The impact of trunk pipelines on the local economy: the case of Russia

Размещено на http://www.allbest.ru/

Правительство российской федерации

федеральное государственное автономное образовательное учреждение высшего образования

"Национальный исследовательский университет

"Высшая школа экономики"

Негосударственное образовательное учреждение высшего образования "российская экономическая школа" (институт)

Курсовая работа

The impact of trunk pipelines on the local economy: the case of Russia

Автор:

К.В. Стефанов

Научный руководитель:

М. Троя-Мартинез

Москва, 2020



Abstract

Giant Russian state natural resource companies often claim to work for the benefit of the Russian people. But the reasons for building major new pipelines are sometimes political: for example, to start exporting gas to China. As we know, a gas pipeline can cause some environmental risks in the region, a "resource curse," or worsen the structure of labor markets. This raises concerns about the real impact of new export pipelines on the local economy and nature. In my thesis I study the impact of a major Gazprom pipeline on the local economy, and in the year of construction I find a high positive correlation between economic activity and the proximity to the pipeline. However, in the post-construction period there is no correlation. Besides, there is no correlation for desert (without initial activity) and highly urbanized territories. Thus, I did not find any connection between the gas pipeline and the increase in long-term economic activity. This puts into question the statements of companies that gas pipelines bring civilization to desert areas.

In turn, the study of harmful emissions found no correlation between the pipeline and additional environmental threat. This result is subject to further verification due to the low reliability of Russian government statistics.

Российские государственные природно-ресурсные компании-гиганты часто утверждают, что работают на благо российского народа. Но часто причины для строительства новых крупных газопроводов носят политический характер - например, начать экспортировать газ в Китай. Как известно, газопровод может вызывать некоторые экологические риски в регионе, "ресурсное проклятие", или ухудшить структуру рынков труда. Это вызывает озабоченность по поводу того, какое реальное влияние новые экспортные трубопроводы оказывают на местные экономику и природу. В статье я провожу исследование влияния крупного газопровода "Газпрома" на местную экономику, и в год строительства нахожу большую положительную зависимость экономической активности от уровня приближенности к газопроводу. Тем не менее, в последующие годы зависимость не наблюдается. Кроме того, нет корреляции для пустынных (без первоначальной активности) и высокоурбанизированных территорий. Таким образом, в результате исследования не обнаружилось связи газопровода с повышением долгосрочной экономической активности. Это ставит под вопрос высказывания компаний о том, что газопроводы приносят цивилизацию в пустынные районы.

В свою очередь изучение уровня вредных выбросов показало, что трубопровод не несет экологической угрозы для региона. Этот результат подлежит дальнейшей проверке ввиду невысокой степени достоверности данных государственной статистики.



Contents

Introduction

1. Literature Review

2. Data

2.1 Economic activity proxy (Night Lights) Data

2.2 Geographic data

3. Identification strategy

3.1 Choice of the Initial distribution form

3.2 Economic Impact

3.3 Ecologic Impact

4. Results

4.1 Economic impact

4.2 Ecologic impact

Conclusion

Appendix

Bibliography



Introduction

Giant Russian state natural resource companies often claim that the building of the major pipelines helps the local Russian economy to develop:

“First of all, (the pipeline) is necessary for Russia. It is the development of regions where no humans have ever set foot. Places where were almost no jobs, electricity or communication. The gas pipeline that comes to the region will let it prosper. Its' construction will be followed by civilization and wellness, industrialization and technical progress”2

But often the reasons for the construction of new major pipelines that happen to go through previously not gasified regions, are political - to start providing gas to China (in Power of Siberia case). As we know, a gas pipeline can cause some environmental risks in the region, a "resource curse," or worsen the structure of labor markets. This raises concerns about the real impact of new export pipelines on the local economy and nature.

In my thesis, I mainly study the short-term and long-term impacts of pipeline construction on local economy, using the Night Lights data. The Night Lights data is derived from the satellite images of Earth: the intensity of lights is proven to be a good proxy for economic development. In an addition, I study the relationship of pipeline with local ecology, using Emissions data from the RosStat database.

The level of research of local economy is municipality, divided into 10-km distance from the pipeline intervals. The period of research of economy is 2004-2013, with Year of Construction being 2010. The level of research of local ecology is 2008-2017, using RosStat database, on municipality level.

The object that I'm researching is “Sakhalin-Khabarovsk-Vladivostok”, 1800km major pipeline of Gazprom, that delivered the gas from Sakhalin to the Far-East Russian regions and other countries as export.

For uses of my research, I divide the timeline into 3 periods:

Initial period: Before 2007, there were no gas pipelines in the region, and the construction of the “Sakhalin-Khabarovsk-Vladivostok” wasn't yet announced.

Construction period: By official info, the Gazprom began to build the “Sakhalin-Khabarovsk-Vladivostok” pipeline in 2009. The construction was ended in 2011. Nevertheless, I concluded from the data, that the whole construction was made during the 2010.

Post-construction: After that, at least until 2017, no new pipelines appeared in the region.

The level of research is years. I had troubles getting precise information of pipelines construction timing. In my main pipeline source - energybase.ru - there are only years of beginning of construction of Pipelines and Stations and their starts of service. I contacted Gazprom, but they were not able to provide what I asked for - the reason was that construction is mainly made by subcontractors, not by Gazprom. So, I have just one period of construction for all the pipeline - 2009-2011 years.

In the year of construction, I find a high positive correlation between economic activity and the proximity to the pipeline. However, in the post-construction period there is no correlation. Besides, there is no effect for desert (without initial activity) and highly urbanized territories. Thus, I did not find any connection between the gas pipeline and the increase in long-term economic activity. This puts into question the statements of companies that gas pipelines bring civilization to desert areas.

In turn, having studied harmful emissions, I did not find a correlation between the pipeline and additional environmental threat. This result is subject to further verification due to the low reliability of Russian government statistics.

So, the positive impact of major pipelines is not universal, doesn't bring civilization to the deserted areas, and brings negative outcomes for the distant areas. The study of emissions levels showed that the pipeline doesn't bring ecological threat to the region.



1. Literature Review

First useful paper is the overview of research on local effects of natural resources (Cust and Poelhekke, 2015). There is extensive description of a mainstream research view and plenty of citates. It's intensively citated, and I will study the papers refer to it and maybe find some more relevant papers. Even though there is only one paper on pipelines there (Carrington, 1996), it's very relevant for me and makes largest part of my literature review.

The paper researching the effect of oil pipeline construction on Alaskan Labor Market (Carrington, 1996) is useful for me also because of the studied region - the large sparsely populated area, similar to the Russian regions I study. There are some important insights in the research that help me construct some hypotheses to examine in my thesis. The object studied is the Trans-Alaska pipeline system built in 2nd half of 1970's. The project at the time was the largest localized labor demand shock after the WWII (800 miles, >50000 ppl and 9$ bln of 1977), so its' role for the Alaskan economy was significant. The authors find the short-term growth of hourly and monthly wages and no long-term effect. No wage stickiness was found, interindustry elasticity of labor was limited.

The author studies some features of Alaskan economy and demography, that would be interesting to look at on Russian regional (and maybe even municipal) levels. The region level gives me more opportunity to take the data from firm statistics: I found that in analytic systems like SPARK firms don't have geographic division of parameters, but it's widely known that all the large firms in Russia work through separate subsidiaries for different regions. I will have to find out if they are formally obliged to do so.

There is also a discussion of the project history: the motives of construction and construction key events. I understood that I should include this section on my thesis to explain the origins of my hypotheses and places and periods of my interest.

The paper introduces, in my opinion, sophisticated theoretical approach to the study of labor market. Authors look at how good the data fits the several models of labor and make conclusion about the character of the market. I thought that I could make the same experiment on my data if I will decide to change my focus of the research or broaden it.

Overall, the paper of Carrington was very useful for me, but I still would like to make an effort to have more variation to judge the changes with more confidence, than the author did. For that I used Night Light Density (Lowe, 2014; Henderson et al., 2012) approach. Lowe paper describes the use of this method extensively. Henderson and coauthors also alter the Night Light data so that it's even better correlated with GDP. A haven't yet completed their approach, yet, but will do it in the future.

To add the ecologic effects to my analysis, I studied the paper with complex analysis of ecological situation among Russian cities (Bityukova and Safronov, 2015). Authors created a novel rating of 1100 Russian cities, using integral index containing many ecological measures. They also made recommendations for improving the situation in different types of cities and regions. Most importantly, the paper used the RosStat database, so I got a confirmation of relative reliability of the source and used the same parameters from the database as the authors.



2. Data

2.1 Economic activity proxy (Night Lights) Data

My main source of economic data is Night Light Density. It's a total yearly Night Lights intensity variable available as a high-definition file (Image 1), that can be used in ArcMap. The data has been collected since 1992, multiply proved to be highly correlated with GDP (Lowe, 2014; Henderson et al., 2012), and especially useful to study the scarcely populated and poorly civilized territories. The Night Light Data after 2013 is collected differently from technical point of view, and is held in another format; in another website. I will add it to my research in future.

2.2 Geographic data

a) Administrative division

The administrative level that I chose was municipal. It's gives me much more objects to study than regional division, and fits very well to my purpose: the pipelines are built not strictly through the urban areas, and these non-urban areas are not included in urban districts in Russia. Therefore, municipal is the most granular administrative division that I could get. For my research, I used the shapefile (Image 2) in the ArcMap.

b) Pipeline and distance intervals around it

I collected the data for pipelines from energybase.ru project. They have exact coordinates of Compressor Stations - constructions for the increase of pipeline power, located along the line in somewhat equal intervals. The data used in the project is being collected manually, by the use of YandexMaps. I checked the data - Stations have specific look and are (in many cases) seen well from the satellite (Image 3).

To have the routes of pipelines, I connected the stations. Here is my map of Sakhalin-Khabarovsk-Vladivostok Pipeline and its scheme from Gazprom website (Images 4, 5).

My next step was to export the route to ArcGis program, and create the distance intervals around it to split my municipalities into municipality-distance areas.



I choose to create 10-km intervals, and to include the whole region in the analysis - 300 km around the pipeline. There are 63 municipalities of 4 regions (Sakhalin, Khabarovsky, Jewish, Primorsky) in the area of my research.

Having intersected 63 municipalities with 30 intervals, I got the result in the shapefile (Image 6): 782 municipality-distance areas. All I had left to do is to find the measure of Night Lights for each area.

c) Resulting municipality-distance areas

After calculating the Night Lights for every area in years 2004-2013 and finding the Density measure, I looked at the distribution of it in 2004 (vertical axis is remoteness from the pipeline, 300 is closest) (Graph 1). I saw the outliers with >100.000 Night Lights Density, that could disrupt my results



Graph 2 depicts the areas with less than 100.000 Night Lights Density. The data is distributed evenly enough, but I also wanted to see the areas with little or no Density.

Graph 3 is the areas with less than 1000 Night Lights Density. The areas with no Density are much more common than those with more than 0 and less than 1000 ones.



Having studied the statistics, I understood that the municipality-distance-interval could be divided in 3 segments:

deserted (0 lights in 2004) - 378 areas

medium - 376 areas

highly-urbanized (>100.000 lights in 2004) - 30 areas

All three segments are interesting for research: the pipelines affect the deserted areas differently than somewhat populated and especially the highly urbanized ones. So I researched the segments separately with the identification strategy below



3. Identification strategy

3.1 Choice of the Initial distribution form

I chose to use quadratic form in pre-construction period.

My logic was:

The pipeline is probably located in the areas of higher economic activity (easier to build and maintain), so 50 km interval is probably much more active in 2004-2009.

The coastal area is 250-300 km interval from the pipeline

The coastal area is usually economically active, and the region is not an exclusion: there are many important ports out there.

It's clear from the map (image 6) that the areas located relatively far, but less than 250-km from the pipeline are also located apart from the sea, in the middle of the region. export pipeline economic emission

This all implies the U-shaped 2004-2009 relationship of economic activity and distance from the pipeline. Which means the quadratic form would fit. And it does, as the results show.

3.2 Economic Impact

Economic Activityymp = б0 + б1 Proximityymp + б2 (Proximityymp)2 + в01 2005 Dummyymp + в02 2006 Dummyymp + … + в09 2013 Dummyymp + в11 2005 Dummyymp * Proximityymp+ в12 2006 Dummyymp * Proximityymp+ … + в19 2013 Dummyymp * Proximityymp + в21 2005 Dummyymp * Proximity2ymp+ в22 2006 Dummyymp * Proximity2ymp+ … + в29 2013 Dummyymp * Proximity2ymp + ?o

Where:

Economic Activity is the Economy size (Night Lights) per geographic unit;

y is Year;

m is Municipality;

p and Proximity are Proximities: 10 for 290-300-km away and 300 for 0-10 km away;

is Intercept;

2005-2013 Dummy are Time fixed effects for every year from 2005 to 2013;

is Standard Error;

3.3 Ecologic Impact

Emissions Densityym = б0 + б1 Proximityym + б2 (Proximityym)2 + в01 2005 Dummyym + в02 2006 Dummyym + … + в09 2013 Dummyym + в11 2005 Dummyym * Proximityym+ в12 2006 Dummyym * Proximityym+ … + в19 2013 Dummyym * Proximityym + в21 2005 Dummyym * Proximity2ym+ в22 2006 Dummyym * Proximity2ym + … + в29 2013 Dummyym * Proximity2ym + ?o

Where:

Emissions Density is the quantity of Emissions per geographic unit;

y is Year;

m is Municipality;

p and Proximity are Proximities: 10 for 290-300-km away and 300 for 0-10 km away;

is Intercept;

2005-2013 Dummy are Time fixed effects for every year from 2005 to 2013;

is Standard Error;



4. Results

4.1 Economic impact

Table 1

Dependent variable: Economic Activity
	Non-highly urbanized	Deserted areas	Highly Urbanized areas
Intercept	6.4k*** (1.4k)		2e+05 (1.4e+05)
Proximity	-108*** (22)	1e-12 (7e-13)	-0.1k (1.8k)
Proximity (squared)	0.51*** (0.07)	-6e-15* (3e-15)	1.7 (4.9)
Pre-Construction (2005)	-0.9k (1.7k)	108 (63)	-6e+04 (2e+05)
Pre-Construction (2005) * Proximity	7 (29)	-2.5 (1.6)	519 (2476)
Pre-Construction (2005) * Proximity (squared)	-0 .02 (0.1)	0.01 (0.01)	-1.2 (7.1)
Pre-Construction (2006)	-1.3k (1.8k)	148 (116)	-5e+04 (2e+05)
Pre-Construction (2006) * Proximity	16 (29)	-3 (3)	0.65k (2.6k)
Pre-Construction (2006) * Proximity (squared)	-0.06 (0.1)	0.01 (0.01)	-1.7 (7.5)
Pre-Construction (2007)	-0.6k (1.9k)	15 (146)	-9e+04 (2e+05)
Pre-Construction (2007) * Proximity	-1 (32)	-1 (3.3)	1.3k (2.6k)
Pre-Construction (2007) * Proximity (squared)	0.06 (0.1)	0.01 (0.01)	-3.2 (7.4)
Pre-Construction (2008)	-0.8k (1.9k)	-124 (206)	-5e+04 (2e+5)
Pre-Construction (2008) * Proximity	15.5 (31)	2 (4)	1.3k (2.7k)
Pre-Construction (2008) * Proximity (squared)	-0.04 (0.1)	0.003 (0.01)	-3.5 (7.5)
Pre-Construction (2009)	-0.4k (2.1k)	-133 (170)	-3e+04 (2e+05)
Pre-Construction (2009) * Proximity	5.5 (34)	2.5 (3.6)	1.4k (2.6k)
Pre-Construction (2009) * Proximity (squared)	0.02 (0.1)	-0.004 (0.01)	-3.6 (7.5)
Year of Construction (2010)	3.7k (2.7k)	14 (237)	-4e+04 (2e+05)
Year of Construction (2010) * Proximity	-77.5 (46)	-1.2 (5.2)	2.8k (2.8k)
Year of Construction (2010) * Proxim (squared)	0.46** (0.16)	0.02 (0.02)	-7.4 (8)
Post-Construction (2011)	0.5k (2.2k)	60 (51)	-1e+04 (2e+05)
Post-Construction (2011) * Proximity	-13 (36)	-1.3 (1.2)	1.3k (2.7k)
Post-Construction (2011) * Proximity (squared)	0.11 (0.13)	0.01 (0.005)	-3.4 (7.6)
Post-Construction (2012)	-0.1k (2.2k)	176 (121)	-4e+04 (2e+05)
Post-Construction (2012) * Proximity	-4 (38)	-4 (3)	1.8k (2.8k)
Post-Construction (2012) * Proximity (squared)	0.06 (0.13)	0.02 (0.01)	-4.5 (8)
Post-Construction (2013)	1.4k (2.3k)	133* (66)	6k (2.2e+05)
Post-Construction (2013) * Proximity	-36 (41)	-1.6 (1.3)	1k (2.9k)
Post-Construction (2013) * Proximity (squared)	0.19 (0.15)	0.007 (0.01)	-2.2 (8)
# observations	7540	3780	280
R2	0.166	0.032	0.173
Heteroskedasticity-robust standard errors (HC3) are reported in parentheses. Regressions include all year fixed effects. All columns are estimated by OLS

We can see from the results, that, out of the three years, officially called the period of construction, that 2010 holds the whole effect. Which means we can call it the Construction year from now on. 2009 is one of the Pre-Construction years, and 2011 is Post-Construction.

Medium segment

There is an interesting relationship (Graph 4) of economic activity (Night Lights) with Proximity. It proves the logic by which the choice of route was endogenous, but still there are active regions far away from it - seaside ones. Endogeneity of route choice is very significant.



For 300-290 km away area Construction Year effect is 3006, which is 56 percent of Pre-Construction Years economy level.

For 260-250 km away area Construction Year effect is 1010, which is 45 percent of Pre-Construction Years economy level.

For 210-200 km away area Construction Year effect is 585, which is 85 percent of Pre-Construction Years economy level.

For 160-150 km away area Construction Year effect is 2460, which is 148 percent of Pre-Construction Years economy level.

For 110-100km away area Construction Year effect is 6635, which is 128 percent of Pre-Construction Years economy level.

For 60-50 km away area Construction Year effect is 13110, which is 116 percent of Pre-Construction Years economy level.

For 10-0 km away area Construction Year effect is 21885, which is 110 percent of Pre-Construction Years economy level.

The proportional construction effect of additional kilometer closer to the pipeline doesn't grow linearly, it's rather negative-parabola shaped: biggest effect (148%) is observed with 160-150 km away areas, and 10-0 km area is 110%. On the distant side there is smallest effect (-45%).

In absolute terms the effect has the shape of a positive parabola: from 3006 in areas 400-290 km away to 585 in areas 210-200 km away to 21885 in 10-km interval areas.

The contradiction of proportional and absolute terms happens because of endogeneity of choice of the pipeline route. The municipalities closer to the pipeline were initially much more economically active, and the pipeline construction didn't have that much of proportional effect on them as it did on relatively distant municipalities.

Such correlation in Post-construction period doesn't exist.

Deserted segment

In 2004, all of them had Economic activity 0. In 2010 and Post-construction years the situation didn't change: deserted areas, that are closer to the pipeline, didn't have benefits in becoming active over those more distant. I can explain this the following way: no construction effect - due to the endogeneity of route choice. The pipeline went through the more active areas, so deserted ones didn't benefit at all. The leftover effect for deserted areas is meant to be caused by some structural economic changes, and deserted areas didn't manage to become populated and economically active. This proves that gas pipeline has no ability to “bring civilization” to the deserted Russian regions. Something more than that is needed to make people start living and working in such places.

Highly-Urbanized segment

The areas in this segment are outliers to the other data - have significantly higher Economic activity levels. Turns out, there was no short-term or long-term effect in highly-urbanized areas. The result is interesting and can be interpreted in many ways, but we should remember that there were only 30 areas of this size on the whole sample, so the absence of effect could be caused by the insufficiency of data.

4.2 Ecologic impact

Dependent variable: Emissions Density
	All municipalities
Intercept	2e+06* (1.6e+06)
Proximity	-4e+04 (3e+04)
Proximity (squared)	150 (113)
Pre-Construction (2009)	-1.1e+06 (1.7e+06)
Pre-Construction (2009) * Proximity	2.2e+04 (3.3e+04)
Pre-Construction (2009) * Proximity (squared)	-80 (118)
Year of Construction (2010)	-9.9e+05 (1.7e+06)
Year of Construction (2010) * Proximity	2e+04 (3.3e+04)
Year of Construction (2010) * Proxim (squared)	-70 (120)
Post-Construction (2011)	-1.1e+06 (1.7e+06)
Post-Construction (2011) * Proximity	2.2e+04 (3.3e+04)
Post-Construction (2011) * Proximity (squared)	-80 (119)
Post-Construction (2012)	-1.3e+06 (1.7e+06)
Post-Construction (2012) * Proximity	2.5e+04 (3.3e+04)
Post-Construction (2012) * Proximity (squared)	-91 (118)
Post-Construction (2013)	-1.2e+06 (1.7e+06)
Post-Construction (2013) * Proximity	2.4e+04 (3.3e+04)
Post-Construction (2013) * Proximity (squared)	-88 (120)
Post-Construction (2014)	-1.2e+06 (1.7e+06)
Post-Construction (2014) * Proximity	2.5e+04 (3.2e+04)
Post-Construction (2014) * Proximity (squared)	-93 (117)
Post-Construction (2015)	-1.3e+06 (1.7e+06)
Post-Construction (2015) * Proximity	2.6e+04 (3.2e+04)
Post-Construction (2015) * Proximity (squared)	-96 (116)
Post-Construction (2016)	-1.5e+06 (1.7e+06)
Post-Construction (2016) * Proximity	3e+04 (3e+04)
Post-Construction (2016) * Proximity (squared)	-113 (114)
Post-Construction (2017)	-1.5e+06 (1.7e+06)
Post-Construction (2017) * Proximity	3e+04 (3e+04)
Post-Construction (2017) * Proximity (squared)	-112 (115)
# observations	570
R2	0.170
Heteroskedasticity-robust standard errors (HC3) are reported in parentheses. Regressions include all year fixed effects. All columns are estimated by OLS

There is no effect of pipeline on the emissions density in the region. The initial distribution, endogeneity of choice doesn't exist. This implies that the rise in economic activity in the region doesn't bring additional ecological issues (at least in the short-run).

One concern should be addressed: the data source is not completely reliable: The State owns the Gazprom, and local officials don't like to report negative changes. So, I can think of several mechanisms of how the ecologic data in the region could be rigged.



Conclusion

In the result of my research, I got few interesting insights into the impact of major pipelines on local areas.

First, I got significant Construction Year correlations on non-highly urbanized areas: positive for all, bigger for the more distant areas and especially more closer areas.

Second, I proved the endogeneity of pipeline route choice.

Third, deserted areas (those that had no economic activity in 2004) didn't get any effect from the pipeline in short and long terms. It raises concerns about the claim of Gazprom officials that their projects bring the civilization into the deserted regions. The results of my research imply no connection of major pipeline construction and appearance of economic activity, not even mentioning the sustainable economy, in the deserted areas.

Fourth, the pipeline doesn't have effect on highly-urbanized areas.

Fifth, major pipeline doesn't bring any ecologic harm to the region, which should be anticipated from any economic activity rise, as I think. This could mean that the economic effect of the pipeline is absolutely ecologically-pure, but the problem lies in the source. RosStat could be formed by people who wouldn't like to report the ecological negative changes.

And last, but not least, is that there is no municipal effect (find the results in the appendix). So, pipeline doesn't help economy through the municipal taxes.

Future Research

There are several opportunities for the further research on this topic.

First, to take a look at different pipelines and research the impact of general motives for construction being export or internal consumption. There could be problems with that, since in my research I found that in Russia all of the major pipelines built in recent decades, were meant to export gas. Internal coverage is happening incidentally. All the internal pipelines were built in USSR times. This data can be collected in the same way that I did for the Sakhalin-Khabarovsk-Vladivostok pipeline

Second, it would be useful to study the impact of gas volumes, going through the pipelines in different time periods. This information should be available at Wood McKenzie.

Third, it's possible to find more precise model, than linear OLS. For example, polynomial or logarithmic, or even discontinuity (since in 10 km-interval the effect could be mainly driven the pipeline itself)

Fourth, it's useful to use different controls to increase the clarity of pipeline effect.

Fifth, I should alter the Night Lights measure so it would better predict the GDP. There are some good economic papers on this theme.

Sixth, I can collect the Night Lights data for the 2014-2018 years to study the long-term effect more precisely.

Seventh, another source of ecologic data for the region should be used to check the truthfulness of the RosStat measures.



Appendix

Effect of pipeline-construction municipality (Municipal effect)

I want to find the effect that area gets if it lies in the municipality where the pipeline was built, in comparison with areas of the same distance from the pipeline interval.

Proximity: 280 (20-30 km away)

Dependent variable: Economic Activity
	20-30 km away municipal-distance areas
Intercept	1.4e+04 (6639.115)
Construction Municipality Dummy	6907 (8761.642)
Pre-Construction (2005)	3915 (1.02e+04)
Pre-Construction (2005) * Construction Municipality Dummy	-5187 (1.28e+04)
Pre-Construction (2006)	-2132 (9139)
Pre-Construction (2006) * Construction Municipality Dummy	370 (1.2e+04)
Pre-Construction (2007)	1.2e+04 (1.33e+04)
Pre-Construction (2007) * Construction Municipality Dummy	-1.04e+04 (1.58e+04)
Pre-Construction (2008)	2965 (1.02e+04)
Pre-Construction (2008) * Construction Municipality Dummy	-804 (1.39e+04)
Pre-Construction (2009)	2982 (1.08e+04)
Pre-Construction (2009) * Construction Municipality Dummy	2685 (1.55e+04)
Pre-Construction (2010)	2e+04 (1.7e+04)
Pre-Construction (2010) * Construction Municipality Dummy	-3931 (2.19e+04)
Pre-Construction (2011)	6702 (1.09e+04)
Pre-Construction (2011) * Construction Municipality Dummy	2910 (1.62e+04)
Pre-Construction (2012)	242 (1.02e+04)
Pre-Construction (2012) * Construction Municipality Dummy	9434 (1.65e+04)
Pre-Construction (2013)	7258 (1.13e+04)
Pre-Construction (2013) * Construction Municipality Dummy	6799 (1.82e+04)
# observations	300
R2	0.046
Heteroskedasticity-robust standard errors (HC3) are reported in parentheses. Regressions include all year fixed effects. All columns are estimated by OLS



As we see from the regressions above, the fact that pipeline goes through the municipality, doesn't bring additional economic rise to the area, taking its' distance from the pipeline fixed.

Additional economic dataset that I used is the DBIMF (БДПМО) - open database, collected by the official state (RosStat), which contains different social and economic data on municipal level. The data starts from 2009 or later. The analysis of has shown some suspicious data, so I found out that the source is not a good fit.

Comparison of Night Lights and DBIMF

Correlation of Population Density Change 2009 to 2013 and Light Density Change 2009 to 2013 is -0.16. It means that Population Density is a bad estimator to study Economic Activity change.

Correlation of Workforce Density Change 2009 to 2013 and Light Density Change 2009 to 2013 is 0.46.



Correlation of Wage Fund Density Change 2009 to 2013 and Light Density Change 2009 to 2013 is 0.63.



Bibliography

1. Cust, J. and S. Poelhekke (2015). The Local Economic Impacts of Natural Resource Extraction. Annual Review of Resource Economics, 7 (1), 251-268.

2. Carrington, W. J. (1996). The Alaskan Labor Market during the Pipeline Era. Journal of Political Economy, 104, 186-218

3. Lowe, M. (2014). Night Lights and ArcGIS: A Brief Guide. Avaliable online: http://economics. mit. edu/files/8945.

4. Henderson, J.V., Storeygard, A. and Weil, D.N. (2012). Measuring economic growth from outer space. American economic review, 102(2), 994-1028.

5. Bityukova, V.R., Safronov, S.G. (2015). Assessment of the Ecological Situation in Russia Using the Method of Potential Surfaces of Human Impact. Regional Research of Russia, 5(4), 367-377.

Размещено на Allbest.ru

...