Abstract:
In this thesis, we examine and forecast the growth cycle in Lebanon through business survey indicators. Survey indicators are combined into a composite index named “Confidence” in each sector to replicate current business conditions on a sectoral level. These indicators are then aggregated into a single composite indicator referred to as ESI to match the reference series and reflect overall economic activity in Lebanon.
On the basis of the quarterly approximation of the Bry and Boschan algorithm, the ESI dating chronology was closely able to match that of the reference series with high degree of concordance between phases.
For more robust findings, short-term forecasts of quarterly economic growth(CI) are compared between time series models incorporating the ESI against their corresponding benchmark (AR, ARIMA) by adopting a pseudo out-of-sample forecasting technique. The results show more than 30% improvement in forecast accuracy across the different horizons, for the AR augmented model compared to its benchmark.
In conclusion , business survey indicators presented by the ESI were able to monitor past economic development of the CI as well as performing short-term forecast for different horizons.