Global climate change is likely to affect the wellbeing of citizens of the Russian Federation, with many regions in permafrost areas and the Arctic Zone, with large forested areas, and with an agriculture adjusted to the current climatic conditions. The question is, whether citizens are sufficiently aware of these challenges in order provide necessary support for related activities of the public administration. In the paper, awareness is introduced by means of a “Multiple-IndicatorMultiple-Causes” (MIMIC) model with “indicators” derived from regional search entries in ⃝cYandex, whereas “causes” result from economic and socio-economic factors. The empirical study investigates dependence of awareness on characteristics of regions with arctic and subarctic zones, the dependence on the devaluation of the Russian Rouble end of 2014, and the influence of the UN Climate Change Conference in Paris 2015.
Are Russians concerned of climate change, which is likely to affect the country in a variety of ways? A
decreasing permafrost area in the arctic zone will require adaptations in urban and industry planning, and global
warming might have serious effects on agriculture
Among the possibilities to conceptualize environmental awareness in order to make stringent use of it in
academic research, [Lin, 2015], by applying Ajzen’s “Theory of Planned Behavior”
This paper is based on a special case of structural equation modeling, the “Multiple Indicators-Multiple Causes” (MIMIC) model that was introduced originally by [J¨oreskog and Goldberger, 1975]. It uses well defined indicators to measure a latent construct and regresses them against certain causes. Along these lines, [Khakimova et al., 2017] use this approach to analyze regional aspects of awareness, whereas [L¨osch et al., 2017] investigate the geographical diffusion of awareness in a time-dependent context.
Regionally stratified search entries in ⃝c Yandex, collected over a period of 28 months (January 2014 – April 2016) are serving as indicators for awareness of climate change, thus revealing interesting differences among Russian regions. End of 2014 started the significant decline of world market prices of oil and gas, and end of 2015 many countries assembled in Paris for the Climate Change Conference. Did these events leave their footprints on awareness, in particular in the vulnerable regions with arctic or subarctic zones?
The paper is structured as follows: the next section addresses the methodology including relevant aspects of the MIMIC model and the data. Thereafter, some empirical results will be discussed with a focus on the arctic and subarctic zone. Some final remarks conclude the paper. 2 2.1
As mentioned, the index of awareness of climate change is derived as a latent variable from search entries in ⃝c Yandex. Economic and socio-economic factors, likely to affect this index, are integrated as “causes”. The MIMIC model with seasonal and trend components is used to estimate the proposed index.
The measurement part includes the indicators y, “the pillars”
˜ + ":
The structural part includes the exogenous causes variables x = (x1; x2 : : : ; xk)⊤, and the model parameters
and . The random errors are assumed to be independent from the other random factors ". Moreover, the
matrix z includes three binary variables for the quarters (reference is the 4th quarter) and two binary variables
for the years (reference is the year 2016):
˜ =
⊤x +
⊤z + :
The parameters as well as the variances of the error terms are then estimated using a ML approach
Relative numbers of queries from ⃝c Yandex, filtered according to approximately 200 climate-related phrases in Russian or English, constitute the indicator variables y. These phrases, and subsequently also the queries, are clustered into the following categories: [Y1] Climate Change Queries, [Y2] Endangered Environment Queries, [Y3] Political Queries, [Y4] Science Queries, [Y5] Renewable Energies and Technologies Queries.
Following the data collection, the numbers of compatible requests in each region and in each category are summarized and divided by the number of all search requests from ⃝c Yandex in these regions. This yields the indicator variables y: yin = number of queries of category i in region n number of all queries in region n ; where i = 1; : : : ; p with p = 5 refers to the categories, and n = 1; : : : ; N with N = 81 to the regions of the Russian Federation. 2.2
As already mentioned, ⃝c Yandex filtered the Internet queries from January 2014 to April 2016 (28 months) according to the environmental phrases provided and according to the Russian regions. Table 1 provides the descriptive statistics.
The data for the regional cause variables are on year level available and provided by the Federal Statistics
Service of Russia
0 0 4 Ixnde 300 e g hanC 200 e t liaCm 100 e g rvea 0 a 0 0 1 −
2e+05 6e+05
8e+05 4e+05
GRP per capita
End of 2014 is marked by the substantial fall of the world market prices of natural oil and gas, leading to the decline of the Rouble-Dollar exchange rate and to other negative consequences for the Russian economy, which is dependent on the export of oil and gas. Whereas this development might adversely affect efforts to mitigate climate change, the Climate Change Conference in Paris end of 2015 inspired many countries to increase efforts to fight climate change and its consequences. Did these milestones leave any footprints in the levels of awareness, particularly in regions with arctic and subarctic zones?
Figure 2 shows the development of awareness (with a 90% confidence interval) in the arctic and subarctic zones in comparison to that in the other regions. The regions with arctic zones comprise Arkhangelsk, Nenets AO, Yamalo-Nenets AO, Chukotka AO, Murmansk, Sakha, and Krasnoyarsk. Moreover, Khanty-Mansi AO, Karelia, Komi, Magadan, Kamachatka, and Leningrad with St. Petersburg are included in the regions with subarctic zones. Interestingly, awareness tends to be higher in these regions in comparison to the rest of the country. ixnde 030 e cnagh 002 ilteam 100 c f sso 0 e renaw −010 a 0 0 2 −
Subarctic Arctic all others
Moreover, both the Rouble devaluation and the climate change conference are followed by a temporary, but at first stronger decline of awareness, also in the regions with arctic and subarctic zones. Whether this decline is a consequence of other developments, remains to be seen and has to be left to a further, more detailed investigation. Nonetheless, the largest differences between the climate change index of the arctic regions and all others are in February and March 2015 as well as, especially, in the same month in 2016.
Rouble Devaluation
Climate Change Conference Paris 01/14 05/14 09/14 01/15 05/15 09/15 01/16 month
Figure 3 details the regions with arctic (blue) and subarctic zones (orange) as well as all others (grey). Actually the region Krasnoyarsk extends from the north to the south of Siberia, which could strongly influence the variability of the arctic climate change index. But if we exclude this region as an arctic area, the results hardly change from Figure 2. [L¨osch et al., 2017] provide more information on the “diffusion” of awareness of climate change, which seems to spread from the eastern parts of the country to the western parts. The paper demonstrates how search requests submitted to ⃝c Yandex can be filtered, structured and used to get some idea on the level and the development of awareness of climate change, depending on the specific situation in the various regions of the Russian Federation.
The focus on regions with arctic and subarctic zones reveals a higher level of awareness, especially in the vulnerable northern parts of the country. This is a signal for accommodating public policies for mitigating climate change and adapting to its consequences. We are very grateful to ⃝c Yandex, which provided us with the necessary empirical data from Russian regions. Research was supported by RSF grant Nr. 15-18-20029 “Projection of optimal socio-economic systems in turbulence of external and internal environment”. public