**4. Empirical results analysis**

We studied the effects of the daily changes in the prices of gold and platinum, two important export items from South Africa, on the country**'**s financial risk indicator (CDS) and the exchange rates (RD). Our empirical findings from the asymmetric TVP-VAR Model connectedness analysis are as follows:

The empirical results of the TVP-VAR Model are obtained based on South Africa**'**s return series for the examined period, and they are given in **Table 4**, and **Figure 2**, **Figure 3**, **Figure 4**, and **Figure 5**, respectively. In general, we monitor that the connectedness level in the network reaches high values between 2014 and 2016 and during early 2018 (**Figure 2**). Considering **Figure 2**, 2014, 2016 and 2018 seem to be prominent as the periods of shocks in South Africa. The country-specific events that occurred during these periods, triggering these effects, are summarised in **Table 5**.

After we make the prediction error variance decomposition, two critical pieces of information emerge. These are called variance shares, which originate from "internal shocks" and "external shocks". External shocks are also called "spillovers" in the literature. The connectedness results in **Table 4** constitute essential details, where all the relationships in the system can be defined to understand the **"**big picture**"**, i.e., the South African case.

In this model, for each variable, Y-return, p-positive, and n-negative returns are shown at **Table 4**. The difference of negative and positive returns of the variables means the divergence of positive and negative effects in this model. Therefore, these effects are defined as "asymmetric" effects.

Notably, there is a spillover from the gold market to the platinum market. The value of (11.16) can be interpreted as a measure of this effect. In general, the positive and negative effects in the gold market do not have an asymmetrical effect on the gold market. This is underpinned by the findings that the gold**'**s positive and negative returns are effectively at similar rates, i.e., gold. Yp (12.65) and gold. Yn (12.22) values are close to each other showing this fact. There is a similar situation occurring from the platinum market to the gold market. However, the gold market appears to be a higher shock receiver than the platinum market because the spread from pla.Y to gold. Y is (13.11), while the reverse spillover is at the level of (11.16).

The cds.Y variables receive the biggest shock from themselves. However, there is a difference between the positive and negative effects, considering the financial markets**'** perspectives of the South African economy. The positive and negative difference here is that the negative shock of cds.Y causes a faster increase than positive shocks, reflecting that the increase in CDSs will have a negative effect on the economy. Since cds.Yp expresses the positive shock increase direction in our CDS variable, it expresses the negative impact on the economy. From this point of view, the impact of the shock in international financial markets on South Africa is asymmetrical in both positive and negative directions. In this context, it is possible to say that the general effect of this

