Determinants of low inflation in emerging, small open economy. Comparison of aggregated and disaggregated approaches

After 2011 headline inflation in many countries has been decreasing excessively raising the question of the determinants of recent unexpectedly low and persistent inflation (European Central Bank, 2016). Excessive disinflation has coincided with a protracted period of an anaemic recovery in many economies as well as plummeting commodity prices. The discussion on the determinants of muted inflation is mainly focused on highly developed economies (e.g. Bobeica and Jarocinski, 2016). Relatively little evidence is presented for emerging economies facing persistently low inflation. For this purpose we employ Polish data and investigate empirically the drivers of inflation in small open economy by estimating a structural vector autoregression model using Bayesian methods. We contribute to the existing literature in two ways. Firstly, we disentangle the structural shocks influencing headline and core inflation and assess the importance of foreign and domestic, supply and demand factors in shaping inflation developments in Poland. These shocks are identified using sign restrictions. Secondly, we extend this analysis on a large number of disaggregated inflation components to investigate whether the disaggregated approach provides a similar picture of determinants of price dynamics. Moreover, this method enables us to isolate components of inflation sensitive to changes in foreign and domestic, supply and demand factors and construct indices of inflation sensitive to the distinguished determinants. To check whether the excessive disinflation period resulted from a structural shift, we split our sample to exclude the period of systematically falling inflation and re-estimate all models. Thirdly, we provide useful evidence for monetary policy conduct in Poland about the determinants of the unravelling phenomenon of low inflation. The substantial strand of empirical literature related to the propagation of global and domestic shocks into the small open economies focuses on analyses performed using aggregated data. The analysis on the disaggregated inflation is relatively scarce. However, research on the disaggregated data reveals that not all inflation components react in the same manner. Bryan and Meyer (2010) point out that the sensitivity of individual prices may differ and that prices of certain goods and services may react to the changes in the domestic economic activity stronger than others. Frohling and Lommatzsch (2011) show that two thirds of the euro area HICP basket reacts to the changes in the domestic slack, however, the results among member states are heterogeneous - larger countries show higher share of output gap sensitive items than smaller ones. Taking into account that disaggregated analysis may show different results than aggregated one we aim to compare results from two approaches. Firstly, we analyse the structural shocks affecting the aggregate level of inflation and we show the historical decomposition of the difference between the long-run level of inflation implied by the Bayesian SVAR model and the observed inflation decomposed into domestic and global, demand and supply shocks in the analysed period. The structural shocks are disentangled by employing sign restrictions and identification strategy put forward by Corsetti et al. (2014) and Bobeica and Jarocinski (2016). Secondly, we run a similar analysis on disaggregated data for around forty major inflation components. The analysis of the shock decomposition can be divided into two periods. The first period comprised the whole sample, which also covers the disinflation period. The second one ends before inflation has begun to decrease excessively, namely 2012. Preliminary results suggest that in the first period cover by the analysis (up to Q2 2104) the domestic demand shock was a main driver of the high inflation what coincided with the high positive domestic output gap. Second source of inflation growth was global demand shock revealed in the faster external GDP growth than the domestic one. Fading domestic demand was the dominant factor leading to a sharp decline in inflation in 2012, reducing CPI in Poland the second and third quarter of 2013 by up to 1 pp. The deflationary period in the Polish economy which started in third quarter of 2014 remains due to the coexistence of the positive global supply and the negative global demand shock. The first of these factors, revealed in the drop in the oil prices, contributed to the average decrease of the CPI by 0.6 pp. The second one, attributed to the weakening of the global economic activity, reduced inflation by an average of 0.7 percentage points. In aggregate terms, throughout the period under analysis the domestic factors remain stronger determinant of price dynamics in the economy, however after the second quarter of 2014 the economy faced growing importance of global factors affecting inflation. At the beginning of disinflation period domestic supply and demand factors prevailed. On the other hand, during deflation these shocks were on average over 50% of the total deviation of the actual inflation from its long-term average. In both analyzed periods, the impact of demand shocks was similar and much larger than the impact of supply shocks.