本研究使用代理人基模型進行模擬,控制代理人與環境,在人工股票市場與設定固定資產的環境下,模擬兩種不同的交易機制,連續雙邊喊價機制(Continuous double auction market, CDA)與集合競價機制(Call market, CAM)下的價格波動差異。 本研究將資產基本面設定成遞增與遞減兩類,代理人分成四組做為對照,每組實驗獨立模擬10000期,共模擬100次。此研究透過將代理人變數控制,排除現實市場裡交易者資訊不對等情況,也排除了在現實中環境的因素,如漲跌幅限制的影響,單純比較在兩種交易機制之下,所產生的價格變動波動。 實驗結果發現到不論是在資產基本面遞增或遞減設定下,結果一致。在各種代理人組合下,連續雙邊喊價機制(CDA)的價格波動顯著大於集合競價機制(CAM)。另外也發現在連續雙邊喊價機制(CDA)與集合競價機制(CAM)中,隨著資產基本面的提高,價格的波動率將趨於平緩。
This study applies agent-based modeling to simulation artificial stock market. Given agents behavior and single asset setting, we compare the price volatilities between two matching mechanisms, continuous double auction (CDA) and call market (CAM). Two fundamental scenarios, increasing and decreasing fundamental values, have been taken into consideration. Under two matching mechanisms and two scenarios, there are total four cases of combination for us to simulate. We conduct 100 simulation runs for each case, and 10,000 periods for each run. In this simple setting, we have 100 artificial traders in each simulation. The market has no information asymmetry and no price limitation during the price matching. We find that the price volatility of CDA is significantly larger than CAM in the four cases. In addition, it is also found that no matter what fundamental scenario is, the price volatility will be converge eventually.