Abstract
This study investigates how repo rates and the nominal effective exchange rates affect stock market capitalization between 2016M1 and 2022M1, with a focus on the threshold level of investment in artificial intelligence. To analyze the data, econometrics techniques were employed. The Augmented Dickey-Fuller test confirmed that while market capitalization and repo rates became stationary after the first difference that of investments in artificial intelligence and the nominal effective exchange rates were stationary at level. A threshold model identified the threshold level of investment in artificial intelligence. The findings indicate that repo rates and the nominal effective exchange rates positively influence stock market capitalization when the threshold level of investment in artificial intelligence is below 7.7647. However, above the threshold, repo rates and the nominal effective exchange rates negatively affect stock market capitalization. This study concludes that the negative impact of the repo rates and the nominal effective exchange rates on stock market capitalization at higher levels of investment in artificial intelligence should be internalized through government subsidies that reduce the production cost of firms. However, the SARB should consistently manage the exchange rates from growing out of proportion.