The Effects of Wildfires on Economic Development in Russia
In Russia wildfires are one of the most common natural disasters and they pose a significant threat to the well-being of many regions of the country. In this paper examines how the 2010 forest fires in Russia affected migration and housing prices.
| Рубрика | Экономика и экономическая теория |
| Вид | статья |
| Язык | английский |
| Дата добавления | 11.10.2020 |
| Размер файла | 3,3 M |
Отправить свою хорошую работу в базу знаний просто. Используйте форму, расположенную ниже
Студенты, аспиранты, молодые ученые, использующие базу знаний в своей учебе и работе, будут вам очень благодарны.
Размещено на http://www.allbest.ru/
УЧРЕЖДЕНИЕ ВЫСШЕГО ПРОФЕССИОНАЛЬНОГО ОБРАЗОВАНИЯ
"НАЦИОНАЛЬНЫЙ ИССЛЕДОВАТЕЛЬСКИЙ УНИВЕРСИТЕТ
"ВЫСШАЯ ШКОЛА ЭКОНОМИКИ"
НЕГОСУДАРСТВЕННОЕ ОБРАЗОВАТЕЛЬНОЕ УЧРЕЖДЕНИЕ
ВЫСШЕГО ОБРАЗОВАНИЯ
"РОССИЙСКАЯ ЭКОНОМИЧЕСКАЯ ШКОЛА" (институт)
ВЫПУСКНАЯ КВАЛИФИКАЦИОННАЯ РАБОТА
The Effects of Wildfires on Economic Development in Russia
Программа Бакалавр экономики
Совместная программа по экономике НИУ ВШЭ и РЭШ
M.V. Filippova
Научный руководитель:
M. Valsecchi
Аннотация
На протяжении всей жизни человечества стихийные бедствия регулярно обрушивались на города и деревни по всему миру, вызывая многочисленные жертвы и нанося огромный ущерб имуществу и инфраструктуре. В России пожары являются одним из наиболее распространенных стихийных бедствий и представляют значительную угрозу благополучию многих регионов страны. В этой исследовательской работе я оцениваю, как пожары 2010 года в России, влияют на миграцию и цены на дома и квартиры. Я нахожу увеличение миграции в город на 1.052 человека на каждые 1000 человек населения и рост цен на дома и квартиры на 1% на каждый дополнительно пострадавший населенный пункт в радиусе до 50 километров от города. Я также получаю уменьшение миграции на 0.6 человек на каждые 1000 человек населения и падение цен на дома и квартиры на 0.5-1.5% на каждый дополнительно пострадавший населенный пункт в радиусе от 50 до 150 километров от города.
forest fires disasters migration housing prices
Abstract
Throughout the life of humanity, natural disasters regularly hit cities and villages all around the world, causing numerous victims and great damage to property and infrastructure. In Russia fires are one of the most common natural disasters and they pose a significant threat to the well-being of many regions of the country. In this research paper I estimate how wildfires of 2010 in Russia affect migration and house prices. I get the result of the increase in in-migration per 1000 people on 1.052 people and 1% increase in price for every additional burnt settlement in the nearest 50 kilometres. I also get the decline on 0.6 people in-migration and 0.5-1.5% drop in prices for every additional burnt settlement at the distance of 50-150 kilometres.
Contents
1 Introduction
2 Related Literature
3 Data
4 Empirical Model
4.1 Burnt settlements
4.2 Burnt wildfire area
5 Main results
5.1 Burnt settlements
5.2 Burnt wildfire area
6 Conclusion
Appendix
1. Introduction
Throughout the life of humanity, natural disasters regularly hit cities and villages all around the world, causing numerous victims and great damage to property and infrastructure. The importance of natural disasters on social and economic development is under big interest. It is well known that natural catastrophes affect death rates, health of people, house prices and products prices, migration rates and a lot of other demographic and socio-economic parameters. However, it is not always clear, how large and robust is the effect of a catastrophe on a particular problem. Economic and social policy programs depend on the scale of the consequences of a catastrophe so it is important to study empirical effects of catastrophes for further policy programs.
The effects of natural disasters are properly studied in other countries. Among the papers, that are dedicated to the analysis of the effect of natural disasters on economic outcomes, there are several research papers which examine the consequences of natural disasters on migration and then look at the link to house prices. The idea is the following: natural disasters provide negative shock and under the catastrophes people move out of the cities decreasing demand on houses. For example, Boustan et al. (2019) found that natural disasters significantly affect housing prices and net-migration between the US counties. There were also found significant effects of particular catastrophes: floods on home prices (Atreya et al. (2013)), fires on home prices (McCoy and Walsh (2014)). Several works even study individual catastrophes of a large scale such as hurricane Katrina (Bleemer and Klaauw (2007)).
In this study I analyze how wildfires in 2010 in Russia affect migration and house prices. In Russia fires and floods are the most common natural disasters - Em-Dat in particular reported that among all natural catasrophes in Russia from 1992 to 2019 there are about 40% of floods and 15% of wildfires. As for my research I decided to explore wildfires of 2010 because it is one of the largest fires in the history of Russia. Wildfires started in July and lasted until September. One of the reasons of the occurence of widespread wildfires is the extremely high temperature. It burned down around 30.500 squared kilometres of forest areas and suffered from fire around 178 settlements: cities, towns and villages in more than 21 regions of Russia.Own calculations The most suffered regions by the number of burnt settlements are Nizhegorodskaya, Ryazanskaya, Voronezhskaya and Lipetskaya oblast. See Appendix To prevent such harm in future and to create effective policy programs we need to estimate the economic effect of wildfires, especially of wildfire of 2010 in Russia.
Like Boustan et al. (2019) I suppose that natural disasters - in my case it is wildfires of 2010, provide negative shocks and people prefer not to move in the suffered cities, reducing net in-migration and providing smaller demand on houses. According to the law of supply and demand, the reduced demand on houses reduces house prices. However, I go further and suppose that the effect is different depending on the distance from a disaster to the city: if the disaster occures in the settlement in the suburban area near the city, then people will move from suffered place to the city and increase migration and house prices. And if the disaster happends in a small town or in a big village on removal from the city, than the attractiveness of this city falls for people from other faraway regions, they don't move in and the in-migration and the house prices fall. Finally, I assume that if the disaster was in forest area just near the city, then it also should decrease the attractiveness of the city, in-migration and house prices. So, the novelty of my research is the deepening of the idea by estimating the effect for specific distances. I contibute to the existing literature a lot by providing the research for Russia.
To check my hypothesis I need datailed data on wildfires of 2010. I build a novel dataset using satellite images and creating a map of wildfires in Russia in 2010. I indicate 178 settlements, which suffered from wildfire, and allocate areas (polygons) of burnt forest. I conduct a city level analysis, taking as obsevations 167 cities with population of more than 100.000 people. Then I take four distances intervals from each city: from 0 to 50 kilometres, from 50 to 100 kilometres, from 100 to 150 kilometres and from 150 to 200 kilometres. For each interval I count how many burnt settlements is in each distance from each city and how large is the burnt forest area in each distance interval from each city. I conduct my econometric analysis in two steps: on the first step I look at the effect of the number of burnt settlements on city outcomes and on the second step I look how burnt forrest area affected city outcomes. I observe two different processes: the effects of suffered populated area and the effect of suffered non-populated area. As for my dependent variables I take net in-migration from Rosstat and house price from real estate agency.
As an econometric approach I use difference-in-difference specification to detect how the migration and house prices changed after wildfires of 2010 for cities that were more affected and cities that were less affected.
Finally, I get the result of the increase in in-migration per 1000 people on 1.052 people and 1% increase in price for every additional burnt settlement in the nearest 50 kilometres. I also get the decline on 0.6 people in-migration and 0.5-1.5% drop in prices for every additional burnt settlement at the distance of 50-150 kilometres.
This research paper has the folowing structure. Section 2 provides the review of the related literature. Section 3 describes the construction of the novel dataset. Section 4 presents the empirical model. Section 5 shows the results. Section 6 concludes.
2. Related Literature
In the recent years, more and more researchers examine the effect of various natural disasters on different measures of economic development such as GDP per capita, migration, housing price index and other variables, that may arouse interest of policymakers.
One very important research in the regard of policymaking is Richardson et al. (2012), who estimate true economic cost of large wildfire in Los Angeles county in 2009 on health. They find that the usual method of counting displays costs equal to $9.5 per exposed person per day and this price underestimates the true effect in $84.42 per person per day calculated using willingness to pay method. This shows us the importance of making research and doing right predictions in the field of consequences of economics disasters.
Among foreign literature which study the question of natural disasters several papers are interested in the same research question as me, exploring the relations between natural disasters and migration and house prices.
One type of papers are those, who try to estimate only prices depending on disasters. Atreya et al. (2013) explore the increase in property price affected by a large flood event at the Flint River in 1994 in Dougherty County in US and try to examine whether the changes in price is temporary or permanent. The authors found the 9% significant discount for properties that is equal $7560 and is associated with the idea that flood risk discount should be larger than the value of the insurance premium. Finaly, they got that the flood risk discount faded away after the flood in 4 - 9 years, which means that the significant effect on prices exists and it is not permanent.
Another paper, which also provides the estimation of price and of duration of price recovery is Mueller et al. (2009). This paper is very interesting because it studies the cumulative effect of repeated forest fires on economic outcomes. Authors find that the effect from the first fire reduces price of nearby house on 10% and from the second fire - on additional 23%. As for the duration of price recovery, authors found that it would take from 5 to 7 years for price to recover after two fires, it is the time of renewing old burned plants, flora to new ones. It is somewhat numerically similar with the previous article of 4 - 9 years effect from one flood. Mueller et al. (2009) also provide the possible explanation, why the effect from second fire on price reduction is bigger: owners and possible buyers perceive the first fire as an occasional and will unlikely move to other place, but after the second fire they perceive the risk of fire as much higher risk and want to move away.
McCoy and Walsh (2014) construct a theoretical model and estimate empirically the effect of severe wildfires on housing pricing and transaction dynamics. They use geospatial data and GIS to find proximity from houses to the wildfires, find the presence of view on the fire burn scars from each house. They find that home prices in the 2km rings fall by 8.3%, 7.8% and 6.8% in the first, second and third years after the wildfire. Houses in high risk zones become cheaper in the first year after the fire on the 6-9%. This paper brings to the idea of multiple ring buffers or distance intervals, where I can find out the decreasing effect of faraway buffers.
One very important research is Champ et al. (2008), who explore the relationship between wildfire risk and home prices. They find that house sales prices increases, if the house is located near the dangerous locations topography while flammable building materials of the house decrease the price. They interpret this result in such way - people want to live in the nice place like near the forest, but these nice places contain more risk (forest may burn). This result is quite unusual and shows that with some probability the effect on prices can be quite unpredictable, which is important to keep in mind when conducting a research.
Other type of papers explore the link between natural disasters and migration. Bleemer and Klaauw (2007) study the consequences of hurricane Katrina, and among other effects they find that out-migration from New Orlean increased on 7% in 10 years after the hurricane. Authors also find the decrease in homeownership rates. This result shows the evidence for a long-term effect on migration from natural disasters.
Boustan et al. (2019) explore how natural disasters affect net out-migration, housing prices and poverty rates in US counties from 1930 to 2010. They find that presence of disaster with more than 25 deaths has positive and significant effect on out-migration and negative and significant decrease on 5.2% on housing prices. It is important that among all natural disasters authors find the significant effect only of fires on migration rates. This even more motivates me to explore my research question and to verify my results with these one's. Also as I said earlier one of my three hypothesis is similar to what Boustan et al. (2019) make, and I expect to get a comparable with these authors result.
There is a lack of empirical findings in Russia that corresponds with my research question. As for the actual literature most russian papers don't contain econometric analysis, however they describe the particular situation and provide us with good descriptive statistics. They also raise the actuality of my research question. For example, Zinovieva (2012) claims that high annual temperature differences and fires are the most widespread natural catastrophes in the country because budget financing of forestry was reduced by government in 1990s and weak forestry negatively affects the work of fire services. Ryazanstev (2011) also claims that it all turned to a bigger harm on the Russian economics. Especially it emerged in 2010, when 27 million Russians lived farther than fire protection standards, it means, that 19% of Russian population couldn't get help from firefighters because firefighters will not have enough time to get there in the case of fire. So, 21 regions in Russian federation with total population of 41 million people seriously suffered from fires in the summer in 2010: number of deaths from different illnesses increased up to 15% in comparison to the number of deaths in 2009. We see that it is crucial to estimate the real effect of wildfires of 2010.
So the novelty of my research is in estimating the effects of wildfires econometrically, to find if they are statistically significant for Russia and based on this make conclusions for policies aimed at reducing the consequences of natural catastrophes. That's why my future research is actual and relevant in the context of particaular Russian case and in general context of contribution to world literature on the topic of natural disasters.
3. Data
Most of research papers take data for their analysis from EM-DAT dataset - it aggregates all types of disasters from 1900 to 2020 all over the world. Unfortunately, there is no enough data that cover the information about wildfire in 2010 in Russia. That's why, to analyze my research question I combine data from different sources and create a novel dataset on wildfires in Russia in 2010.
Firstly, I collect geospatial data on wildfires using dataset collected by Russian engineering and technology center Scanex. As the data is incomplete, I add images from space satellites: LandsatNASA program with sevaral satellites providing the information on Earth's surface, Terra and Aqua Two NASA satellites using MODIS technology, regularly monitoring land situation, http://www.scanex.ru/data/satellites/terra-aqua-modis/. In particular, I take two types of objects: the point layer, consisting from settlements which completely or partially suffered from fire, and the poligons with the forest territories under the fire during the summer of 2010.
I manage data in QGIS - open source software for working with geospatial data. I take 167 Russian cities with population more than 100.000 people, for which I conduct my econometric analysis. For each of these cities I construct the distance matrix to burned settlements. Because of my final goal is to understand how the closeness from the burnt settlement affects the city, I split the distance on several intervals: from 0 to 50 kilometres, from 50 to 100 kilometres, from 100 to 150 kilometres, from 150 to 200 kilometres and calculate the number of burnt settlements in each interval for each city. This is some kind of intensity of the fire - I suppose, that the more is the number of burnt settlements and the closer are these settlements to the city, the bigger is the effect on economic outcomes of the city.
Moreover, I construct buffer zones around each city on several intervals: from 0 to 50 kilometres, from 50 to 100 kilometres, from 100 to 150 kilometres, from 150 to 200 kilometres and calculate the area of burnt forests in each buffer zone. I again assume, that the bigger and the closer is the burnt forest area to the city, the bigger is the impact of fire on this city. On the Figure 2 we can see an example of buffer zones with the interval from 0 to 50 kilometres around each city with population more than 100.000 people. Descriptive statistics on fires is in the Appendix.
My contribution to the compilation of fire map with data is that there was no single
Note: blue dots represent cities with population more than 100.000 people; red dots represent burnt settlements; green polygons represent burnt forests comprehensive map of the fires of 2010. Previously the data consisted of separate parts for specific fires for a specific day, and to find out the total number of burnt settlements and the total area of burnt forests, I had to collect data from several sources: I collected data from satellites for all period of interest and eventually got a map that shows the number of fires and burning areas during the summer of 2010. The contribution of my work is that I got this single map with all the information on the fires of 2010, which I then used for my analysis.
Secondly, for my dependent variable I take annual net migration for cities with population more than 100.000 people from Rosstat for 2005-2015 years. Because the cities have different size of population, I divide migration on the population of base 2004 - the year preceding the beginning of my analysis. Finally, I have the variable - net migration per 1000 population.
Thirdly, for another dependent variable I take house price index from Rosrealt website - the aggregator of advertisements for rent and sale real estate, that provides the data on average apartment sale prices per square meter. I take average monthly prices from April 2009 to December 2012 in all cities with population with more than 100.000 people. To conduct my analysis, I wanted to take data two years before and two years after the occasion of wildfires of 2010, but I don't have data on house prices earlier than April 2009, so I restrict my analysis by the avaliable data. The house prices are corrected on the inflation to the base month April 2009.
Note: blue circles are the area of radius 50km from cities with population more than 100.000 people.
I also made similar circles with other radius for data collection.
To sum up, I have 167 units of observations - cities with population more than 100.000 people. For these cities I have the data on migration, which is collected annually, and the data on house price, which is collected monthly. I conduct city-level analysis.
4. Empirical Model
The goal of my analysis is to estimate the effect of wildfires of 2010 on economic development of Russian cities. I look at such economic outcomes as the net migration in the city per 1000 people and the average house price per sqare meter. I choose these variables because I want to analyze the following idea: the negative shock of wildfire should encourage city residents to move from this are to more favorable and safe places increasing out-migration. The idea is that real estate prices should fall because out-migration has increased, i.e. the number of people who left this city, and in-migration has fallen, i.e. the number of people who want to come to this city. This all should lead to a drop in demand for home purchases, which should increase the price according to the law of demand and supply.
My identification strategy is to compare the migration and house price index in thecities affected by the wildfire in 2010 and not affected before and after my treatment period - the year of 2010. So, I get cities, affected by the wildfire, as the treatment group and cities, not affected by the wildfire, as the control group. To conduct my analysis I use difference-in-difference approach.
As I have two independent variables: the number of burnt settlements and the area of wildfires - I split my analysis into two parts. Firstly, I explore the effect of burnt settlements with different distance from the city on the net migration and on the house price index. Then, I repeat the same actions with another independent variable - the area of wildfire.
4.1 Burnt settlements
To estimate the effect of burnt settlements on the dependent variables I use the following specification:
where Yi,t is in the set of my two dependent variables: the net migration per 1000 people for city i in period t and the logarithm of average house price per square meter for for city i in period t. NumFireSji is the number of burnt settlements for city i for four intervals denoted by j: from 0 to 50 kilometres, from 50 to 100 kilometres, from 100 to 150 kilometres, from 150 to 200 kilometres. As I have different periodicity of data: migration is collected annually and house price is collected monthly, the variable After 2010t is calculated differently. After 2010t is a dummy equal to 1 for the years starting from 2010 and later if Yi,t is migration. After 2010t is a dummy equal to 1 for the months starting from July 2010 and later if Yi,t is house price. ai is city fixed effects, yt - time fixed effects. As I have 167 cities (much more that necessarily 50 units for clustering) I cluster errors by city.
According to my hypothesis small distances to the burnt settlements are associated with high in-migration to the city. This implies that people, who live in suburban zone, are inclined to move to the nearest big city if their settlement burned down. More precisely, the nearest 50 kilometres from the town is the rural area of this city, and people, who live there, are country people. In the case of disaster such as fire in the territory of their villages, people will be forced to move to the nearest urban center, which is the city - my unit of analysis.
Since I assume that migration and house prices are related, I expect to see that increased migration has led to increased demand for real estate, which in turn has increased house prices. That is, these same country people in 50 kilometres from the city, for whom the nearest major point of salvation is this city, will provide additional demand on housing. Therefore, I expect to find a positive effect on prices affected by the number of burnt settlements in the area of 50 km.
For the burnt settlements located in 50 - 150 km from the city I expect to find the opposite results. Settlements located at this distance are not small villages, but large district centers or small towns. The fire in these settlements should reduce the attractiveness of migration from other cities and regions to this city, because an unfavorable risk zone is created around this city.
These burnt settlements located in 50 - 150 km from the city should reduce the attractiveness of this city and decrease in-migration to the city. Then, according to law of demand and supply the house prices should decrease because of decreased demand.
For burnt settlements at a distance of more than 150 km I do not expect to see any result, because they should not have any impact. Burnt settlements located at distance more than 150 km will most probably affect another city, which is nearer to them.
4.2 Burnt wildfire area
My next step is the estimation of the effect of burnt wildfire area on dependent variables, so I use the following specification:
The same as for previous evaluation Yi,t is in the set of my two dependent variables: the net migration per 1000 people for city i in period t and the logarithm of average house price per square meter for for city i in period t. AreaFireSj,i is the area of burnt forest in corresponding interval, one of four intervals denoted by j: from 0 to 50 kilometres, from 50 to 100 kilometres, from 100 to 150 kilometres, from 150 to 200 kilometres. I again have different periodicity of data: migration is collected annually and house price is collected monthly, the variable After 2010t is calculated differently. After 2010t is a dummy equal to 1 for the years starting from 2010 and later if Yi,t is migration. After 2010t is a dummy equal to 1 for the months starting from July 2010 and later if Yi,t is house price. ai is city fixed effects, yt - time fixed effects. As I have 167 cities (much more that necessarily 50 units for clustering) I cluster errors by city.
My last hypothesis is that burnt forest area just near the city (at 0-50 kilometres distance or maybe at 0-100 kilometres distance) should decrease the attractiveness of the city, the net in-migration should also fall, and the demand and house prices fall. Burnt forest areas, which are located far from city, shouldn't affect city outcomes.
5. Main results
5.1 Burnt settlements
The results of my first specification are represented in Table 1. Columns (1)-(4) report the coefficients for the specification, including the dependent variable migration, columns (5)- (8) report the coefficients for the specification, including the dependent variable logarithm of house prices. The presence of controls and fixed effects is represented at the bottom of the table. All interpreted coefficients I take from columns (4) and (8), because they have all types of fixed effects.
Table 1: Estimation of how the number of burnt settlements affected migration per 1000 people and house price per sq. m.
Dependent variable:
|
Migration per 1000 people |
Log(House price per sq. m.) |
||||||||
|
(1) |
(2) |
(3) |
(4) |
(5) |
(6) |
(7) |
(8) |
||
|
Number of Fires in 0 - 50 km * After 2010 |
1.052* (0.564) |
1.052* (0.537) |
1.052** (0.417) |
1.052*** (0.374) |
0.010 (0.007) |
0.009 (0.007) |
0.010*** (0.003) |
0.010*** (0.002) |
|
|
Number of Fires in 50 - 100 km * After 2010 |
-0.602 (0.401) |
-0.602 (0.382) |
-0.602** (0.297) |
-0.602** (0.266) |
-0.016*** (0.006) |
-0.015*** (0.006) |
-0.016*** (0.002) |
-0.015*** (0.002) |
|
|
Number of Fires in 100 - 150 km * After 2010 |
-0.222 (0.266) |
-0.222 (0.253) |
-0.222 (0.197) |
-0.222 (0.177) |
-0.006* (0.003) |
-0.007** (0.003) |
-0.005*** (0.001) |
-0.005*** (0.001) |
|
|
Number of Fires in 150 - 200 km * After 2010 |
0.355 (0.216) |
0.355* (0.205) |
0.355** (0.159) |
0.355** (0.143) |
0.010*** (0.003) |
0.009*** (0.003) |
0.006*** (0.001) |
0.006*** (0.001) |
|
|
After 2010 |
-0.098 (0.968) |
-0.098 (0.716) |
-0.014 (0.011) |
0.017*** (0.004) |
|||||
|
Number of Fires in 0 - 50 km |
0.873** (0.417) |
0.873** (0.397) |
0.015** (0.006) |
0.016*** (0.006) |
|||||
|
Number of Fires in 50 - 100 km |
-0.134 (0.297) |
-0.134 (0.282) |
-0.038*** (0.005) |
-0.039*** (0.005) |
|||||
|
Number of Fires in 100 - 150 km |
-0.243 (0.197) |
-0.243 (0.187) |
-0.002 (0.003) |
-0.001 (0.003) |
|||||
|
Number of Fires in 150 - 200 km |
0.129 (0.162) |
0.129 (0.154) |
0.024*** (0.003) |
0.024*** (0.002) |
|||||
|
Geographic controls City FE Time FE Observations |
YES NO NO 1,716 |
YES NO YES 1,716 |
NO YES NO 1,716 |
NO YES YES 1,716 |
YES NO NO 5,138 |
YES NO YES 5,138 |
NO YES NO 5,138 |
NO YES YES 5,138 |
Note: *p<0.1; **p<0.05; ***p<0.01. Robust standart errors in parentheses. Errors are clustered at the city level
We see that the number of burnt settlements in treatment period (starting from 2010) positively and significantly affects migration for distance under 50 kilometres for all specifications. The coefficient is 1.052, which means that one additional burnt settlement in 50 kilometres from the city after the treatment inceases net in-migration per 1000 people on 1.052 people for periods after the treatment.
The number of burnt settlements in 50 kilometres from the city significantly affects the price only after adding fixed effects. Nevertheless the coefficient is positive and significant in (7) and (8) and equal to 0.01, which means that one additional burnt settlement after the treatment in 50 kilometres from the city increases house price per sq. m. on 1% after the treatment. Both of these results correspond with my hypothesis that burnt settlements near the city increase the migration and house prices in the city.
The coefficient for the number of burnt settlements in treatment period for distance from 50 to 100 kilometres significantly and negatively affects migration for (3) and (4), where added all fixed effects. The coefficient is -0.602, which means that one additional burnt settlement in 50 kilometres from the city after the treatment decreases net inmigration per 1000 people on 0.602 people for periods after the treatment. The coefficient for the number of burnt settlements in treatment period for distance from 100 to 150 kilometres doesn't significantly affect migration.
The coefficients for the number of burnt settlements in distances 50-100 kilometres and 100-150 kilometres from the city significantly and negatively affect the price just as I expected. The coefficient for the interval 50-100 kilometres is -0.015 and for the interval 100-150 kilometres, which means that one additional burnt settlement after the treatment in 50-150 kilometres from the city decreases house price per sq. m. on 0.5 - 1.5 % after the treatment.
Both of these results correspond with my hypothesis that burnt settlements located in 50 - 150 km from the city should reduce the attractiveness of this city, decreasing in-migration and house prices for the city.
Finally, for the distance interval of 150-200 kilometres I unexpectedly get positive and significant coefficient: each additional burnt settlements in the area of 150-200 kilometres increases migration per 1000 people on 0.355 people and increase house prices on 0.6%. Such result may be explained in the following way: the settlement at a distance of more than 150 km from this city A is possibly located near another city B with population more than 100.000 people. This burnt settlement negatively affects that city B, which induces people to move from city B to city A.
5.2 Burnt wildfire area
I report the results of my second specification in Table 2. Columns (1)-(4) report the coefficients for the specification with the dependent variable migration, columns (5)-(8) report the coefficients for the specification with the dependent variable logarithm of house prices. The presence of controls and fixed effects is represented at the bottom of the table. All interpreted coefficients I take from columns (4) and (8), because they have all types of fixed effects.
Table 2: Estimation of how the area of burnt forests affected migration per 1000 people and house price per sq. m.
|
Dependent variable: |
|||||||||
|
Migration per 1000 people |
Log(House price per sq. m.) |
||||||||
|
(1) |
(2) |
(3) |
(4) |
(5) |
(6) |
(7) |
(8) |
||
|
Log(Area 0-50 km |
-0.659 |
-0.659 |
-0.659* |
-0.659** |
-0.027*** |
-0.023*** |
1 О о ю |
1 о о 00 |
|
|
+ 0.1) * After 2010 |
(0.481) |
(0.458) |
(0.354) |
(0.318) |
(0.007) |
(0.006) |
(0.002) |
(0.002) |
|
|
Log(Area 50 - 100 km |
-0.228 |
-0.228 |
-0.228 |
-0.228 |
t>- O Ц |
0.015** |
о о со |
о о ю |
|
|
+ 0.1) * After 2010 |
(0.500) |
(0.476) |
(0.368) |
(0.330) |
(0.007) |
(0.007) |
(0.003) |
(0.002) |
|
|
Log(Area 100 - 150 km |
1.061* |
1.061* |
1.061** |
1.061*** |
0.008 |
0.010 |
-0.003 |
-0.002 |
|
|
+ 0.1) * After 2010 |
(0.578) |
(0.551) |
(0.425) |
(0.382) |
(0.007) |
(0.007) |
(0.002) |
(0.002) |
|
|
Log(Area 150 - 200 km |
-0.285 |
-0.285 |
-0.285 |
-0.285 |
-0.008 |
-0.011* |
-0.001 |
-0.003 |
|
|
+ 0.1) * After 2010 |
(0.502) |
(0.478) |
(0.369) |
(0.331) |
(0.006) |
(0.006) |
(0.002) |
(0.002) |
|
|
After 2010 |
-0.589 |
-0.589 |
-0.030** |
0.001 |
|||||
|
(1.069) |
(0.786) |
(0.013) |
(0.005) |
||||||
|
Log(Area 0 - 50 km |
-0.155 |
-0.155 |
-0.011** |
1 О о 4^ |
|||||
|
+ 0.1) |
(0.357) |
(0.340) |
(0.006) |
(0.005) |
|||||
|
Log(Area 50 - 100 km |
-0.008 |
-0.008 |
-0.008 |
-0.006 |
|||||
|
+ 0.1) |
(0.370) |
(0.353) |
(0.006) |
(0.006) |
|||||
|
Log(Area 100 - 150 km |
-0.014 |
-0.014 |
0.008 |
0.006 |
|||||
|
+ 0.1) |
(0.427) |
(0.407) |
(0.006) |
(0.006) |
|||||
|
Log(Area 150 - 200 km |
0.293 |
0.293 |
0.007 |
0.009* |
|||||
|
+ 0.1) |
(0.373) |
(0.355) |
(0.005) |
(0.005) |
|||||
|
Geographic controls |
YES |
YES |
NO |
NO |
YES |
YES |
NO |
NO |
|
|
City FE |
NO |
NO |
YES |
YES |
NO |
NO |
YES |
YES |
|
|
Time FE |
NO |
YES |
NO |
YES |
NO |
YES |
NO |
YES |
|
|
Observations |
1,716 |
1,716 |
1,716 |
1,716 |
5,138 |
5,138 |
5,138 |
5,138 |
Note: *p<0.1; **p<0.05; ***p<0.01. Robust standart errors in parentheses. Errors are clustered at the city level
As I predicted I get the negative and significant coefficient for the area of burnt forest at the distance 0-50 kilometres in columns (3) and (4). This means that in-migration in the city negatively depends on the forest fires near city. And respectively I get negative and significant effect on house prices - demand falls, price falls. With the growth of burnt forest area at distance 0-50 kilometres on 1% the in-migration per 1000 people falls on 0.0065 people. With the growth of burnt forest area at distance 0-50 kilometres on 1% the house price per sq. m. decreases on 0.018%. This corresponds to the intuition that a fire near the city creates smog and smoke - bad environmental situation that encourages people to leave the city. Accordingly, the demand for houses falls and house prices decline.
6 Conclusion
This research based on the novel dataset has shown that there is an evidence of relationship between wildfires in 2010, migration and house prices. I got three main effects that confirm my hypothesis. I found the growth of 1.052 people for in-migration per 1000 people and growth on 1% in house prices for every additional burnt settlement in the nearest 50 kilometres, which may be the evidence of reallocation of villagers during the wildfires. The increase in prices may reflect the growth of demand according to the law of supply and demand. I also find the reduction of 0.6 people in-migration for every 1000 people and drop of 0.5 - 1.5 % in house prices for every additional burnt settlement at the distance of 50 - 150 kilometres. This may mean, that fires on a removal from the city reduce the attractivness of the city, and so they decrease the in-migration, demand for houses and house prices.
In addition, I explore how the area of burnt forests affected migation and house prices. I find, that forestfires in the nearest 50 kilometres from the city decrease in-migration per 1000 people on 0.66 people and the house prices drop on 0.018 % with the growth of burning area on every 1%.
My results are comparable with foreign literature that migration and house prices are affected by fires on period of 2-5 years. My results confirm the logic of Boustan et al. (2019) that out-migration, caused by a disaster, decrease demand and house prices.
The novelty of this analysis is the deepening analysis and providing the estimation of the effect in different intervals from the city. Moreover I am the fist, who joined different sources to construct the novel dataset, using GIS data, the data from Rosstat and the data from real estate agency. Also I am the first, who estimated econometrically the effect of wildfires of 2010 in Russia. I find that the effect depend on whether the fire is in the settlements or only unpopulated forest areas.
This research provides interest for understanding the effects of fires on large cities in Russia. This work can be useful for policy makers who are engaged in constructing fire prevention measures as well as the programs, reducing the negative consequences of fires.
References
Atreya, A., S. Ferreira, and W. Kriesel (2013). Forgetting the flood? an analysis of the flood risk discount over time. Land Economics 89 (4).
Bleemer, Z. and W. Klaauw (2007). Disaster (over-)insurance: The long-term financial and socioeconomic consequences of hurricane katrina. Federal Reserve Bank of New York Staff Reports.
Boustan, L., M. Kahn, P. Rhode, and M. Yanguas (2019). The effect of natural disasters on economic activity in us counties: a century of data. National Bureau of Economic Research (Working Paper 23410).
Champ, P., G. Donovan, and C. Barth (2008). Homebuyers and wildfire risk: A colorado springs case study. Society and Natural Resources.
McCoy, S. and R. Walsh (2014). W.u.i. on fire: Risk, salience and housing demand. National Bureau of Economic Research (Working Paper 20644).
Mueller, J. M., J. B. Loomis, and A. Gonzalez-Caban (2009). Do repeated wildfires change homebuyers' demand for homes in high-risk areas? a hedonic analysis of the short- and long-term effects of repeated wildfires on house prices in southern california. Journal of Real Estate Finance and Economics 38 (2).
Richardson, L., P. Champ, and J. Loomis (2012). The hidden cost of wildfires: Economic valuation of health effects of wildfire smoke exposure in southern california. Journal of Forest Economics.
Ryazanstev, S. (2011). Demographic and socio-economic consequences of abnormally high temperatures and forest fires in the regions of european russia in the summer of 2010. Today and Tomorrow of the Russian economy.
Zinovieva, I. (2012). Assessment of damage from forest fires. Bulletin of the Moscow University.
7 Appendix
Table 3: Number of burnt settlements in regions of Russia
|
Region |
Number of burnt settlements |
|
|
Nizhegorodskaya oblast |
49 |
|
|
Ryazanskaya oblast |
24 |
|
|
Voronezhskaya oblast |
21 |
|
|
Lipetskaya oblast |
16 |
|
|
Tambovskaya oblast |
9 |
|
|
Kirovskaya oblast |
8 |
|
|
Moskovskaya oblast |
8 |
|
|
Tulskaya oblast |
8 |
|
|
Republic of Mordovia |
7 |
|
|
Vladimirskaya oblast |
5 |
|
|
Ivanovskaya oblast |
5 |
|
|
Volgogradskaya oblast |
3 |
|
|
Sverdlovskaya oblast |
3 |
|
|
Ulyanovskaya oblast |
3 |
|
|
Altaiskii krai |
2 |
|
|
Saratovskaya oblast |
2 |
|
|
Belgorodskaya oblast |
1 |
|
|
Kalugskaya oblast |
1 |
|
|
Penzenskaya oblast |
1 |
|
|
Samarskaya oblast |
1 |
|
|
Smolenskaya oblast |
1 |
|
Statistic |
N |
Mean |
St. Dev. |
Min |
Max |
|
|
Number of burnt settlements in 0 - 50 km |
167 |
0.611 |
2.145 |
0 |
13 |
|
|
Number of burnt settlements in 50 - 100 km |
167 |
1.240 |
3.463 |
0 |
21 |
|
|
Number of burnt settlements in 100 - 150 km |
167 |
2.737 |
5.758 |
0 |
39 |
|
|
Number of burnt settlements in 150 - 200 km |
167 |
3.820 |
6.695 |
0 |
31 |
|
|
Area of wildfires 0 - 50 km |
167 |
28.931 |
109.261 |
0 |
1085 |
|
|
Area of wildfires 50 - 100 km |
167 |
77.084 |
207.794 |
0 |
1617 |
|
|
Area of wildfires 100 - 150 km |
167 |
167.200 |
399.384 |
0 |
2233 |
|
|
Area of wildfires 150 - 200 km |
167 |
251.912 |
483.888 |
0 |
2863 |
Note: Area of wildfires is calculated in square kilometres
Размещено на Allbest.ru
...Подобные документы
Issues about housing prices formation process. Analytical model of housing prices. Definition a type of relationship between the set of independent variables and housing prices. The graph of real housing prices of all Russian regions during the period.
курсовая работа [1,6 M], добавлен 23.09.2016Stereotypes that influence on economic relations between the European Union countries and Russia. Consequences of influence of stereotypes on economic relations between EU and Russia. Results of first attempts solving problem. General conclusion.
реферат [19,0 K], добавлен 19.11.2007The Human Capital Theory. External Migration in Kazakhstan. The major causes of out-migration in Germany. Migration in Kazakhstan during 2004-2010. Internal Migration in Kazakhstan. The major factors determining the nature of the migration to Russia.
реферат [2,2 M], добавлен 14.04.2012Evolutionary and revolutionary ways of development of mankind. Most appreciable for mankind by stages of development of a civilization. The disclosing of secret of genome of the man. Recession in an economy and in morality in Russia. Decision of problems.
статья [12,1 K], добавлен 12.04.2012The stock market and economic growth: theoretical and analytical questions. Analysis of the mechanism of the financial market on the efficient allocation of resources in the economy and to define the specific role of stock market prices in the process.
дипломная работа [5,3 M], добавлен 07.07.2013The influence of the movement of refugees to the economic development of host countries. A description of the differences between forced and voluntary migration from the point of view of economic, political consequences. Supply in the labor markets.
статья [26,6 K], добавлен 19.09.2017Analysis of the causes of the disintegration of Ukraine and Russia and the Association of Ukraine with the European Union. Reducing trade barriers, reform and the involvement of Ukraine in the international network by attracting foreign investment.
статья [35,7 K], добавлен 19.09.2017The major structural elements of economic safety of a national economy branches. The structural analysis of economic activity. Share of wages in ВВП, of productivity of Russia and western countries. The essence of the economic taxes and their purpose.
статья [166,3 K], добавлен 12.04.2012Establishing a favorable environment for investments, removing administrative barriers. Establishing high-technology parks. Formation of financial mechanisms to attract and support investments, tax stimulation measures. Brand promotion of Russian regions.
реферат [15,9 K], добавлен 04.06.2013Antitrust regulation of monopolies. The formation and methods of antitrust policy in Russia. Several key areas of antitrust policy: stimulating entrepreneurship, the development of competition began, organizational and legal support for antitrust policy.
эссе [39,2 K], добавлен 04.06.2012Negative consequences proceeding in real sector of economy. Social stratification in a society. Estimation of efficiency of economic safety. The parity of the manufacturers of commodity production. Main problems of the size of pension of common people.
статья [15,4 K], добавлен 12.04.2012The air transport system in Russia. Project on the development of regional air traffic. Data collection. Creation of the database. Designing a data warehouse. Mathematical Model description. Data analysis and forecasting. Applying mathematical tools.
реферат [316,2 K], добавлен 20.03.2016Prospects for reformation of economic and legal mechanisms of subsoil use in Ukraine. Application of cyclically oriented forecasting: modern approaches to business management. Preconditions and perspectives of Ukrainian energy market development.
статья [770,0 K], добавлен 26.05.2015The influence of corruption on Ukrainian economy. Negative effects of corruption. The common trends and consequences of increasing corruption. Crimes of organized groups and criminal organizations. Statistical data of crime in some regions of Ukraine.
статья [26,7 K], добавлен 04.01.2014Характеристика направлений иностранных инвестиций в российскую экономику. Наиболее привлекательные для иностранных инвесторов сферы. Российские города в рейтинге Doing Business in Russia. Анализ опыта развивающихся стран и стран с переходной экономикой.
курсовая работа [94,4 K], добавлен 29.01.2014Concept of competitiveness and competition, models. Russia’s endowment. Engendered structural dominance and performance. The state of Russian competitiveness according to the Global Competitiveness Index. Place in the world, main growth in detail.
курсовая работа [1,2 M], добавлен 28.05.2014Assessment of the rate of unemployment in capitalist (the USA, Germany, England, France, Japan) and backward countries (Russia, Turkey, Pakistan, Afghanistan). Influence of corruption, merges of business and bureaucracy on progress of market economy.
реферат [15,5 K], добавлен 12.04.2012Mergers and acquisitions: definitions, history and types of the deals. Previous studies of post-merger performance and announcement returns and Russian M&A market. Analysis of factors driving abnormal announcement returns and the effect of 2014 events.
дипломная работа [7,0 M], добавлен 02.11.2015Natural gas market overview: volume, value, segmentation. Supply and demand Factors of natural gas. Internal rivalry & competitors' overview. Outlook of the EU's energy demand from 2007 to 2030. Drivers of supplier power in the EU natural gas market.
курсовая работа [2,0 M], добавлен 10.11.2013The essence of Natural Monopoly. The necessity of regulation over Natural Monopoly. Methods of state regulation over the Natural Monopolies. Analysis and Uzbek practice of regulation over Monopolies. Natural Monopolies in modern Economy of Uzbekistan.
курсовая работа [307,7 K], добавлен 13.03.2014
