Влияние экономики краткосрочных контрактов на зависимый труд Великобритании
Исследование влияния экономики краткосрочных контрактов на величину наёмного труда в Великобритании с использованием метода инструментальных переменных. Появление новых участников рынка краткосрочных контрактов, их влияние на количество работников.
| Рубрика | Экономика и экономическая теория |
| Вид | дипломная работа |
| Язык | английский |
| Дата добавления | 19.08.2020 |
| Размер файла | 2,3 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/
Федеральное государственное автономное образовательное учреждение высшего образования
"Национальный исследовательский университет
"Высшая школа экономики"
Международный институт экономики и финансов
Выпускная квалификационная работа
по направлению подготовки 38.03.01 "Экономика"
образовательная программа "Программа двух дипломов по экономике НИУ ВШЭ и Лондонского университета"
Влияние экономики краткосрочных контрактов на зависимый труд Великобритании
(Influence of the gig economy on the dйpendent labour of the Great Britain)
Юкляева Дарья Александровна
Научный руководитель Кривошея Егор Артемович, преподаватель
Москва 2020
Аннотация
В настоящей работе оценивается влияние экономики краткосрочных контрактов на величину наёмного труда в Великобритании с использованием метода инструментальных переменных, а именно - двухшагового метода наименьших квадратов. Также, для подтверждения стабильности полученных результатов, был применен метод моментов Ареллано-Бонда, включающего в себя рассмотрение лага зависимой переменной в модели. В ходе работы было установлено, что появление новых участников рынка краткосрочных контрактов оказывает положительное влияние на количество работников в наёмном труде. Результаты исследования являются идентичными при рассмотрении и анализе двух разных моделей, а также различных спецификаций. наемный труд контракт
Данная работа вносит большой вклад в современные исследования, касающиеся эффекта рынка краткосрочных контрактов, который в данный момент развивается в очень стремительном темпе, на традиционную форму занятости. Ранее подобные исследования не проводились для рынка Великобритании.
Abstract
This paper considers the impact of the gig economy on the dependent employment in the United Kingdom with the help of the instrumental variable method, namely, the two-step least squares method. Further, for the justification of the stability of results, the paper considers the Arellano- Bond method of moments, which includes considering the lags of variables in the model and different specifications of the two-stage least squares method. During the research, it was found that the emergence of new participants in the gig economy contributes to an increase in the number of employees in the dependent employment. The results of the study are identical when considering and analyzing two different models.
The current work makes a great contribution to the modern research sphere, connected to the effect of the emergence of the gig economy, which is now developing at a very high pace, on the dependent labor. The researches like this on the effect on the dependent labor of Great Britain has never been carried before.
Introduction
This paper aims to determine the relationship between the gig economy and dependent employment in the United Kingdom, using the data from 2014 to 2017 for the analysis.
That research carries an important role as the topic of the gig economy is new and very developing in the world. More and more people start entering digital platforms to sell their labor, so that could have an influence on the traditional form of employment - dependent labor. For example, there was a 54-fold increase in the number of workers, who use digital platforms to sell their labor in the US since the year 2012. (Farell and Greig, 2012) Moreover, from 2005 to 2015, the share of people who sell their labor using the alternative types of work has raised from 10,7% to 15,8%. (Katz and Krueger, 2016)
However, nowadays there is a little number of researches that considers the effect of the emergence of the gig economy on the dependent labor. It is important to consider that research to understand the economic and social implications of the emergence of the gig economy for the future modification of existing rules and policies in the economy.
The paper by Schwellnus et al. (2019) considers the effect of the gig economy platforms on the overall employment, dependent employment, and wages in the dependent employment, using the information on the US counties. It concludes that when the regression does not account for the controls, such as the business cycle and the initial condition, there is a positive relationship between the growth of the firms in the gig economy and the dependent employment, which is found to be weekly significant.
Nowadays, technology has an increasing impact on every aspect of our lives. This opens the opportunity for dramatic changes in the ways how people work and communicate. In the future the structure of the economy and the nature of employment would be modified by digital innovations, such as the introduction of the big data application, using cloud technologies. (Kenney and Zisman, 2016) Some time ago, the main types of labor were dependent employment, part-time employment, self-employment. However, recently, there has been an emergence of a new type of employment - the emergence of the gig economy.
The first mentioning and recording of the gig economy were seen in the time of the emergence of the Great Financial crisis because people who lost their jobs during the severe crisis times were forced to look for jobs that were part-time and usually short in the duration. (Geraint, 2019)
It is essential to draw the line between the definitions of the freelancer, temporary worker, and a gig worker. Freelancers are independent workers that run their business and have the responsibility for all the tasks that usually perform different departments in the company, such as marketing, financing, communicating affairs. In comparison, gig economy workers do not own their own business, marketing is done by the digital platform in which they offer their services and they usually randomly receive orders. The temporary worker is a type of work that lies between the full-time employee and a freelancer. Temporary workers offer their services usually through temporary staffing agencies, where special people responsible for recruitment match them with their customers.
The practical contribution and the peculiarity of the current study are that it would allow us to analyze the correlation between the gig economy and the dependent labor, taking into account different industries, without concentrating on the influence of the particular platform, or industry. The study by Berger, Chen, and Frey (2017) concentrated on the particular platform, Uber, showing that there is no negative employment effect of the Uber introduction, however, there is a decrease in the earnings of the drivers who receive wages, that is balanced by the increase in the earnings of the drivers, who are self-employed. This study does not concentrate on the sole platform but considers the gig economy as a whole.
The paper is divided into two main parts - the first one considers the theoretical framework of the study, chapter 1, and the second part is devoted to the empirical research of the current paper, chapter 2.
The first chapter describes the definition of the gig economy as people that use digital applications in order to be able to sell their labor. (Taylor, 2017) Then it turns the view of the mechanism by which the development of the gig economy could influence the dependent employment. Overall, there are two main effects of the gig economy on dependent employment: the substitution effect and the market expansion effect.
The substitution effect implies a negative correlation between the gig economy and the dependent employment, as the new participants in the gig economy substitute the existing traditional employment forms. The market expansion effect implies a positive effect on the dependent employment, as the appearance of the gig firms extends the range of services available in the market, decreasing the price of the service, generating higher demand for the service, influencing the demand for the employees. The theoretical framework extends the mechanisms of the correlation between the gig economy and the dependent labor and highlights the peculiarities that exist in the gig economy.
The second chapter is devoted to the empirical investigation of the research topic under consideration. It starts with a description of the data sources and samples used in the research. The research data uses the period from 2014 to 2017 and the information on 216 counties of the United Kingdom. Moreover, this paragraph highlights the reasons for choosing particularly the United Kingdom for the investigation, noticing that the gig economy in the United Kingdom is one of the most developing nowadays. This part of the chapter also gives some more detailed statistics on the counties which serve a role of a leader in the number of employees, for the size of the gig economy, and the percentages of the internet users for the years 2014-2017.
The chapter also describes the main method that is used in the paper to investigate the desired relationship. The main method that is used in the model is the two-stage least squares model. That model is an instrumental variable model that is used to solve the problem of endogeneity.
The problem of endogeneity may arise in the model due to the reason that there could be a factor omitted from the regression, the factor which could influence the dependent variable (dependent labor), and the variable under consideration (the number of firms in the gig economy). The paper by Schwellnus et al. (2019) gives an example of a positive shock of the demand that could influence both the dependent labor and the development of the gig economy. Moreover, as could be seen from the analysis, there is a high correlation between the variable related to the dependent labor and the variable on the data on gig firms. To avoid the bias in the regression results, the following analysis uses the instrumental variable approach and adds to the regression the number of control factors.
Then the chapter discusses the obtained results about the correlation of the gig economy with the dependent labor. It has concluded that the gig economy has a small but positive effect on the dependent employment, concluding that the market expansion effect dominates.
After the results section, there is a part related to the robustness check of the achieved results. The robustness check section consists of using different specifications in the different stages of the two-stage least squares model and using the Arellano Bond model to check the stability of the results. After conducting the robustness check, the paper concluded that the results remain stable and that there is a small, but positive correlation between the gig economy and the dependent employment.
Overall, the research question of the current paper is to determine the relationship between the gig economy and the dependent labor in the United Kingdom.
Chapter 1. Gig economy
1. Definition and types of the gig economy
Rise in the development of internet connections, mobile applications has helped people to find their jobs through digital platforms, rising the gig economy.
The gig economy is defined as people that use digital applications in order to be able to sell their labor. (Taylor, 2017). Gig economy platforms are defined as digital platforms that match workers on the one side with consumers on the other side on the gig basis. (OECD Working papers, 2019) Another definition could be stated as "The "Gig Economy" can be characterized by digital platforms facilitating services between freelancers and customers" (Mastercard Gig Economy Industry Outlook and Needs Assessment, 2019)
The gig economy is mainly divided into two big types: the crowd work and the work on demand. (De Stefano, 2016). The main difference between these two types is that crowd work is the cloud-based work that could be done remotely - at a worker's place (home). Crowd work also could be divided into two main parts, such as microtasks and specific professional expertise to deliver work online. Microtasks are small tasks that use human intelligence to perform tasks that computers could not perform. (Scholz, 2016) Specific professional expertise consists of the work that requires special technological knowledge, such as data entry or software programming, that could be easily done at home. However, work on demand could not be performed at the worker's home, these are "real-world" services, such as taxi services, cleaning services, food delivery services, etc. Examples of these services in the modern world are Uber, TaskRabbit, Deliveroo.
The number of people working in the gig economy is growing from year to year. To the recent research that was conducted by Mastercard, it was founded that the gig economy all over the world generates about 204 billion dollars in gross volume. Mastercard projected that gig economy transactions would acquire a 17% compound annual growth rate and a gross volume would rise to 455 billion dollars to 2023.
2. The economic impact of the gig economy
This paper considers the correlation between the gig economy and the dependent labor, so it is essential to look at the channels of influence of the gig economy on the labor market. There are two main channels of influence: market expansion effect and the substitution effect. This paper would consider the mechanism of both effects in some detail.
Market expansion effect
The paper considers the mechanism of the market expansion effect step by step.
The first step would be to examine the role of the rise of the digital technologies that affect the development of a gig economy. The rise of the internet opened the possibility of changing the initial form of labor. Feldman etc. claimed that informational technology (IT) made possible services transformation with the application of newly developed computable algorithms that apply to the wide variety of activities in the economy, both from the producer and consumer perspectives. When those algorithms were made easy to have access to (by introducing clouds) it became possible to create infrastructure with platform-based markets (Kenney and Zysman (2016)). So, as digital technologies develop, digital platforms appear.
There are several ways in which the platforms influence the mediation of the work process. Firstly, some platforms enable the transformation of the work of previously independent professionals. As an example, the LinkedIn platform helps enterprises by making available information that is openly provided by members to human resources departments. This paper has focuses on other platforms, such as Task Rabbit, which makes accessible the crowdsource of the performance of tasks performed by a human. Digital platforms generate the existence of the gig economy.
The second step would be to consider the main peculiarity of the gig economy platforms is that they help to match the workers with clients. The paper by Cramer and Krueger (2016) in its analysis uses the Uber platform, comparing it to the traditional taxi drivers. It claims that Uber possesses "more efficient driver passenger matching technology", moreover, as Uber operates on a larger scale, the driver-passenger matching is performed even faster. Moreover, the study shows that capacity utilization is higher (for about 50%) for drivers in the gig economy (particularly, Uber) than for divers in the traditional taxi industry.
The next step of the chain is that as the development of matching workers with clients progresses, the variety of services in the market increase as there is a more attractive opportunity to offer services on the market.
As the variety of services increases, the competition in the market also increases, and the prices on the service decrease.
Forder and Allan (2014) have estimated that the competition in the English care/nursing home market decreases prices in this market. Increased competition benefits the consumer. Hausman and Leibtag (2005) estimated the consumer benefits from the entry and expansion of Wal-Mart and other supercenters into geographic markets. They concluded that the benefits were substantial. Moreover, Hausman in his previous researches has found that there is an increase in consumer welfare from the introduction of a new brand or product.
As the prices on service decrease, it becomes more attractive to consumers, making the demand bigger.
As there is a higher demand for the service, more workers are needed to satisfy that demand, so the labor demand increases. Increased labor demand would imply the increase in the dependent labor in the market.
Summing up a bit, the chain in short looks like this: digital innovations allow the rise of the gig economy, the gig economy has a peculiarity of increased matching efficiency, so the attractiveness of the service rises in the market. The variety of services increases, decreasing the price, and increasing demand. Increased demand generates the need for the employees and the labor demand increases, rising the dependent employment.
That chain has considered market expansion effect and it could be seen that this effect implies a positive impact of the gig economy on the dependent labor.
Furthermore, lower prices in one industry that is connected with the gig employment (for example, prices on the cleaning services) would have a reducing influence on the cost in the firms that use these services. The cost reduction may generate a reduction in prices of that related industry's products that influences an increased income of the households, impacting the demand in other industries. As the demand in other industries is positively affected, that would imply a positive influence on the overall employment statistics. (Bessen, 2018)
Moreover, labor productivity in the market could be increased because of the increase in the efficiency of matching consumers with workers, however, the effect on productivity could be under doubts due to reasons related to low barriers to enter the gig economy, allowing low productive people enter the market.
The paper considers the mechanism of the substitution effect step by step.
As was stated before, the development of digital platforms enhances the development of a gig economy. As the gig economy increases, the variety of services on the market is also increased.
As the variety of services increases, gig economy platforms may also generate a negative effect - they may act as a substitute for services offered by the traditional producers.
Frey and Osborne (2013) in their research estimated how susceptible are jobs to computerization using a Gaussian process classifier and estimating how probable would be the computerization for 702 occupations. They stated that because of the nowadays developing informational technologies, a huge part of employment shortly would be at risk (about 47per cent of the employment in the US).
That has effects on the direct industry where the gig economy increases and also on the related industries. As the gig economy substitute for the service offered by traditional producers, less labor is needed in the traditional industry, dependent labor decreases. That means the substitution effect implies that there is a negative effect of the development of a gig economy on dependent employment.
The existence of two effects implies that there are two opposing influences on dependent employment. The expansionary effect has a positive influence on dependent labor, and the substitution effect has a negative influence on dependent labor. The overall effect on dependent labor depends on the size of the market expansion and substitution effects. Market expansion effect could dominate if the worker-client matches are highly sensitive to changes in demand for services.
3. Another peculiarity of the gig work
Another important peculiarity of the gig economy that should be mentioned is the difference between the conditions of entering the gig economy and entering the traditional labor and selfemployment.
Generally, it is considered that there are several barriers for the worker entering the traditional and self-employment types of labor. The traditional labor requires the worker to have a professional degree in the sphere that he or she enters, which means a worker should have a diploma that would confirm his or her knowledge. Entering into the self-employment industry also has barriers - that is the cost of creating the business, the development of the client network, licensing requirements, etc.
Comparing the gig economy platforms to these two forms of labor, it could be noted that entering the gig economy does not require a formal qualification and does not imply incurring business costs, instead, gig workers attract customers by the reputation rating mechanisms. In these reputation rating mechanisms workers are given points for their work from customers and future customers before choosing which worker to employ, could base their decision on the publicly available information from the rating systems.
4. Factors that drive the growth of the gig economy Demand-side
To have a broader understanding of the topic, it is important to look at the factors that drive the growth of the gig economy. For convenience, the demand side and supply side of the gig services should be considered separately.
From the demand side, at first, it should be noted that nowadays there is a lot of venture capital invested in the gig services that allow attracting customers quickly to the newly developed services.
Secondly, with the rapid emergence of the gig economy platforms people became used to the service that could be offered to them immediately, so the growth of the variety of gig services is expected because that would allow the sphere to stay competitive. With the growing technological impact, people expect that any information or service they need could be reached in seconds.
Consumers are found to expect instant satisfaction of their needs with the increasing information that is available to them (Deloitte, Consumer Report 2020). Moreover, two important trends in consumer behavior in 2020 due to the research are "Catch Me in Seconds" and "Multifunctional Homes" (Euromonitor International, 2020). "Catch Me in Seconds" trend means that consumers nowadays expect that the service they need would be available to them as faster as possible, which is connected to the factor of an enormous amount of information that is reaching modern consumers. "Multifunctional Homes" trend is connected to the idea that consumers nowadays expect that all the services they need could be available from the places, so they do not have to leave the place where they live. These modern trends have their influence on the development of the gig economy.
The third factor from the demand side that could be highlighted is the demand from business units. Small and medium-sized businesses to reduce their overhead costs may decide to use the work of the gig worker instead of hiring full-time employees, generating the demand for the gig workers. In the report that is called 'The Future of Work is Anywhere - Gig Workforce' by Noble House stated that about 45 percent of Human Resource Heads that took part in the survey would substitute skills of the existing workers with gig workers, 39 percent of which would undertake this to reduce costs. (The Economic Times, 2019)
Supply-side
From the supply side, several important factors generate growth in the gig economy.
Firstly, nowadays we live in a world where the digitization rates are growing, and internet access appears in an increased number of countries and regions throughout the world. That enables more people to enjoy the opportunity of gig working, the percentage of internet users worldwide has increased from 30% in 2010 to 53,6% in 2019. (Telecommunication Development Bureau, 2020)
Secondly, nowadays a cultural shift towards more flexible work desire has happened. Some people want to spend more time with their families and devote more effort to their hobbies, so they prefer a flexible type of work. The New Paper magazine in May 2019 has posted results of the research that finds that "Men have become as interested as women in more flexible work options". It stated that there is a cultural shift in the minds of nowadays "dads". (Lee, 2019)
The third reason for the growing gig economy is that the cost of living nowadays is rising and the middle class that could afford rising costs is decreasing, forcing people to look for the additional sources of income, such as working in the gig economy after working at their main place of work. The Guardian in the article claims that the cost of living for the poorer families rises with the higher pace than for the richest. It based the article on the report made by the Office for National Statistics. (The Guardian, 2019)
Hypothesis setting
The research in this paper intends to check the hypothesis:
H1: The rise of the gig economy platforms positively/negatively correlates with the dependent labor of the United Kingdom
Chapter 2. Empirical Set-up
1. Data sources and sample
Table 11. The description of the main variables used in the analysis
|
Variable |
Description |
|
|
internetuse |
Percentage of people in the region who has used the internet last 3 months |
|
|
annual pay (lnpay) |
Median pay in the region |
|
|
median age (medianage) |
Median age in the region |
|
|
jobs density (jobsdensity) |
Number of jobs per resident aged 16-64 |
|
|
education |
The percentage of students entering high education after school |
|
|
employees (lnemp) |
Number of employees in the region |
|
|
gig firms (lngig) |
Number of non-employee firms |
|
|
weekly allowance (lnweek) |
Size of the weekly payments to people who do not have a job |
|
|
GDHI |
Gross Disposable Household Income in the county, amount of money available for spending after income distribution measures |
> Main sources
The study analyzes the relationship between the gig economy and dependent labor in the United Kingdom. Data used for the estimation was obtained from the statistics published on the site of the Office of National Statistics of the United Kingdom and from the statistics published in the source called Nomis, that includes more detailed statistics on the labor market of the United Kingdom obtained from the Office of National Statistics.
Area
The database used in the research contains the data from 2014 to 2017 on the 216 counties of the United Kingdom, that comprise more generally 12 regions: North East, East Midlands, Yorkshire and the Humber, South West, West Midlands, East of England, North West, London, South East, Scotland, Wales, Northern Ireland.
To be able to estimate the relationship between the dependent employment and the rise of the gig economy for the United Kingdom as a whole, the paper would consider the statistics from all the regions of the United Kingdom. However, the information on some counties of the regions was not available. Due to some of the information was excluded from the analysis. Moreover, some information included in the initial statistics could include the problem of double counting. To avoid that problem some counties were excluded from the regression.
Year
The analysis takes into account the recent years from 2014 to 2017. These years are representative of the question connected with the gig economy, as the gig economy has been recently developing and coming into force.
The years included in the analysis were based on the availability of official and reliable data. Outliers for all variables were also excluded from the regression analysis. The data used in the analysis is available not for every region due to several factors: in some statistics, regions are divided differently, some statistics were collected only for special years.
Below there is a summary and the description of the panel data under analysis:
***Insert Table 1 here***
***Insert Table 2 here***
> Reasons to consider the United Kingdom for the research There are several reasons considering the topic, why it would be interesting to take the gig economy of the United Kingdom into consideration for the future analysis:
1) The gig economy in the UK follows a great growth dynamic. It has doubled between 2016 and 2019 (The Guardian, 2019)
UK's freelance market in 2019 is in top-2 based on year-on-year revenue growth (Payoneer's Global Gig-Economy Index of Q2 2019, Forbes)
> Summary of statistics
Below there is a summary statistic of regions-leaders in the number of employees in the region:
Figure 1. Counties-leaders in the number of employees, thousands, 2014-2017, source: Office of National Statistics
It could be seen that two years in a row (in 2014 and 2015) the greatest number of employees were in the region - City of Edinburgh. After 2015, the leader region was the Tower Hamlets, and after, in 2017 it was the Leicestershire.
Figure 2. Counties-leaders in the number of gig firms, 2014-2017, source: Office of National Statistics
As for the number of gig firms in the regions, in different years there are different regionleaders. As could be seen from the diagram above, in 2014 Buckinghamshire was the leader for the number of gig firms. In 2015, the place of the leader was given to the Lambeth, whereas in 2016 the Oxfordshire became a leader and in 2017 - Lancashire.
Figure 3. Counties-leaders in the percent of internet users in the last 3 months, % 2014-2017, source: Office of National Statistics
2. In the diagram above, the counties that had the maximum percentage of internet users in the different years are depicted. In 2014 the leader was the county, Milton Keynes, with 94,2% of recent internet users. Then, in 2015 the county Thurrock became a leader with 94,9% of recent users in the county. In 2016 the greatest percentage of the recent internet users, 96%, was in the county Brighton and Hove, afterwards, in 2017, the greatest percentage, 97,8%, was in the Isle of Anglesey.
3. Methods
The research in order to investigate the relationship between the increase of the number of gig firms in the economy and the dependent employment considers two models: two-stage least squares model and the Arellano-Bond model with lags of the dependent and exogenous variables. The research would afterward compare the results of both models.
1. Two stage-least squares model
The model used in the research for the estimation considers the instrumental variable model - two-stage least squares model, that uses the panel data for the analysis. The model in the research is built with the use of panel data model. The panel data is a set of observations on the same units obtained in several different time periods. (Kennedy, 2008) That could be observations of different countries, individuals, locations. In the current research, the units of observation are the county areas in the United Kingdom.
Assumptions of the two-stage least squares model:
1) The model is correctly identified
2) Errors are normally distributed
3) The variance of errors is equal for all variables
4) There are no outliers in the analyzed data
5) Independent observations
From the theoretical background, the method of instrumental variables is used to solve the problem of endogeneity in the regression.
The problem of endogeneity may arise due to several issues:
• there could be a problem of omitted variables in the regression
• there could be a measurement error in the estimation of one of the explanatory variables
The two-stage least squares method includes estimating the regression that includes the instrumental variable. There are four types of variables in that method: endogenous, exogenous, dependent variable, and an instrument variable. The model is performed in two stages. At the first stage, the role of the dependent variable plays the variable that is suggested to be endogenous.
So, at the first stage, the endogenous variable (for example, Xi) is regressed on the number of instruments (for example, R1i...Rni) and on the number of exogenous variables (for example, P1i...Pni) by which it could be influenced, using the Ordinary Least Squares model (OLS). Then after the regression is made, the predicted values of Xi are computed and are usually called XЈ.
In the second stage of the model, the main dependent variable is regressed on the exogenous variables and the predicted value(s) of the endogenous variable(s).
Two-stage least squares model would be used to determine the relationship between the gig economy in the United Kingdom and the dependent labor, allowing to answer the main research question in the current paper.
2. Arellano-Bond model
The research considers the possibility that there could exist an effect on the dependent employment from the lagged variables. To take into account lagged variables of the dependent variable and some independent variables in the panel data model - dynamic models should be used. This paper would implement the Arellano Bond model to consider the described effects, and then it would compare the results from the two-stage least squares model and the Arellano- Bond Model.
Arellano Bond estimator is an example of a dynamic panel data model, that includes lags of the dependent variable as regressors and takes into account unobserved random or fixed panel level effects. This estimator was first proposed to overcome the problems of endogeneity, serial correlation, heteroskedasticity in static panel data models (Arellano and Bond, 1991).
This model accounts for a generalized method of moments (GMM), that is suitable for data collections with a lot of panels and a small number of periods, and that is used for estimating the parameters. Using the GMM, the one can construct estimators, similarly the same as the Maximum Likelihood estimator, however, the estimator of the generalized method of moments would be more robust since GMM assumes specific moments of the random variables, while ML assumption applies to the whole distribution.
The generalized method of moments estimator is more efficient than the Method of Moments (MM) as the number of moments conditions that it takes into consideration is greater than the number of parameters.
3. In the Arellano Bond model, individual effects are removed by taking the first difference of the regression equation, after that, higher lags of the dependent variables are used as instruments for differenced lags of the dependent variables.
4. Theoretical background for the choice of variables
Dependent variable:
Number of employees
The research aims to estimate the effect of the gig economy on dependent labor. Dependent labor is the employees who work in the enterprise with the official long-term contract. The research would use the information on the number of employees in all the industries in the United Kingdom from the year 2014 to the year 2017. In order for the estimate to be more economically explainable, the analysis takes into account the logarithm of the number of employees, as it would be easier to interpret the result. The number of gig firms in the economy is not included in the statistics for this variable.
Endogenous variable:
Number of gig firms
Careful measurement of the growth of the gig economy is essential for analyzing current influence on the labor market. Unfortunately, the question relating to the gig economy is of the recent nature, consequently, there is no official statistics on the number of the gig economy workers in the economy. There is a survey that was undertaken to measure the number of gig economy workers. However, this research would not take into consideration survey results due to the reason that survey could contain several problems. Modern surveys could not answer questions about labor market trends as they are not appropriately suited. (Fox, 2014)
Problems that the survey may obtain:
1. Platform workers may classify themselves as employees especially when the platform work is a second or third source of income
2. There is evidence that survey participants could answer differently depending on the way the data is collected (face-to-face, online, by telephone)
3. Data is gathered infrequently and is often incomplete (Rozzi, 2018)
Moreover, online platforms that provide the ability for the gig economy to exist, are private companies, that are not obliged to disclose their number of employees.
In order to be able to estimate the effect of gig firms, the paper would use the proxy for the number of gig firms. The proxy would be the number of non-employee firms in the industries, where the development of gig firms is more widespread.
Industries that are chosen for the statistics purposes: transportation and storage, accommodation and food service activities, information and communication, administrative and support service activities, other services activities.
Considering the question in more detail, the data used in the research is the data for the nonemployee firms. Non-employee firms are the types of businesses that have no paid employees, and which pay federal income tax (Holtz-Eakin et al., 2017). The data was captured for the years 2014-2017 and was obtained from Nomis, Official Labour Market Statistics. It includes information on the firms with the number of employees from 0 to 4. Non-employee statistics do not account for flexible workers. That data set might not an ideal measure for the estimation of the gig economy as there could be some discrepancies. However, the measure, overall, could be used as a proxy for the size of the gig economy as was shown in several researches.
The choice of the non-employee firms to be the proxy for the number of gig firms in the economy was also based on several investigations that were undertaken previously. In a study undertaken by Holtz-Eakin et al. (2017) the data on the non-employer firms were used to measure the overall growth of the gig-economy workforce. Moreover, the research conducted by Schwellnus et al. (2019) to estimate the impact of gig economy platforms employment, taking into account the evidence from US counties, uses the number of non-employer businesses as a proxy for the gig economy. The research made by.. .tested the hypothesis that "the number of non-employer firms increases more in metropolitan areas where Uber comes in", after the investigation, the null hypothesis of no increase was rejected and the initial hypothesis of the increase was justified.
Exogenous variables:
Annual pay
The annual pay variable is the statistics on the median annual pay that is received by the fulltime employees in the region. The logic here follows the same pattern as with the variable of total jobs. The higher is the value of the median annual pay in the region, the less is the desire of residents to enter the gig economy with the incentive to earn for living. The statistics in the research uses the median annual pay, not the mean, because using the median is more reliable, as it is not affected so hardly by extreme values, as a mean figure.
Control variables:
Jobs density
By definition, job density is the number of jobs per resident aged 16-64. As an example, a job density of 1 means that there is one job for every resident in the county of working age.
The higher is the job density in the region, the larger there are opportunities that people have in the case of employment. That means the dependent employment could tend to increase with a higher job density. Moreover, there is evidence that there is a strong local positive correlation between job density and wages. (Hakansson and Isacsson, 2013) Higher wages could also be the motivator to enter the dependent employment.
The opposite effect would have job density on the gig firms. One could suppose that the higher is a job density in the region, the less is the desire of residents to enter the gig economy with the desire of earning money for living, because residents have several alternatives where they would like to work.
The data on the variable was obtained from the Office of National Statistics.
Median age
The median age in the region is a statistic showing the median age of the residents in the region. The prediction is that the higher is the median age, the lower would be the dependent employment in the region, county. Higher median age, in terms of aging, has a negative effect on the labor market. (Serban, 2012)
Moreover, as the paper stated before, nowadays a cultural shift towards more flexible work desire has happened. Some people want to spend more time with their families and devote more effort to their hobbies, so they may prefer a flexible type of work.
The data on the variable was obtained from the Office of National Statistics.
The variable week allowance indicates the money size of the allowance that is paid to the jobseekers of a special group.
The paper considers including the effect of this variable, as intuitively, there could be the case that the higher is the job seekers allowance in monetary equivalent, the less actively would people search for the job. The sign expected before this variable is negative. The paper by Tatsiramos (2014) claims that the higher the benefits, the larger is the unemployment period. Moreover, the paper states that with higher unemployment benefits, people undertake less effort to find a job. Models created by Bover et al. (1997) showed that when people receive unemployment benefits, that reduces the desire to leave unemployment.
The data on the variable was obtained from the Office of National Statistics.
Instrumental variables:
Internet use
The model uses the percentages of people using the internet in different regions in different years. As gig platforms are functioning due to the existence of the internet, the more some people are using the internet, the greater is the size of the gig economy. That means, the regression model with the dependent variable the number of gig firms, should include the percentage of internet users as one of the regressors. The paper by Schwellnus et al. (2019) considers the growth in the use of the internet connection to be a strong predictor of the emergence of the gig firms.
The data used in the analysis is the percentage of residents in the region who have used the internet connection in the last three months.
Dummy variables The year 2014
In the first stage of the two-stage least squares model the paper includes a dummy variable of the year 2014, as in this particular year, the changes related to the employment law had taken place, and these changes had an influence on the workers in the gig economy.
4. ***Insert Table 3 here***
5. Model
Two-stage least squares model
In the current research, the two-stage least squares model has the form of: 1st stage:
lngigt = Yo+ YilnPayt + Y2medianagei + y^Jobsdensityi
+ Y^lnweeki+Yseducationi + Yeinternetusei + Y7lnGHDIiet
2nd stage:
Inempi = a0 + a^ngigt + a^nPayi + a2medlanagei + a^Jobs density i + a^lnweeki + st
On the first stage, the endogenous variable (number of gig firms in the economy) is regressed on the number of instrumental variables (the logarithm of the people who receive the allowance and the percentage of internet users), and on the exogenous and control variables (the logarithm of the median pay in the economy, the median age, the jobs density, the logarithm of the weekly payments to jobseekers).
On the second stage, the dependent variable (the number of employees in the economy) is regressed on the predicted value of the InGigi, then on the same exogenous and control variables.
The choice of the model is based on the existing endogeneity problem in the model. One of the main issues in the research is limiting the problem of endogeneity that could exist due to the factor omitted from the regression analysis, which influences both the dependent variable (number of employees in the economy) and the regressor (number of firms in the gig economy). Moreover, it could be seen that the correlation between the number of employees and the number of firms in the gig economy is large, so the instrumental variable is needed. Instrumental variables should be such that they have a low correlation with the dependent variable (number of employees).
Instruments used in the model
The instrumental variable method in the paper considers the instruments for the analysis that could influence the gig economy in the county, and that is exogenous to the dependent employment.
The paper by Schwellnus et al. (2019) uses the variable related to the share of internet users as an instrument in the two-stage least squares model to determine the relationship between the growth in the gig economy and the growth in the dependent labor. It suggests that it could be used as an instrument because the variable is exogenous to the dependent labor growth.
In the current paper, the statistics on the percentage of recent internet users are considered, the paper suggests that the instrument is valid, as it is exogenous to the dependent employment. Moreover, the increased use of the internet in the county makes it affordable to the greater number of people the opportunity to enter the gig employment.
Education
That is the variable showing the percentage of students, which have entered high education after school. That instrument has not been studied before in the researches. The paper considers that instruments because no variable could directly influence the percentage of people entering high education and the number of employees at the same time, so, the instrument is exogenous.
However, as was noticed in the theoretical framework, the emergency of the gig economy platforms reduces the barriers to enter the labor market - the person is not obliged to obtain a higher education, special rating mechanisms exist. So, the instrument education may influence the gig economy, such as fewer people entering higher education might drive the growth of the gig economy.
Gross Household Disposable income (GHDI)
Another instrument considered in the model is the value of the gross household disposable income. There seems to be no variable that has a direct influence on the GHDI and the dependent employment at the same time, so the variable is exogenous. The GHDI also could influence the gig economy development, as the low income of the households forces them to search for other ways of earning income.
What is important for the instruments, when they are chosen is that there is no correlation between an instrument and a dependent variable for which the instrument is built. It could be seen from the correlation matrix below that there is no significant correlation between the dependent employment and education variable, dependent employment and the GHDI, dependent employment, and the internet use variable.
> Hausman test
Before running the model on the panel data, it is important to choose between the fixed effect and random effect model. The null hypothesis of the Hausman specification test is that there is no correlation between individual effects and any other regressor in the model under consideration. (Hausman, 1978)
***Insert Table 9 here***
From the results of the test conducted on the models with fixed and random effects, it could be concluded that the null hypothesis is failed to be rejected. Consequently, the random effect model is preferred.
> Breusch-Pagan Lagrangian multiplier test for random effects Next, it is important to choose between the random effect model and the pooled model. It could be done using the Breusch-Pagan Lagrangian multiplier test for random effects. The null hypothesis of this test is that the specific variance component is equal to zero.
***Insert Table 6 here***
From the table above it is seen that the null hypothesis is rejected, concluding that the randomeffects model is preferred to the pooled model.
6. Results and appropriate tests
For the interpretation of regression results, the paper uses the method of multiplication of the regression coefficient by the standard deviation of the variable. That would provide a more meaningful interpretation.
1) Two-stage least squares model > First stage
As the problem of endogeneity could arise in the model, the paper used the two-stage least squares model. Moreover, the test has shown that it would be more appropriate to use a random effect model. At the first stage of the variable lngig were regressed on the number of instrumental, exogenous, and control variables.
Below there is a result of the estimation of the first stage, that is depicted by the regression (2), without dummy variable:
***Insert Table 4***
As could be seen from the regression model, several variables have a significant correlation with the appearance of the firms in the gig economy (number of non-employee firms taken as proxy).
Firstly, it could be seen that there is a significant negative correlation between the median annual pay in the region on the number of people entering the gig economy. It could be seen that there is a significant negative relationship: a 1 percent increase in the median annual pay in the county decreases the number of participants by 0,105 percent. Intuitively, it seems to be logical, as the higher is the median pay, fewer people would enter the gig economy to earn for living purposes.
Secondly, there is a significant correlation between the number of people who receive job seekers allowance and the number of firms in the gig economy. When the number of people who receive job-seekers allowance increase by 1 percent, the number of gig forms increases by 0,0224 percent, meaning that there is a positive relationship between two variables. That dependency could be explained by the fact that people who leave the dependent employment and start receiving jobseekers' allowance, maybe entering the gig economy, influencing the number of gig firms in the region.
...Подобные документы
Характеристика источников информационного обеспечения анализа краткосрочных активов. Изучение их состава, структуры, динамики. Расчет общей рентабельности организации и показателей эффективности краткосрочных активов. Оптимизация производственных запасов.
курсовая работа [427,4 K], добавлен 01.12.2016Сущность и общая характеристика модели неблагоприятного отбора. Особенности информативных сигналов. Понятие и значение подоходного налога, капитала и трудовых контрактов. Динамические аспекты теории контрактов. Репутация и карьерные соображения.
курс лекций [517,7 K], добавлен 28.02.2011Контракт как основополагающей формой человеческого взаимодействия. Минимизация издержек явного (письменного) внесения. Сущность, типы и особенности контрактов. Отношения управления поведением исполнителя. Теоретические подходы к объяснению контрактов.
контрольная работа [28,6 K], добавлен 02.04.2009Понятие и типы контрактов. Основные принципы контрактных отношений. Полные и неполные, явные и неявные контракты. Направления развития теории контрактов. Теория агентских отношений. Решение проблемы ухудшающего отбора. Проявления морального риска.
презентация [518,2 K], добавлен 30.08.2013Теория проведения аудита, сущность и типы краткосрочных активов. Этапы и практика проведения аудита краткосрочных активов. Проверка касс и счетов. Аудит инвестиций и оформление результатов. Проверка расчетов с дебиторами. Методы проверки запасов.
курсовая работа [37,8 K], добавлен 10.09.2010Понятие и элементы контракта, параметры измерения. Классификация и типы контрактов, отличительные особенности совершенных и несовершенных. Ex ante контрактный процесс: проблема неблагоприятного отбора, морального риска и вымогательства, их решение.
презентация [270,7 K], добавлен 12.05.2013Нефтяные контракты и их характеристики. Углеводородное законодательство. Соглашение о разделе продукции. Типичное сочетание групп платежей инофирмы в соглашениях. Особенности отдельных видов нефтяных контрактов, заключенных зарубежными государствами.
курсовая работа [44,7 K], добавлен 22.11.2013Экономическая характеристика и содержание оборотных активов и краткосрочных финансовых обязательств предприятия. Политика оперативного управления оборотными активами. Динамика состава и структуры краткосрочных финансовых обязательств ООО "Центрострой".
дипломная работа [752,0 K], добавлен 10.02.2012Объекты, информационное обеспечение, методика анализа долгосрочных и краткосрочных активов промышленной организации. Состав, динамика и структура долгосрочных и краткосрочных активов в составе всех активов организации. Структура дебиторской задолженности.
курсовая работа [522,0 K], добавлен 29.11.2014Исследование макроэкономического равновесия на рынке труда. Анализ рынка труда и заработной платы в условиях реформирования экономики. Механизм формирования спроса на труд и его предложения. Методы государственного регулирования занятости и рынка труда.
курсовая работа [454,2 K], добавлен 03.04.2015Особенности взаимодействия спроса и предложения на рынке труда. Различия в зарплате высококвалифицированных и низкоквалифицированных работников. Тенденции в области защиты занятости стран ОЭСР. Сравнительный анализ рынков труда Великобритании и Франции.
курсовая работа [87,7 K], добавлен 03.03.2011Анализ абсолютных показателей финансовой устойчивости предприятия, наличия оборотных капиталов, объема краткосрочных кредитов и заемных средств. Сравнение активов и пассивов по степени ликвидности, их влияние на платежеспособность предприятия "Эфест".
курсовая работа [358,8 K], добавлен 08.01.2015Общая характеристика экономической системы Великобритании, история и основные этапы ее формирования, повлиявшие факторы. Политика в отношении присоединения к евро. Анализ основных макроэкономических показателей, оценка уровня инфляции в государстве.
контрольная работа [424,2 K], добавлен 15.12.2014Методика анализа влияния изменений структуры российской экономики на динамику производительности труда в экспортируемых и неэкспортируемых секторах экономики в рамках TNT-модели. Темпы роста производительности выше в секторе экспортируемых товаров.
контрольная работа [45,2 K], добавлен 28.10.2010Анализ рынка труда и социально-трудовых отношений. Исследование сущности, причины безработицы и ее социальных последствий. Классификация безработицы. Основные направления политики занятости. Количество безработных в Нижегородской области и Н. Новгороде.
контрольная работа [38,1 K], добавлен 26.02.2010Различные интерпретации модели ненаблюдаемых действий и области их применения. Адаптация модели к рынку труда. Оптимизация реального трудового контракта. Причины ограниченного использования модели при разработке договоров о трудовом сотрудничестве.
курсовая работа [310,1 K], добавлен 26.11.2013Институты, их классификация и виды. Роль институтов экономической системе. Основные понятия теории прав собственности. Понятие и виды тансакций. Понятие и классификация контрактов. Сущность и природа экономических организаций. Государство как институт.
шпаргалка [78,4 K], добавлен 27.05.2010Процесс формирования цены на товар или услугу. Определение сущности ценообразования на товарных и фондовых биржах. Методы исчисления стоимости акций, облигаций и краткосрочных ценных бумаг. Анализ ценовой политики на товарных биржах финансового рынка.
курсовая работа [47,5 K], добавлен 28.11.2012Сущность рынка труда, инфраструктура и особенности. Характеристика рынка труда в современной России. Виды и формы занятости. Перспективы выхода российской экономики из кризисного состояния. Регулирование рынка труда, устранения диспропорций и деформаций.
курсовая работа [351,1 K], добавлен 15.01.2009Сущность и назначение корреляционного анализа в статистике, основные этапы его реализации. Краткая экономическая характеристика Великобритании и Венгрии. Корреляционный анализ экономики данных государств, показателей прироста иностранных инвестиций.
курсовая работа [181,4 K], добавлен 25.06.2010
