The Balassa‐Samuelson Hypothesis in Estonia: Oil Shale, Tradable Goods, Regulated Prices and Other Culprits

This paper analyses the Balassa‐Samuelson (B‐S) framework for the case of Estonia using a unique dataset that consists of a 15‐sectoral breakdown of GDP and a five‐digit level CPI disaggregation with 260 items over the period from 1993 to 2002. Unlike the existing literature, the paper focuses on the following four aspects of the phenomenon: (a) data disaggregation, (b) definition of goods tradability, (c) price regulatedness in services and (d) possible heterogeneity across transition countries. It turns out that the first three aspects do matter and, in addition to this, Estonia appears to bear very specific characteristics when compared with other transition countries. A battery of cointegration techniques (DOLS, ARDL, Johansen) shows that productivity is strongly related to relative prices only when regulated prices are controlled for appropriately in the consumer price index and when country‐specific classification is applied to the open and closed sectors. The B‐S effect contributed to CPI by 1 to 1.5 per cent at the outset of the period and by 0.4 to 0.6 per cent in 2002, whereas its potential long‐run impact is estimated to be 1 to 1.2 per cent. Although real appreciation due to the B‐S effect seems higher in the early 1990s, it explains that better real appreciation occurred in recent years.