Page 39 - Azerbaijan State University of Economics
P. 39

Fatih Chellai: Regime-Dependent Effects of Public Spending in Algeria: A Structural VAR and
                                                            Markov-Switching Approach


                    Estimation and identification procedure
                    The SVAR model estimation and identification procedure will follow a structured
                    approach:  First,  stationarity  will  be  tested  using  unit  root  tests  (e.g.,  ADF),  with
                    differencing applied if necessary. Next, the optimal lag length (  ) will be selected
                    using  information  criteria  (AIC,  BIC,  or  HQ),  followed  by  the  estimation  of  the
                    reduced-form  VAR  model  via  OLS.  Structural  shocks  will  then  be  identified  by
                    imposing  restrictions  on  the  contemporaneous  (  )  or  long-term  impact  matrix,
                    drawing  on  economic  theory  and  Algeria-specific  conditions—such  as  assuming
                    public spending shocks do not instantly affect GDP or treating oil price shocks as
                    exogenous. Cholesky restrictions will be explored initially, with alternative schemes
                    considered  for  robustness.  Subsequently,  impulse  response  functions  (IRFs)  will
                    analyze  dynamic  responses  to  shocks,  complemented  by  variance  decomposition
                    (VD)  to  quantify  each  shock’s  contribution  to  forecast  error  variance.  This
                    methodology  will  elucidate  the  dynamics  of  public  spending’s  impact  in  Algeria,
                    offering insights for economic policy formulation.

                    RESULTS
                    This section presents and analyzes the main empirical results from the estimation of
                    the SVAR model applied to the Algerian economy over the period 2000-2023. The
                    aim is to assess the dynamics of interactions between public spending, real GDP,
                    inflation and unemployment, while taking into account the exogenous influence of oil
                    prices.  After  checking  the  statistical  properties  of  the  series  and  identifying  the
                    optimal structure of the model, we successively present the results of the diagnostic
                    tests, the impulse response functions and the variance decomposition, in order to better
                    understand  the  transmission  channels  of  fiscal  shocks  and  their  impact  on
                    macroeconomic equilibria.

                    Descriptive data analysis
                    Over the period 2000-2023, the data reveal continuous growth in Algeria's real GDP,
                    rising from 4,454 billion dinars in 2000 to over 8,726 billion in 2023. This growth,
                    although steady, appears to be strongly correlated with the evolution of oil prices,
                    which have gone through marked cycles. The spectacular rise in oil prices between
                    2003 and 2008 (from 28.85 to 96.94 USD) coincided with an acceleration in GDP,
                    supported by the expansion in oil revenues. After a period of stabilization between
                    2011 and 2014 at around 100 USD per barrel, a sharp fall was observed in 2015 (52.32
                    USD), moderately impacting growth,  without reversing it. It  should be noted that
                    public spending followed a continuous upward trend, rising from around 20 billion to
                    over 44 billion, reflecting an expansionary fiscal policy, probably to support demand
                    and cushion external shocks linked to the volatility of oil revenues.


                                                           39
   34   35   36   37   38   39   40   41   42   43   44