Market integration is of interest not only in the analysis of cross-border integration (considered in this report), but also in studies into the integration within separate countries (where integration means the establishment of relationships between isolated regional markets) or even in the understanding of the interaction between particular stock markets. It is not by chance that this area of research produced many approaches to the quantitative assessment of integration. It is practical to discuss four main methods of measuring market integration.
1. Integration can manifest itself in the cross-border flows of goods, services, and capital. This approach is usually the simplest one to use when analysing interaction between separate markets, although one should remember that it can produce overoptimistic results. So, the study of the global economy in the 1-18th centuries shows that the growth of sales between regions
was not accompanied by any price convergence between markets, which suggests that the arbitration opportunities were not actually used (O’Rourke and Williamson, 2002). However, an advantage of this method is the relative accessibility of statistical data. At the same time, the interpretation of quantitative indices can be difficult and produces a variety of alternative characteristics.
The classical index to measure trade between countries is the share of the intra-regional trade in the total foreign trade turnover (see, for example, Osterkamp, 2008). A drawback of this method is that when the share of an integration grouping’s countries in the global GDP grows, the share of intra-regional trade rises as well irrespective of whether integration has actually deepened. This distorts the assessment of the integration level for larger groupings. Likewise, the number of countries in a region under consideration influences the index. Hence the need to use a variety of alternative indices that “modify” the index of intra-regional trade. The following alternatives deserve mention:
(1) regional trade concentration indices (that help calculate the analogue of well-known indices such as the Herfindahl-Hirschman Index, the Gini coefficient, and others);
(2) the indices of intra-regional trade intensity for export and import (today, there are several versions of these indices) that are, simply said, a “weighted” index of the share of intra-regional trade where the “weights” are represented by the aggregate trade of the respective partner countries;
(3) the indices of export absorption capacity and import saturation capacity that help determine the complementarities of trade between the countries and their combinations and are, in essence, a modification of the revealed comparative advantages indices, which are actively used to characterise foreign trade as a whole (Vollrath, 1991; Floerkemeier, 2002; Iapadre, 200; Asian Development Bank, 200).
Most of the existing systems of indicators focused on the measurement of cross-border market integration use both indices of the dynamics of mutual trade (including aggregate trade) and indices of the intensity of intra-regional trade. Thus, for the Asia and Pacific region these calculations are given in the systems of integration indices of the Asian Development Bank (ARIC, 2009) and UNESCAP (UNESCAP, 2009). Another integration index that can be used is the variability of intermediate imports resulting in an increasing variety of imported products from the previous stage of a production chain becoming available for national industry (Madani, 2001). However, there are two effects that should be distinguished: the effect of competition with national producers of the same goods as those imported, and the complementary effect that has a positive impact on national industry.
An alternative to the measurement of the intensity of trade is the network analysis of trade flows. In this case, standard network characteristics (such as the centrality or closeness of connections) are used to quantify regional economic integration. This approach was used by, for example, (Iapadre and Tironi, 2009) to assess regional trade integration in East and Southeast Asia. And, finally, of the most interest is the assessment of integration through the comparison of factual and contra-factual results of gravitational regressions4. To this end one should first
calculate the volume of mutual trade between countries using a “theoretical” model (primarily standard gravitational regression, according to which mutual trade is directly proportional to GDP and inversely proportional to the distance between countries). In this case the integration index is the remainder in the assessment of regression. For example, if the remainder in the measurement of trade between a country and a region is a significant negative number for a particular country, then the volume of trade between that country and the region is subtantially greater than the “forecasted” theoretical trade and, consequently, the attained level of integration is worthy of note (Bussiere et al., 200). However, this approach is rather time-consuming and cannot really be used to measure integration systemically.
The integration indices for various factor (capital, workforce) markets can be calculated, in principle, by analogy with the afore-mentioned indices. So, in the regional integration measurement methodology proposed by (Dennis and Yussof, 2003) for ASEAN, the components of the integration index include the measurement of intra-regional trade and investment. The problem is however that the data on the global factor migration is, at best fragmentary; in addition, there is no accurate data on the “sectoral” specialisation of flows, and data cannot be generated for migration in principle. Another factor that needs to be taken into account in the analysis of the factor market integration is the necessity to compare (if possible) the indices of status (such as accumulated investment or the total number of labour migrants) and dynamics (the inflow of investment or migrants) that would appear to be of no interest to the analysis of trade. For this reason, when assessing integration in the area of capital or workforce, the “simplest” indices are normally used (such as the dynamics of the share of investment inflows or accumulated investments), which naturally leaves room for criticism.
For separate functional areas, market integration can be described with the help of specific indices used in a particular sector, such as the number of phone calls (integration in communications) or the trade in food commodities (integration in agriculture or other sectors).
2. Integration is reflected in the structure of prices: the law of one price governs the integrated markets, i.e. prices of similar goods in various countries or regions of a country should be the same. This approach is usually preferred when internal integration within a country is analysed, and this is connected to the level of development of econometric tools (for example, based on an analysis of the cointegration of series of prices), including those for post-Soviet states (Gluschenko, 2008). However, in studying cross-border integration, the analysis of price dynamics has limited application because of the accessibility of data (both in space and in time, cf. Dreger et al., 2007). Three approaches to the analysis of integration based on the law of one price can be distinguished. First, price convergence can be analysed at the micro-level for particular markets, for example in particular consumer goods (Gil-Pareka and Sosvilla-Rivero, 200) or raw materials (e.g., Findlay and O’Rourke, 2001). Second, financial markets can be studied (interest and exchange rate correlation, cf. Babetskii et al., 2007). Third, subject to particular assumptions, aggregated market indices can be the basis for analysis, as in the related analysis of the purchasing power parity (cf. Qin et al., 2007; Kim and Lee, 2008).
3. Integration can manifest itself in consumer behaviour in various countries. When the level of integration is high, the players can “insure” themselves against specific shocks by buying assets and products in other countries or regions, as a result of which consumption in countries or regions should correlate better than production (Christelis et al., 2008). This approach is obviously interesting, first of all, for the research into capital market integration. An indirect method to assess integration in the latter can be the β-regression coefficient, which describes the share of investment in country i against the percentage of savings in that country: the higher the level of integration, the lower coefficient, i.e. the correlation between internal savings and investment (Bilas, 2007).
4. The level of regional specialisation can be considered as an indirect integration index: the higher level of market integration, the higher the motivation of the various regions for specialisation. This approach has, however, at least two drawbacks. Firstly, it ignores the findings of the New Trade Theory that emphasise the role of intrasectoral trade. Secondly, in its essence, it contradicts the logic of the economic convergence indices discussed below.