Subnational economic complexity analysis: case-study of the Kaliningrad region

. Currently, the economic complexity (EC) theory is of considerable relevance. Developed and rapidly developing countries invest heavily in research and development to increase their products complexity as it brings an economy’s competitiveness and revenues to a higher level. The article presents the main results of the 2017-2019 EC analysis of a Russian exclave, the Kaliningrad region, whose trade and production specialization have changed dramatically. The study relies on the data of the Atlas of Economic Complexity, the Federal Customs Service of Russia, the Kaliningrad Regional Customs. It applies the author's method for "cleaning" the data. The key feature of the study is the incorporation of the regional data into the global trade statistics. The analysis reveals general trends towards an increase in capabilities in low complexity products. The paper emphasizes that the regional government needs to pursue an active sectoral policy aimed at increasing the economic complexity.


Introduction
According to the UN 2030 Agenda, sustainable development implies the balance between social, economic and environmental spheres.One of the ways to achieve this balance is to increase the economic complexity (EC) of a country or a region [1].According to studies, it is accompanied by reduced air pollution [2,3], decreased economic inequality [4,5] and increased human development [6], all of which is in line with the UN sustainable development goals.
However, most studies on economic complexity are conducted at the national level, which primarily stems from the availability of trade statistics.In recent years, there have been works on subnational economic complexity (for example, [7,8,9,10]), however, it is not always possible to apply the methods they propose to regions in other countries due to national specifics of statistical recordings.

Materials and Methods
Figure 1 shows the general framework for measuring and assessing changes in subnational economic complexity.The study of foreign trade flows of the Kaliningrad region was carried out using the customs statistics of the Federal Customs Service of Russia and the Kaliningrad Regional Customs, which made it possible to combine international and interregional export data.The exclave position of the region results in a high level of openness of the economy comparing to the other regions of Russia [11].In 2019, it exceeded the average regional level 10-fold.The key issue in preparing data for the EC analysis is "cleaning" them, i.e. processing both exports and imports data to exclude commodity flows that are not related to the result of the region's economy.These are, first of all, re-export flows, i.e. the export of foreign goods from the territory of the region, temporary exports, goods that initially entered the region as inter-regional imports, etc.The study covers the period of 2017-2019 since the EC analysis relies on the global trade data with the latest being currently available for 2019.
4-digit HS codes.Number of lines in the database: 1,719.Source : Database of the Kaliningrad Regional Customs (Customs office records interregional trade data as the region is an exclave)
The data on the products coming to the region as a part of interregional import is removed.

Assesing the results
Calculations rely on Harvard Growth Lab software : https:// github.com/cid-harvard/py-ecomplexity.The Kaliningrad region s data are incorporated into global trade statistics.

Collecting Kaliningard region s trade data
Source : Open database of the Federal Customs Service: https:// stat.customs.gov.ru/unload.

Fig. 1.
A structural and logical framework for assessing changes in subnational economic complexity.Source: [12] Preparing the data for measuring subnational EC level, the authors used their calculations and methodology [13,14], as well as the Harvard Growth Lab software (https://github.com/cid-harvard/py-ecomplexity),and global trade data of the Atlas of Economic Complexity.
The study applies general scientific research methods (statistical, comparative, descriptive, etc.) and EC analysis methods.
Table 1 shows the general structure of the combined export of the region for the study period.Source: Authors' calculations The combined export of the region was formed mainly by the groups "84-90" and "01-24".The largest share of groups "84-90" in the export structure was recorded in 2018.The data analysis indicates a low degree of export diversification.
The use of the Harvard Growth Lab software to process "cleaned" aggregate export data allowed us to assess the changes in the level of economic and product complexity of the region (Tab.2).An indicator of a sustainably developing economy is a decrease in low PCI products and an increase in high PCI products.In the Kaliningrad region, only 36% of the export growth is accounted for by the products with an increased PCI, which is not satisfactory.At the same time, there is a 92% decline in exports of products with a decreasing PCI, which is a positive trend.
The implied capability density has increased for all 158 product categories for which data are available for the entire study period (Table 4).This means that the economy is increasing its capabilities underpinning the products that its exports.Over time, the economy will increase its export performance in product categories where its revealed comparative advantages have increased.Table 4 shows that the net export increase in product categories with increasing RCA is 2,695 mln.rub.and the net export decrease in product categories with decreasing RCA is 229 mln.rub.
The most desirable categories are those where both PCI and RCA are increasing.These include 8701 "Tractors", 9403 "Furniture and parts thereof", 8309 "Stoppers, caps and lids", 8418 "Refrigerators, freezers and other refrigerating or freezing equipment", 5703 "Carpets and other textile floor coverings", 7308 "Structures and parts of structures of iron or steel", 7326 "Articles of iron or steel", 8413 "Pumps for liquids; liquid elevators; parts thereof", 2923 "Salts and hydroxide", 4814 "Wallpaper", etc.

Conclusions
Economic complexity analysis is a suitable tool to understand the competitive development of regional economies provided that there is interregional export and import statistics available.The Kaliningrad region's data meet this requirement, thus a subnational economic complexity analysis can be carried out.
Structural changes in an economy normally take some time, however, the three-year analysis allows identifying some general trends.These include the decline in the export volume-weighted product complexity and the export weighted revealed comparative advantage.
Combined with the increase in implied capability density, this indicates that the economy is deepening its capabilities underpinning the export of low complexity products resulting in increased revealed comparative advantage in these product domains.In times of global structural change (greening and digitalisation of the economy), this is not necessarily providing a foundation for success during and after this structural change.
Given that the smaller an economy, the greater the extent to which market failure becomes its feature, to ensure the future success for the Kaliningrad region, it is essential for the government to conduct an active industry policy.Its aim shall be building capabilities for future and developing competitive export industries in higher complexity products founded in currently exported products and present capabilities (given that economic development is path-dependent).
The analysis indicates that most of the industry policy recommendations presented in the original analysis [14] still stand.

Table 1 .
Commodity structure of the combined export of the region in 2017-2019, %

Table 2 .
General EC analysis of the Kaliningrad region, 2017-2019There are 158 product categories for which the data are available for 2017 -2019.Table3provides a visual representation of the development for these categories.

Table 1 .
PCI for the Kaliningrad region's exports in 2017 -2019