Inflow of Capital and its Sudden Stop in the 2000's: The Case of Serbia

Abstarct

The paper explores and assesses impacts of large capital inflows and subsequent sudden stop and capital reversal on aggregate demand, exchange rate, inflation and economic activity in an emerging Europe transition country – Serbia. The examination covers the period of the 2000s when capital flowed extensively downhill into emerging Europe as well as the following financial crisis. Thus results obtained for Serbia may be relevant for some other emerging Europe countries.

A stylized pattern suggest that capital inflows, with relatively fixed exchange rate (e.g. ‘fear of floating’) would increase money supply, and hence aggregate demand thus leading to inflation and rise in output. If however exchange rate is flexible, the main impact would be nominal currency appreciation. And the other way round in the case of sudden stop and capital reversal.      

Thus we estimated relation between price level (core prices pc), nominal exchange rate (epe) and real broad money (bmr), but also price of oil (poil) is included as it varied substantially in the 2000s.

A clear-cut cointegrating relationship (Johansen, 1996) for the price level is obtained:           

       pc = 0.53 epe + 0.28 bmr + 0.17 poil + 0.19 dum_oct2008               (1)
                (15.38)         (9.00)    (5.64)         (8.05)
_{Note: t-ratios are in parentheses; Dummy variable that takes take non-zero value 1 for the period of financial crises (from October 2008 to the end of the sample); Monthly: December 2002 to August 2009.}

Thus the price relation above is stable across both inflow and sudden stop episodes. As to the inflow episode, it captures two distinct sub-periods: one with relatively fixed exchange rate and the consequent inflation, and a subsequent one when currency was allowed to appreciate thus curbing inflation. Upon the crisis eruption and depreciation of 20%, inflation has not increased. Namely, changes in real broad money, affected by a sudden stop, kept it under control, as predicted by the price relation above.

Reduced Phillips curve corresponding to the price equation above (1) is estimated by GMM procedure and forward vs. backward looking behavior of the public assess (cf. Gali et al., 2005):        

Dpc_t = 0.05 Dbmr_t-1 - 0.03 Dbmr_t-2 + 0.32 Dpc_t-1 + 0.64 Dpc_t+1   ;              R² = 0.41           (2)

          (4.15)              (-2.49)            (2.80)         (5.01)                     J-stat. = 0.03

Preliminary estimates for Serbia suggest that the public is more forward than backward looking.

Relation (1) enables one to assess also the size of pass-through from exchange rate to price level of tradables (core prices in Serbia almost completely coincides with tradables):

1. Pass-through is (0.53) modest for the tradables in a small open developing country – result recently obtained for other developing countries.
2. Pass-through is relatively stable, i.e. has not changed neither during nominal appreciation period nor upon abrupt depreciation  during the recent crisis.
3. Pass-through into tradables is approximately twice as large the one into retail price index (i.e. both tradables and non-tradables). Thus nominal depreciation can engineer real depreciation, that will be badly needed for post-crisis adjustments of large current account deficits in a number of emerging Europe countries.