECM lag 7 found negative long-run relationship between Lnldrus and Ln]KSE. It suggests that Idrus - IKSE relationship in the long-run has more exposure to inflation disturbance. Granger causality runs only one way from LnJKSE to Lnldrus. It suggests that JKSE is a leading indicator in the short-run. DJIA differential or D(LnDJIA) has negative effect on D(Lnldrus), which is rupiah - US dollar exchange rate is free floated, D(LnDJIA) has no effect on D(LnlHSG) because of difference in determination method. lt is consistent with the result of before economic crisis period.