Context:

The currency market is known as the most liquid market in the financial system. Its strong repercussion in the economy is tied to the capitalization and the impulse that this market offers through the increase of investments and therefore of macroeconomic effects that is found in the development of the economy, For these reasons, it is relevant to study models that predict the behavior of the main Latin currencies.

Method:

Two models are proposed for the prognosis and the identification of factors implicit in the behavior of these currencies. The first proposed model corresponds to the Black-Scholes, which allows obtaining the future price of the currency based on integrating price processes, a Wiener process and a static volatility is recognized. The second model is Heston, which describes the evolution of the volatility of an underlying asset and assumes a stochastic volatility.

Results:

After the application of the proposed mmodels, their efficiency are evaluated by means of the Diebold- Mariano Test, in order to identify the forecast model that best adapts models, their efficiency areto the real behavior of the parities.

Conclusions:

It is identified that the Heston model offers a better fit to the forecast, assuming random volatility in the short term for each of the currencies, while the long term presents the largest adjustment of the Black

& Scholes model, It is evident that the longer the forecast time, the greater the uncertainty and the greater the prediction error.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )