• Initiative -
  • State Active Project

Project Aim

Leading indicators of economic activity are used widely in OECD countries, although not in Saudi Arabia.  To the best of our knowledge, a composite leading indicator of economic activity has not been developed by the Saudi Central Bank or GASTAT.  ‘Heat maps’ for Saudi Arabia have, however, been constructed by private sector entities such as Bloomberg Economics.  Heat maps are essentially tables which report on the trends in constituent variables, such as those presented in the list below:

  • Purchasing Manager’s Index (PMI) for Saudi Arabia – overall index or sub-indices.
  • Monthly usage of cement.
  • Money supply (M3)
  • Cash withdrawals from ATMs
  • The value of cheques cleared – a possible proxy for consumer spending.
  • Point of sale transactions
  • Stock market index reading (for TADAWUL all share index).

A ‘trend’ may simply be measured as a year-on-year growth rate deflated by the consumer price index. The constituent variables are examined singly, or in conjunction with one another, and are used to gauge qualitatively what the result might be for quarterly private sector GDP in the next quarter.

The current project will apply the more commonly used leading indicator variables, and other data series that are produced with high frequency. The time-series data is often non-stationary. A top-down model of vector auto-regressions will be employed. Forecasts which extend out by several quarters will be prepared.

Projections will also be prepared for the institutional, non-oil government sector GDP. A different set of leading indicator variables will be applied in this context.

Seasonal adjustment of the quarterly GDP data is a core requirement. GASTAT reports that it applies the TRAMO-SEATS method of seasonal adjustment of quarterly GDP data, although the full historical, seasonally adjusted series of GDP and its components has not been released. Under TRAMO-SEATS, there are at least two steps involved in the seasonal adjustment process. The second step uses an ARIMA-based decomposition of an observed time series into unobserved components: The trend cycle, a seasonal component, and an irregular component.

We will apply TRAMO-SEATS and other methods for seasonal adjustment to the institutional non-oil GDP (private sector and government sector). Other co-variates which appear to exhibit patterns of seasonality will also be subject to seasonal adjustment.

Key questions

  • To what extent can we identify turning points in the trajectory of growth of real private sector GDP in Saudi Arabia?
  • Can we use leading indicators in a quantitative framework to formulate projections for the institutional, non-oil government sector GDP?
  • Does the evidence support the notion of a de-coupling of institutional non-oil sector activity from institutional oil sector activity?
  • How do the results vary when different methods of seasonal adjustment of the quarterly GDP data are trialed?
  • How useful are Bayesian econometric methods for the development of macro-economic forecasts in Saudi Arabia?
  • How do the different leading indicators of economic activity perform when assessed within a Bayesian econometric framework?

Stay informed

I'm interested in

Select the updates you'd like to receive from us


A bit about you