Forecasting Foreign Direct Investment to Zimbabwe: A Time Series Analysis

Author

Sibanda, Methembeni. M

Title

Forecasting Foreign Direct Investment to Zimbabwe: A Time Series Analysis

Abstract

This thesis presents an empirical study which attempts to model and forecast time series data of
Zimbabwe net Foreign Direct Investment spanning from 1970 to 2021, yearly time series data was
used. This study is built upon the fundamental that the researcher is tempted to feel perched among
the pioneers of research of this kind, narrowing the gap on FDI, by assessing the performance of
the ARIMA and GAM models in terms of accurate predictions of inflows. Probability sampling
was applied for data collection and the Box- Jenkins ARIMA methodology was applied for
forecasting. The diagnostic checking has shown that ARIMA (0, 1, 1) model as the optimal model
to forecast FDI in Zimbabwe based on the AIC. The ADF test also indicates that the residuals of
the model are stationary, thus confirming its adequacy. The results of the ARIMA (0, 1, 1) model
forecast showed that the FDI inflows will remain constant. Due to the dynamic world GAM model
was utilised and the results showed that the forecasts exhibited excellent properties of being best
linear unbiased estimates (blue) with least Mean Squared Error (MSE) compared to the traditional
ARIMA (0, 1, 1) and they are highly recommended in this research. These forecasts will help
policy makers in Zimbabwe to sustain their effort to expand the tax base, reduce red tape, and
strengthen the regulatory framework to investment and also investor’s friendly policies in order to
attract the much needed FDI. The researcher because of the limited time zoomed on assessing two
models and recommending areas for further studies that other significant models can be used which
were not considered in the research to come up with the best model.

Date

2022

Publisher

BUSE

Keywords

Forecasting

Foreign Direct Investment

Supervisor

MR. Barisa