The objective of this study is to investigate the impact of accounting methods for transaction gains (losses) on earning response coefficients (ERCs). This study investigates whether investors respond differently to three accounting methods of the transaction gains (losses). The first method, based on benchmark treatment of Statement of Financial Accounting Standard (Pernyataan Standar Akuntansi Keuangan) No. 10, treats transaction gains (losses) as revenue (expense). The second, based on alternative treatment of Statement of Financial Accounting Standard No. 10 (Interpretation of Financial Accounting Standard No. 4), treats gains (losses) as partially capitalized accounts. The third, based on Capital Market Supervisory Agency regulation No. VIII G10, treats them as fully-capitalized accounts.This study uses the sample of 225 firms listed in the Jakarta Stock Exchange (JSE), during 1993-1999. The hypotheses are tested by applying two empirical models. The first, cumulative abnormal returns are regressed on unexpected earnings and annual return to find ERCs variable, that is, the magnitude of coefficients of unexpected earnings. The second, earnings response coefficients are regressed on transaction gains (losses), earnings persistence, earnings growth, earnings predictability, beta risk, capital structure, firm size and industry effect. Transaction gains (losses) are measured by using absolute value of the average of total transaction gains (losses).The results of this study show that the impact of transaction gains (losses) on earnings response coefficients is statistically significant; and that investors respond indifferently on the firms recognizing different methods of transaction gains (losses). This study, hence, primarily contributes to regulations of Capital Market Supervisory Agency (No. VIII G10) and Financial Accounting Standard Committee of the Indonesian Institute of Accountant (Statement of Financial Accounting Standard No. 10 and Interpretation of Financial Accounting Standard No. 4).